This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Clark, R. Search for Related Content PubMed PubMed Citation Articles by Clark, R.

The Somatogenic Hormones and Insulin-Like Growth Factor-1: Stimulators of Lymphopoiesis and Immune Function

Endocrinology Group, Genentech, Inc., South San Francisco, California 94080

	Abstract

Top Abstract I. Introduction II. Local GH Axis... III. GH Administration IV. PRL V. Insulin-Like Growth Factors VI. Actions of IGF-I VII. IGF-I in Different... VIII. GH/IGF-I as Antistress... IX. Therapeutic Potential X. Conclusions References

 

I. Introduction

II. Local GH Axis in Lymphoid Tissue

A. Background: extrapituitary production?

B. GH expression

C. GH regulation

D. GHRs

E. GHR signaling

III. GH Administration

A. Effects on the thymus

B. Pattern of GH exposure

C. GH and immune function in humans

IV. PRL

A. PRL expression in lymphoid tissues

B. PRL receptors

C. Administration of PRL, anti-PRL antibodies, or bromocriptine

V. Insulin-like Growth Factors

A. Background

B. IGF peptides

C. Regulation of lymphocyte IGF-I

D. IGF receptors on lymphocytes

E. IGFBPs

VI. Actions of IGF-I

A. Bone marrow

B. Effects on lymphoid organ size

C. Effect of in vivo treatment on lymphocyte number and in vitro function

D. Functional effects of IGF-I in vivo: antibody responses

E. Immune reconstitution

F. Mechanism of action: apoptosis

VII. IGF-I in Different Physiological States

A. IGF-I in pregnancy

B. IGF-I in diabetes

C. IGF-I in gastrointestinal disorders

D. IGF-I action in polycythemia vera

VIII. GH/IGF-I as Antistress Hormones

IX. Therapeutic Potential

X. Conclusions

	I. Introduction

Top Abstract I. Introduction II. Local GH Axis... III. GH Administration IV. PRL V. Insulin-Like Growth Factors VI. Actions of IGF-I VII. IGF-I in Different... VIII. GH/IGF-I as Antistress... IX. Therapeutic Potential X. Conclusions References

 
THE chief messenger systems between organs and tissues are the nervous, endocrine, and immune systems, which must be integrated on all levels to maintain homeostasis. The central hypothesis of this review is that the anabolic hormones (Fig. 1) GH, PRL, and the insulin-like growth factors (IGFs) that regulate whole body growth, metabolism, tissue repair, and cell survival also play an integrating role (Fig. 2) in the growth, maintenance, repair, and function of the immune system.

View larger version (28K): [in this window] [in a new window]   Figure 1. The GH/IGF-I axis. This figure depicts the sites of production of GH and IGF-I, the feedback loops regulating their secretion, binding proteins, and main metabolic actions. The hypothalamic hormones GHRH and somatostatin control GH secretion from the pituitary. GH circulates in the blood, in part bound to a GH-binding protein (GHBP), to inhibit its own secretion and to stimulate the production of insulin-like growth factor 1 (IGF-1) in the liver. IGF-I circulates in the blood bound primarily to IGF-binding protein-3 (IGFBP-3), which is in turn complexed to a third protein, the acid-labile subunit (ALS), but a total of six IGFBPs are known. IGFBP-IGF complexes are subject to protease attack, which assists the dissociation of IGF-I. The actions of GH are exerted either directly or indirectly, via the generation of IGF-I.

View larger version (18K): [in this window] [in a new window]   Figure 2. Endocrine GH secreted by the pituitary stimulates the production of endocrine IGF-I by the liver and local IGF-I in many other tissues, including the stromal cells of hematopoietic tissues. GH and PRL also act directly on lymphocytes or their precursors in hematopoietic tissues such as bone marrow, spleen, and thymus.

To dissect the effects of GH, PRL, and IGF-I on the immune system, it is important to understand the relative importance of the mediating role of IGF-I generation to the effects of GH (Fig. 1). This is not a simple exercise. The history of the mechanism of action of GH is a long one, strewn with many theories. For example, fragments of GH were once proposed to mediate all the effects of GH (18). In 1953, Salter and Best (19) described body growth in hypophysectomized rats treated with insulin, a result that was not confirmed by others (20). Such data led to the hypothesis that insulin mediates many of the growth-promoting effects of GH (21). This hypothesis was proposed almost contemporaneously with the discovery, in the serum of hypophysectomized rats treated with GH, of a sulfation factor activity (22) that was different from insulin (23). The hormones with this activity were later renamed the somatomedins (24). Purification (25, 26) led to the finding that somatomedins and nonsuppressible insulin-like activities (27) were identical and they were then renamed (25, 26) IGF-I and IGF-II.

