| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Biology, University of Victoria, Victoria, British Columbia V8W 2Y2, Canada
 |
Abstract |
|---|
 Top Abstract I. IntroductionII. Evolution of the...III. Superfamily Evolution by...IV. Superfamily...V. Conservation of PACAP—A...VI. Conclusions and Future...References |
|---|
I. Introduction
II. Evolution of the Superfamily—Two Special Members
A. PACAP as the most tightly conserved family member
B. GRF with rapid structural changes in evolution
C. GRF/PACAP gene - a duplication late in evolution
III. Superfamily Evolution by Exon and Gene Duplication
A. Six single genes in human superfamily
B. Members of superfamily in many vertebrates
C. Key to origin of superfamily in protochordates
D. Extension of PACAP ancestry uncertain in insects
E. Exon duplication as original step in ancestry
F. Gene duplication crucial in creating gene families
G. Peptide elongation important in some families
H. Alternative splicing for expansion of individual families
I. Alternative promoters for expansion of family functions
J. Posttranslational processing for increase of peptide families
IV.Superfamily Members—Overlapping Functions, Expression, and Receptors
A. PACAP and VIP
B. PHM
C. Glucagon
D. GLP-1
E. GLP-2
F. GRF
G. GIP
H. Secretin
V. Conservation of PACAP—A Clue to Function
A. A regulator of the cell cycle and development
B. A regulator of smooth and cardiac muscles
C. An immune system regulator
D. A regulator of bone metabolism
E. An endocrine/paracrine regulator
F. An exocrine regulator
G. A regulator in the nervous system
VI Conclusions and Future Directions
|
I. Introduction |
|---|
|
TopAbstractI. IntroductionII. Evolution of the...III. Superfamily Evolution by...IV. Superfamily...V. Conservation of PACAP—A...VI. Conclusions and Future...References |
|---|
and include PACAP, glucagon,
glucagon-like peptide-1 (GLP-1), glucagon-like peptide-2 (GLP-2),
GH-releasing hormone (GRF), vasoactive intestinal polypeptide (VIP),
peptide histidine methionine (PHM), secretin, and glucose-dependent
insulinotropic polypeptide (GIP). We have just begun to understand the
origin of this rapidly evolving family of peptides and its many
important functions.
|
Meanwhile, there has been an explosion of information about the other glucagon superfamily members in regard to sequences of the peptides, cDNAs, and genes. The location of many of the family members in the gut or pancreas was only a beginning to finding the peptides widely distributed in several tissues and species. For example, secretin was considered to be a gastrointestinal peptide but is now found in the gonads, brain, and developing pancreas (Section IV). Wide distribution is a recurring theme for the superfamily in which all but one of the members have been isolated from the brain in vertebrates; in addition, many are present in the gastrointestinal, pancreatic, and gonadal organs. Even the newly discovered (1989) family member named PACAP (6) has been identified in both brain and gonads; it is given special consideration in this review (Section V) as it is the most likely ancestral molecule for the superfamily.
The multiple functions of the peptides in the PACAP/glucagon superfamily overlap and continue to grow in number. For neuropeptides in general, Leslie Iversen (7) argues that the coordinated functions of single neuropeptides remain unclear despite our progress in identifying family members, receptor subtypes, and specific antagonists. Iversen states "To understand the wider biological significance of neuropeptides we need, perhaps, to look at some of their functions in simpler organisms." He suggests that one of the best-known examples of a simple peptide system is the egg-laying behavior in Aplysia in response to the release of several peptides from the bag cells; these peptides act on both the nervous system and gonads. In vertebrates, many of the PACAP/glucagon superfamily members also have actions in several organs. Thus, it seems appropriate to investigate the origin of the PACAP/glucagon superfamily with a view to identifying both the structures and functions, especially coordinated functions that have evolved in invertebrates and vertebrates.
|
II. Evolution of the Superfamily—Two Special Members |
|---|
|
TopAbstractI. IntroductionII. Evolution of the...III. Superfamily Evolution by...IV. Superfamily...V. Conservation of PACAP—A...VI. Conclusions and Future...References |
|---|
). The sequence conservation is shown by
the complete identity of the amino acid sequence for mammalian PACAP
peptides. The chicken and frog forms of PACAP each have only one amino
acid change in comparison with the mammalian form, whereas fish PACAP
peptides have three or four changes (89–92% amino acid identity) with
mammalian PACAP. With one exception, amino acid substitutions in fish
PACAP peptides are in the C-terminal region beyond amino acid 27.
|
). The tunicate PACAP-1
cDNA encodes a PACAP of only 27 amino acids, whereas the vertebrate
PACAPs are present as both a 38-amino acid form and a truncated
27-amino acid form that is identical to the N-terminal portion of
PACAP-38. In the tunicate the 27-amino acid form is the result of a
stop codon in the cDNA encoding PACAP-1 (23). There is a striking
conservation of the tunicate PACAP-1 amino acids: 26 of 27 amino acids
are conserved in comparison with the mammalian, frog, salmon, catfish,
and stingray PACAPs. The sequence conservation is not limited to the
peptides as shown by a nucleotide identity of 96% between the human
and tunicate PACAP-1 cDNAs. In addition to the tunicate PACAP-1 cDNA, PACAP-2 cDNA was isolated and found to encode a 27-amino acid peptide. The difference between PACAP-1 and -2 is that the latter has 4 amino acid changes compared with the mammalian form. Nevertheless, the high identity that is found between the two tunicate cDNAs and peptides suggests that the two PACAPs in tunicate probably originated from a gene duplication. This duplication could have occurred in the tunicate lineage at any time over the last 700 million years or in an organism that originated before the phylogenetically ancient tunicate. A peptide (amnesiac) has been identified in Drosophila with 18% sequence identity to human or tunicate PACAP-27; another peptide (maxadilan) in the sand fly (Lutzomyia longipalpis) has 15% identity with PACAP-27 and 16% with PACAP-38. Evidence to date does not clearly support these peptides as homologous to PACAP (see Section III.D. below). It is clear from the tunicate data that the evolutionary pressure to maintain the PACAP amino acid sequence is high; other known members of the PACAP/glucagon superfamily do not have such high conservation of amino acid sequences. As discussed in Section V, evidence as to why the primary structure of the PACAP family is so tightly conserved in terms of function is beginning to emerge.
B. GRF with rapid structural changes in evolution
Not quite 20 yr have passed since a GH-releasing factor (GRF) was
isolated and sequenced from human pancreatic tumors. Rivier and
co-workers (24) in 1982 found a 40-amino acid GRF peptide with a free
carboxy terminus within their tumor extract. In addition to a 40-amino
acid form, Guillemin et al. (25) in 1982 found a 44-amino
acid, amidated GRF peptide as well as a 37-amino acid peptide (Fig. 2)
from a different single pancreatic tumor.
In 1984 the hypothalamic form of the GRF peptide was sequenced and
reported to be 44 amino acids and identical to the pancreatic tumor
sequence (26). In subsequent years, the primary sequences of GRF from
16 vertebrate and 1 protochordate species have been identified (Fig. 2). A total of 15 distinct sequences were identified from human, pig,
cattle, goat, sheep, hamster, rat, mouse, chicken, salmon, carp,
catfish, and tunicate (Figs. 2 and 3).
Of the 15 GRF sequences known, the GRF peptide sequence was determined
by protein chemistry from 7 of the species. In addition, the cDNA
and/or gene was isolated and the GRF peptide was deduced for the human
(28, 29, 30) hamster (34), mouse (37, 38), chicken (15), salmon (18, 19),
catfish (20), and tunicate (23).
|
). Little or no conservation is seen in the remaining
amino acids (Fig. 2). The tunicate GRF-1 and GRF-2 are both 59%
identical in amino acids to human GRF and 89% identical to each other.
Not only sequence but also the length of GRF peptides varies among
species. The human GRF peptide is 44 amino acids in healthy tissue.
Other mammalian GRFs show a full-length peptide varying from 44 amino
acids in cow, goat, and sheep to 43 amino acids in rat and 42 amino
acids in mouse. All seven fish sequences encode a protein of 45 amino
acids (Fig. 2). Some of the changes in the GRF sequence are due to use
of different splice sites in intron 4, which interrupts GRF; this
mechanism partially explains differences in the C-terminal portion of
human and rat GRF (2). Likewise, the chicken gene produces two peptides
of different lengths, 43 and 46 amino acids due to alternative
processing at the intron 4-exon 5 boundary. Only the human, pig, cow,
goat, and sheep GRF peptides are amidated at the carboxy terminus.
The most ancient GRFs isolated to date are those of the protochordate in which a GRF-like peptide is shorter and encoded on a single exon. The tunicate cDNAs that encode PACAP-1 and -2 also encode the tunicate GRF-like peptides (termed tunicate GRF hereafter) (23). Similar to tunicate PACAP-1 and -2, the tunicate GRFs are only 27 amino acids in length; this length is much shorter than all other known GRFs, but is similar to the 29-amino acid bioactive core of vertebrate GRFs (140, 141). The tunicate 27-amino acid GRF-1 and GRF-2, as deduced from the cDNAs, are the result of posttranslational processing of the precursor proteins at the dibasic site between the tunicate GRF and PACAP. Longer forms of GRF cannot be translated because the nucleotides that encode PACAP follow immediately after the GRF sequence. The function of tunicate GRF is not known and indeed, GH has not been identified in tunicates to date.
C. GRF/PACAP gene—a duplication late in evolution
It was suggested previously that the DNA encoding GRF evolved into
a distinct gene about 750 million years ago (142). This idea was based
on protein biochemistry, peptide sequence comparisons, and the
demonstration that the mammalian genome encodes GRF and PACAP on
separate genes and chromosomes. However, recent evolutionary studies
have shown that both peptides are encoded in the same gene in birds
(15), fish (18, 19, 20), and tunicates (23). Therefore, the evidence that
GRF and PACAP are encoded in one gene supports the hypothesis that a
separate GRF gene evolved only about 250 million years ago. The current
theory is that the GRF-PACAP gene duplication was followed by
nucleotide substitutions giving rise to different genes for GRF and
PACAP. This is thought to have occurred after separation of the avian
and mammalian lineages. Neither the exact evolutionary time nor taxa
are known for the GRF-PACAP gene duplication, but further studies on
extant species will help to delineate when the gene duplication
occurred (see Section III.B, paragraph 4).
The function of the two peptides encoded in one gene is interesting. In fish, GH is released by both PACAP and GRF, provided the correct fish form of the peptide is used. In salmon and goldfish, synthetic PACAP (salmon, ovine, frog, or zebrafish forms) and GRF (salmon and carp forms) were compared for their ability to release GH from pituitary cells in vitro. In salmon, GRF-45 (the salmon form) released GH at concentrations between 10-12 and 10-8 M, and PACAP-38 (salmon form) was effective at 10-10 to 10-7 M, but the dose-response curve was more consistent for PACAP (19). In goldfish, several peptides released GH at 10-10 to 10- 6 M; the potency (derived from ED50) was carp GRF-45 > zebrafish PACAP-38 > ovine PACAP-38 > ovine PACAP-27 > zebrafish PACAP-27 = frog PACAP-38 >> mammalian VIP (143). In cultured pituitary cells of eels, human PACAP-27 and PACAP-38 were effective in releasing GH, but human GRF was not effective (144), as might be expected from the rapid change in GRF structure in evolution. This evidence suggests that PACAP is a hypophysial factor in nonmammalian vertebrates. Even more intriguing is the observation that alternative splicing, which occurs in brain tissue, results in a short transcript encoding PACAP without GRF, in addition to a full-length transcript with GRF and PACAP encoded. The short transcript implies that the brain can selectively produce more PACAP than GRF, but the reason remains unknown. In mammals, GRF or PACAP can release GH, but there is debate as to whether PACAP acts directly on somatotropes (Ref. 145 and see Section V.E.1).
|
III. Superfamily Evolution by Exon and Gene Duplication |
|---|
|
TopAbstractI. IntroductionII. Evolution of the...III. Superfamily Evolution by...IV. Superfamily...V. Conservation of PACAP—A...VI. Conclusions and Future...References |
|---|
. The six
genes encode at least nine bioactive peptides (PRP has not been shown
to be bioactive). The VIP and glucagon genes encode two or three
bioactive peptides, whereas the PACAP, GRF, secretin, and GIP genes
encode only a single bioactive peptide. Nonetheless, the structural
organization of all the genes is similar, and the precursor molecules
are even more striking in their similarity. Each gene encodes a signal
peptide, an N-terminal peptide (also named a cryptic peptide), one to
three bioactive peptides, and a C-terminal peptide. Each of the six
family members in humans is a single copy gene. Specific receptors of
the seven-transmembrane type have been identified for each of the
bioactive peptides except PHM in the superfamily (Section
IV).
|
), and the
organization of the precursor and gene compared with that of PACAP is
likewise similar (Fig. 4). The structure of the human PACAP gene was
reported by Hosoya and colleagues (8) in 1992. The complete human
VIP gene sequence has been reported by four different groups (41, 42, 43, 44).
Although the VIP gene has seven exons compared with PACAP’s five
exons, the two genes have a marked similarity. Only the VIP and PACAP
genes have an additional exon between the signal peptide and one of the
biopeptide genes (Fig. 4). The function of this cryptic or spacer
protein is not clear, but the extended protein may be useful for
folding of the precursor before cleavage. Exons 4 and 5 in both genes
are thought to result from exon duplications (Section
III.E). It is the peptides (PACAP or VIP) encoded on exon 5 that have well documented and potent effects. Also, exon 4 in the VIP gene encodes PHM, which is reported to have fewer, but similar, functions to VIP in regard to the gut and to pituitary and pancreatic hormone release (see Ref. 67). In contrast, the nucleotides of exon 4 in the PACAP gene of mammals encode PACAP-related peptide (PRP), which is not a functional peptide in studies done to date (146), most likely due to the substitution of Asp for His or Tyr in the N-terminal position (147). However, in animals that evolved before mammals, the PACAP gene does encode a functional peptide (GRF) on exon 4. The appearance of a nonfunctional peptide on exon 4 may be the result of a gene duplication in the early mammalian lineage (15) resulting in separate genes for PACAP and GRF. Finally, the VIP gene encodes on exons 5 and 6 a C-terminal 15-amino acid peptide with no known function and encodes on exons 6 and 7 the 3' untranslated region. These double exons beyond the exons encoding bioactive peptides are also present in the GIP and secretin genes.
The glucagon, GRF, GIP, and secretin genes form a separate group of
genes compared with PACAP and VIP genes because they lack an exon
between the exons encoding the signal peptide and bioactive peptide
(Fig. 4). However, it seems clear that all six genes are in the same
line of evolution because each of the four genes lacks exactly one
exon. The N-terminal peptides encoded in the four genes results from a
small region at the 3'-end of the exon with the signal peptide and the
5'-end of the exon with the bioactive peptide (Fig. 4). Hence, each of
the six precursors has a cryptic or N-terminal protein 5' to the
bioactive peptide. It is an open question whether the PACAP and VIP
genes gained, or the other genes lost, an exon. It is also noteworthy
in the GRF, GIP, and secretin genes that only one bioactive peptide is
encoded. A C-terminal peptide is encoded on the exon adjacent to the
bioactive peptide exon(s), but a function is not known.
The glucagon gene is unique in that it encodes three bioactive
peptides: glucagon, GLP-1, and GLP-2 (Figs. 3 and 4). Each peptide is
encoded on a separate exon, which suggests that there were exon
duplications (Section III.E). The posttranslational
processing of the precursor varies in specific tissues and species and,
in addition, alternative splicing of the transcript is tissue specific
in vertebrates other than mammals.
The GRF gene is best compared with the PACAP gene as they are thought
to be duplicates from early mammals. Although the GRF and PACAP genes
have five exons, the PACAP gene has gained exon 3, creating coding for
a longer N-terminal peptide. Instead, the GRF gene has an additional
exon at the 3'-end encoding the 3'-untranslated region (UTR). The
advantage of either exon is not clear but may be related to folding of
the precursor or stability of the mRNA. The comparison between the GRF
and PACAP genes is important as both GRF and PACAP appear to be encoded
on a single gene in birds, fish, and tunicates (Fig. 4). The GRF gene
can be also viewed as having six exons because exon 1 in brain
transcripts is distinct from exon 1 in the placental or testicular
transcript (148).
The GIP gene has six exons with the GIP hormone encoded on exons
3 and 4 (Fig. 4) (129, 130, 136). Thus, the biologically active core
(30 amino acids) is encoded on exon 3 (35 aa) with the remaining 7 aa
encoded on exon 4. The N-terminal peptide spans exons 2 and 3; the
C-terminal peptide spans exons 4–6.
In terms of evolution, it is interesting that the two gut hormones, GIP
and GLP-1, in humans have 41% identity of amino acids between the two
peptides and between the receptors. This percent identity suggests the
peptides have been separated for a long time. However, the origin of
GIP has been traced only to birds to date (Fig. 3), whereas GLP-l is
present even in jawless fish. The separation of the two genes is
difficult to deduce until fish have been carefully analyzed for the
presence of GIP. Within mammals, GIP is highly conserved in that rat
and human GIPs are 95% identical in amino acids (137).
Finally, the secretin gene is the shortest gene of the superfamily in
the sense that only four exons are present (Fig. 4). The first exon in
the other glucagon family genes is always one that contains the 5'-
untranslated region exclusively. The secretin gene lacks this exon,
which means the gene lacks an intron between the transcriptional and
translational start sites. Compared with the vertebrate PACAP and VIP
genes, the secretin gene also lacks the exon that occurs between the
exon encoding the signal peptide and the exon encoding secretin (4, 5).
The secretin hormone is encoded on the second exon, whereas an
N-terminal peptide spans exons 1 and 2, and a C-terminal peptide is
encoded on exons 2, 3, and 4 (Fig. 4). The C-terminal peptide can be 44
or 72 amino acids depending on whether alternative splicing removes
exon 3 (Section III.H). The splicing of this exon leaves
only three exons and leads to speculation that the ancestral secretin
gene was three exons. Meanwhile, neither the N-terminal peptide (10
amino acids) nor the C-terminal peptide is known to have a function
unless it is to maintain a folding pattern that allows cleavage of
secretin from the precursor.
One of the earlier rules for the superfamily was that position 6 is
Phe. However, evolutionary studies show that rabbit secretin has
Leu6, chinook salmon and catfish GRF have Leu6,
and several GLPs have Tyr6 (Fig. 3). These are
not radical substitutions (139). In PRPs, Ile6 is
a radical substitution, but the peptide does not appear to be
functional.
A comparison of the promoter regions for the six superfamily genes is of considerable interest to determine how regulation of gene transcription has evolved. These studies are in their infancy, although the promoter region responsible for tissue-specific expression has been analyzed in the VIP and glucagon genes.
For the VIP gene, the transcription initiation site has been determined (42). In the promoter region of the VIP gene, there are three TATAAA boxes (44, 149), AP2 consensus sites (149), a cAMP-responsive element (CRE) within 100 bases from the start site (44, 149, 150), a cytokine-responsive area at about 1 kb upstream (see Ref. 151), and an area containing repeated DNA sequences, which are not Alu sequences (43). A tissue-specific element is 4–5 kb upstream from the transcription start site (150). The importance of the tissue-specific element was shown indirectly by inappropriate expression in tissues when only 2 kb of the VIP promoter was present (151).
For the PACAP gene, analysis of promoter locations supports the idea that the mammalian GRF and PACAP genes are products of a duplication. Thus, in rat the PACAP and GRF genes each contains a testis-specific promoter and first exon that are 13.5 kb (PACAP) or 10.7 kb (GRF) upstream from the transcription start sites used in the hypothalamus (148, 152, 153). In addition, there are several transcription start sites for PACAP in the hypothalamus in rat (153) and mouse (14), but all are part of the proximal promoter region. In the human PACAP gene, the transcription start site was not definitively identified (8), but at least one start site was identified for the chicken and salmon genes (15, 19). For promoter analysis, specific binding sites have not been tested, but consensus sequences that resemble response elements have been noted. For example, several potential Sox5/SRY sites were noted in the upstream testis-specific PACAP promoter (153).
The glucagon gene also contains a TATAAA sequence (82, 154) in the
5'-flanking region. The regulation of the glucagon gene results in
tissue-specific expression in the pancreas, intestine, and brain (see
Ref. 155 for a recent review). In experiments in which the glucagon
promoter was fused to the coding sequences for SV40 large T antigen,
only 850 bp of proximal promoter were sufficient for expression in the
pancreas and brain (156), whereas 2.0 kb were needed for additional
expression in the intestine (157). However, a fusion gene with 1.3 kb
of the glucagon promoter and the coding region of luciferase resulted
in expression in the intestine (158, 159). In more detailed experiments
using a variety of glucagon promoter segments fused to the coding
sequence of luciferase, it was shown that the promoter region between
-1.3 and -2.2 kb affected expression in both the islet and intestinal
cells, but the regions that enhanced or suppressed expression in the
two tissues were not always the same. Other tissue-specific elements
have been identified (see Ref. 160), including four DNA control
elements (G1 to G4) within the first 300 bp of the rat glucagon
promoter (161). A number of responsive elements that bind transcription
factors in the glucagon promoter have been studied: the G1 element
includes nucleotides between -60 and -118 that bind proteins
restricting expression of glucagon to the 
-cells in the islets
(162); the G1 and G3 elements bind the Pax-6/Cdx-2/3 heterodimer for
gene activation (163); a G2 element from -165 to -200 is a calcium
response element that binds hepatic nuclear factor (HNF)-3ß
(winged-helix family) and NFATp (164). Also, G2 binds an
Ets-like protein (164) and HNF-3 (165); the G3 element (-234 to
-274) includes an insulin-responsive element that inhibits
glucagon gene transcription (166); the G3 element binds a PISCES
protein and a winged helix protein to activate transcription (167); an
E box in the G4 region binds a heterodimer E47/BETA2 that
transactivates glucagon gene transcription unless E47 is overexpressed
(161); a cAMP-responsive element (CRE) of eight nucleotides binds cAMP
response element binding protein (CREB) (161); and TCATT motifs that
are adjacent to the CRE sites bind protein factors that modulate CREB
(168). Further proof that HNF-3 is an important activator of
transcription for the glucagon gene is shown in that loss of HNF-3
by a null mutation resulted in hypoglycemic mice and a marked reduction
in plasma glucagon (165).
For the GRF gene, regulation of the transcription has been difficult to study because there is a lack of suitable cultured neurons. Lack of a consensus TATA box in the promoter is thought to be the basis for multiple transcription initiation sites. One important factor that controls the expression of GRF is GH: transcripts of GRF mRNA increase if GH is eliminated by removal of the pituitary, and GRF mRNA transcripts decrease if GH levels are high as in transgenic mice that overexpress GH (169). Specific analysis of response elements in the promoter of the GRF gene have not been reported, although it is known that an upstream (10 kb) promoter is used as the transcription initiation site in placenta compared with the one used in the brain resulting in different exon 1 sequences for the 5'-UTR (170, 171). Also, like PACAP, the GRF gene has a testis-specific promoter that is 10.7 kb upstream from the promoter used in the hypothalamus (148). Again, in the testis transcripts, exon 1 is not the same as the one used in the hypothalamus.
For the GIP gene, two or possibly three transcription start sites were found, but all were within three bases of the other (130). In the flanking region, a TATA and CAAT motif were found as were several consensus sequences of Sp1, AP-1, AP-2, and CRE within 400 bp of the start site. An interesting observation by Higashimoto and Liddle (136) about the GIP gene was the presence in intron 1 of putative TATA and CCAAT boxes and several translational cis elements. Indeed, they found minor amounts of a short transcript that starts from exon 2 during development in the rat intestine. The use of exon 2 for the initiation of both transcription and translation is of interest because the tunicate PACAP genes and vertebrate secretin genes lack the usual exon 1 found in other superfamily members and also use a single exon as the initiation site for both transcription and translation.
The secretin gene promoter contains a TATA box, an E box, and consensus sequences for transcription factors AP-2 and Sp1 (4, 172). A promoter region of only 174 bp of flanking sequence was sufficient for maximal expression of secretin mRNA in HIT cells (173). Within this region is an E box (CAGCTG motif), which is similar to the core of enhancer elements found in the rat insulin I gene; this motif may explain expression of secretin mRNA in the developing pancreas. Mutations in the CAGCTG region decreased transcription to less than 20% (173). Evidence suggests that a transcription factor BETA2 (a basic helix-loop-helix factor) forms a heterodimer with either E12 or E47 in activating transcription of the secretin, glucagon, or insulin gene by binding to their E boxes (174). Secretin is not expressed in the S cells in the gut if the BETA2 gene is disrupted (174). In addition, BETA2 has a coactivator p300 that enhances secretin gene transcription; BETA2 and p300 together block cell division in the S cells (175). In normal mice, it is not known why the secretin gene ceases to be expressed in the pancreatic ß-cells in adults, but continues to be expressed in gut S cells. In summary, only a few transcription factors or response elements that are shared in the superfamily have been identified to date for the six genes.
B. Members of superfamily in many vertebrates
If the hypothesis is correct that the glucagon superfamily evolved
from a single ancestral gene, then it is theoretically possible to
identify each member of the family in extant vertebrates or
invertebrates and to deduce information about the distinct origin of
each hormone. The phylogenetic distribution of each glucagon family
member has advanced rapidly due to chemical and molecular sequencing.
Immunohistochemical identification has been helpful, but not
definitive, due to cross-reactivity of antibodies.
In mammals, all nine bioactive hormones (PACAP, VIP, PHM, glucagon,
GLP-1, GLP-2, GRF, GIP, and secretin) have been identified, and the
structures are shown in Fig. 3 where most of the references are listed
for this section. Current evidence suggests that the PACAP-related
peptide (PRP) is not bioactive. Therefore, PRP will not be discussed
further.
Although the PACAP family is highly conserved with identical sequences for both PACAP-27 and PACAP-38 in the four mammalian species studied to date, there is also high conservation of VIP in mammals where VIP-28 is identical in 10 mammalian species, but differs in guinea pig and opossum. For glucagon-29, eight mammalian species share the same sequence, but guinea pig, degu, and opossum each have a distinct sequence. In contrast to PACAP, VIP, and glucagon, there are four distinct sequences for GLP-1 in the four mammalian species studied. Also, four sequences are identified for GIP-42, one for each of the four species. PHM/PHI-27 has five distinct sequences for eight species. Considerable sequence variation is seen in secretin-27 in most mammals: six sequence variations are present in the eight species for which sequences are identified. At the other end of the conservation spectrum from PACAP is GRF, with rapid changes in sequence, as noted above in Section II.B. Seven distinct sequences have been identified for GRF (42–44 amino acids) in eight mammalian species. Thus, it is likely that the strongest evidence for phylogenetic ancestry for the molecules will come from PACAP, VIP, and glucagon. GRF and secretin vary so rapidly in sequence during the evolution of the mammals that ancestral tracking will require study of more species at closer intervals. Variation in structure can occur rapidly in the C-terminal portion of the peptide if the coding is interrupted by an intron as for GRF and GIP, but tracking of the ancestral N-terminal portion of the hormone should be possible.
In birds, reptiles, or amphibians, it is difficult to generalize
about the PACAP/glucagon superfamily ancestry as very few sequences are
known for these three classes of vertebrates (Fig. 3). Chicken and
turkey have been used exclusively among birds for sequence
identification with two exceptions in which duck and ostrich were
studied for glucagon. GIP has not been identified in any bird, and only
one species has been used to identify several of the hormones: PACAP,
GLP-1, secretin, and GRF. Two distinct sequences in birds were
identified for VIP (two species), PHI (two species), and glucagon (four
species). For each avian hormone, the sequence is different compared
with the mammalian form, but the differences are small enough to
clearly identify the family lineage. The organization of the avian
genes for family members has been very useful in understanding the
evolution of the superfamily. The avian gene that encodes both PACAP
and GRF (Fig. 5) supports the idea that a
gene duplication occurred in early mammals leading to separate genes
for PACAP and GRF (Fig. 6a). In contrast,
the avian gene for glucagon/GLP-1/GLP-2 has the same organization as
the mammalian gene, which is additional proof that the organization of
this gene has remained stable in the vertebrates, including fish.
|
|
An interesting variation in the PACAP/glucagon superfamily occurs in
reptiles in that the gila monster has a novel peptide family named
exendin (Fig. 3). The peptides are clearly related in structure to the
superfamily, but the peptides are found exclusively in the venom made
in the salivary gland of the gila monster. Four variant peptides of
35–39 amino acids have been identified and named exendin-1
(helospectin), exendin-2 (helodermin), exendin-3, and exendin-4. In
peptide structure exendin-2 is related to human PACAP with 53%
sequence identity, whereas exendin-4 is related to human GLP-1, also
with 53% identity. Exendin-2 interacts with the VIP and secretin
receptors resulting in an increase in cAMP in rats (73). In contrast,
exendin-4 binds with the GLP-1 receptor, increases cAMP, stimulates
glucose-dependent insulin secretion if injected, and increases insulin
gene expression in cultured islet cells (76). However, extensive
studies suggest that exendins do not have mammalian homologs (16).
Rather, evidence suggests that one or two gene duplications, possibly
within the taxon of lizards including the gila monster, have resulted
in separate genes for PACAP, glucagon, exendin-2, and exendin-4. The
cDNAs and distribution pattern of the latter two are much closer to
each other than to the cDNAs for PACAP or glucagon, which have been
partially sequenced for the gila (16). The origin of the venom peptides
will be easier to discern when the gene for the gila PACAP can be
compared with that for exendin-2 and the gene for glucagon can be
compared with that for exendin-4. Alternative splicing, which is common
within the superfamily, may explain the differences in cDNA
organization.
In amphibians only five peptides have been identified. They are PACAP, VIP, glucagon, GLP-1, and GLP-2.
In fish, six out of nine families (PACAP, VIP, GRF, glucagon,
GLP-1, GLP-2) have been found. Only secretin, GIP, and PHI have not
been isolated and sequenced from any fish species. Secretin bioactivity
has been reported from intestinal extracts of bony and cartilaginous
fish, but the structure of secretin is not known (176). PHI is assumed
to be present as it occurs in tandem with VIP on one gene in reptiles
(16), birds, and mammals. A fish VIP cDNA or gene has not been isolated
to date but is likely to be identified soon, as VIP peptides have been
sequenced from dogfish shark, bowfin, trout, and cod (Fig. 3).
Of the superfamily peptides identified in fish, the most conserved
compared with human is PACAP followed by VIP, and then glucagon. The
least conserved are the GLP and GRF families. PACAP-27 is identical in
salmon and catfish compared with human PACAP-27. There is more
variation with PACAP-38 between fish and human. Of the hormones
identified (Fig. 3), the amino acids of fish peptides compared with
human peptides (100%) have an identity of 82% for VIP, 69–86% for
glucagon, 35–68% for GLP-1, 40% for GLP-2, and 30–45% for GRF.
Hence, fish PACAP-27 has the highest sequence identity with the human
form; GRF has the least identity, but the percentage of GRF would be
higher if only the biologically active cores were considered. Secretin
has not been identified by structure except in mammals and birds, but
intestinal extracts of reptiles, amphibians, bony fish, and
cartilaginous fish have been shown to have secretin bioactivity (176).
In jawless fish, which to date includes only studies of lamprey,
the peptide sequences for glucagon and GLP-1 have been reported (Fig. 3) and are 72% and 48% identical, respectively, to their counterparts
in humans. Recently, two cDNAs encoding glucagon and GLPs were isolated
from sea lamprey (Petromyzon marinus) intestine (Ref. 116
and Fig. 3). One precursor encoded glucagon and GLP-1, but the
GLP-2-like molecule was lacking five critical amino acids near the N
terminus, making functionality unlikely. The other precursor encoded
glucagon-II that was 72% identical to glucagon I and did not encode
GLP-1, but did encode a GLP-2-like molecule (116).
Proof that the PACAP/glucagon superfamily members were present in the earliest vertebrates, the jawless fish, has been established only for the glucagon/GLP-1/GLP-2 trio. However, the superfamily peptides, with the exception of secretin and GIP, have been shown to be present in representatives of most other vertebrate classes. The origin of these two families will remain unknown until an exhaustive search has been done for the peptides in nonmammalian vertebrates. The rapid evolution of this superfamily includes gene duplications and alternative splicing, leaving open the possibility that the origins of secretin and GIP were within the vertebrates or preceded the vertebrates. The latter possibility is favored by the fact that the sequence of these two peptides in humans is quite distinct from the highly conserved PACAP, suggesting a long separation time.
C. Key to origin of superfamily in protochordates
The protochordates are the major group from which the vertebrates
are thought to have arisen. Thus, they are a logical group in which to
investigate the origin of the PACAP/glucagon superfamily members in
invertebrates. The Garstang theory states that the key group of
protochordates from which the vertebrates arose was the tunicates
(Urochordata). One possibility is that the tunicate’s mobile larvae,
which have a fish-like appearance, went through sexual maturation
without metamorphosis. This process could result in a sexually
mature adult without transformation into the sessile adult that
normally occurs in tunicates (although one group of urochordates, the
larvaceans, remain as mobile adults). This ancestral tunicate may have
given rise to the amphioxus, another protochordate, and to the
vertebrates. Other scenarios for the origin of the vertebrates vary as
to the taxon for the ancestral group, but tunicate still remains
closely related to vertebrate ancestors (177).
It is now clear that several of the superfamily members are present in
tunicates (Figs. 2, 3, and 5). PACAP-27 was identified encoded in a
cDNA isolated from a tunicate; the percent identity of amino acids was
96% with human PACAP-27 (23). An extended peptide (PACAP-38) was not
encoded, suggesting that PACAP-27 is the original form of PACAP in
evolution. A second peptide was encoded on the same gene with PACAP-27.
We called it GRF-27 because it was on the exon that was 5' to the PACAP
exon as in the vertebrates. Again, this may be the original form of
this peptide as only the biologically active core is encoded and not
the extended peptide found in vertebrates. However, the identity of the
GRF-27 peptide was 59% with human GRF and 61% with human glucagon.
This closeness in identity makes the point that the PACAP/glucagon
peptides may be closer in sequence as the ancestry is traced in animals
that evolved earlier than vertebrates.
A further interesting twist to the tunicate PACAP story is that a second cDNA was isolated from the same tunicate library and was shown to encode two peptides also. One peptide was 85% identical to the deduced amino acids of PACAP-27 in humans and the other peptide was 59% identical to human GRF-27. The question is whether this cDNA represents a duplication of the PACAP gene in the tunicates in the 700 million years since they separated from the stem line leading to vertebrates or whether the duplication occurred in the stem line that may have continued into vertebrates. The second form of PACAP-27 is 67% identical to human VIP and could be a precursor. Further studies in animals whose ancestors evolved before the protochordates may help to distinguish the two possibilities.
The presence of PACAP-27 in the tunicate establishes the origin of the PACAP/glucagon superfamily in the invertebrates. It also adds weight to the idea that PACAP was the original peptide in this family and the one most tightly conserved. The isolation of glucagon from a tunicate will be an important step in understanding the origin of the super-family.
D. Extension of PACAP ancestry uncertain in insects
In Drosophila a neuropeptide gene was identified that
has some identity to PACAP (178). This gene, named amnesiac,
encodes a signal peptide followed by several possible peptides
depending on the cleavage sites. One of the peptides deduced from the
gene had 10% identity with human PACAP-38 or 18% with PACAP-27 (see
sequences below).
This identity is too low to claim that amnesiac is homologous to PACAP in tunicates or vertebrates. However, the authors show that an inserted space in PACAP after both amino acids 23 and 27 would increase the identity to 21% for PACAP-38 and 30% for PACAP-27. If amino acid similarity is used for the calculation, then the relationship is higher. However, the amnesiac peptide has four cysteines, which are usually highly conserved, and PACAP-27 and PACAP-38 do not have any cysteines. Also, it is not clear why only PACAP-38 and not PACAP-27 has electrophysiological effects and why cross-reactivity in Drosophila occurs only with antisera raised against PACAP-38 and not PACAP-27 (179) because the Drosophila peptide is only 32 amino acids and identity to PACAP is only within the first 27 amino acids. Therefore, with respect to PACAP homology, amnesiac will have to remain in a gray area until other invertebrate PACAP-like peptides are identified.
Another intriguing story is that the sand fly (Lutzomyia longipalpis), which obtains blood from vertebrates, has a peptide (maxadilan) that activates the PAC1 receptor resulting in vasodilation (180). Like the fruit fly’s amnesiac peptide, maxadilan has a low identity with hPACAP-38 (16%) or hPACAP-27 (15%) (see sequences at top of page 631).
Again, the low sequence identity and presence of four cysteines in maxadilan compared with no cysteines in PACAP suggests a decision about homology is not possible at this time. The authors (180) suggest maxadilan is an example of functional convergent evolution.
E. Exon duplication as the original step in ancestry
The suggestion that each peptide in the PACAP/glucagon superfamily
is encoded on a single exon is based on three facts. First, the four
peptides isolated from tunicates, the most ancestral forms to date, are
each encoded on one exon (Fig. 5). Second, most of the superfamily
peptides isolated from vertebrates, including humans, are encoded on
one exon. Third, the two peptides (GRF and GIP) that are not contained
on one exon do have their bioactive cores encoded on one exon. Tunicate
GRF-like peptide is an example of a peptide that was originally only 27
amino acids as in tunicates and was extended later in the 3' direction.
Exon duplication of the ancestral peptides is the most logical
explanation for some of the peptides present in mammals to date (Fig. 6B). Three of the six superfamily genes are examples of possible exon
duplication: GRF/PACAP (in nonmammals), PHI/VIP, and
glucagon/GLP-1/GLP-2. The gene for which we have the longest
evolutionary story is PACAP. The two PACAP genes can be clearly
identified in tunicates because of the high sequence identity (96% or
85%) of the deduced PACAP-27 peptides with human PACAP-27. The
argument that an exon duplication is responsible for PACAP and the
GRF-like exons is that the sequence of the two exons has remained close
in the tunicates. Only a 15% decrease is apparent in recently evolved
mammals if the first 27 amino acids are compared as explained below and
in Fig. 8. In tunicates the percent identity between PACAP-27 on one
exon and GRF-like peptide-27 on the adjacent exon is 44% or 48% (Fig. 7). After the transition between
invertebrates (tunicates) and vertebrates, the divergence between the
exons encoding PACAP and the associated peptides (a GRF-like peptide)
continued to increase. In fish such as sturgeon that have survived from
some of the earliest bony fish, the percent identity between the
peptides on exons 4 and 5 are almost as high (4l%) as in the tunicates
(D. W. Lescheid and N. M. Sherwood, unpublished). In fish,
such as salmon and catfish, in which the ancestor evolved more recently
than sturgeon, the identity is reduced (26–33%); the identity between
exons for chicken is similar (33%). But it is in mammals such as
humans where a drastic reduction in peptide identity (7%) occurs
between the peptides on exons 4 and 5 (Fig. 7A). One possibility is
that there was a complete gene duplication in early-evolving mammals
resulting in one gene in which PACAP is conserved and another gene in
which GRF is conserved (Fig. 7B). One piece of evidence that supports
this hypothesis is that human PACAP-27 has only 7% identity with the
peptide (PRP) on the adjacent exon and human GRF-27 has only 19%
identity with the adjacent peptide (a cryptic peptide without known
function), but if PACAP-27 on one gene is compared with GRF-27 on the
other gene, the identity leaps to 33%, which is the same as that in
chicken and salmon.
|
|
The evolution of the VIP gene has not been traced to invertebrates. In the human and chicken VIP genes, the identity between PHM/PHI (exon 4) and VIP (exon 5) is 44% at the amino acid level. Linder et al. (43) argue that the VIP/PHM exon duplication was an ancient event because there is low conservation between VIP and PHM and a lack of conservation in the flanking sequences around VIP/PHM. Yamagami et al. (44) contend that there is enough sequence similarity in the introns flanking the PHM and VIP exons to suggest that a broad region, and not just an exon, was duplicated. If true, the phrase intragenic duplication rather than exon duplication might be more accurate. Although the protein sequence of VIP is known for four fish, the cDNA or peptide structure of the peptide on the adjacent exon is not known. Hence, the origin of the VIP gene or the time of the exon duplication is not clear but appears to be more recent than the PACAP gene. Another possibility is that the VIP gene resulted from a duplication of the PACAP gene as there is 70% identity between the VIP and PACAP peptides in humans. In tunicates we did not find a VIP gene, but the two PACAP peptides in tunicates were both 67% identical to human VIP, which is similar to the identity between human PACAP and human VIP (70%).