The finding that human IGF-I purified from serum caused significant whole body growth in the rat (28) seemingly confirmed that many, if not all, of the growth-promoting effects of GH were mediated via the systemic generation of IGF-I in the liver: the somatomedin theory of growth regulation (29). It was then discovered, using recombinant human IGF-I (rhIGF-1) and recombinant human GH (rhGH), that rhIGF-1 and rhGH have very distinct differential effects on the size of different body organs in the rat (30, 31). If the effects of GH were all mediated by IGF-I generation, then this would not be the case. Earlier it had been proposed (32) that some of the endocrine effects of GH were a result of its direct action on tissues (Fig. 1) rather than being indirect via the generation of IGF-I in the liver. The much greater effect of rhIGF-1 than of rhGH on the weight of the spleen and thymus in the rat (30, 31) was one of the first indications that IGF-I and GH had different growth- promoting activities in vivo. That IGF-I, GH, and PRL can have different activities on the lymphoid tissue will be discussed throughout this review. The review will also focus on the evidence for an autocrine or paracrine GH/PRL/IGF system in lymphoid tissues that produce PRL and GH, contain PRL and GH receptors (GHRs), express IGF-I, contain IGF-I receptors, and secrete IGF-binding proteins.

This review is timely in that rhIGF-1 is being tested in large clinical trials in several human diseases (33), and it is possible that the immunological activities of GH/IGF-I may be found useful in the treatment of immune-deficient states in humans (34). For related information not included in this review, particularly the effects of GH/IGF-I on other hematopoietic cells, several related recent reviews (9, 35, 36, 37, 38, 39, 40) and a monograph (41) are available.

	II. Local GH Axis in Lymphoid Tissue

Top Abstract I. Introduction II. Local GH Axis... III. GH Administration IV. PRL V. Insulin-Like Growth Factors VI. Actions of IGF-I VII. IGF-I in Different... VIII. GH/IGF-I as Antistress... IX. Therapeutic Potential X. Conclusions References

 
A. Background: extrapituitary production?
Human GH is a protein of 191 amino acids produced and released by the anterior pituitary gland to circulate as an endocrine hormone (42). Until the placental GHs were discovered, GH was considered to be an exclusively pituitary hormone (42). It is now apparent that cells and tissues other than the pituitary and its somatotrophs also produce GH. Very recently, normal mammary tissue and mammary tumors, particularly under steroid stimulation, have been shown to produce surprisingly large amounts of GH (43, 44) that can be so large as to be detectable in blood.

This recent demonstration of GH production at an extrapituitary site may shed new light on the following discussion of the production of GH by lymphoid tissue. The total production of GH by extrapituitary tissues must be low compared with the production by the pituitary because hypophysectomy has such profound effects on body growth in a young animal. These findings, i.e. in some tissues GH is produced locally and probably acts locally, reinforce the idea (32) that GH has many actions on tissues that are not mediated by endocrine IGF-I.

B. GH expression
The first evidence of GH production by lymphoid cells came from direct staining using fluorescent-labeled anti-GH antibodies (45) and showed that about 10% of unstimulated human peripheral blood mononuclear cells (PBMCs) were positive for GH, whereas after mitogen stimulation 20% were positive. This surprising result was then independently confirmed (46) and extended to show that GH mRNA is expressed in lymphocytes (47) and that anti-GH antisense oligonucleotides inhibit lymphocyte proliferation (48). A plaque assay confirmed that the GH produced by human PBMCs is biologically active GH (49). Two human cell lines, the B cell lymphoma line IM-9 (50) and the Burkitt lymphoma cell line sfRamos (51), have been shown to synthesize and release human GH (hGH). These results suggest that GH is synthesized de novo and perhaps continually secreted from lymphocytes. By in situ hybridization and immunocytochemistry, GH was shown to be expressed in human and rat bone marrow, spleen, thymus, lymph nodes, and in human tonsil (41). A recent study (52) of human tissues, using in situ hybridization and RT-PCR, has also shown GH mRNA in spleen, lymph node, tonsil, and thymus. The GH mRNA was present not only in lymphocytes but also in the supporting structures, e.g. in the thymus GH was expressed by epithelial cells and reticular cells (52). There are reports that granulocytes, rather than lymphocytes, are the main cell type producing GH in peripheral blood (41). In marked contrast, another study in the rat showed that although GH was expressed in developing lymphoid tissues, it was not expressed in adult tissues (53). The consensus of all the literature data, however, is that GH is expressed locally in lymphoid tissues (41).

C. GH regulation
The transcription of the GH, PRL, and TSH genes in the pituitary depends on the activity of the transcription factor Pit-1 (54). The presence of this transcription factor in lymphoid tissues, and its colocalization with GH and PRL, strongly supports the idea of regulated extrapituitary GH production in spleen, bone marrow, and thymus (41, 55).

GH production in the pituitary gland is directly regulated by the hypothalamus, which produces the inhibitory factor somatostatin and the stimulatory factor GHRH (56). GHRH peptide and mRNA have been detected in human lymphocytes (57). Somatostatin has been identified in lymphoid tissue (58) as has the somatostatin receptor (59). However, whether somatostatin and GHRH are involved in the local production of GH is less clear (60). Physiological concentrations of GHRH and somatostatin have been reported to have no effect on the secretion of GH from human lymphocytes (60). The effects of GHRH and somatostatin on the activity of cultured lymphocytes have produced conflicting data: various assays have reported stimulation (61), inhibition (61, 62), or no effect (62). It is possible that somatostatin has a direct effect on IGF-I production because in the liver somatostatin exerts a direct inhibitory effect on IGF-I generation (63). A recent review suggests that, compared with GHRH, there is more evidence for somatostatin having an immune- modulatory role (64). However, there seems to be no published data testing for a direct effect in vitro of somatostatin on IGF-I mRNA or peptide in lymphocytes. The regulation of local GH may be different from that in the pituitary because the addition of IGF-I to lymphocyte cultures has been reported not to affect their secretion of GH (65). The most convincing evidence of PRL/GH release from both T and B lymphocytes has been that seen after direct mitogen stimulation in vitro (66). Small amounts of GH (0.2–0.6 pg/well, measured by RIA) are released by nonstimulated human PBMCs, but after phytohemagglutinin (PHA, a T cell mitogen) or pokeweed mitogen (a B cell mitogen), a dose-related increase in GH to several picograms per well has been shown (60). It appears likely that the regulation of GH in the immune system differs from that in the endocrine system.

D. GHRs
The finding that GH is produced locally by lymphoid tissues has been given more of a functional significance by evidence that lymphocytes also express GHRs. GH binding was first detected on a human B cell lymphoma (IM-9) lymphocyte cell line (67) and subsequently identified (68) on human PBMCs. The development of specific monoclonal antibodies against the GHR allowed flow cytometry to confirm that GHRs are present on human IM-9 cells (69) and in human PBMC with the highest expression on B cells (70). The GHR on lymphocytes has been sequenced (71, 72) and found to be identical to the GHR cloned from liver (73). A more recent study confirmed that the hGH receptor is present on more than 90% of B lymphocytes and monocytes, but only variably present on T lymphocytes. B lymphocytes and monocytes had approximately 6000 GHRs per cell, and this number was not affected by a donor being GH-deficient (74).

E. GHR signaling
The purification and cloning of the GHR (73), the discovery of the dimerizing stochiometry of the GH-(GHR)2 complex (75), and the crystallization of this complex (76) have revolutionized understanding not only of GH but of the new family of the helix bundle peptide (HBP) cytokine receptors (40). The placing of the GH and PRL receptors in the family of hematopoietic cytokines (77), which includes erythropoietin, granulocyte-colony-stimulating factor, granulocyte/macrophage colony-stimulating factor, and the interleukins, has provided a theoretical basis to the experimental results showing that GH and PRL have significant activity as hematopoietic cytokines (40).

Like other members of the HBP receptor superfamily, GH signals through the JAK kinase/STAT cascades. However, because many of these intracellular signaling mechanisms appear to share several receptor/ligand systems, it has been difficult to see how specificity is maintained (78). If signaling pathways do overlap, then the responses of lymphocytes to GH/PRL could be viewed as minor epiphenomena, with other ligands or receptors mediating the same responses with more potency and specificity. However, mice with a disrupted STAT1 gene show defects in interferon signaling but normal responses to other cytokines, including GH (79). Such specificity was not predicted from in vitro studies and suggests that GH and PRL do have unique effects in lymphocytes.

GH action involves the sequential binding of an initial site on hGH (site 1) to a GHR molecule to form a GH/GHR monomer complex, followed by the binding of another GHR molecule to the second site on hGH (site 2), to form a receptor dimer complex (75). It is this complex that activates the JAK2/STAT cascades. GH analogs with mutations in the second binding site, but with an intact site 1, prevent dimer formation and can act as antagonists (80). Because the affinity of these mutant GH molecules for GHR remains high, and can be engineered to increase the affinity for site 1 selectively, highly potent hGH antagonists have been produced. A GH antagonist (80) is potentially a useful tool for elucidating the importance of GH to lymphoid tissues. Unfortunately the mutant hGH molecules that act as antagonists at the hGH receptor show no evidence of antagonist activity when administered to rats (81). The reason for the species specificity of hGH antagonists is unclear because native hGH binds the rat GHR with high affinity and stimulates body growth. A GH antagonist fully active in the rat is eagerly awaited because it would help explain the importance of endogenous GH, the importance of local GH production, and especially the importance of GH to lymphoid tissues.

	III. GH Administration

Top Abstract I. Introduction II. Local GH Axis... III. GH Administration IV. PRL V. Insulin-Like Growth Factors VI. Actions of IGF-I VII. IGF-I in Different... VIII. GH/IGF-I as Antistress... IX. Therapeutic Potential X. Conclusions References

 
A. Effects on the thymus
The thymus grows rapidly postnatally in mammals, peaks in size at around sexual maturity, and then slowly involutes with age (82). In humans, maximal thymic size is attained at puberty and then by 45–50 yr of age involutes so that only 5–10% of the cellular mass remains. In mice, maximal size (70 mg) is attained at puberty (6 weeks) while by 9 months of age the thymus has declined in weight to only 20 mg. Involution of the thymus with age was recognized (83) well before its immunological role (84). These changes in thymic structure and function follow the rising activity of the GH system, with serum IGF-I levels peaking at puberty and declining gradually with advancing age (85, 86, 87) suggestive of a causative relationship. The age-related decline in immune function is poorly explained. A portion of this decline could be related to the lack of activity, or resistance to the actions, of the anabolic hormones.