The glucagon gene is thought to have had an exon duplication twice,
resulting in three exons encoding one peptide each (Fig. 6B). If all
three peptides are compared with each other in the human, chicken,
bullfrog, and trout genes, then glucagon tends to be closer to GLP-1
than GLP-2; the latter is closer to GLP-1. The percent identity among
the three peptides in each of four species is 28–59% (Fig. 3). Hence,
there is not enough information to identify the origin of glucagon. A
recent paper shows that glucagon, GLP-1, and GLP-2 are encoded in
lamprey, but the three intact peptides are not encoded in a single cDNA
(116). The authors speculate that glucagon originated about 1 billion
years ago, whereas GLP-1 and GLP-2 diverged from each other about 700
million years ago (116).
Isolation of invertebrate genes encoding glucagon and its associated peptides is needed to narrow the evolutionary period for the origin of the gene and exon duplication within the gene. In tunicates we have not found a glucagon gene, although we found two cDNAs encoding a peptide with 63–67% identity to human glucagon. However, we judged these peptides to be a GRF-like peptide because each was in the same gene with PACAP and was in the exon that is 5' to PACAP coding. The peptide identity of the tunicate peptides was 59% with human GRF. Functional studies may clarify the true identity of the tunicate peptides.
F. Gene duplication crucial in creating gene families
In evolution several gene duplications in the superfamily have
occurred, resulting in six distinct genes in mammals as shown in Fig. 8. The two GRF-PACAP genes in tunicates
could have arisen in the tunicate stem line or before tunicates
separated from other invertebrates. In vertebrates it is clear that
fish (at least bony fish) have two genes in which one encodes GRF/PACAP
and the other glucagon, GLP-1, and GLP-2. It is likely that a PHI/VIP
gene exists in fish, although the cDNA has not been isolated. GIP and
secretin genes have not been isolated from fish. The most
evolutionarily recent gene duplication that has been suggested for the
superfamily is the duplication of the PACAP-GRF gene in early mammals
(15, 19).
These several gene duplications have not resulted in gene clusters on a chromosome. Rather, each of the superfamily genes is located on a different chromosome, although the location of the secretin gene is not reported. In humans the glucagon gene is on chromosome 2 (2q36-q37); VIP gene is on chromosome 6 (6q16-q22d); GIP is on 17 (17q21.3-q22), PACAP is on 18 (18p11), and GRF is on chromosome 20 (20q) (from Refs. 8, 130). The separate location of the genes suggests that 1) the duplication events did not occur at the same time; 2) the genes have been evolving for a long time; and 3) the function of the genes is not interdependent. This is in contrast to Hox genes where the genes remained clustered in a continuous linear sequence after tandem duplication in invertebrates. Only complete genomic duplications in vertebrates resulted in copies of the clusters on more than one chromosome (181).
At least one gene duplication in the superfamily has resulted in the novel exendin family identified to date only in a few lizards (gila monster). There are at least two scenarios. The PACAP gene may have duplicated to produce exendin-2 (helodermin), and the glucagon gene may have duplicated to produce exendin-4, a scenario based on the sequence identity between PACAP and exendin-2 and between GLP-1 and exendin-4. The second possibility is that either PACAP or the glucagon gene duplicated and the new gene later duplicated again. This concept is based on the organization of the cDNAs of exendin-2 and -4, which are closer to each other than to the genes in the original family. Also, the expression pattern shows that the exendins are present only in the venom produced in the salivary glands (see Refs. 16, 76).
G. Peptide elongation important in some families
In the six superfamily genes of mammals, there are two peptides
that are not encoded on a single exon. GRF and GIP extend onto the
following exon. In addition, GRFs in vertebrates may be extended
because the GRF-like peptides in tunicate are known to be 27 amino
acids and to be encoded on one exon (Fig. 5). In vertebrates GRF is
42–46 amino acids, but one exon encodes the first 32 amino acids and
the next exon encodes the remaining amino acids. Thus, the first exon
encodes a complete biologically active core of 29 amino acids with full
biological activity and about 50% potency compared with the
full-length hGRF of 44 amino acids (182). The exon-intron organization
in the middle of the human GRF gene shows splice sites of gt/ag at the
start/end of each of the four introns. There is no question that the
extension of GRF in mammals makes the peptide more effective for the
release of GH; the fragment of GRF (1–29 amide) has full biological
activity but reduced potency (25, 140, 182).
Because an ancestral GIP has not been detected, we don’t know whether a shorter peptide on one exon evolved earlier. We do know that human GIP has 35 amino acids encoded on one exon and the remaining seven amino acids on the adjacent exon (130, 136) and that the bioactive portion of GIP is encoded on one exon. The evidence is that a synthetic peptide for GIP-30 is as potent as GIP-42 in stimulating insulin secretion (183, 184) and proinsulin gene transcription (184), although other functions, such as somatostatin release and inhibition of acid secretion from the stomach, are reduced (185). Further evidence that the shorter GIP-30 might be relevant is that a cleavage site exists at amino acid positions 31–33 (Gly-Lys-Lys), although GIP-30 has not yet been isolated from the blood. Also, transfected GIP receptors bound both GIP-42 and GIP-30 with high affinity in one study (186), but lower affinity for GIP-30 in another study in which GIP-30, nonetheless, was as effective as GIP-42 in stimulating cAMP accumulation (187). The extension of the GIP peptide onto another exon does not involve a novel mechanism, as the human GIP gene is very similar to those of other glucagon superfamily members in that gt/ag is present at the start/end of each of the five introns (130), including the intron that interrupts GIP coding.
H. Alternative splicing for expansion of individual families
There are several types of alternative splicing that have been
documented in the PACAP/glucagon superfamily: 1) exon deletion (exon
skipping); 2) intron sliding; and 3) splicing within exons encoding the
5'-UTR.
Exon skipping is a dominant characteristic of this superfamily (Fig. 8). PACAP/GRF mRNA is expressed as a long and short transcript in the
brain in fish and chicken; the short mRNA is lacking exon 4 (15, 19).
The functional significance is that exon 4 encodes the bioactive core
of GRF. When PACAP and GRF are encoded together on one gene, the ratio
of GRF to PACAP can be changed. The exon encoding PACAP is not deleted.
Mammals achieve the same effect with more control due to the separation
of PACAP and GRF on two genes. A similar effect is observed with the
VIP gene in which the PHI exon is deleted, although this has only been
reported for chicken (60) and turkey (61); the most abundant (98%)
form of mRNA in several tissues is the shorter version with PHI
deleted. There is no evidence of alternative splicing for the mammalian
VIP transcript (43). In the glucagon gene, the GLP-2 exon is spliced in
a tissue-specific manner. GLP-2 is not present in the pancreatic mRNA
in salmon and chicken, but is present in the gut mRNA (91) where the
GLP-2 peptide is known to affect intestinal growth (188, 189, 190). The
secretin gene is another example of exon skipping in specific tissues
and species (4, 5, 122), but the peptide encoded on exon 3, the deleted
exon, does not have a known function.
Exon deletion could be an alternate way to regulate a gene that encodes more than one bioactive peptide. It seems less sophisticated than encoding each peptide on a separate gene with a distinct promoter. In the course of evolution, the mammals now have four superfamily genes (GRF, PACAP, secretin, GIP) with only one biopeptide. The proposed duplication of the GRF and PACAP gene in early mammals is a striking example of a process that could lead to finer regulation of two important peptides. It will be of great interest to determine whether secretin and GIP originated on genes in which solitary or multiple peptides were encoded.
Intron sliding occurs particularly in the GRF gene. In the human GRF gene, a downstream splice site results in a C-terminal peptide that is one amino acid shorter than if the upstream site is used (30). This is unlikely to have any functional consequences. More important in explaining the rapid changes in GRF structure during evolution of mammals is the observation that there is intron sliding in the rat GRF gene compared with the human gene. This change in splice donor site between intron 3 and exon 4 results in a change in the C terminus in the rat GRF by which it no longer matches the human GRF. Intron 3 interrupts the coding of GRF, which spans exons 3 and 4. Also, there is intron sliding when the human and rat GIP genes are compared. In the rat gene, the splice site for intron 2 and exon 3 is 24 nucleotides downstream compared with the human gene. This splice site in GIP is within the N-terminal peptide and in rat results in an N-terminal peptide that is eight amino acids shorter than the human form (136, 137).
One of the most interesting uses of alternative splicing in the superfamily is within the receptors. The PACAP receptor has at least five different inserts (191, 192), and the GRF receptor has at least one (169, 193); both sets of receptor inserts are in the third intracellular loop and are thought to control coupling to different intracellular signaling pathways. This alternative splicing is clearly functional. In addition, alternative splicing is reported for the glucagon receptor (194), GIP receptor (195), and VIP receptor (196), although the biological function of these variants is not known. Any member of this subfamily of receptors that has multiple signaling pathways (e.g., GLP-1) may be shown eventually to have variant forms of its receptor.
I. Alternative promoters for expansion of family functions
For several family members, alternative promoters are used in a
tissue-specific way. Upstream promoters can lead to changes in the
5'-UTR and may alter the stability of the mRNA before translation. An
example is the GRF gene transcript. In the placenta and testis,
different upstream promoters are used compared with the brain
transcription start site. This results in a different exon 1, which
encodes the 5'-UTR (148, 171).
For PACAP at least three distinct mRNAs have been detected. The
dominant one in rat neural tissue is 2.2 kb and is the same as the one
shown in Fig. 4 (PACAP, human). The other two transcripts are shortened
(0.9 kb) but not identical to each other. One is a testis-specific
transcript with a shortened 5'- and 3'-UTR plus an extra 126-bp
sequence in the 5'-UTR compared with the 2.2-kb transcript (152, 153).
The other short transcript was detected in the superior cervical
ganglion after stimulation by depolarization; the sequence was not
determined, but primers were used to show that it was not identical to
the testis-specific transcript (197). Hence, PACAP alternative
promoters have been identified and result in tissue-specific (testis,
placenta, and sympathetic nervous system) transcripts.
The GIP gene may use two promoters: the most abundant mRNA begins at
exon 1 as shown in Fig. 4; a minor mRNA begins at exon 2, as suggested
by the ribonuclease protection assays. In this case it appears that
intron 1, which contains a TATA box and cis-acting elements,
is the alternative promoter (198). The purpose of mRNA lacking exon 1
and most of the 5'-UTR is not clear.
J. Posttranslational processing for increase of peptide families
Several of the superfamily peptides can be cleaved from their
protein precursors in more than one way, creating new peptides. The
enzymes that cleave PACAP and possibly other hormones from the
precursor are important and have been reviewed (145). In most cases,
biological activity resides with the shorter peptide.
The glucagon precursor can be cleaved variously depending on the tissue. In the mammalian pancreas, glucagon is cleaved from the precursor, whereas GLP-1 and GLP-2 are not cleaved from the larger fragment except to a minor extent (199). In contrast, GLP-1 and GLP-2 are cleaved from the precursor in the small intestine; GLP-1 is further cleaved in mammals to GLP-1(7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37) or GLP-1(7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36) amide before it becomes biologically active. From the intestinal precursor, glucagon is not released; instead the precursor is processed so that glicentin (glucagon that is extended at the N and C termini) and oxyntomodulin (glucagon that is extended at the C terminus only) are released. The fish glucagon precursor lacks the GLP-2 coding region in the pancreas, but not in the intestine; this is a result of alternative splicing and not posttranslational modifications. In fish, GLP-1 is 31 amino acids as in mammals, but the six-amino acid N terminus that is cleaved to make a mature peptide in mammals is not present in fish who use GLP-1(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31). The initial cleavage from the fish precursor results in an active GLP-1 of only 31 amino acids. Finally, in primitive fish including lamprey, the glucagon precursor is cleaved further in the intestine to produce glucagon from glicentin (200). Glicentin (69 amino acids) does not release insulin, whereas the shorter oxyntomodulin (37 amino acids) is almost as effective as glucagon (29 amino acids) in releasing glucose-induced insulin from the pancreas (201). Likewise, GLP-1 and GLP-2 are active only as cleaved peptides and not as the extended product of GLP-1 and GLP-2 connected by a small peptide as released from the pancreas.
PACAP is processed as a 38- or 27-amino acid peptide (10). Both peptides are equally effective in some functions, but the extended PACAP-38 is more effective in others (see Section V). VIP is a 28-amino acid peptide, but extended forms of VIP have been identified as VIP(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28) Gly-Lys-Arg (69). PHI and PHM are found as both PHI/PHM-27 and in the extended form known as PHV-42 (70, 202). Also, PHI(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27) Gly, secretin(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27) Gly, and secretin(1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27) Gly-Lys-Arg exist (69).
|
IV. Superfamily Members—Overlapping Functions, Expression, and Receptors |
|---|
|
TopAbstractI. IntroductionII. Evolution of the...III. Superfamily Evolution by...IV. Superfamily...V. Conservation of PACAP—A...VI. Conclusions and Future...References |
|---|
First, eight of nine of the superfamily members are found in the brain and, hence, properly classified as neuropeptides. The exception is GIP, which has not been identified in the brain as a protein or mRNA. In addition, the presence of secretin in the brain is controversial with reports on both sides. The location in the brain varies widely among the members of the superfamily.
Second, all of the nine superfamily members are found in the gut. GLP-1, GLP-2, secretin, and GIP are traditional gut peptides, but small amounts of glucagon (larger amounts of oxytomodulin and glicentin) are released. PACAP and VIP/PHM are in the gut; both PACAP and VIP are localized in nerve endings in the gut (203). GRF can be measured in gut extracts. All the hormones are predominantly in the small intestine. Their functions as gut hormones vary, but most of them release insulin in response to a meal. PACAP, GLP-1, and GIP are the most potent insulin releasers (release is glucose dependent), but secretin, GRF, VIP, and PHM also release insulin. Glucagon opposes the actions of insulin, but at supraphysiological levels, glucagon releases insulin, probably by binding the GLP-1 receptor. GLP-2 is not known to affect insulin.
Third, several of the hormones are expressed in the pancreas: glucagon,
GLP-1 (small amounts), GLP-2 (small amounts in mammals), PACAP, GRF,
and secretin; the latter is expressed only transiently during
development. GRF and VIP/PHM are both present in pancreatic tumors such
as GRFomas and VIPomas. GIP is not detected in the pancreas. Although
PACAP, VIP, and PHM affect the secretion of glucagon, most of the other
peptides do not. GLP-1 and GIP do not appear to have receptors on
-cells in the islets and do not act at physiological doses on
glucagon.
Fourth, several of the peptides are found in the testis, ovary, and placenta. Both the peptide and its receptor are usually found in the same organ, suggesting a paracrine effect. In the testis PACAP, VIP/PHM, GRF, glucagon, and secretin have been identified in Leydig or germ cells or both. In ovary PACAP, VIP/PHM and GRF are reported, but in the placenta, only GRF is reported to be present and thought to be a growth factor for the embryo. Although GRF was found in the rodent placenta, it was not found in human placenta.
Fifth, one of the common functions that is primarily shared by PACAP and VIP/PHM is vasorelaxant activity. PACAP and VIP are potent vasodilators, whereas the other peptides appear not to affect the vascular system to any special degree.
Sixth, several superfamily members affect anterior pituitary secretion: PACAP was shown in some reports to release GH, PRL, LH, and ACTH from anterior pituitary cells in vitro (see Section V.E.1). VIP/PHM release PRL and some GH and LH; GRF releases GH. In addition, at least five of the peptides act as growth factors. GRF acts on the somatotropes in the anterior pituitary to increase proliferation and hypertrophy, whereas secretin in the developing pancreas acts in the same way to increase ß-cell division and growth. PACAP and VIP are both important in the early development of the brain and other tissues. Inhibition of VIP in early embryos led to microcephaly in mice.
Finally, specific receptors have been identified for eight superfamily members. A specific receptor for PHM-PHI has not been isolated and sequenced. VIP is unusual in that it shares its two receptors with PACAP, whereas PACAP has an additional specific receptor with eight subtypes. All the superfamily receptors are seven-transmembrane, but belong to a subset of G-coupled receptors that include the receptors for calcitonin, CRF, and PTH. The subset, referred to as the class II family (204), are distinct from the seven-transmembrane receptors in the rhodopsin-like members of the class I family.
The class II family of receptors shares characteristics such as a long
N-terminal extracellular domain containing a number of cysteines that
are important in ligand binding, and a complex exon-intron organization
in the genes (205). The class II receptor group has a low overall
sequence similarity of less than 20% with other receptors in the G
protein-coupled class I family. The presence of a distinct receptor
group for the PACAP-glucagon superfamily and a few other ligands may
eventually provide important information about the coevolution of
peptides and their receptors. It is reasonable to speculate that the
class II subset of receptors evolved from a common ancestral gene in
invertebrates, although no such receptors have been isolated. In an
evolutionary sense, a PAC1 receptor has been
isolated and sequenced from chicken (206), frog (207), and goldfish
(143). A full-length VPAC receptor in frog has common features with
both VPAC1 and VPAC2 in
mammals (208), whereas the full-length goldfish VPAC receptor (209) has
not been characterized by pharmacology or tissue distribution as to
whether it is like VPAC1 or
VPAC2. These sequences suggest that only one type
of VIP receptor existed earlier in evolution. It is not clear where the
specific PAC1 receptor separated from the VPAC
receptors. However, we know that PACAP exists in invertebrates (23).
Although the PACAP or VIP receptors have not been identified, other
G-coupled receptors are present in invertebrates. Even yeast have a
G-coupled receptor for -mating factor (210). One scenario is that an
ancestral receptor(s) (at least for PACAP and glucagon) was present for
the PACAP-glucagon receptor superfamily in protochordates that were
ancestral to vertebrates. Early in chordate evolution, two genomic
doublings are thought to have led initially to four copies of each gene
that encoded a class II family receptor. Structural alterations may
have resulted in the eight receptor types (including two VPAC
receptors) identified to date. Invertebrate receptors in the class II
family must be isolated and sequenced to verify this suggestion of
receptor evolution.
Details of the functions of the superfamily members are discussed below.
A. PACAP and VIP
PACAP is discussed in great detail in Section V and
only a brief description is given here as a comparison to VIP. In
addition to structural similarities, the function of VIP is closer to
that of PACAP than the other family members. This is expected because
VIP and PACAP have equal affinity for two receptors,
VPAC1 and VPAC2, both of
which have been cloned and sequenced (196, 211, 212, 213, 214). In addition,
PACAP binds to a PACAP-specific receptor, PAC1
(191, 192, 215). Both VIP and PACAP have a widespread distribution and
are known to affect the neural, circulatory, gastrointestinal,
endocrine, and immune systems.
The function and location of VIP have been reviewed a number of times since its discovery (216, 217, 218, 219, 220, 221). Basically, VIP is localized primarily in nerve cells. In the gut, VIP and PACAP are in nerve fibers or ganglia; their release has potent effects on vasorelaxation of smooth muscles, absorption of water and ions, and secretion from the pancreas and intestine (216, 217, 218, 219, 220, 221). VIP stimulates vasorelaxation in a number of tissues in addition to the gut (216), including the brain (222). VIP is a neuromodulator in the brain (217, 218) and has been suggested to be associated with the "visceral forebrain system" that influences and at times overrides brainstem control of cardiovascular, respiratory, and gastrointestinal functions (223). In the endocrine system, VIP releases PRL (223, 224, 225), LH, and GH (226) from the pituitary. VIP acts on the pancreas to release either insulin or glucagon, depending on the glucose level; VIP also acts on the exocrine pancreas to increase bicarbonate output (227). PACAP also releases PRL and other pituitary hormones (6) and is a potent releaser of insulin (see Section V.E.6) and glucagon (228). Both VIP and PACAP are present in nonneuronal tissues such as the gonads (229, 230) and some immune cells (217, 231).
During development, VIP is expressed in the embryonic brain (232, 233, 234) and has trophic and mitogenic effects (220, 232, 235, 236). If VIP is inhibited in the early embryo, severe microcephaly occurs (236) and, vice versa, VIP (10-10 to 10-7 M) stimulates growth in whole mouse embryos at E9.5 in culture (237). Also, VIP acts on sympathetic neuroblasts in the superior cervical ganglion (SCG) in culture to increase mitosis, survival and neurite outgrowth (238). The role of VIP appears to be limited to a critical fetal period around embryonic day (E) 15 in the rat; the expression of VIP protein in the SCG also peaks at this time (238). The role of VIP in development is undercut by the fact that PACAP is 1000 fold more potent than VIP in producing the same effects (mitosis, survival, neurite outgrowth) on sympathetic neuroblasts of the SCG (238). However, there are reports that both VIP and PHM were antiproliferative in a neuroblastoma cell line (239).