The first studies suggesting that treatment with GH could affect the thymus in nonrodents was a study in hypopituitary dwarf Weimaraner dogs (88). In follow-up studies, young, middle-aged, and aged dogs were treated with GH. Thymic growth was observed in middle-aged, but not in aged, dogs and the blood level of thymic hormone increased (89). The thymus glands of the treated dogs were described (90) as "resembling thymic tissue of young dogs." In the aged rat, the implantation of GH3 pituitary cells reversed age-related thymic atrophy and increased the number and function of T cells in the thymus (91). The next section discusses in more detail the effects in rodents of continuous infusions of hGH, which are very effective at stimulating thymic growth (92). It should be noted that GH3 cells may produce GH, but little PRL, in vivo (93), and their release of GH would be continuous, perhaps accounting for the clear effects reported above (91) on the thymus.

B. Pattern of GH exposure
In several tissues the pattern of GH administration or exposure can have major quantitative and qualitative tissue-specific effects (94, 95). For example, different patterns of endogenous GH exposure, best illustrated by differences in GH-secretory profiles in male and female rats, cause a number of sexually dimorphic responses (96). The administration of GH to rats either by daily injection (male pattern) or by continuous infusion (female pattern) can replicate these dimorphic responses, which include influencing hepatic enzymes (97), hepatic growth (98), and lipid metabolism (99). In mice that overexpress bovine GH, there is an enlargement of the internal organs, particularly of the spleen, which may be due to both the high levels of GH and the continuous pattern of GH exposure (100).

As described above, it is likely that GH is produced locally in lymphoid tissues. The production of this GH is probably by constitutive expression and release, giving a continuous pattern of local GH exposure. To explore this idea, GH was given by injection or infusion, and the effects on different body tissues were compared (98). It was found that in rats (Fig. 3A), injections of GH are much less potent than GH infusions at stimulating lymphoid tissue growth (98). Figure 3A shows that in hypophysectomized rats the spleen more than doubles in size after GH infusions, whereas the same doses of GH given by injection elicit no splenic growth. It has been stated that in mice, 20- to 40-fold higher doses of GH and PRL are needed by injection, compared with minipump infusion, to reverse corticosterone-induced suppression of splenic lymphocyte responses to mitogens (101). Recent data in young castrate and intact genetically obese pigs also suggest that treatment with porcine GH can affect thymic size and thymosin concentration in serum (102). When porcine GH was given by injection or by a slow release depot formulation, there was a greater effect on the thymic parameters of the slow release formulation (102). A different form of long-acting GH, i.e. that made by coupling polyethylene glycol (PEG) to hGH, has now been described (103). The administration by infrequent injection of PEG-rhGH to hypophysectomized rats also induced lymphoid organ overgrowth compared with daily injections of nonmodified rhGH. This difference was probably due to the more continuous pattern of GH exposure caused by the long plasma half-life of the PEG-rhGH (103).

View larger version (18K): [in this window] [in a new window]   Figure 3. Spleen weight (A) and serum IGF-I concentrations (B) in hypophysectomized rats treated with excipient, or three doses of hGH given by injection or infusion. At these doses, infusions, but not injections of hGH, increased serum IGF-I concentrations, perhaps explaining the selective effect of hGH infusions on spleen weight. Means and SDs are shown. [Derived from Ref. 98.]

C. GH and immune function in humans
The strong evidence from studies in animals that the pituitary (3), GH (35), PRL (16), and IGF-I (36) affect hematopoietic and lymphoid tissues is generally accepted. However, the evidence in humans is perceived by recent reviewers as being unconvincing (39, 105). The most cited evidence (39, 105) is that GH-deficient children are not clinically immunodeficient and therefore replacement therapy with hGH would not be expected to have significant effects on immune function. This perception of a general lack of effect of GH deficiency or GH replacement treatment on immune function has been taken as evidence that the GH/IGF axis has a lesser effect on lymphoid tissues in humans (39). A lesson can be taken here from animal studies with other HBP cytokines, such as interleukin-6 (IL-6), which has a range of pleiotropic actions on T cell and B cell proliferation that it shares with IL-1, IL-2, IL-4, and IL-5 (106). The overexpression of IL-6 in mice has only a mild effect on B cells (107). Another example of the pleiotropic actions of cytokines is that disruption of the IL-2 gene allowed almost normal hematopoiesis (108). These cytokines do have important actions on the immune system, but the presence of multiple cytokines ensures that homeostasis is maintained. Therefore, the apparent lack of effect of an endocrine GH deficiency in humans should not be taken as evidence that GH has no effects on immune function in humans.

Human studies have concentrated on discovering the immunological phenotype of patients who lack pituitary GH. There has been almost no effort toward discovering whether these patients are deficient in local GH, local IGF peptides, or receptors or have a changed local IGF-binding protein (IGFBP) status. Even if pituitary GH is disturbed, it is likely that in many patients the local paracrine/autocrine axis in lymphoid tissue is intact. In GH-deficient humans it is possible that the GH produced locally in the immune system compensates for the lack of endocrine GH. The much higher endogenous levels of GH (96) in the rat, about 10-fold higher than in the human (85), would be expected to cause locally produced GH to be less important in the rat that in humans. In addition, rats, particularly female rats, have relatively high GH-binding protein levels that may act to enhance the activity of GH (109). This may help explain why in humans a deficiency in pituitary GH or endocrine IGF-I appears to have minor effects on immune function compared with the effect of such deficiencies in the rat. One group of patients who may be immunologically impaired are patients with defective GHR function who were once termed Laron dwarfs but are now described as having GH insensitivity syndrome (110). The largest cohort yet described, that localized in Ecuador, may have a significantly higher pediatric mortality (110), perhaps indicating an impaired immune system. A recent monograph includes a lengthy discussion of the relative importance of hormones to immunological status in rodents and humans (41). A proposal by these authors that should be considered, but for which there is little or no evidence, is that IGF-I expression is less dependent on GH in humans than in rodents (41) and that this may account for the apparent lack of effect on immune function of GH deficiency in humans. Another reason could be that IGF-II concentrations in adult rodents are very low compared with the concentrations in humans. Therefore, a deficiency of GH in rodents, leading to a fall in IGF-I concentrations, has a greater impact than in GH-deficient humans in whom the maintained IGF-II concentrations may preserve blood IGF concentrations and activity.

	IV. PRL

Top Abstract I. Introduction II. Local GH Axis... III. GH Administration IV. PRL V. Insulin-Like Growth Factors VI. Actions of IGF-I VII. IGF-I in Different... VIII. GH/IGF-I as Antistress... IX. Therapeutic Potential X. Conclusions References

 
A. PRL expression in lymphoid tissues
The first cytokines in the HBP family to be identified were GH and PRL. Because hGH, unlike rat GH, also binds with high affinity to PRL receptors, it was unclear, until human PRL was purified by Friesen and colleagues (111), whether humans had a separate PRL. PRL has always been viewed as having a very broad range of activities (40); therefore the subsequent demonstration that it had effects on lymphoid tissue was not surprising. The discovery of PRL expression in lymphocytes was a surprise. A sensitive bioassay, based on immunostaining of Nb2 cells, showed PRL-like activity to be present in the culture fluid from concanavalin A (ConA)-activated murine splenocytes (112). ConA-stimulated thymidine incorporation was reduced by adding an antibody against PRL, suggesting that a PRL-like molecule was produced by the splenocytes and was essential for lymphocyte proliferation (112). Using an enzyme-linked immunoplaque assay, human PRL secretion was found after ConA or PHA stimulation of PBMCs, but not in unstimulated PBMCs (113). Murine and human T- and B cell mitogen proliferation has been shown to be inhibited by antibodies to PRL, due to a block in the G1 to GS transition in the cell cycle (114). In situ hybridization showed the presence of an mRNA in murine splenocytes that hybridized with a rat PRL cDNA probe (115). The PRL protein in human lymphocytes appears to be similar to pituitary PRL, i.e. the multiple forms present in the pituitary are also present in lymphocytes (116). The PRL gene is also expressed in rat thymus (117). As described above for GH, the extrapituitary production of PRL in lymphoid tissue is not unique; the PRL gene, like the GH gene (43, 44), also appears to be transcribed in the mammary gland (118).

B. PRL receptors
Lymphocytes not only produce PRL but also possess PRL receptors. The use of the rat Nb2 T cell line for bioassaying PRL clearly suggested the presence of PRL receptors on lymphoid cells (119). PRL binding on normal lymphocytes was first demonstrated on human T and B lymphocytes (120). Since then, biotinylated monoclonal antibodies against the human PRL receptor (121) have been used to show PRL receptors in the mouse (122), particularly on B cells, and the presence of receptors throughout human hematopoietic tissues including bone marrow and thymus. B cells were the most strongly labeled, while T cells showed an increased labeling upon activation (121). In the mouse and rat, PRL receptors are present in bone marrow, thymus, spleen, lymph nodes, and on peripheral blood lymphocytes (122), with receptor number increasing in a draining lymph node after foot pad immunization (123) and on T cells after ConA administration (124). In the thymus, PRL receptor number is greatest in the cortex and in thymic epithelial cells (125). Two forms of the PRL receptor, which differ in the length of their cytoplasmic domains, are present in lymphoid tissues in the mouse and rat (117). The PRL receptor can be detected by PRL binding, by antibodies, and by PCR in many lymphoid cell lines (both T and B cell) and hemopoietic cell lines (41). The signaling of PRL, especially in Nb2 cells, has been studied intensively and reviewed recently (40).