B. PHM
Peptide histidine-methionine (PHM) and VIP generally exert similar
biological effects, but PHM is less potent. (67). Like VIP and PACAP,
PHM has been reported to release prolactin (240), insulin and glucagon
(241). Evidence from binding studies suggests that the PHM/PHI molecule
may have a specific receptor (239, 242) but the receptor has not been
isolated or characterized. The VPAC1 and
VPAC2 receptors bind PHM/PHI, but to a lesser
extent than they bind VIP or PACAP (196, 211).
C. Glucagon
The processing of the transcript and precursor of glucagon occurs
in different ways depending on the species and tissue in which they are
expressed. In mammals glucagon is the main product of the cells in
the pancreatic islets; GLP-1 and GLP-2 are released primarily as an
extended precursor fragment without function in which the GLPs are
connected by a small peptide (Fig. 4) (243). In the L cells of the
intestine, GLP-1 and GLP-2 are the main products; extended forms of
glucagon (e.g., glicentin and oxyntomodulin) are released
here (159, 243). In the brain all three peptides are found, but appear
to be synthesized only in the hypothalamus and brain stem (159, 244).
In addition, the GLP-1 peptide requires a further cleavage at the
N-terminus to become biologically active in mammals (see below). Each
of the three peptides (glucagon, GLP-1 and GLP-2) has a specific
receptor that has been isolated and sequenced. In other vertebrates,
tissue processing may differ from that in mammals. In birds and fish,
alternative splicing removes the GLP-2 exon from mRNA in the pancreas
but not the intestine (91) resulting in the secretion of only glucagon
and GLP-1 from the pancreas (245). In the intestine of birds and fish,
all three peptides are encoded in the precursor (91).
The function of glucagon is exerted through a 7-transmembrane receptor expressed in the liver, pancreas and brain (246, 247, 248). Glucagon, like several other members of the superfamily, is an important regulator of glucose metabolism. The binding of glucagon to its receptor in the liver results in glycogenolysis, gluconeogenesis and other processes that increase the level of glucose in the blood (249). The action of glucagon on pancreatic ß cells is to increase glucose-induced insulin release (247). Also, glucagon is reported to affect lipid, urea and amino acid metabolism and to inhibit feeding behavior (245). In the brain, glucagon induces hyperglycemia if injected into the cerebral ventricle (244).
During development, glucagon/glicentin transcripts and peptides are expressed in a tissue-specific manner. The sensitive technique of PCR was used to show that glucagon (and insulin) mRNA can be detected in the mouse embryonic foregut (E 9.0 or 20 somite state) in the region of the duodenum where the pancreas will arise by evagination 10–12 h later (250). This early expression of the hormone precedes the morphogenetic phase of development. In contrast, the protein measured as immunoreactive glucagon was low in the rat fetal pancreas, intestine, and brain (251). After birth, immunoreactive glucagon in the rat pancreas rose to a peak concentration at 7 postnatal days followed by a gradual return to adult levels. In the intestine and hypothalamus, glucagon (expressed also as glicentin) began to rise to adult levels beginning on postnatal day 7. However, in the brainstem the hormones decreased from fetal to adult levels with time. Proglucagon mRNA was present in all three tissues (pancreas, gut, brain) in the fetus and adult (251). The authors suggest that glucagon may have a role in the developing pancreas.
D. GLP-1
The function of this hormone, which is encoded on exon 4 of the
glucagon gene, is primarily a gut hormone that helps to regulate blood
glucose and feeding behavior once it is converted to GLP-1(7–36 amide
or 7–37) in mammals (252). Hereafter, GLP-1 will refer to the
biologically active form of GLP-1(7–36 amide or 7–37). GLP-1 is
released from enteroendocrine L cells when glucose enters the gut
during a meal. In turn, GLP-1 enters the bloodstream, potentiates
insulin release from the pancreas in a glucose-dependent manner (253, 254), induces transcription of the proinsulin gene, and increases
insulin content in the pancreas (184, 199, 252, 255). At the same time,
GLP-1 is reported to inhibit glucagon secretion and gastric emptying
(255, 256), but to increase somatostatin release, lipolysis, and
lipogenesis (199). The suppression of glucagon may be an indirect
effect as the GLP-1 receptor is not found on -cells and the insulin
released by GLP-1 is known to decrease glucagon release (199). Specific
GLP-1 receptors have been identified in mammals, including humans, as
seven-transmembrane molecules (257, 258, 259, 260, 261). The GLP-1 receptors are
present in pancreas (ß-cells only), lung, stomach, brain (248, 262),
heart, and kidney (263). The presence of GLP-1 and GLP-1 receptors in
specific brain cells is thought to be the basis of the effect of GLP-1
on inhibiting food intake (264, 265) and decreasing body temperature
(266) when injected into a brain ventricle.
The importance of GLP-1 in mammals has been examined in a mouse where the GLP-1 receptor gene was disrupted (knocked out), making the receptor nonfunctional (262, 267, 268). The resulting mice did not express a GLP-1 receptor, but appeared to be physically normal including body weight. The main defects were a moderate increase in blood glucose levels after glucose was given orally or intraperitoneally and a decrease in insulin blood levels (267, 268). The conclusion from the mice with the disrupted gene was that GLP-1 has a role in controlling glycemia and in potentiating glucose-stimulated insulin secretion, but is not essential for regulating food intake, body weight, or ß-cell competence (262, 267, 268). Transcription of the insulin gene in mice with disrupted GLP-1 receptors is reported to be unaffected (262, 268) or reduced (269).
In contrast to the function of GLP-1 in mammals, GLP-1 in fish does not oppose the action of glucagon but has similar actions to those of glucagon. GLP-1 peptides isolated from several different fish species have the same effect as mammalian GLP-1 when tested in mammalian tissues, but the same GLP-1 peptides in fish act on the liver rather than the pancreas. In the fish liver, GLP-1 increases glycogenolysis, gluconeogenesis, and lipolysis (245, 270). Differentiation of function between glucagon and GLP-1 is thought to have occurred during evolution of tetrapods.
E. GLP-2
For some time it was not clear whether GLP-2 was biologically
active in any vertebrate. Recently, a specific GLP-2 receptor was
isolated, sequenced, and found to be expressed in the gut and
hypothalamus (271). A function has not been found for the GLP-2
receptor in the brain, but GLP-2 has been shown to stimulate growth in
the intestine (255). GLP-2 not only stimulated proliferation of crypt
cells, resulting in villi of greater height, but also inhibited
apoptosis of enterocytes (188, 189, 190). In addition, GLP-2 was found to be
present not only in mammals, but in other vertebrates. The lack of
detection of GLP-2 in fish was due to alternative splicing of the
transcript in the pancreas, resulting in the absence of GLP-2; in
intestinal transcripts, GLP-2 is retained (91).
F. GRF
With regard to function, GRF is secreted from nerve cells within
the hypothalamus, and then binds to receptors in the anterior pituitary
to stimulate the secretion and synthesis of GH (33, 140, 169, 272, 273, 274)
and to stimulate somatotrope cell proliferation, differentiation, and
growth (30, 193, 275). Transgenic mice that overexpress GRF grow at a
faster rate than controls (276).
The GRF receptor is specific for GRF but is related to a subset (class II family) of seven-transmembrane receptors that include the VIP, PACAP, and secretin receptors (277). GRF receptors have been isolated from human (278, 279, 280), pig (281), rat (278), mouse (193), and goldfish (282); it appears there is both a long and short receptor due to alternative splicing in mammals (169, 278). GRF actions and receptors are not exclusively in the anterior pituitary. In rat brain, GRF receptors were found only in the hypothalamus in one study (283), but were found in the cerebral cortex, cerebellum, and brainstem in another study (284). The low abundance of the receptor required the highly sensitive method of RT-PCR Southern blot hybridization for detection. GRF axon terminals within the brain synapse onto neurons containing GRF (285), somatostatin, or unidentified factors (286). The neural connections may be the basis of central regulation of GH release. Also, release of GRF within the brain increases food intake (287) and enhances sleep (288, 289, 290).
In nonneural tissues, GRF is secreted from lymphocytes and may be immunomodulatory (291). GRF is found in ovary (284, 292), placenta (38, 170, 171, 284, 293, 294, 295, 296), testis (38, 148, 284, 297, 298, 299, 300), pancreas (301, 302), gastrointestinal tract (284, 301, 303, 304), several other tissues (284), and in carcinoid tumors especially in the pancreas and lung (24, 25, 303, 305). GRF at a very low dose (10-12 M) released insulin from pancreatic islets in vitro, but in dispersed islet cells, a dose of 10-9 M was needed to release insulin (306). There are reports that plasma GRF increases significantly after meals with the suggestion that GRF might be another gut hormone that releases insulin (307, 308). As expected, GRF receptors are present in the anterior pituitary, but are found only with sensitive RT-PCR methods in other tissues such as placenta, kidney, and gonads (169). The goldfish GRF receptors were in similar tissues, as well as in gill and liver, in addition to brain and pituitary (282).
The GRF transcripts in placenta and gonads have distinct exon 1 sequences (5'-UTR) compared to those in the brain (148, 170, 171, 297); the transcript in placenta results from the use of an alternative upstream promoter (171), whereas the testis transcripts use the placental promoter or one even further upstream (148). GRF may be a paracrine factor in tissues where it is synthesized outside of the brain (295, 296). Also, placental GRF may act as a growth factor in the fetus (38, 169) or regulator of growth factors in the placenta (293). In ovary and testis, GRF enhances steroidogenesis and cAMP accumulation that is induced by gonadotropins in granulosa cells (309) or Leydig cells and enhances FSH-induced cAMP in Sertoli cells (299). In the testis, GRF is present in Leydig cells and spermatids, but not in mature sperm or Sertoli cells, suggesting GRF acts as a paracrine factor (299).
During development GRF in median eminence fibers was first detected in the mouse fetus at E16.5, which is the same day that the GRF receptor mRNA and GH mRNA are detected in the pituitary. The fetal pituitary first becomes responsive to GRF on this day, and by E18.5, the GH gene is activated (193). Placental GRF mRNA levels peak on days 16–17 of gestation in the mouse (only days 14–19 were tested) (38). The importance of GRF as a growth factor on somatotropes was noted above. Immunoreactive GRF was not detected in the rat until fetal day 18, the same day that GH was detected in the plasma (274). In the testis, GRF was not detectable (dot blot hybridization) until postnatal day 2. The levels remained low until day 21 and then increased to adult levels by day 30 (297).
In fish as in mammals, GRF acts on pituitary cells to release GH; salmon GRF (19) and carp GRF (310) each released GH from cultured salmonid pituitary cells. In addition, GRF/PACAP mRNA is expressed in fish in the brain, ovary, testis, and gut (19, 20).
G. GIP
Based on function, GIP was originally named gastric inhibitory
peptide, but later it was found that GIP does not inhibit gastric acid
secretion if given in a physiological dose (311). Subsequently, GIP was
named glucose-dependent insulinotropic polypeptide to reflect its most
important function as a gut hormone. Like GLP-1, GIP is released in
response to the presence of glucose or fat in the gut (312) and then
acts on the pancreas to release insulin (312, 313). GIP is clearly
important as it appears to compensate for GLP-1 in regard to insulin
release if the GLP-1 receptor is knocked out or disrupted (269).
For location, GIP is secreted from K cells in the small intestine
(314), enters the blood, and then binds to specific receptors on the
pancreatic islet cells (- and ß-cells) (248), resulting in the
release of insulin. In addition, glucagon and somatostatin are reported
to be released by GIP (315, 316, 317) although the release of glucagon is a
partial reversal of glucagon suppression due to high blood glucose;
also, the level of GIP (10-9 M) was
somewhat above the physiological level (315). The GIP receptor has been
cloned in humans (186, 195), rats (187, 318), and hamsters (319); the
human receptor has two forms that differ by a 27-amino acid insertion
in the C-terminal cytoplasmic tail (195). The tissue distribution of
the GIP peptide and mRNA in the gut was exclusively in the small
intestine; it did not occur in the pancreas, liver, or other
gastrointestinal structures (130, 320, 321). In addition, GIP mRNA has
not been detected in the brain, although very sensitive methods have
been used (135, 137, 318). However, GIP mRNA is present in the salivary
glands (137, 322) and in a tumor cell line (STC 6–14) in which GIP
mRNA increased if incubated with a glucose stimulus (138).
The developmental pattern for GIP expression in the small intestine shows that GIP mRNA was first detected at fetal day 20 and remained at a low level of expression until postnatal day 3. After birth a rapid increase in GIP mRNA occurred between days 3 and 5 followed by a steady increase to the adult level (198). The increase between postnatal days 3 and 5 may be stimulated by suckling because of the presence in milk of carbohydrates and fats, the natural stimulants for GIP release. A somewhat different pattern of expression of GIP mRNA was measured by Tseng et al. (322) in which the mRNA was first detected in an 18-day fetus and reached maximum levels at birth.
H. Secretin
Secretin is one of the superfamily members that have retained the
basic 27-amino acid length. For location it is usually classified as a
gut hormone because it is secreted from endocrine S cells of the small
intestine, but it is also found in the developing pancreas in ß-cells
(173). Secretin may be expressed in other tissues, but the evidence is
not consistent. Secretin was reported to be in brain (5, 323, 324),
testis (324, 325), heart, lung, and kidney (324). In contrast,
expression in transgenic mice with a construct containing only 1.6 kb
of the secretin promoter resulted in a change in tissue expression.
Expression was in the small intestine, pancreatic ß-cells, and testis
as expected, but was additionally in colon, spleen, liver, and thymus.
The construct did not produce expression in the brain or kidney (326).
High expression of the transgene in additional tissues compared to
secretin may be due to the lack of additional 5'-flanking region in the
construct.
A specific receptor has been cloned and characterized for humans (327, 328). Also, the secretin receptor was isolated from two cell lines: human pancreatic adenocarcinoma (329) and rat/mouse hybrid NG108–15 cells (330).
Eighteen functions for secretin were published more than 15 years ago (216, 331) and more have been added. Among the important functions and ones that are within a physiological dose range are the actions of secretin on the pancreas, liver, and gut. Secretin is released from cells in the small intestine in response to gastric acid and fat; in turn, secretin acts on the pancreas to release a bicarbonate-rich fluid into the small intestine to neutralize acidity (216, 331). This secretin action is on the ductal cells and not the acinar cells of the pancreas where the enzymes are synthesized. Also, secretin stimulates gastric pepsin secretion and inhibits gastric acid secretion and gut motility (216, 331). Recently, secretin was found to cause a 3-fold increase in somatostatin release in the small intestine; somatostatin-14 (rather than -28) was selectively released and is known to inhibit acid secretion from the parietal cells (332). In the bile ducts, the ability of secretin to stimulate bile secretion is by facilitation of osmotic water transport in ductal cells in response to the intracellular transport of aquaporin (AQP1) water channels to the ductal plasma membranes (333). Secretin acts on the large bile ducts where the secretin receptor is expressed to increase Cl-/HCO3- exchanger activity and water flux resulting in more alkaline fluid in the ducts (334). This is similar to the action of secretin on the pancreatic ducts. In human adults, secretin injected in physiological amounts released insulin (335) and improved glucose tolerance (336). In dogs, secretin (injected into the endoportal system) rapidly released insulin, but not glucagon, from the pancreas (337).
Outside of the gut and related organs, secretin and its receptor are expressed in the testis in spermatids (325) and in the pituitary (5). In brain, secretin-like bioactivity is present (338) with widespread distribution of secretin mRNA and immunoreactivity in the cortex, limbic system, cerebellum, midbrain, hindbrain, and pineal gland (5, 123, 323). In contrast, the secretin receptor could not be detected in the brain (329, 330), opening the question of whether secretin acts in the brain. However, there are reports that secretin injected into the ventricle decreased open field activity and approaches to novel objects (339). Secretin did not affect analgesia in mice but depressed the analgesic effect of a challenge dose of morphine (340). Also, secretin acted on PC12 cells to stimulate tyrosine hydroxylase, which is important in the biosynthesis of catecholamines (341). Clinical trials are in progress to test the effect of secretin on autism, but to date the reports that behavior is improved in autistic children by treatment with secretin are mainly anecdotal (342).
In developing animals, secretin expression in the intestine (4) and pancreas (173) occurs late in gestation and is coincident with rapid pancreatic growth (4). Indeed, injection of secretin stimulates the growth of the pancreas if measured by weight or DNA content (343). Secretin synthesis is thought to be controlled by factors other than milk fat or gastric acid because feeding and gastric acid secretion occur after birth. In the pancreas a different expression pattern occurs for secretin mRNA; the highest expression occurs at fetal day 19 followed by a decline through postnatal day 14 and complete disappearance in the adult (173).
Secretin has some structural similarity to two novel neuropeptides, named hypocretins (344). Hypocretin-1 is 38 amino acids of which nine are identical to secretin (9/36 = 25% identity) and hypocretin-2 is 28 amino acids of which 14 are identical to secretin (14/36 = 39%). The two peptides, found only in the dorsal and lateral hypothalamus, are probably not members of the PACAP/glucagon superfamily as they do not have His or Tyr at the N terminus and their identity, unlike the superfamily members, is in the mid- and C-terminal region. Nonetheless, there is sufficient identity to suggest there is some structural relationship with the superfamily.
|
V. Conservation of PACAP—A Clue to Function |
|---|
|
TopAbstractI. IntroductionII. Evolution of the...III. Superfamily Evolution by...IV. Superfamily...V. Conservation of PACAP—A...VI. Conclusions and Future...References |
|---|
PACAP, like VIP, is reported to have an array of functions that involve the nervous, endocrine, cardiovascular, muscular, and immune systems. The puzzling aspect of PACAP is that the physiological event that triggers its release is not clear. For example, we know that high blood glucose triggers the release of insulin; high blood calcium triggers calcitonin; and excitement triggers adrenalin. But we do not know whether PACAP’s effects are a coordinated response to a single event in the body. One approach to understanding the major role of PACAP is to consider the effect of complete removal (gene knockout) of PACAP or its receptor from the time of fertilization in mice. The first gene knockout paper has just been published in which the specific PAC1 receptor was knocked out (345). In the pups, 60% of the PAC1-R null mice died in the first 4 weeks after birth, showing the importance of PACAP even though other superfamily members may compensate some of its functions. The knockout mice showed reduced glucose-stimulated insulin release and marked glucose intolerance after administration of either intravenous or gastric glucose. Another approach is to consider the basic functions of PACAP in the context of evolution. In vertebrates other than mice, it is not technically possible to do gene knockout, but in amphibians and fish we know that PACAP is as effective as GRF (provided the correct form is used in a species) in releasing GH from cultured pituitary cells. This observation suggests that PACAP is important in the metabolism of proteins (via GH), and the knockout studies in mice show that PACAP has a role in carbohydrate metabolism. Thus, experiments are beginning to answer the question of why PACAP is so tightly conserved in evolution: PACAP is a regulator of metabolism. Although the research on PACAP functions began only about 10 yr ago, the regulation of metabolism by PACAP is a workable hypothesis and creates a context into which the multiple functions described below may eventually be placed.
An overview of PACAP’s expression and receptors shows that PACAP mRNA expression and protein production have been localized to the following areas: the central nervous system (CNS), especially the hypothalamus, brainstem, and spinal cord (15, 17, 18, 19, 20, 143, 144, 346, 347, 348, 349, 350, 351), the peripheral nervous system innervating the eye (351), pituitary gland (17, 144, 346, 352), respiratory tract (351, 353, 354, 355), salivary glands (351), gastrointestinal tract (19, 348, 351, 356, 357, 358), reproductive tract (359, 360, 361), pancreas (358, 362, 363), urinary bladder (351, 358), and swimbladder (358). PACAP is also produced in several nonneural tissues such as the adrenal gland (348, 350, 356, 358, 364, 365, 366), gonads (15, 348, 356, 367), immune cells (231), and pancreas (368). Upon its discovery, PACAP was proposed to be a hypothalamic releaser of anterior pituitary hormones. This definition has since been expanded to include regulation of cell cycle, smooth muscle and cardiac muscle function, immune response, endocrine and paracrine secretions outside of the anterior pituitary, and exocrine secretions. All studies cited in this review involving exogenous exposure to PACAP have been conducted at physiologically relevant concentrations (1 x 10-7 M or less).