C. Administration of PRL, anti-PRL antibodies, or bromocriptine
Hypophysectomized rats are almost devoid of a primary antibody response after the injection of sheep red blood cells (126). Replacement with lactogenic hormones (40 µg/day) at the time of immunization restored antibody titers to those of a normal rat (126). After complete hypophysectomy an animal should be completely PRL-deficient if the pituitary were the only source of PRL. However, Nagy and Berczi (127) used the Nb2 cell bioassay to show that immediately after hypophysectomy rats have unexpectedly high blood PRL concentrations, 10–20% of normal, which then rise to 50% of normal 8 weeks after hypophysectomy. In normal rats, PRL had been described as having an effect on hematopoiesis; however, in the hypophysectomized rat, red cell count, although low, was compatible with life. It was therefore reasonable to conclude that PRL is of marginal importance to red cell biology. In an important set of experiments Nagy and Berczi (127) gave anti-PRL antibodies to hypophysectomized rats to neutralize this residual PRL. After the anti-PRL sera was given, severe anemia developed and all the animals were dead within 6 weeks (127). This is very compelling evidence that significant amounts of PRL are made by extrapituitary sources and that this local PRL has vital functions, especially for hematopoietic tissues. Comparable studies using anti-GH antibodies would be of great interest. The sensitivity of hypophysectomized rats to the lethal effects of estrogens also needs reevaluating (128).

In humans, the administration of hGH, as it binds to hGH and human PRL receptors, will activate both GH and PRL receptors and their signaling pathways. The apparent lack of effect of hGH administration in humans on immunological function may be due to a lack of selectivity. The administration in humans of hPRL, a specific hPRL receptor ligand, or V-gene hGH (42), a more specific hGH receptor ligand, may show different effects than hGH administration. The prospective availability of a GH antagonist (80, 81) provides the opportunity to antagonize both local and systemic hGH. Treatment with a GH antagonist may therefore cause a different immunological phenotype than a deficiency of pituitary GH.

Published data (66) describe the immunological effects of bromocriptine, a dopamine ergot alkaloid that inhibits the release of PRL from the pituitary. A bromocriptine-treated rat has been reported to show a reduced mixed lymphocyte reaction in vitro and a reduced graft-vs.-host reaction in vivo (66), presumably due to a suppression of pituitary PRL. This may not be the case. Bromocriptine appears to have direct effects in vitro on lymphocyte proliferation in the absence of added PRL (129, 130). Animal studies in transplantation (131) and a clinical trial using bromocriptine, in combination with cyclosporine, as an immunosuppressive regimen, have met with some success (132). The production of genetically engineered PRL-deficient mice or PRL receptor-deficient mice may help answer the question of the importance of PRL to lymphocytes and to immune function.

	V. Insulin-Like Growth Factors

Top Abstract I. Introduction II. Local GH Axis... III. GH Administration IV. PRL V. Insulin-Like Growth Factors VI. Actions of IGF-I VII. IGF-I in Different... VIII. GH/IGF-I as Antistress... IX. Therapeutic Potential X. Conclusions References

 
A. Background
The chief regulators of the IGF-I levels in blood are GH status (133) and nutrition (134, 135). GH and IGF-I, acting together, ensure ordered body growth and therefore are involved in complex interactions with most organ systems, tissues, cell types, and also with many growth factors (133, 136). Due to these multiple effects, GH, PRL, and IGF-I can affect diverse physiological processes, including immune function, in many ways, both directly and indirectly (133, 136). The GH system regulates and coordinates whole body growth to ensure that different tissues grow in unison and are then maintained in an optimal proportion to the rest of the body. The immune system may be one such tissue. For example, GH and IGF-I stimulate cartilage growth and lengthen bones causing statural growth (136). By controlling the size of the bones in the growing animal, GH and IGF-I therefore indirectly control the volume of bone marrow and thus the production of hematopoietic cells (137). Recent evidence shows that IGF-I differs from insulin in that, at physiological concentrations, it also plays a direct and significant role in regulating hematopoiesis, especially lymphopoiesis and immune function (36).

B. IGF peptides
IGF-I and IGF-II, peptides of 70 and 67 amino acids, respectively, were named because of the similarity of their actions to that of insulin (138) and their chemistry to that of proinsulin (25, 26). A major difference between these hormones and insulin is that the IGF peptides are expressed almost ubiquitously (139, 140, 141). In adult humans, IGF-I and IGF-II are both present in large amounts in blood, as they are in fetal rodents, but in adult rodents IGF-II concentrations in blood are very low (142). It is unclear whether these differences between IGF-I and IGF-II status result in species-specific effects on immune function. IGF-I concentrations in blood are controlled by GH status; IGF-II levels are much more GH-independent (142). Exons 1 and 2 of the IGF-I gene contain two distinct promoters that give rise to IGF-I mRNAs containing either exon 1 or exon 2 (143). Exon 1 mRNA is the form in fetal tissue, whereas the exon 2 form appears postnatally when GH responsiveness is acquired (143). These different forms of IGF-I mRNA perhaps supply either GH-dependent endocrine IGF-I (exon 2) or local GH-independent paracrine or autocrine IGF-I (exon 1) (143). In myeloid cells, IGF-I transcripts have been found to be exclusively initiated within exon 1, characteristic of extrahepatic IGF-I mRNA (144). IGF mRNA and peptides are produced by myeloid cells, particularly by macrophages, in relatively large amounts (36, 144) and by human peripheral lymphocytes in small amounts (36, 144, 145). Bone marrow stromal cells also release IGF-I (146, 147) as do thymic epithelial cells (93), which can be stimulated by GH in culture (148). Therefore, there are ample data showing the local production of IGF-I in lymphoid tissues. However, there is scant evidence describing the regulation of this locally produced IGF-I or the relative importance to lymphoid tissues of local or endocrine IGF-I.