Several PACAP receptors have been identified to date. They are members
of the secretin/glucagon subfamily of receptors that are
seven-transmembrane receptors coupled to a G protein (204, 369). The
PACAP receptors have traditionally been known as type I receptors that
bind PACAP with greater affinity (100–1000x) than VIP and type II
receptors that bind PACAP and VIP with equal affinity (196, 211).
Recently, these receptors have been reclassified as PACAP-1 receptors
(PAC1-R) and VIP/PACAP-1 or -2 receptors
(VPAC1-R or VPAC2-R) (370).
Other names for the PAC1 and the two VPAC
receptors are listed in Table 1. The
PACAP receptor is a most interesting example of a product from a single
gene that has multiple signaling paths due to its variant forms. The
wide distribution of this receptor and the receptors shared with VIP
provide clear evidence that PACAP has many target sites and functions.
|
The eighth variant in the PAC1 receptor group is PAC1-R-TM4. This receptor is also a G protein-linked receptor, but unlike the PAC1-R and either of the VPAC receptors, the G protein is not linked to AC or phospholipase. Instead, this receptor appears to affect an L-type calcium channel. It differs from the PAC1 receptor variants by amino acid substitutions and deletions in the II and IV transmembrane domains. The receptor has been cloned in the rat and is found along with other PAC1-R variants in the rat cerebral cortex, cerebellum, brainstem, vas deferens, and lung. Interestingly, the PAC1-R-TM4 receptor is the only PACAP receptor expressed in rat pancreatic ß-islet cells (386).
Activation of VPAC1-R by PACAP or VIP stimulates an increase in cAMP via AC. The AC activation is likely achieved through coupling with a Gs protein. Originally, VPAC1-R activation was thought not to affect the inositol phosphate (IP)/PLC system (196, 211, 387). Although not observed in normal tissues, the VPAC1-R also couples to a Gi or Go protein when transfected in Chinese hamster ovarian (CHO) cells and does have a stimulatory effect on IP production in cells that express ß2 and ß3 PLC isoforms (388). In addition to its effects on the AC and PLC systems, in stably transfected CHO cells and HT29 human intestinal epithelial cells, VIP induces the VPAC1-R to increase intracellular calcium (389). The VPAC1-R has been cloned in humans (196, 387, 390) and rats (211). Distribution of the receptor in rats occurs in the following locations: lungs, small intestine, thymus, heart, aorta, liver, vas deferens, pancreas, kidney, adrenal gland, uterus, and the brain (especially the cerebral cortex, hippocampus, and several amygdaloid nuclei) (211, 213, 387). In frogs a VPAC-R with a sequence identity closest to the VPAC1-R and a distribution and binding pharmacology closest to VPAC2-R has been cloned. In goldfish a VPAC receptor has been sequenced (209) but has not been tested pharmacologically to determine whether it is a VPAC1, VPAC2, or a hybrid like the frog receptor. These studies may indicate that duplication of the VPAC receptors occurred in vertebrates.
The VPAC2-R has been cloned in humans (214, 391), rats (212), and mice (392). Like the VPAC1-R, the VPAC2-R binds PACAP and VIP with equal affinity. Both VIP and PACAP stimulate, with approximately equal potency, the activation of AC (212, 213, 214, 392, 391). There is also some suggestion that this receptor is linked to the PLC-IP system (392, 393). However, there are many instances in which no IP turnover is stimulated (394, 395). A clear picture of the link to the PLC-IP system remains to be found. The VPAC2-R has been identified in human skeletal muscle, heart, pancreas, placenta, and the brain (391). In rats and mice it has been localized to the stomach, colon, spleen, kidney, thymus, adrenal gland, heart, lung, pancreas, testis, ovary, uterus, pituitary, and the brain (especially the thalamus, hypothalamus, midbrain, brainstem, and olfactory bulbs) (212, 213, 392).
A. A regulator of the cell cycle and development
PACAP is reported to regulate cell division, cell cycle arrest,
differentiation, and cell death. These fundamental functions can affect
development and cell cycle dysfunction. The following section surveys
these actions of PACAP and how they relate to normal development,
particularly of the CNS, and to abnormalities resulting from improper
PACAP or PACAP receptor expression.
1. Cell division. PACAP regulates division of several cell types. For example, PACAP stimulates proliferation of a folliculostellate-like cell line (396), primordial germ cells (397, 398), chromaffin cells (399), clonal lactotrope and gonadotrope cell lines (400, 401), astrocytes (402), and peripheral sympathetic neuroblasts (403). In contrast, PACAP can inhibit proliferation of cerebral cortical precursors (403), corticotrope cells (404), and murine splenocytes induced to divide by concavalin A (405); PACAP in rats also inhibits DNA synthesis in aortic smooth muscle cultures stimulated by arginine vasopressin (AVP) (406), growth factor-stimulated chromaffin cells (399, 407), and cortical precursors (408). Thus, it would appear that PACAP regulates two opposing actions, both the stimulation and inhibition of cell proliferation. To explain these actions several scenarios have been studied, three of which are related below.
The first explanation is that PACAP can use different receptors to facilitate opposing outcomes (403). PACAP has been shown to stimulate proliferation in sympathetic neuroblasts and inhibit proliferation in cerebral cortical precursors by using different signaling pathways in these tissues. For example, sympathetic neuroblasts express PAC1-R-hop and have a measurable increase in cAMP and IP after PACAP stimulation, whereas cerebral cortical precursors express PAC1-R-s and, to a much lesser extent, PAC1-R-hop. Cerebral cortical precursors exhibit an increase in cAMP only after exposure to PACAP (403, 409). Therefore, the distribution of PACAP receptors can dictate opposing functions in different tissues. Although both PAC1-R-s and PAC1-R-hop can stimulate cAMP and IP accumulation, some difference in intracellular signaling efficiency appears to exist depending on the cellular system employed.
The second explanation is that PACAP can selectively activate intracellular pathways through concentration differences. In cultured rat astrocytes, PACAP causes proliferation at concentrations that are below those that stimulate cAMP. At these low concentrations (10-14-10-12 M) PACAP activates mitogen-activated protein (MAP) kinase, which is associated with DNA synthesis and proliferation. In fact, a cAMP analog actually suppresses MAP kinase activation in cultured rat astrocytes (402). Thus, PACAP can trigger different intracellular pathways by concentration-dependent pathway selection.
Finally, PACAP can have both mitogenic and antimitogenic effects on the same tissue. For example, PACAP has mitogenic effects on adult rat chromaffin cell cultures but has antimitogenic effects on nerve growth factor (NGF)-stimulated proliferation of chromaffin cell cultures in rats (399, 407). These effects do not seem to be regulated by either dosage effects or differential receptor activation. One possible explanation is that inhibition of the mitogenic effects of NGF by PACAP might be a mechanism by which neurally derived signals override growth factor-stimulated proliferation during development. PACAP might accomplish this by commandeering portions of the same intracellular pathway used by the growth factor (399).
In summary, the evidence points to an important role for PACAP in fine tuning the cell division of various neuronal and nonneuronal cell types.
2. Differentiation. Not only does PACAP regulate proliferation, but it also appears to regulate differentiation. The presence of multiple PACAP receptor variants makes these opposing actions seem plausible. PACAP activation of the PAC1-R is known to stimulate neurite outgrowth, a sign of differentiation, in rat pheochromocytoma PC12 cells (410, 411), some human neuroblastoma cell lines (412), cortical precursor cells (408), immature cerebellar granule cells (413), and a corticotrope cell line (404). In addition, the coincident expression of both PACAP and the PAC1-R in the ovary (granulosa cells) and testis suggests that PACAP may be involved in germ cell maturation (191, 367, 381, 414, 415, 416). In the testis PACAP is processed and expressed in maturing spermatids in a stage-specific manner at a critical point in spermatogenesis (367, 414, 416). PACAP may also affect Cl- secretions in the epididymal epithelium. Cl- secretions are thought to help maintain a stable microenvironment, which is important for the promotion of maturation and storage of spermatozoa (417).
3. Apoptosis. PACAP is involved in both protecting cells from apoptosis and in triggering apoptosis, depending on the system (413, 418, 419, 420, 421, 422). Apoptosis, or programmed cell death, is a tightly controlled suicide program initiated by the cell. It has two main functions: to act as a developmental regulator and to kill cells that have been damaged. The latter role prevents a potentially dangerous phenotype from being propagated. PACAP affects apoptosis by regulation of gene transcription, perhaps by transactivation of cAMP response elements and/or through activation of particular PACAP receptor variants (419, 420).
The neuroprotective role of PACAP has been studied in rat cerebellar cells. In these cells PACAP acts on PAC1-R to induce the MAP kinase pathway via AC and protein kinase A (PKA) activation (419, 420). This activation may be required to protect cells from apoptotic events, ensuring proper cerebellar development. Cerebellar granule cells express PAC1-R-s and PAC1-R-hop, whereas cerebellar glial cells express only PAC1-R-s. PACAP stimulates c-fos gene expression in cerebellar granule cells through a cAMP/protein kinase A-dependent mechanism involving the PAC1-R. The protein kinase A pathway is a major mediator of the neurotrophic actions of PACAP. It is likely that both c-fos gene transcription and one or both of the PAC1-R variants are involved in the anti-apoptotic effects of PACAP on cerebellar granule cell cultures (413, 418, 419, 420, 423). Also, the neuroprotective actions of PACAP have been noted in the induced cell death of the following: rat embryonic cortical neurons (424), hippocampal neurons (425), sympathetic neurons (426), rat thymocytes (427), and PC12 cells (407). In contrast, in chick sympathetic neuroblasts PACAP rescues cells from death but does not act through the usual pathways. Instead, PACAP appears to operate through an unidentified receptor to decrease the concentration of a death protein by stimulating destruction of the protein (428). Although a direct link for protection from apoptosis has not been established, PACAP has been shown to promote cell survival in the following cultured primary neurons: cortical, hippocampal, septal cholinergic, mesencephalic dopaminergic, and dorsal root ganglion (429). Clearly, PACAP has a neuroprotective function in many neural and some nonneural cell types.
The ZAC1 protein, a recently discovered zinc-finger protein, induces apoptosis and G1 cell cycle arrest in tumor cells. ZAC1 and the tumor suppressor, p53, inhibit tumor cell growth in vitro by different pathways. Interestingly, the two proteins also induce expression of the PAC1-R gene. Thus PACAP has a protective function in neurons but it appears to promote apoptosis in tumor cells (421, 422).
4. Development. There are several lines of evidence to suggest that PACAP may have a role in development of the nervous system and several other organs in mammals. The presence or effect of PACAP and its receptor has been examined in the developing CNS (350, 376, 403, 408, 430, 431, 432, 433, 434, 435, 436, 437, 438), eye (439), liver (440), adrenal glands (350, 364), and pancreas (441). The data to date do not necessarily agree on a common function for PACAP among these organs or even on a common function within different regions or cell types of the same organ. However, both cAMP and IP signaling pathways are activated by PACAP in developing tissues. The predominant receptor expressed during development in the tissues examined is the PAC1-R with the exception of rat fetal hepatocytes, which appear to express a VPAC-R, and human fetal retinal cells, which express mRNA for both the PAC1-R and a VPAC-R (376, 403, 408, 432, 433, 434, 435, 436, 437, 439, 440, 441). In short, the functions of PACAP in development are not well understood, but several examples are highlighted to demonstrate the possible significance of PACAP in the developmental process.
The role of PACAP in the developing nervous system has been examined in greater detail than in any other system during development. Evidence for the presence of PACAP and its receptor is found during embryogenesis in the autonomic and sensory ganglia and the spinal cord (350, 429), glial and neuronal cells (432), and in the following brain regions: the neocortex, cortex, cortical plate, thalamic and hypothalamic nuclei, habenular nucleus, hippocampus, septum, trigeminal ganglion, amygdala, olfactory bulbs, inferior colliculus, solitary nucleus, inferior olive and other pontine nuclei, midbrain, and hindbrain, particularly the cerebellum and neural tube (376, 403, 429, 431, 433, 434, 435, 438, 442). In the developing brain the PAC1-R is the dominant receptor, but there is differential distribution of the variants. In the postnatal rat brain, the PAC1-R-s variant is found in the cortex, hippocampus, cerebellum, and hypothalamus, whereas the PAC1-R-hop variant is found only in the cerebellum and the hypothalamus, not the cortex and hippocampus. Differential distribution of PAC1-R variants is further confirmed by experiments revealing that PACAP-induced cAMP production occurred in all four of the brain regions examined, but [3H]inositol monophosphate accumulation occurred only in the cerebellum and hypothalamus (376, 419, 423, 431, 436, 437). Although both PAC1-R-s and PAC1-R-hop have been shown to trigger IP turnover, it appears that PAC1-R-s is not linked to the IP path in the cortex and hippocampus. The significance of this distribution pattern in the cerebellum is discussed below.
The cerebellum, a division of the hindbrain, is one of the first brain regions to express PACAP and the PAC1-R variants short and hop. The immature cerebellum is composed of four layers, the external granule cell layer (EGL), the molecular layer, the internal granule cell layer (IGL), and the medulla. Immature neurons are generated early in development in the EGL. These neurons migrate through the molecular layer to reach their destination in the IGL. The development of the cerebellum is a complex process that involves proliferation, differentiation, migration, and massive cell death in both the EGL and IGL. PACAP immunoreactivity is present early in the cerebellum of postnatal rats (432, 433). At birth (postnatal day 0, P0), PACAP receptors are present in the EGL and medulla. Later, these receptors disappear (P8-P25) in concert with the involution of the EGL. Concurrent with the involution of the EGL, PACAP receptors appear in the IGL and the molecular layer. Once the cerebellum matures, receptors appear only in the granule cell layer (434). This period of intense PACAP and PACAP receptor expression is coincident with a period of neurogenesis in the rat brain (443). The cerebellum has two PAC1-R variants, PAC1-R-s and PAC1-R-hop (419, 431, 434, 435, 436, 437). The differential distribution of PACAP receptors in the immature brain may allow PACAP to play several roles in the postnatally developing cerebellum. Experiments on cerebellar neuroblasts in culture indicate that PACAP acts through PAC1-R to increase cell survival and differentiation (413). The cell survival actions of PACAP appear to be mediated by the cAMP second messenger pathway. In addition, the PACAP-induced cAMP accumulation leads to the transcription of important regulatory factors such as, c-fos. c-fos Acts as a transcription factor for many genes. Therefore, PACAP-induced expression of c-fos provides an indirect pathway for PACAP to promote the expression of an array of genes in these cells (419). The effects of PACAP on cell survival and differentiation have been confirmed in vivo (444). PACAP injections in P8 rats cause an increase in the thickness of the EGL, the molecular layer, and the IGL, but PACAP has no effect on the medulla. It is known that during the first two postnatal weeks the EGL undergoes apoptosis. Given the cell survival effects of PACAP in vitro it can be supposed that the increase in EGL thickness after in vivo PACAP administration probably results from apoptosis protection rather than proliferation. An increase in cells migrating through the molecular layer to the IGL and an increase in the number of neurites from the IGL projecting into the molecular layer have also been observed. These factors could account for the increase in the volume of the molecular layer and the IGL after PACAP injections. There is also some suggestion that PACAP accelerates the migration of granule cells to the bottom of the IGL. In summary, both the in vivo and in vitro findings indicate that PACAP has a role in protecting cerebellar cells from apoptosis and in promoting differentiation (i.e., neurite outgrowth) (413, 444). Given the fact that we know the cerebellum expresses two PAC1-R variants, it will be interesting to determine whether these variants are used to regulate the distinctly different effects of PACAP within this region of the developing brain.
Several studies indicate that PACAP may function in many other regions of the CNS during development. PACAP mRNA and PAC1-R mRNA are detected as early as embryonic day 9.5 (E9.5) in mice (430) and at E14 in rats (433). The following PACAP receptors have also been found: PAC1-R-s and a longer transcript (representing one or more of the variants, PAC1-R-hip, PAC1-R-hop1, and PAC1-R-hop2 (423, 431). In situ studies in mice ages E9.5 and onward localized the receptor and peptide mRNA to the neural tube, hindbrain, trigeminal ganglia, dorsal root ganglia, and the developing sympathetic chain. PACAP mRNA was also found in the hypothalamus and the nuclei of the pons and medulla. The presence of labeled cells in the dorsal root ganglia and in autonomic structures suggests that neural crest-derived structures may express PACAP and PAC1-R during development (431, 438). A recent study suggests a role for PACAP in the patterning of the nervous system (438). PACAP and PAC1-R were present by in situ hybridization in the mouse neural tube on E10.5. PACAP was expressed throughout the neural tube, and PAC1-R mRNA was expressed in the alar and floor plate region of the underlying ventricular zone. Distribution of PACAP mRNA in two columns of cells in the ventromedial portion of the neural tube places PACAP in the same region as developing autonomic motoneurons. In addition, PACAP down-regulated expression of two PKAdependent patterning genes (sonic hedgehog and gli-1) in cultured neuroepithelial cells (438). Human fetal retinal cells also express mRNA for PACAP and PACAP receptors (PAC1-R, VPAC-R). In these cells PACAP stimulates AC activity, although, the exact function (i.e., proliferation, differentiation, apoptosis, etc.) is not clear (439). Also, PACAP is involved in cerebral cortical neurogenesis through initiating inhibition of proliferation in cortical precursors. It is this action of PACAP through the PAC1-R-s receptor that is thought to elicit cell cycle exit and differentiation of the developing cerebral cortex (403, 409).
PACAP is involved in development outside of the brain as well. In fetal rat hepatocytes, PACAP stimulates an increase in cAMP levels through a VPAC-R. In addition, exposure to PACAP results in an increase in corticosteroid-binding globulin mRNA, suggesting it may participate in the regulation of gluconeogenesis during development (440). In the adrenal gland of newborn rats, immunoreactive (ir)-PACAP nerve fibers have been observed on chromaffin cells in the adrenal medulla. In addition, mRNA for the PAC1-R has been localized to adrenal medullary cells by in situ hybridization (350, 364). In human fetuses at 14–20 weeks of age, PAC1-R binding sites have been localized to chromaffin cells. PACAP induced a dose-dependent increase in cAMP production and a modest increase in IP formation in human fetal adrenal cell suspensions and in cultured cells. These data prove that a functional PAC1-R is present in the human adrenal medulla during a phase of organization characterized by the migration of chromaffin cells from the periphery to the central part of the gland (445). The presence of both PACAP and its receptor in the rat and human adrenal gland suggest a possible role for this peptide in adrenal gland development. In addition, PACAP affinity studies have revealed the presence of a PACAP binding site in the postnatal calf pancreas and PACAP activates AC in this tissue, suggesting a role in postnatal pancreatic function (441).
Thus, like some other superfamily members, PACAP has a role in development, particularly brain development. VIP has considerable effects on brain development (i.e., proliferation, cell survival, and differentiation; see Section IV). However, due to the widespread and abundant presence of the PAC1-R, PACAP may overshadow VIP. Other factors, such as glucagon and GRF, have a more localized expression in the brain during development (Section IV). This may suggest a more discrete function for these hormones. Outside of the brain PACAP may play a role in liver, adrenal gland, and pancreatic development. Other superfamily members, such as glucagon, GIP, and secretin, are found in the gut or pancreas but do not appear to have a major role until late gestation or postnatally. It is noteworthy that, so far, PACAP and VIP are the only superfamily members found in both the brain and gut during fetal development.
5. Dysfunction. PACAP and its various receptor forms are found in many cancers. PACAP is expressed as a protein or mRNA in pheochromocytomas (446), neuroblastomas (447, 448, 449), human ovarian cancers (450), nerves innervating parathyroid adenomas (451), and in gliomas (452). The receptors (PAC1-R and/or VPAC-R) have an even wider distribution among cancerous tissues and can be found in glial tumors (452, 453), breast cancer (454), intestinal cancer (454), pancreatic cancer (454, 455), non-small-cell lung cancer (456), retinoblastomas (457), lymphoid tumors (458), pituitary adenomas (459), adenocarcinomas (460), tumorous adrenal cells (461), prostate cancer (462), and neuroblastomas (401, 448, 449).
In these tissues PACAP is involved in both proliferation and differentiation, making study of this peptide useful for both determining how cancers develop and how proliferation can be hindered or stopped. In glioblastomas and some human colonic adenocarcinoma cell lines, PACAP reduces proliferation (460, 463), and in neuroblastomas (NB-OK-1), exposure to PACAP stimulates cAMP accumulation, arrests cell growth, and induces morphological changes such as neurite outgrowth (412, 457). PACAP stimulates proliferation in a pancreatic acinar tumor cell line (464, 465), non-small-cell lung cancer cells (466), and a prostate cancer cell line (462). Differences in receptor subtype expression may explain the different actions of PACAP listed above (460). In addition, the presence or absence of retinoic acid may alter the mitogenic effects and binding capacity of PACAP (449). The proliferative actions of PACAP may be regulated in some cancers through PACAP-induced transcription of c-fos and c-jun, two important nuclear oncogenes. The products of these two oncogenes heterodimerize to form the transcription factor AP-1 (462, 465, 467). Moreover, PACAP induces MAP kinase activity (465). Given the above information, PACAP’s role in the cell cycle appears to extend to a role in cancer as well. However, a major role in tumorigenesis cannot yet be supported by these data because tumor cells are known to express many peptides and receptors compared with normal cells.
Holoprosencephaly is a developmental dysfunction, which may in some cases be caused by chromosomal abnormalities affecting PACAP and PACAP receptor expression. The holoprosencephaly phenotype is characterized by incomplete cleavage of the forebrain and several facial abnormalities, which range from mild (microcephaly, mild hypotelorism, and single maxillary central incisor), to severe (cyclopia, a primitive nasal structure and sometimes midfacial clefting). Four loci have been identified as sites of chromosomal rearrangements leading to holoprosencephaly. One of the sites identified is associated with PACAP and involves a chromosomal rearrangement mapped to the location of the human PACAP gene at 18pter-q11 (468). Another of the four loci maps to the human chromosomal region 7q36, which includes the site for the VPAC2-R (7q36.3). These genetic studies suggest that PACAP is an important gene involved in patterning of the midline ventral CNS (468, 469).
The actions of PACAP on proliferation, differentiation, and apoptosis present a complex and interesting story. Although studies on cell cycle regulation, development, and dysfunction are now only scratching the surface, the data collected so far suggest that there is an intricate balancing act between the death-promoting, protective, and proliferative actions of PACAP.
B. A regulator of smooth and cardiac muscles
PACAP functions as a neurotransmitter or neuromodulator of smooth
muscle tone. Consequently, PACAP may affect many systems in the body
that are composed in part of smooth muscle. Its effects on the vascular
system, respiratory system, digestive system, and reproductive system
are discussed below. Again, the studies have been done primarily in
mammals so that the evolutionary trends are not known.
1. Effects on vascular system smooth muscle.
a. General circulation.
PACAP appears to play an important
role in neural and hormonal regulation of systemic circulation. The
primary action of PACAP on circulation may be accomplished through
vasorelaxant effects on vascular smooth muscle. PACAP-ir nerve fibers
have been found around blood vessels in the respiratory tract (353, 354), and PACAP can bind to the membranes of blood vessels (470).
However, localization of PACAP to vascular smooth muscle is not the
only evidence of its action here. Intravenous or intraarterial
injection of PACAP into humans, sheep, dogs, cats, and rats results in
a decrease in blood pressure (6, 10, 470, 471, 472, 473, 474, 475, 476). However, different
studies have shown that PACAP causes a variety of other responses in
the same animals. For example, in cats and dogs, intravenous
administration of PACAP induces a biphasic change in arterial pressure
characterized by an initial decrease followed by an increase (471, 472, 477, 478). Other studies in rats and dogs contradict the above results
and show that after intracerebroventricular (rats) or intracisternal
(dogs) injection of PACAP, there is an increase in blood pressure (479, 480). Although it is true that PACAP plays a role in regulation of the
vascular system, it is evident from the literature that this
role is unclear. Some of this confusion is likely due to the
number of species studied, receptor differences, modes of
administration of the peptide, concentration of the peptide, and other
aspects of the experimental methods used. In general, in the above
in vivo studies it is difficult to separate the direct
effects that PACAP has on vascular smooth muscle from its other
possible actions on the vascular system. Therefore, in vitro
studies have attempted to isolate the direct actions of PACAP on
vascular smooth muscle tone.
First, it is clear that PACAP directly causes relaxation of vascular smooth muscle. In vitro studies have revealed that PACAP is a potent vasorelaxant of arterial segments (354, 473, 481, 482, 483). These vasorelaxant effects of PACAP appear to be mediated by the AC/cAMP signaling pathway (473, 482). In porcine coronary arteries, the vasorelaxant effect of PACAP was equal to that of VIP, but in the rabbit aorta, PACAP was a 100-fold more potent vasorelaxant than VIP. In both cases the effects of PACAP and VIP were endothelium independent. This is unlike the situation in humans and guinea pigs where the removal of the endothelium from the pulmonary arteries abolished the vasorelaxant effects of PACAP, but not those of VIP. This suggests some receptor or tissue differences may exist between the species (354, 481, 482, 483). Those in the "endothelium-dependent" group suggest that PACAP’s vasorelaxant effect is mediated by nitric oxide (NO) synthase, which is released from the endothelium. Support for this theory is derived from the fact that a NO synthase inhibitor, N(G)-monomethyl-L-arginine, inhibits PACAP-induced vasorelaxation of endothelium-intact human and guinea pig pulmonary artery segments (481). In contrast, in vivo analysis of vascular responses in cats indicates that nitro-L-arginine methyl ester (also a NO synthase inhibitor) has no effect on PACAP-induced vasorelaxation of the pulmonary vascular bed (472). Perhaps information on receptor distribution in tissue layers, among vessel types, and between species will explain these results.
Second, a perplexing aspect of PACAP’s action on vascular smooth muscle is the biphasic change in arterial pressure that sometimes occurs after PACAP administration. The initial decrease in blood pressure is logical, considering the vasorelaxant properties of PACAP on vascular smooth muscle. The increase in blood pressure seen in the latter phase of the biphasic response may be due to less direct actions of PACAP. One of the key factors that may determine the vascular response is dosage. In cats and dogs, a low dose of PACAP administered into the external jugular vein (0.1 nmol/kg) or femoral vein (0.01 nmol/kg) induces a decrease in arterial pressure, and a higher dose (3.0 nmol/kg) triggers a biphasic change in arterial pressure (471, 472, 478). The initial phase probably results from the vasorelaxant effects of PACAP. The second phase is thought to result from a large increase in cardiac output and/or some central actions of PACAP (478). Two studies examine the second phase of the response. In the first experiment, intracerebroventricular injection of PACAP caused a dose-dependent (0.1–0.5 nmol/rat) increase in mean arterial pressure in rats. This reaction can be explained in part because intracerebroventricular injection of PACAP raises plasma AVP, a vasoconstrictor (479). In a second experiment the same increase in AVP was found after intracisternal injection of PACAP in dogs (480). This effect of PACAP on AVP release appears to be mediated by both a VPAC and a PAC1-R in rats (479). However, the increase in mean arterial pressure is greater than can be accounted for by AVP alone; therefore it has been suggested that, in addition, PACAP may influence the central cardiovascular control system through stimulation of the sympathetic nervous system (479, 480). A recent study (484) confirms this suggestion. Intrathecal injection of PACAP into anesthetized rats has an excitatory effect on sympathetic preganglionic neurons. One of the effects of this excitation is an increase in spinal sympathetic outflow, which, in turn, leads to an increase in blood pressure. Like intracerebroventricular and intracisternal injections, intrathecal administration limits access of the peptide mainly to neurons in the spinal cord. It is also possible that the pressor response after intravenous injections of high doses may allow PACAP access to sympathetic neuron excitation. The receptor type mediating the effect on the spinal cord has not yet been determined, but immunocytochemical and mRNA studies have confirmed localization of PACAP to sympathetic preganglionic neurons (197, 485).
Catecholamine release may be another factor that mediates the PACAP-induced pressor response. In cats, when PACAP is administered intravenously at a high dose (3.0 nmol/kg), vasorelaxation is followed by vasoconstriction in the hindquarter vascular beds. In this case the pressor response is thought to result from a PACAP-induced stimulation of the release of catecholamines from the adrenal gland or by the release of norepinephrine from adrenergic terminals in the vascular bed (472). Also, PACAP has been noted to cause an increase in plasma epinephrine after intracisternal administration in dogs (480).
The different pressor responses also raise questions regarding receptor variants. So far, studies in the cat suggest that two receptor types are present due to different responses to PACAP and VIP (471, 472). In addition, the endothelium-dependent actions of PACAP vs. the endothelium-independent actions of VIP in guinea pig and human pulmonary artery relaxation may reflect the use of PAC1-R and VPAC-R types. One receptor (likely a VPAC-R) appears to be present in the rat aorta, tail artery, iliac and femoral veins, guinea pig pulmonary artery, guinea pig aorta, and porcine coronary artery (470, 472, 483).
The studies summarized above identify PACAP as having a role in vascular regulation of the general circulatory system. Its primary action is vasorelaxation, and its vasoconstrictor actions are likely to result from a secondary reaction and include sympathetic regulation of central cardiovascular control, differential receptor use, PACAP-induced AVP, or catecholamine release. Other experiments have revealed that PACAP can have a more localized vasorelaxant effect on blood flow and hence affect the function of particular organs and systems, such as the respiratory system, digestive system, reproductive system, visual system, pancreas, and brain. The central question remains as to whether PACAP’s effect on vasorelaxation is related to its effect on metabolism. If PACAP triggers storage of glucose (via insulin) or proteins (via GH), is vasorelaxation part of a coordinated response?
b. Reproductive tissue circulation.
In the reproductive
system, PACAP can act as a vasorelaxant in the human-uteroplacental
unit (486), rabbit ovarian artery (487), and the vasculature of the
female genital tract (360, 488). In males PACAP has been localized to
nerve fibers in the human cavernous tissue where it may play a role in
the vasorelaxant response necessary for engorgement of the penis with
blood, leading to penile erection (359).
c. Respiratory tissue circulation.
PACAP-immunoreactive nerve
fibers have been localized around blood vessels found in the nose and
tracheobronchial wall of the respiratory tract of the rat, guinea pig,
ferret, pig, sheep, and squirrel monkey (353) and in the esophagus of
sheep and man (489). Localization of PACAP-ir fibers to the respiratory
system suggests it may play a role in vascular regulation of these
tissues.
d. Digestive tissue circulation.
PACAP
(10-11 to 10-7
M) has vasorelaxant effects on endothelium-denuded rat
mesenteric arteries (483) and causes an increase in blood flow to the
submandibular gland (490). It decreases blood flow to the duodenum when
administered at 10 nmol/kg body weight (474). In addition, PACAP-ir
nerve fibers innervate small arteries and arterioles of the fish and
mammalian gastrointestinal tract, which suggests it may be involved in
the regulation of blood flow in the digestive system (355, 358). This
also suggests that the vasorelaxant properties of PACAP may have arisen
at least as early as the emergence of fish.
e. Pancreatic tissue circulation.
PACAP increases pancreatic
blood flow at lower doses (10-11 to
10-9 M), but at higher doses
(10-7 M) PACAP provokes a transient
decrease of blood flow followed by a return to a basal flow rate. The
effect of PACAP on pancreatic blood flow may be mediated by a VPAC
receptor (491). In situ hybridization studies in the rat
pancreas identify the VPAC2-R subtype in
pancreatic blood vessels (213). However, the vasoconstrictor effect of
PACAP at a high concentration (10-7
M) and the failure of VIP to induce the same
response indicate that PAC1-R receptors are
involved as well. The increase in blood flow to the pancreas caused by
PACAP results in redistribution of the blood flow in favor of the
exocrine pancreas so as to stimulate blood perfusion and exocrine
secretions for the gut (474).
f. Ocular tissue circulation.
Intraocular tissues are supplied
with nutrients from two vascular beds, the retinal and uveal blood
vessels. In rabbits, infusion of PACAP results in a decrease in uveal
vascular resistance and an increase in blood flow to the uveal blood
vessels of the choroid. PACAP receptors are present in the anterior
uvea, choroids, and retina. In the choroid, it is likely that these
receptors are localized to blood vessels. Both
PAC1-R and VPAC-R are present in the intraocular
tissues although PAC1-R predominate in the
choroid. Exposure to PACAP causes an increase in cAMP concentrations in
these tissues. PACAP also results in an increase in blood flow to the
eyelids and nictitating membrane (492, 493).
g. Cerebral circulation.
PACAP may also have a role as a
vasorelaxant of canine cerebral arterioles, rat intracerebral
arterioles, cat pial arteries in vitro, and cerebral
arterioles in newborn pigs in vivo (494, 495). Innervation
of cerebral arteries by PACAP-ir nerve fibers has been observed in cat
pial arteries (496). Topical administration of PACAP through a cranial
window in newborn pigs caused vasorelaxation and a dose-dependent
increase in cerebrospinal fluid cAMP levels. Similar results were
observed in the pial artery after intravenous administration of PACAP
during hypoxia in the pig (495, 497). It appears, therefore, that PACAP
is a regulator of vasorelaxation in the cerebral circulatory system.
2. Effects on nonvascular smooth muscle.
a. Respiratory smooth muscle.
PACAP-ir fibers and receptor
sites (possibly a VPAC-R and a PAC1-R) have been
localized in the rat lung (353, 354, 381, 382, 498). In particular,
PACAP-ir fibers have been observed close to small bronchioles in
squirrel monkeys, pigs, sheep, ferrets, guinea pigs, and rats (353, 354). The current hypothesis is that PACAP is involved in bronchial
relaxation as shown in primates and guinea pigs (499, 500).
Outside of the lung, PACAP-ir fibers are located throughout the respiratory tract. Immunoreactive fibers have been observed beneath the epithelial surface among bundles of smooth muscle in man, squirrel monkeys, pigs, sheep, ferrets, guinea pigs, and rats (353, 354, 489). In vitro studies revealed that PACAP has a concentration-dependent relaxant effect on guinea pig and rabbit tracheal segments (354, 501, 502, 503). Removal of epithelium from guinea pig tracheal segments reduced the relaxant properties of PACAP but did not entirely inhibit relaxation (501). In rabbit tracheal segments, PACAP-induced relaxation was followed by an increase in intracellular cAMP levels (503). The relaxant effects of PACAP may also involve activation of a Na+-K+-ATPase (503). These studies suggest that PACAP is a neurotransmitter/neuromodulator regulating smooth muscle tone in airways. PACAP’s relaxant effects may be useful for treating bronchial asthma. Studies in guinea pigs show that PACAP inhibits histamine-induced respiratory resistance (504).
b. Digestive tract smooth muscle.
PACAP-ir fibers have been
identified innervating smooth muscle in the gastrointestinal tract of
humans, pigs, sheep, cats, ferrets, guinea pigs, hamsters, rats, mice,
chickens, and fish (355, 357, 358, 505). Several experiments have shown
that PACAP induces a concentration-dependent relaxation of
gastrointestinal tract smooth muscles in humans, cats, rabbits, guinea
pigs, and rats (483, 506, 507, 508, 509, 510, 511, 512, 513, 514, 515). The actions of PACAP are mediated by a
PAC1-R in the human sigmoid colon and rat distal
colon (506, 513), whereas inhibition of contractions is mediated by a
VPAC-R in rabbit pyloric muscles and in gastric cells (511, 516). A
study on rat ileal longitudinal muscles suggests three receptors induce
relaxation: 1) a PACAP-27-preferring receptor coupled to an
apamin-sensitive Ca++-dependent
K+ channel; 2) a PACAP-27 and -38 preferring
receptor; and 3) a VIP-specific receptor (508). Neither the
PACAP-27-preferring receptor nor the VIP-preferring receptor has yet
been cloned. However, other researchers have proposed the existence of
a PACAP-27-preferring receptor (517). In addition, an apamin-sensitive
PACAP receptor coupled to potassium channels has been proposed to exist
in gastrointestinal tract smooth muscle, but the preference for
PACAP-27 has not been observed (509, 514, 515). To date, studies have
linked PACAP’s relaxant effects to cAMP, NO synthase, and
K+ channels in the digestive tract (506, 508, 512, 513, 516).
In addition to its relaxant effects, PACAP appears to mediate central nervous system effects on gastric smooth muscle. PACAP causes an increase in rat intragastric pressure via excitatory actions on the vagal nerve, located in the hindbrain. Thus, PACAP can have direct relaxant effects via local interaction with PACAP receptors in the gut smooth muscles or indirect effects via stimulation of excitatory neurons located in the brain. These neurons synapse with gastric smooth muscle and release excitatory neurotransmitters (518). Taken together, these studies suggest PACAP is a potent modulator of stomach muscle relaxation and contraction, colonic motility, and motor activity in the gut. Also of note, PACAP has both excitatory and relaxant effects on guinea pig gallbladder muscles. The contractile effect in the gallbladder is mediated through a PAC1-R and the relaxant effect is mediated through a VPAC receptor (519).
c. Reproductive tissue smooth muscle.
In the reproductive
system, PACAP has been localized to fibers associated with smooth
muscle bundles in the rat genital tract (360). PACAP has also been
shown to induce relaxation of urethral smooth muscle in the female pig
(361).
3. Effects on cardiac muscle. It is difficult to elucidate PACAP’s direct effect on heart rate due to its concurrent effect on blood pressure. However, PACAP has been observed to increase heart rate in sheep and cats (472, 475), although an intravenous injection of PACAP into beagle dogs slowed heart rate after a transient increase. The changes in heart rate in dogs resulted in hypotension after transient hypertension (478). In isolated and blood-perfused dog heart preparations, PACAP directly increased the sinus rate and atrial and ventricular contractile force (520). The direct effects of PACAP on heart rate and contractile force have been investigated in vitro on rat cell cultures of cardiac myocytes. In these cultures PACAP stimulates cAMP production in a dose-dependent manner (10-12 to 10-6 M) (477). These results suggest that in the first phase of the in vivo biphasic response, PACAP directly stimulates cardiac myocytes to induce an increase in heart rate and contractile force. The second phase (decrease in heart rate and contractile force) is mediated by interaction of PACAP with cardiac parasympathetic nerves via the PAC1-R (520, 521, 522). PACAP immunoreactivity has been localized in the guinea pig heart to neuronal fibers and a subpopulation of intrinsic postganglionic cardiac neurons. In addition, cardiac ganglia expressed the PACAP transcript and PAC1-R transcripts. The predominant receptor expressed is the PAC1-R-vs., but low levels of the short and hop2 isoforms are also expressed. The majority of parasympathetic neurons were immunoreactive for PAC1-R. Exposure of atrial whole-mount preparations to PACAP results in depolarization of cardiac ganglia and increased neuronal excitability. These results support the evidence that PACAP and a PAC1-R modulate the parasympathetic-mediated inhibition of cardiac output (522).
It should be noted that VIP is another superfamily member that is a potent relaxant of vascular and nonvascular smooth muscle in the gut (Section IV). Therefore, its effects on the common VPAC-R may be dominant compared with PACAP. However, in this review we show that PAC1 receptors are also expressed in these tissues, indicating that PACAP has a unique function separate from VIP. Knockout experiments involving either one or both of the receptors or hormones would be helpful in determining unique vs. shared responses to these hormones. Other than VIP, the rest of the superfamily members do not play a significant role in smooth muscle regulation.
C. An immune system regulator
PACAP immunoreactive cells can be found in various tissues
associated with the immune system in rats: bone marrow, thymus, spleen,
lymph nodes, and duodenal mucosa (231). PACAP receptors are also
associated with many immune cells. VPAC receptors have been localized
to leukocytes, mononuclear cells, lymphoblasts, lymphocytes, monocytes,
lymphoid cells, macrophages, myeloma cells, T cells, and B lymphocytes
(523). Several other papers have noted that PACAP receptor expression
is mainly in lymphocytes, macrophages, and astrocytes. In the
lymphocytes of rats and mice, a VPAC1-R gene is
expressed (458, 524, 525) and in rat macrophages both immunoreactive
and mRNA data identify a VPAC-R and a PAC1-R
(524, 526, 527). Evidence also suggests the existence of a VPAC-R and
perhaps a PAC1-R in human mononuclear cells
(528). In rat astrocytes a PACAP receptor exists, although the type of
receptor has not yet been determined (529). The function of PACAP in
these cell types is unclear but there are connections between PACAP and
immune cell maturity, mobility, and the inflammation response as
described below.