C. Regulation of lymphocyte IGF-I
Considerable amounts of IGF-I are regulated by GH-independent pathways, as seen by the presence of significant serum IGF-I levels in GHR-deficient humans (110) and sex-linked dwarf chickens (149). Treatment with GH increases IGF-I mRNA in many tissues (140), whereas in other tissues IGF-I generation is controlled by factors or hormones other than GH. For example, in the rat uterus IGF-I may be regulated chiefly by estrogen, rather than by GH (150). Cytokines other than GH affect IGF-I synthesis in lymphoid tissues, e.g. in macrophages tumor necrosis factor- has been shown to regulate IGF-I production (151). Tumor necrosis factor- and prostaglandin E₂ (PGE₂) stimulate IGF-I synthesis in macrophages by two separate pathways, with PGE₂ stimulating IGF-I synthesis through a cAMP/protein kinase A pathway (152). The colony-stimulating factors also induce the expression of IGF-I mRNA in macrophages (153), whereas the T cell-derived cytokine IFN- reduces macrophage IGF-I mRNA in a time- and dose-dependent manner (154).

It is therefore likely that lymphocytes are exposed to endocrine IGF-I from the circulation, their own autocrine IGF-I, and perhaps most importantly, in lymphoid organs and bone marrow, a third source of IGF-I from epithelial cells (148) and stromal cells (35). The proliferation of thymic epithelial cells can be stimulated in vitro by both hGH and IGF-I (155) and the effect of hGH blocked by either an anti-IGF-I or an anti-IGF-I receptor antibody (93). This effect of hGH may be via the PRL receptor, rather than the GHR, as rat and bovine GH (which do not bind to the PRL receptor) appear to be inactive in this model (156). However, in vivo in the mouse, although hGH and ovine GH exerted positive effects on thymic development, ovine PRL was described as having "the opposite effect of GH on the thymus" (157). Careful comparisons of GH and PRL activity using homologous systems are needed to discover their relative activities on lymphoid tissues. The effects of GH on the thymus may be due to local IGF-I generation (148), which would fit with the in vivo data of IGF-I administration having larger effects than GH on thymic growth (31). However, it should be remembered that GH and PRL may have effects, including on lymphoid tissues (158), that are not mediated via IGF-I generation.

Normal human PBMCs express very low amounts of the IGF peptides, which can be increased after mitogen stimulation (36, 144, 145). In contrast, macrophages are reported to produce much more IGF-I especially when they are differentiating to the mature phenotype (36). Human IM-9 lymphocytes (B lineage cells) also express IGF-I mRNA, but this message is not sensitive to treatment with hGH, despite evidence of GH-induced tyrosine phosphorylation (159). In contrast, transformed B cells (160) release IGF-I in response to GH. In human T lymphoblast cell lines, the stimulation of colony formation produced by GH may be mediated by IGF-I (161). The growth of T-acute lymphoblastic leukemic (ALL) cell lines can be slowed by antibodies against either IGF-I or the Type 1 IGF-I receptor, suggesting autocrine or paracrine activity of IGF-I in T-ALL cell lines (162).

D. IGF receptors on lymphocytes
The insulin receptor and the Type 1 IGF receptor are both tyrosine kinase receptors, have similar structures, and both bind insulin and the IGFs, albeit at lower affinities for the heterologous ligands (163). The receptors are so similar that their subunits are believed to be interchangeable so that they can naturally form so-called hybrid receptors (163, 164). Both the IGFs and insulin therefore have similar powerful metabolic effects, but the Type 1 IGF receptor also possesses many of the differentiating and mitotic effects found for the ligand/receptor complexes of other tyrosine kinase receptors such as c-kit/KL and c-fms/colony-stimulating factor-1. Such receptors are important regulators of the differentiation of hematopoietic cells (165). As described above, the GH and PRL receptors are members of another family, the HBP family of cytokine receptors, which also regulate many processes in hematopoietic cells (40). As would be predicted from this discussion, the IGFs and the Type 1 IGF receptor have differentiating and mitogenic activities, including in hematopoietic cells (35, 133).

A third IGF receptor, termed the IGF-II receptor, is also the mannose-6-phosphate receptor, whose role in IGF biology is unclear (166) although it does not appear to transmit a direct intracellular signal (167). This receptor binds IGF-II with high affinity but binds IGF-I with about 100-fold less affinity (168). The IGF-II receptor may act mainly as a functional IGF-II "antagonist" to regulate local IGF-II cell exposure and may have tumor suppressor-like properties (167).