1. Immune cell protection. In rats, thymus lymphocytes (thymocytes) have increased cell survival when exposed to PACAP. The protection results from exposure to PACAP causing a dose-dependent (10-13 to 10-6 mol/liter) inhibition of spontaneous apoptosis and a decrease in proliferation (405, 427). It has been suggested that PACAP is involved in T cell maturation. As thymocytes reach maturity, proliferation decreases. However, one of the first processes in thymocyte maturity is rescue from negative selection by inhibition of apoptosis. Because PACAP can regulate inhibition of both thymocyte proliferation and apoptosis, it may be an important regulator of thymocyte maturity (427). This effect appears to be mediated by a VPAC1-R. The antiproliferative actions of PACAP have been linked to the stimulation of a cAMP-dependent protein kinase (405, 427, 458, 525). Unlike rats, mice do not have PACAP-binding sites on thymocytes (381).
PACAP is also involved indirectly in lymphocyte maturation by stimulating the release of interleukin-6 (IL-6) from folliculostellate cells in the pituitary. The release is triggered by a PACAP-induced dose-dependent increase in Ca++. IL-6 stimulates B cell growth and differentiation and increases synthesis and secretion of immunoglobulins by B lymphocytes (530, 531).
2. Immune cell phagocytosis. When exposed to PACAP, rat peritoneal macrophages were observed to increase both phagocytosis and production of superoxide anion (required for digestion of ingested cells) in a dose-dependent manner. phagocytosis and cell digestion are the two most characteristic functions of macrophages. This process appears to be mediated by PAC1-R in rat peritoneal macrophages via phosphokinase C stimulation (532). Recently, the presence of PAC1-R in these cells was confirmed (527). At present, the role of VPAC receptor types remains unclear, although VPAC1-R expression has been confirmed in rat peritoneal macrophages (524, 526). In mice, PACAP is also involved in increasing phagocytosis in peritoneal macrophages, probably through a PAC1-R variant (533).
PACAP (10-13 to 10-6 M) increases adherence of both macrophages and lymphocytes. In addition, it increases motility of macrophages while decreasing the motility of lymphocytes. The data suggest these processes are regulated by PAC1-R in macrophages via PLC activation and by VPAC-R in lymphocytes via AC activation (534, 535).
3. Immune response. PACAP is reported to both suppress and activate inflammation through regulation of certain interleukins (IL-6 and IL-10). The PACAP-induced inhibition of IL-10 production by lymphocytes is likely mediated via cAMP. PACAP appears to regulate transcriptional expression of IL-10. In turn, IL-10 suppresses the action of certain cytokines that are involved in the local immune response (536). PACAP can inhibit IL-6 production from stimulated macrophages through the PAC1-R (537). This action of PACAP suppresses inflammation. In contrast, PACAP also has proinflammatory effects. In unstimulated macrophages PACAP acts through VPAC1-R and PAC1-R to up-regulate IL-6 transcription and release (538). In addition, PACAP stimulates IL-6 production in rat astrocytes (539). Thus, it appears that PACAP both initiates and inhibits the inflammatory response. The findings suggest that PACAP may be involved in immune system homeostasis in the absence of stimulation. However, in the presence of intense stimulation or toxicity, the PACAP-regulated inhibition of IL-6 transcription may help protect the tissue from excessive IL-6 release to reduce inflammation or shock (537, 538). Finally, PACAP induces extravasation in rat skin by a cAMP-independent mechanism. Part of the extravasation response may be mediated by PACAP-induced histamine release from mast cells in the skin, which also contributes to inflammation (540).
In the brain PACAP may modulate the immune response through regulation of outward potassium currents in microglial cells. The regulation of ion channels is a means by which PACAP may modulate microglia activity (541).
D. A regulator of bone metabolism
Traditionally, PACAP has not been thought of as a hormone that has
direct effects on the skeletal system. Any effect of PACAP on bone and
cartilage has been considered indirect through the release of GH and
subsequently IGF-II, which is involved in skeletal growth. However,
recent evidence suggests PACAP has a direct modulatory role in bone
tissue. PACAP has been identified by immunocytochemistry in nerve
fibers innervating the cartilage canals of newborn pigs. Receptors for
PACAP can be found in human osteoblast and osteosarcoma cells
(VPAC1-R), rabbit stromal cells
(VPAC2-R), and rat bone marrow-derived stromal
cells (VPAC2-R >> PAC1-R
or VPAC1-R mRNA) (393, 542, 543, 544). In osteoblasts
and osteoblast-enriched bone cultures, PACAP stimulates cAMP
accumulation (545, 546). Several of these studies suggest PACAP may
play a role in bone formation (542), bone resorption (544), and
hematopoeisis through stimulation of IL-6 production in bone
marrow-derived stromal cells (393). Although these findings are not
conclusive, it seems that PACAP functions in some capacity as a
neuroendocrine or paracrine regulator of bone metabolism.
E. An endocrine/paracrine regulator
1. Anterior pituitary secretions. PACAP was discovered in
ovine hypothalamic extracts because it was able to increase cAMP in
static rat anterior pituitary cell cultures and cause the release of
GH, PRL, ACTH, and LH from superfused rat pituitary cells (6, 10). This
original work paved the way for numerous studies into the role of PACAP
as a hypophysiotropic factor. Several criteria need to be met for a
substance to be classified as a hypophysiotropic factor. These criteria
are discussed in detail in an review by Rawlings and Hezareh
(378) on PACAP’s actions on the anterior pituitary. In summary, the
substance must be 1) present in neurons in the hypothalamus that
project to the hypophysial portal blood system, 2) present in the
hypophysial portal blood, 3) able to interact with receptors on cells
of the anterior pituitary, and 4) able to regulate anterior pituitary
cell function. Ironically, sheep, the animal in which PACAP was
discovered, appear not to use PACAP as a hypophysiotropic hormone.
Unlike rats and humans, infusion of PACAP into sheep has no effect on
the release of anterior pituitary hormones (547, 548, 549). However, PACAP
is a hypophysiotropic factor in many other vertebrates.
First, PACAP is present in the hypothalamus. PACAP immunoreactivity has been demonstrated in the hypothalamus of primates (including humans), sheep, rats, frogs, and fish (346, 144, 143, 356, 432, 433, 549, 550, 551, 552, 553, 554). In humans and spider monkeys, the immunoreactive fibers were localized in the supraoptic and the paraventricular nuclei, the periventricular region, and preoptic area. Both the supraoptic and paraventricular nuclei send fibers to hypophysial portal blood vessels. In monkeys PACAP-ir fibers were found in the external zone of the tuber cinereum around hypophysial portal capillaries; this region is close to the transition of the pituitary stalk (median eminence). A similar distribution of PACAP immunoreactivity has been observed in the rat, frog, and fish (144, 346, 551, 554). PACAP is also detectable as mRNA in human and rat hypothalami (348). Together, these results give evidence that PACAP is present in the hypothalami of several vertebrates, with a distribution suitable for a hypophysiotropic factor.
Second, PACAP has been detected in the portal blood of rats at a concentration significantly higher than in the peripheral blood (555). The presence of PACAP in the portal blood is a crucial piece of evidence supporting the role of PACAP as a hypophysiotropic factor.
Third, PACAP interacts with the anterior pituitary. Immunoreactive PACAP is found in the human, rat, amphibian, and fish anterior pituitary (21, 144, 348, 556, 557). mRNA studies reveal it is not synthesized in the anterior pituitary of humans; the presence of it here may be the result of transport via nerve terminals and portal blood from the hypothalamus (348). In fish, PACAP-containing nerve terminals grow into the anterior pituitary from the brain. Several immunoassay and mRNA studies have identified PACAP receptors in rat anterior pituitary cells or whole pituitary extracts (191, 394, 372, 381, 382, 459, 558, 559). PACAP receptors have also been identified in human pituitary adenomas (gonadotropin-producing, GH-producing and ACTH-producing adenomas) by binding assays (560). The most frequently identified receptor type in the anterior pituitary appears to be the PAC1-R (381, 382, 559, 560). In frog pituitaries a PACAP-preferring receptor has been proposed that is pharmacologically different from mammalian PAC1 receptors (63), and recently PAC1 receptors have been cloned in chicken, frog, and goldfish pituitaries (143, 206, 207). Other studies also detect the presence of VPAC2-R and VPAC1-R (weakly) (394). The abundance of the receptor subtype mRNA in normal rat anterior pituitary cells is as follows: PAC1-r = VPAC2-R and VPAC1-R is expressed weakly (394). Although all the splice variant forms of the PAC1-R may be present in the anterior pituitary of rats, the dominant forms are PAC1-R-s and PAC1-R-hop (394).
Fourth, PACAP regulates anterior pituitary function. The rat anterior pituitary releases several pituitary hormones into the blood in response to intravenous PACAP administration (561). Each pituitary cell type reacts to PACAP differently. In mammals there has been considerable debate as to whether PACAP is a modulator of pituitary hormone release or a true hypophysiotropic hormone. The picture should be clarified eventually in knockout mice. In the meantime, several studies are outlined below that summarize what is known about PACAP’s actions on anterior pituitary cells.
a. Gonadotrope secretions.
PACAP affects LH and FSH
synthesis/release on two levels. It can modulate the actions of GnRH at
the level of the hypothalamus, or it can act alone or in concert with
GnRH to modulate gonadotropes in the pituitary. A recent study on rats
has localized PACAP to LH- and FSH-producing cells in the anterior
pituitary (557). This suggests PACAP may also have some paracrine or
autocrine effects on anterior pituitary cells. The following paragraphs
will discuss the actions of PACAP on gonadotrope function both alone
and in conjunction with GnRH exposure in the anterior pituitary. The
actions of PACAP on GnRH function in the hypothalamus are discussed in
Section V.E.3.
Many of the experiments that have attempted to assess gonadotrope
response to PACAP have used the T3–1 clonal gonadotrope cell line.
It is clear that exposure to PACAP has an effect on these cells, but it
is important to note that this cell line is immortalized and may not
reflect normal cell function (400). The T3–1 cell line was
generated by targeted oncogenesis in mice. Like normal gonadotropes,
these cells express the -subunit gene and synthesize and secrete the
-subunit protein, making them a good model. However, they do not
express the ß-subunit gene and therefore cannot synthesize LH or FSH
(562).
PACAP regulates an increase in the release of gonadotropins and the
-subunit. At physiological concentrations, continuous PACAP exposure
stimulates the release of LH and FSH from normal rat gonadotropes (10, 400, 563, 564, 565, 566, 567). After the initial stimulation, secretion of both LH and
FSH drops. However, continued PACAP exposure causes a steady increase
in LH release back to maximum levels, whereas FSH secretion gradually
decreases (564). In addition, PACAP triggers the release of the
-subunit protein from normal rat gonadotropes (564, 565, 567). The
-subunit protein is also released from the T3–1 clonal
gonadotrope cell line in response to PACAP exposure (400, 568). If
PACAP is administered in a pulsatile fashion, it likewise increases
gonadotropin secretion in normal gonadotropes (565).
In addition to its releasing powers, PACAP also elicits an increase in
gonadotropin subunit transcription. Continuous exposure to PACAP causes
1) an increase in LHß half-life by increasing transcript length; 2)
an increase in -subunit transcription; and 3) a decrease in FSHß
subunit transcription (400, 564, 568). These results correlate well
with the measures of LH and FSH release noted above. Pulsatile delivery
of PACAP increases the level of LHß mRNA and -subunit mRNA but has
no effect on FSHß mRNA. The native pattern of exposure is unknown but
is thought to be continuous (565). PACAP regulation of gonadotropin
gene transcription appears to be regulated via the cAMP/PKA pathway
(568, 569), but the exact mechanism by which PACAP affects LHß and
-subunit gene transcription is not clear. However, the decrease in
FSHß mRNA appears to be correlated with a PACAP-induced increase in
follistatin mRNA. Follistatin is produced by a subset of gonadotropes
and folliculo-stellate cells. It is thought that PACAP (via cAMP/PKA)
stimulates follistatin transcription, which, in turn, neutralizes
activin activity. Once neutralized, activin is no longer able to
increase FSHß mRNA and the level drops (570). Therefore, the action
of PACAP on FSHß mRNA transcription is likely indirectly mediated
through PACAP’s stimulation of follistatin gene transcription.
Long-term exposure to PACAP results in down-regulation of PAC1-R concentration in the cell membrane and homologous desensitization of the AC and PLC pathways (564, 571). PACAP dissociation from its receptor is slow, and gonadotropes become desensitized to continuous PACAP exposure through changes in receptor-effector coupling and PAC1-R down-regulation at the cell surface (571). The main effector of gonadotropin regulation is GnRH, but PACAP and GnRH synergistically increase gonadotropin release and regulate transcription differently and more efficiently than either hormone alone (563, 564).
PACAP is thought to regulate gonadotropin release through changes in
cystolic calcium (394, 566, 572, 573, 574). Due to the presence of two PACAP
receptor types (PAC1-R and
VPAC2-R) in gonadotropes, PACAP can potentially
regulate several intracellular signaling mechanisms (394). In rat
gonadotropes, PACAP alters gonadotropin mRNA levels as well as LH, FSH,
and -subunit release, probably via the PAC1-R
(564, 568). Gene transcription is mediated by the cAMP/PKA pathway.
However, cAMP has no effect on the observed
PAC1-R-mediated changes in
Ca++ concentration (572, 573, 575). Typically,
Ca++ is an important molecule involved in
regulation of exocytosis. It is thought that, as is the case with
GnRH-induced gonadotropin secretion, the actions of PACAP on PLC and
cystolic calcium are involved in exocytosis of the gonadotropins (576).
Because PACAP and GnRH both affect gonadotropin secretion and
production and, in some cases, use the same intracellular pathways,
there is the potential for cross-talk and coincidence signaling (571, 577, 578). Together, continuous PACAP infusion in the presence of genre
pulses stimulates LH and FSH release, increases -subunit and FSHß
mRNA, and lengthens LHß mRNA transcripts (564). In addition, in
humans, infusion of PACAP causes no effect on gonadotropin release,
even when administered along with GnRH pulses (579, 580). These
differences between rats and humans may reflect species differences or
experimental conditions. In addition, PACAP may increase proliferation
of T3–1 cells (400). Further research is required to clarify these
results.
b. Somatotrope secretions.
Somatotropes are GH-producing cells
found in the anterior pituitary. In mammals the primary releaser of GH
is GRF, but PACAP also appears to have a role in GH release. PACAP
(10-11 to 10- 8
M) increases Ca++ concentrations in
rat anterior pituitary cell cultures in a dose-dependent manner, and
some of the cells experiencing a rise in Ca++
concentration have been identified as somatotropes (531). PACAP also
stimulates cAMP accumulation in rat static pituitary cell cultures
(581). Several researchers have shown that PACAP induces GH release and
synthesis (567, 581, 582, 583, 584, 585, 586). The proposed mechanism is that PACAP
increases Ca++ levels via a cAMP/PKA-mediated
pathway (572, 587) and that the release of GH by PACAP does not require
PKC activation. This points to the involvement of cAMP in the process
of GH release and not IP (581). Also, it has been suggested that PACAP
activates a sodium channel via an AC/PKA pathway (587). The result is
membrane depolarization and, in turn, calcium channel activation that
triggers the increase in cytosolic calcium necessary for GH release.
PACAP stimulates GH release from GH3 and
GH4C1 cells through a
VPAC2-R (394, 588). However, in a rat
GH-producing enriched cell population, the only PACAP receptor
expressed, as determined by RT-PCR, is PAC1-R-hop
(459). More work is required to determine whether this is the only
receptor that PACAP activates to cause GH release. In addition to
stimulating GH release, PACAP induces an increase in rat GH gene
transcription in static cell culture (582). The effects of PACAP on GH
may involve increasing the number of GH-secreting cells (583). It
should be noted that PACAP infusion into human males does not cause any
increase in serum GH levels (580, 588). Although the actions of PACAP
on GH release in mammals have not been fully elucidated, PACAP
functions have been studied across vertebrates and an evolutionary
perspective is clear. In fish, at physiologically relevant
concentrations, PACAP stimulates GH release from cultured pituitary
cells (19, 143, 144). Therefore, it appears that, at least in some
nonmammals, PACAP is a hypophysiotropic releaser of GH.
c. Lactotrope secretions.
Lactotropes are PRL-producing cells
in the anterior pituitary. PRL-producing enriched cells from rat
anterior pituitaries express both PAC1-R and
VPAC2-R receptors. There are three
PAC1-R splice variants expressed in these cells:
PAC1-R-s, PAC1-R-hop, and
PAC1-R-hiphop (weakly expressed) (459). The
action of PACAP on PRL release is not yet clear. PACAP increases cAMP
concentrations in a dose-dependent manner and also increases PRL mRNA
in GH3 cells and in total pituitary cell
suspensions by direct action on the PRL promoter (582, 589). The
effects of PACAP on PRL gene transcription may be controlled by a
cAMP-independent pathway (589). In contrast, the action of PACAP on PRL
release is not yet clear. In vitro experiments on dispersed
anterior pituitary primary cell cultures indicate that PACAP
(10-7 and 10-6
M) inhibits PRL release in a dose-dependent
manner (584, 586). These results are obviously in conflict with the
actions of PACAP on gene transcription. A theory explaining this
paradox suggests that PACAP may trigger the secretion of a paracrine
factor that stimulates PRL release (584). In intact pituitaries the
paracrine factor is more potent than PACAP and overrides PACAP’s
inhibitory action on PRL. In cell culture the paracrine factor is
diluted by the medium and is unable to override the inhibitory effects
of PACAP. These results have since been confirmed and the proposed
paracrine factor is IL-6 (590). In vivo experiments in
humans show that infusion of high doses of PACAP (4–10 pmol/kg/min)
results in an increase in plasma PRL. This effect appears to be
mediated by both VIP and PACAP through a VPAC-R (580, 588).
d. Corticotrope secretions.
Corticotrope cells are responsible
for the release of several POMC hormones, including ACTH. PACAP
(0.1–100 nmol) causes an increase of ACTH release from dispersed rat
pituitary primary cell cultures. The release of ACTH is only
significant after a 24-h time delay. The time delay between cAMP
stimulation and ACTH release suggests other factors are involved in
ACTH secretion, perhaps IL-6 (567). Also, PACAP induces a
dose-dependent secretion of ACTH from an AtT-20 pituitary cell line and
in human males infused with PACAP (580, 586, 591). The effect of PACAP
in humans appears to be mediated by PACAP-specific receptors (not
VPAC-R) (580). In contrast, in the mouse corticotrope tumor cell line
(AtT-20), PACAP was found to stimulate POMC gene transcription, cAMP
accumulation, and cellular differentiation through the
VPAC2-R (404, 591). These receptor differences
may reflect changes between normal and tumorous cells or among species.
PACAP can act alone or in combination with hypothalamic CRH to regulate
POMC transcription. CRH activity on POMC is PKA dependent. In contrast,
PACAP acts in a PKA-independent manner. These results suggest that POMC
gene transcription can be induced by multiple signaling pathways, one
of which is regulated by PACAP (591).
e. Folliculo-stellate secretions.
Folliculo-stellate cells are
agranular cells that differ substantially from the other pituitary
glandular cells. Folliculo-stellate cells have long processes and
phagocytic ability. The majority of PACAP receptors in the anterior
pituitary are found on folliculo-stellate cells (558). PACAP is
suspected to cause IL-6 release from these cells through AC stimulation
(530).
2. Intermediate pituitary secretions.
a. Melanotrope secretions.
Melanotropes, located in the pars
intermedia, are another cell type that transcribe the POMC gene.
-MSH production is associated with POMC transcription in these
cells. In the rat, PACAP-like immunoreactivity is exhibited by cells
located in the pars intermedia of the pituitary (352). Also, cells of
the pars intermedia express PACAP and PAC1-R mRNA
(191, 592, 593). Mouse melanotropes express two isoforms of the
PAC1-R (PAC1-R-s and
PAC1-R-hop), and PACAP stimulates both cAMP and
IP accumulation in melanotropes (592). In melanotropes, PACAP and CRH
were additive in increasing cAMP levels. PACAP has also been reported
to stimulate -MSH secretion in rats and mice (592, 593, 594). In rats,
the signal transduction mechanism triggered by PACAP is PKA activation
of voltage-dependent Ca++ channels and PKC
activation of nonselective cation channels (593). The expression of
both PACAP and its receptor in the pars intermedia suggests that PACAP
may be an autocrine/paracrine regulator of melanotrope secretion.
3. Hypothalamic secretions. The presence of PACAP receptors in the hypothalamus suggests the existence of paracrine regulation and/or the possibility of a feedback loop. In hypophysectomized rats the rate of PACAP gene transcription and hormone levels decreased 1 to 2 weeks after hypophysectomy. Replacement therapy with GH, PRL, T4 , and corticosterone significantly restored these levels, suggesting a feedback loop between pituitary hormones or pituitary-dependent factors and PACAP gene transcription (595). Regarding paracrine regulation, the GT1–7 neuronal cell line (an immortalized hypothalamic cell line) expresses PAC1-R splice variants and VPAC2-R mRNA. Both VIP and PACAP stimulate cAMP in these cells, but neither affect IP turnover. Likewise, in the chick hypothalamus, PACAP stimulates cAMP accumulation, probably through the PAC1-R (596). In addition, PACAP and VIP stimulate GnRH release from these cells, probably by VPAC2-R action on cAMP (597). Intracerebroventricular injections of PACAP into rats cause an increase in GnRH and somatostatin mRNA levels. However, intravenous administration causes a decrease in GnRH mRNA levels and no change in somatostatin mRNA expression. Although not conclusive, these results suggest that PACAP may be a neurotransmitter/neuromodulator of hypothalamic hormone secretion (598).
PACAP also causes the release of AVP in human males and may have the same effect in rats (479, 580). This action of PACAP may be involved in the control of osmolarity and blood pressure. Recently, PACAP has been localized to nerve terminals in the hypothalamus that innervate the AVP-containing neurons of the hypothalamic supraoptic nucleus. These AVP-rich neurons also express the PAC1-R. PACAP acts on AVP-containing neurons to increase cytosolic calcium via a cAMP/PKA pathway. This influx of calcium may trigger AVP secretion (599).
4. Gonadal secretions. PACAP mRNA is found in the developing germ cells of the rat testis (367, 414, 416). Both the PAC1-R and VPAC-R are expressed in the rat testis (191, 381). In situ hybridization and Northern analysis reveal that PACAP mRNA is present in round spermatids in stages III–VII, spermatogonia, and primary spermatocytes. PACAP was not found in mature spermatids, testicular spermatozoa, or epididymal spermatozoa (367, 414). PACAP’s action in these cells is mediated through a VPAC2-R (600). PACAP can regulate synthesis of secreted and intracellular proteins by spermatids and spermatocytes in vitro. However, the effects of PACAP on these stages of maturing germ cells differ. In spermatocytes PACAP increases production of certain germ cell-secreted proteins, whereas in spermatids PACAP inhibits production of the proteins (601).
Neither PACAP nor PACAP receptors have been found on Sertoli or Leydig cells (367, 602). However, it has been suggested that PACAP can stimulate cAMP accumulation and secretion of lactate, estradiol, and inhibin in Sertoli cells (603). In addition, PACAP expression in the rat testis may be positively regulated, at least in part, by FSH released from the pituitary. Pituitary gonadotropes, such as FSH, directly affect other cells in the testis (Sertoli and Leydig cells), but not the germ cells. However, by affecting PACAP expression, FSH may have an indirect effect on germ cell function (595).
Although the receptor has not been identified, PACAP causes a dose-dependent secretion of testosterone from rat Leydig cells. This effect is independent of Ca++ and cAMP but is dependent on an influx of Na+ from the extracellular medium. These results suggest that PACAP may exert its effect through stimulation of a new PACAP receptor subtype that is Na+-dependent and cAMP- and Ca++-independent (604). A different study suggests that the cAMP-independent actions of PACAP on testosterone are mediated by a known PAC1-R variant. Also, this study determined that the presence of human CG is required for testosterone production (603).
In the female, PACAP causes a dose-dependent (10-11 to 10-10 M) stimulation of cAMP accumulation, estradiol secretion, and progesterone secretion from cultured rat ovarian granulosa cells. Also, PACAP may be involved in the induction of LH responsiveness in granulosa cells. These actions may be mediated by the PAC1-R (605, 606). Both PACAP and PAC1-R mRNA transcripts are found in the rat ovary, specifically granulosa and luteal cells (415). The splice variants found are PAC1-R-s, PAC1-R-hip, or -hop (the hip and hop variants are not distinguishable by the methods employed in this experiment), and PAC1-R-hiphop. These expression data combined with the aforementioned study suggest that PACAP may be an autocrine and/or paracrine regulator of ovarian function (415, 607).
5. Adrenal gland secretions. PACAP and PACAP-binding sites have been detected in the adrenal gland of mammals, amphibians, and fish by immunoreactive methods, binding assays, and mRNA detection. In particular, chromaffin cells of the adrenal medulla are innervated by PACAP-ir nerve fibers (191, 348, 356, 365, 366, 381, 382, 407, 608, 609). In cultured chromaffin cells (PC12), PACAP increases cAMP production (608, 610) and inhibits proliferation (407). PACAP also increases cell survival but does not significantly promote neurite outgrowth. PACAP may be involved in overriding growth factor mitogenic signals and maintaining the postmitotic state of chromaffin cells. During development (P8–P12) in the rat, innervation of the adrenal gland with preganglionic sympathetic cholinergic fibers occurs at about the same time as a decline in cell division (407). Not only is PACAP expressed in fibers innervating the adrenal gland, but after sympathectomy PACAP is expressed in some adrenal chromaffin cells (611). These results indicate that PACAP expression may be suppressed under normal conditions.
In the adrenal gland, PACAP is involved in the secretion of a catecholamine (adrenalin) from chromaffin cells of the adrenal medulla. The PACAP-induced secretion of adrenalin modifies cardiac function (477). Also, release of catecholamines and leu-enkephalin from chromaffin cells is stimulated by PACAP, likely through a PAC1-R (610, 612, 613, 614). The PACAP-induced release of catecholamines is Ca++ dependent (608).
Not only is PACAP involved in catecholamine release, it also causes an increase in the mRNA for tyrosine hydroxylase in a neuron-derived cell line (CATH.a) and in porcine chromaffin cell cultures. In addition, the porcine chromaffin cultures increase dopamine ß-hydroxylase gene expression in response to PACAP. Both tyrosine hydroxylase and dopamine ß-hydroxylase are involved in catecholamine synthesis. PACAP activates tyrosine hydroxylase and dopamine ß-hydroxylase transcription through cAMP accumulation and activation of PKA (613, 615, 616).
Not only does PACAP act on the chromaffin cells of the adrenal medulla, it is also active in the cortex. Another adrenal gland hormone, aldosterone, is synthesized by the adrenal cortex. PACAP stimulates aldosterone secretion directly through VPAC receptors in human tumorous adrenal cells, via PAC1-R-hop or PAC1-R-s in normal bovine adrenal cells and in unidentified PACAP-preferring receptors in frog adrenal cells (366, 461). PACAP-stimulated aldosterone secretion is inhibited by atrial natriuretic peptide (ANP) in both bovine and human cells. However, ANP acts via a different mechanism in the two cell types. This suggests that there may be species-specific factors involved in aldosterone production as regulated by PACAP (461). In addition, PACAP stimulates the renewal of intracellular pools of the neuropeptides, ANP and brain natriuretic peptide. PACAP is probably involved in the biosynthesis of these peptides via the PAC1-R. The process appears to involve the intracellular messengers, PKA and PKC (612). Another study suggests that PACAP acts indirectly on aldosterone secretion. PACAP-induced catecholamine release stimulates aldosterone synthesis and secretion in humans by enhancing aldosterone synthase activity (614). In frogs, PACAP induces an elevation in cAMP and Ca++ in adrenal slices. A PACAP-preferring receptor is present on both chromaffin and adrenocortical cells. In addition to aldosterone, the adrenocortical cells also secrete corticosteroid in response to PACAP (10-8-10-5 M) (609).
6. Pancreatic secretions. The two main hormones involved in the regulation of blood glucose are insulin and glucagon. Surprisingly, PACAP regulates the secretion of both hormones. In fact, PACAP is a more potent releaser of insulin than glucagon (362, 368). In mammals, PACAP-like immunoreactivity is found in pancreatic nerve fibers, islets, and capillaries and PACAP mRNA has been localized to ß-islet cells (362, 368). In fish the insulin and glucagon cells are innervated by PACAP-containing fibers (363). In vitro experiments in the isolated perfused rat pancreas show that PACAP releases insulin in a glucose-dependent manner. When the pancreas is perfused with 8.3 mM glucose, PACAP (10-11 to 10-10 M) elicits a concentration-dependent biphasic release of insulin. However, in the presence of nonstimulating glucose levels (2.8 mM), no change in basal insulin secretion is observed (362, 491, 617, 618). In contrast, glucagon secretion is stimulated by PACAP (10-10 to 10-8 M) in the presence of nonstimulating glucose levels (2.8 mM) but not in the presence of stimulating glucose levels (8.3 mM) (491). One proposal is that the effect of PACAP on insulin and glucagon secretion is mediated by a VPAC receptor (491) and involves PKA and PKC, but not Ca++ (617). However, recent findings suggest otherwise. In fact, the PAC1-R-TM4 variant is the only PACAP-specific receptor expressed in the ß-islet cells of the pancreas (386). PACAP is known to stimulate insulin secretion from ß-islet cells through activation of L-type Ca++ channels (362, 368). The PAC1-R-TM4 is not coupled to a G protein and does not act on AC or PLC, but it is coupled to L-type Ca++ channels (386). Thus, PACAP appears to be a potent releaser of glucose-induced insulin secretion from ß-islet cells through activation of the PAC1-R-TM4. The action of PACAP on insulin has been confirmed in a knockout mouse in which the PAC1-R was disrupted. The PAC1-R null mice had reduced glucose-stimulated insulin secretion in vivo and in vitro (345).
In contrast to rats, an intravenous PACAP infusion (100 pmol/kg/min) in sheep produced no effect on insulin or glucagon plasma concentration whether the system was glucose stimulated or not. This may be due to species differences (619).
7. Cardiac secretions. The heart is not traditionally thought of as an endocrine or paracrine organ but it does produce and secrete hormones. In cultured rat neonatal cardiomyocytes PACAP is involved in the regulation of one of the hormones secreted by the heart, ANP. Incubation with PACAP causes a dose-dependent increase in cAMP content in cardiac myocytes and a concomitant dose-dependent increase in ANP content (620).
8. Gastrointestinal secretions. The enterochromaffin-like cells of the gastrointestinal tract secrete histamine and pancreastatin in response to gastrin. However, PACAP has been shown to stimulate histamine and pancreastatin secretions from enterochromaffin-like cells with a greater potency than gastrin (621). The distribution of PACAP in the gastrointestinal tract was discussed in Section V.B.2. In addition, PACAP-ir endocrine cells are present in the chicken gut. This provides further evidence that PACAP may be a gut hormone or paracrine factor (357).
F. An exocrine regulator
1. Gastrointestinal secretions. PACAP-ir nerve fibers
innervate the gut wall of humans, pigs, sheep, cats, ferrets, guinea
pigs, hamsters, rats, mice, and chickens (355, 357, 505). The following
are examples of PACAP regulating exocrine secretions in the gut.
Intravenous exposure to PACAP inhibits gastric acid secretions from the
parietal cells. PACAP also inhibits both histamine-stimulated and
pentagastrin-stimulated gastric acid secretions (622, 507). Although
VIP is also active in this system, PACAP-27, but not VIP or PACAP-38,
inhibits pentagastrin-stimulated gastric acid secretion (507). The
receptors active in these responses have not been determined. Also,
PACAP causes an increase in pepsinogen release from chief cells via a
VPAC receptor (623).
2. Pancreatic secretions. PACAP not only increases blood flow in the exocrine pancreas and regulates insulin and glucagon but also regulates exocrine function. VPAC2-R binding sites have been identified in the pancreatic acini of rats (624). PACAP and VIP have been shown to increase acinar lipase release, amylase release, and cAMP accumulation (624, 625). In sheep, pancreatic juice flow and protein secretions increase after PACAP or VIP (1, 3, 10 pmol/kg/min) infusion. Bicarbonate secretion also increases due to PACAP or VIP (1, 3, 10 pmol/kg/min) infusion. The effect of PACAP and VIP on bicarbonate secretion may result from a direct action of the hormones on the duct. PACAP (1, 3, 10 pmol/kg/min), but not VIP, increases amylase output in a dose-dependent manner (619). PACAP’s action on protein and amylase secretion may be indirectly mediated via cholinergic nerve stimulation (619). Similar results were observed in the dog exocrine pancreas after intravenous injection of PACAP (476).
PACAP stimulates calcium signaling in AR-4–2J cells, a rat acinar-like pancreatic cell line. The calcium signaling has two phases. First, an initial peak is dependent on an influx of external calcium and mobilization of internal calcium. Second, a plateau phase is dependent on influx of external calcium only. The peak is PLC dependent. The above effects are evoked at levels that stimulate amylase release (626). Perhaps the unique actions of PACAP that regulate amylase output are mediated by PAC1 receptors. This would account for the inability of VIP to elicit a change in amylase secretion. It should be noted that, whereas PAC1 receptors dominate in AR-4–2J cells, the VPAC2-R dominates in normal pancreatic acini (624).
3. Seromucous secretions. PACAP also appears to affect mucous secretions in various tissues. In several mammals PACAP and VIP were coexpressed in nerve fibers observed around seromucous glands in the nose, trachea, and lungs. This suggests that along with VIP, PACAP may have a role in mucous secretions (353, 354). Also, in the human vagina, PACAP and VIP were coexpressed in fibers running parallel to the mucosal epithelium. The distribution of the fibers suggests that these hormones may play a role in lubrication of the vagina during sexual stimulation (488).
G. A regulator in the nervous system
1. Circadian rhythm. The suprachiasmatic nucleus (SCN) is
located in the hypothalamus and acts as the primary pacemaker in
mammalian systems. The SCN is responsible for generating circadian
rhythms. The actions of the SCN are linked to environmental cues such
as light/dark cycles. Light/dark cues are transmitted from the eyes
back to the hypothalamus via the retinohypothalamic tract. The SCN
regulates synthesis and release of melatonin by the pineal gland and
melatonin feeds back to the SCN (627). PACAP immunoreactivity has been
observed in retinal ganglion cells that send axons to the SCN where
PACAP immunoreactivity has also been observed (628). Through its action
on the SCN, PACAP may stimulate melatonin synthesis and secretion in
the pineal gland. In the SCN, PACAP regulates phosphorylation of CREB,
likely through the cAMP intracellular signaling system. In turn, the
CREB transcription factor stimulates melatonin synthesis. Melatonin
appears to feed back and inhibit PACAP- induced phosphorylation of CREB
(629, 630, 631). The level of PACAP in retinal ganglion cells is low in
the day and high at night. Thus, the presence and location of PACAP
suggest it is a neurotransmitter/neuromodulator of the
retinohypothalamic tract (627). Experiments show that PACAP acts as a
neurotransmitter/neuromodulator late in the day (628, 629). The phase
of the circadian rhythm can also be advanced by PACAP during the day
but not at night. The phase shift is mediated through the
PAC1-R (628, 629). In a recent publication, PACAP
and VIP were shown to directly stimulate melatonin synthesis in the rat
pineal gland through the VPAC2-R (632).
2. Autonomic and sensory nervous systems. The neurotransmitter/neuromodulator actions of PACAP have already been discussed regarding smooth muscle regulation, hormonal secretions, exocrine secretions, and circadian rhythm. Additionally, PACAP appears to be involved in regulation of autonomic functions associated with the medulla oblongata, otic ganglia, sphenopalatine ganglia, and the jugular nodose ganglia (349, 633). The involvement of PACAP in sensory neurotransmission has been suggested due to PACAP or PACAP receptor presence in the ciliary ganglion neurons, jugular nodose ganglia, dorsal horn neurons, spinal cord, mesencephalic trigeminal nucleus, and the vagus nerve (633, 634, 635, 636, 637, 638). In addition, PACAP has been noted to have a role in synapse formation and neurotransmission in the limbic system (639, 640). PACAP is involved in neurotransmission in several sympathetic neurons (197, 485, 641).
3. Behavior. A limited number of studies have examined the effect of PACAP on behavior. Distribution of PACAP and VIP immunoreactive fibers in the rat forebrain and the formation of synapses with CRF-ir neurons have led to the suggestion that PACAP may have a stress-related function (640). PACAP may also be involved in behavior associated with increased grooming, pain responses, reduced food intake, motor activity, and body temperature (642, 643, 644, 645, 646, 647). Finally, PACAP may enhance rapid eye movement (REM) sleep. This action of PACAP on REM may be mediated, in part, by PACAP-induced PRL release from the hypothalamus. PRL is also known to stimulate REM sleep (648, 649).
|
VI. Conclusions and Future Directions |
|---|
|
TopAbstractI. IntroductionII. Evolution of the...III. Superfamily Evolution by...IV. Superfamily...V. Conservation of PACAP—A...VI. Conclusions and Future...References |
|---|
The gene with the shortest known evolution is GIP, which has not been identified in birds or in other animals that evolved before mammals. The sequence of GIP is quite distinct from other families and hence GIP may be more difficult to detect and purify. Also, secretin has not been identified in reptiles, amphibians, or fish, but extracts of the gut in these animals have secretin bioactivity, suggesting that the peptide exists. Other new hormones may exist, but have not yet been identified. In addition, alternative splicing has greatly extended this superfamily of peptides and receptors.
In general, the nine hormones in this superfamily are related in terms of distribution and function. All members of the superfamily are present in the brain (except GIP) and gut, including the pancreas. The majority of family members function in reproductive organs. In addition, all known hormones in this superfamily elicit their actions through binding to a subset of seven-transmembrane receptors. However, some family members share a receptor: PHM/I does not have an identified receptor and probably uses the VIP receptors; and VIP has two receptors that it shares with PACAP. PACAP additionally has its own receptor. The hypothesis that PACAP is the ancestral molecule, or is directly in line with the ancestral molecule, in the PACAP/glucagon superfamily is bolstered by functional evidence showing that PACAP is pivotal in a functional sense in all organs or tissues in which this superfamily is expressed.
The array of effects triggered by PACAP is intriguing. The effects span almost all systems in the body including nervous, endocrine, muscular, bone, and immune. Recent work showing the three types of PACAP receptors, their variants, and wide distribution convince us that PACAP could trigger many functions through several signaling paths. However, these functions are strange in that they do not fit together easily. We know neither the signal that triggers PACAP release nor the subset of PACAP functions that are coordinated to respond to the signal. In addition, research that examines the evolution of PACAP’s functions is meager except in relation to the endocrine system. Thus, the functions of GRF and PACAP in a protochordate, which lacks a pituitary and closed circulatory system, are not clear.
In the near future, there are three areas that seem to be important. First, more structures of superfamily members are needed to show the origin of secretin, VIP, and GIP. The abbreviated story of these peptides may reflect a recent origin due to a gene duplication from one of the other family members or merely technical problems in isolating the peptides or cDNAs. Certainly, more structures of superfamily members in invertebrates will help clarify the origin of each peptide family.
Second, studies on the evolution of function may help identify basic functions of each peptide. If more than one peptide is encoded in a single gene due to exon duplication, then closely related functions of the peptides may be evident in animals closer to the duplication event.
Third, the event that triggers PACAP release needs to be identified. The generation of knockout mice in which individual genes encoding superfamily members or their receptors are disrupted is crucial. Only then can we tease apart functional overlap, at least for mouse as a representative of mammals.
 |
 |
|
Footnotes |
|---|
1 Supported by the Medical Research Council of Canada. 
2 Present address: Biotechnology Laboratory, University of British
Columbia, Vancouver, British Columbia V6T 1Z2, Canada
|
References |
|---|
|
TopAbstractI. IntroductionII. Evolution of the...III. Superfamily Evolution by...IV. Superfamily...V. Conservation of PACAP—A...VI. Conclusions and Future...References |
|---|
cells. Neuron 1:605–613[CrossRef][Medline]
cell-specific expression. J Biol Chem 270:3046–3055
affects glucose homeostasis and islet glucagon gene expression in
vivo. Genes Dev 13:495–504-factor receptor of Saccharomyces cerevisiae. J Biol
Chem 263:10836–10842 cells. Regul Pept 32:65–73[CrossRef][Medline]
2-adrenoreceptors and potassium channels. Br J
Pharmacol 121:1605–1612[CrossRef][Medline]
T3–1 cell line. Endocrinology 134:315–323