Twenty years ago the binding of insulin to resting (169) and activated (170) lymphocytes was an active area of research. This was followed by the discovery of IGF-I binding to human leukemic lymphoblasts (171), PBMC, resting and activated T cells (172, 173), and the cross-linking of IGF-I to activated T cells (172, 173). This strongly suggested the presence of IGF-I receptors on lymphocytes. The functional importance of these receptors to lymphocyte activation by mitogens was shown by the peak receptor number occurring at the same time as maximal thymidine incorporation (172). The use of two-color flow cytometry, staining with antibodies against the human Type 1 IGF-I receptor, the insulin receptor, and lymphocyte markers, showed that both IGF-I and insulin receptors are present on most monocytes and B lymphocytes, but on only 2% of T lymphocytes (174). Using similar techniques, IGF-I receptors were found in high numbers on monocytes, natural killer cells, and CD4+ cells, an intermediate number on CD8+ cells, and a relatively low number of receptors on B cells (175). Using flow cytometry and biotinylated des(1, 2, 3)IGF-1, IGF-I receptors were detected on rat T cells, B cells, and monocytes with the expression on resting CD4+ cells being greater than on CD8+ cells and increasing severalfold after ConA stimulation (176). Why there are discrepancies between these studies is unclear. In another study, T lymphocyte activation, by PHA or the OKT-3 monoclonal antibody (which binds to the CD-3 antigen of the T-cell receptor), led to peaks in IGF-I receptor mRNA after 20–60 h and IGF-I receptor content after 48–72 h (177). If the increased IGF receptor number caused by mitogens is physiologically relevant, the addition of IGF-I should increase proliferation caused by the mitogens. This has been shown for human peripheral lymphocytes (178) and thymocytes (179). Freshly isolated human peripheral lymphocytes have been shown to express IGF-I receptors by RT-PCR (145), and human T lymphoblast cell lines possess IGF-I receptors (180). Some of the differences between the data sets for IGF receptors on peripheral lymphocytes could be methodological, either due to differences between rodents and humans or to lymphocyte receptors being atypical IGF receptors. For example, it has been claimed that the majority of the IGF-I receptors on human IM-9 lymphocytes (B lineage cells) are atypical IGF-I receptors (150).

IGF-I receptors have also been identified on rodent (181) and human thymocytes (179, 182). The signal transduction in normal human thymocytes and T cells appears to be similar to that in other tissues, e.g. it involves the phosphorylation of insulin-receptor substrate-1 (IRS-1) (182). In human thymocytes, DNA synthesis can be stimulated directly by IGF-I in vitro, and DNA synthesis initiated by the mitogen PHA can be potentiated by adding physiological concentrations of IGF-I (179).

E. IGFBPs
The IGFs also differ from insulin in that in vivo they are bound (Fig. 1) to a family of at least six specific, soluble, high-affinity IGFBPs, termed IGFBPs 1–6, which are unrelated structurally to the IGF receptors or the insulin receptor (133). It is possible that novel IGFBPs remain to be discovered. In fact, mac25 (183) and PSF (184) or ESM-1 (185) show marked structural similarity to the IGFBPs. The binding proteins differ in their modes of regulation and perform a variety of functions. For example, the majority of the IGF in blood is bound to IGFBP-3 (Fig. 1), and this complex is bound by a third protein, the acid-labile subunit, to form a large stable 150-kDa complex (186). IGFBP-3 and acid-labile subunit are regulated primarily by GH, have a slow clearance from blood, and provide an accessible pool and reservoir of IGF in the blood (133). In contrast, IGFBP-1 concentrations in blood can change rapidly, are regulated by insulin, and may serve an acute metabolic role to bind and inactivate unbound IGF (187).

Many tissues and cell types secrete IGFBPs, including hematopoietic cells. By RT-PCR, normal human peripheral lymphocytes express mRNAs for the IGF-I receptor, the IGF-II receptor, IGFBP-2 and -3, but not the IGF peptides (145). After stimulation with PHA they express IGF-I, IGF-II, and IGFBP-4 and -5, in addition to IGFBP-2 and -3 (145). Ligand blotting of lymphocyte-conditioned media with labeled IGF-I revealed 34-, 43-, and 49-kDa IGFBPs. The addition of estrogen, progesterone, IGF-I, or GH did not affect secretion of IGFBPs by lymphocytes (145). IM-9 cell-conditioned medium has also been shown to contain a 30-kDa IGFBP (141). Murine stromal bone marrow cells, which support developing hematopoietic cells, not only produce IGF-I but also secrete IGFBPs (147). By ligand blotting (188), the most prominent IGFBPs were IGFBP-4 and IGFBP-5 whereas, by RNase protection assay, murine stromal cells expressed IGFBP-2 to IGFBP-6 mRNAs, with IGFBP-4, IGFBP-5, and IGFBP-6 mRNAs being predominant. These authors (188) suggest that IGFBPs 4–6 are released by stromal cells to modulate the hemopoietic response to IGFs. Sheep thymus cells also produce IGFBPs in culture, secretion is increased by mitogen stimulation, and medium from these cells also degrades recombinant human [¹²⁵I]IGFBP-3, suggesting IGFBP-3 protease production (189).

IGFBP-2 is the predominant IGFBP during fetal life (as IGFBP-3 is in the adult) and is expressed in a range of tumor cell lines (190). Significantly increased serum levels of IGFBP-2 have been detected in the sera from ALL and non-Hodgkin-lymphoma patients (191). At the time of diagnosis with leukemia, non-Hodgkin’s lymphoma, or solid tumors the serum concentrations of IGF-I, IGF-II, and IGFBP-3 are very low in children (192). Such low concentrations of these proteins are normally seen only in patients with GH deficiency or during starvation (192). Somewhat surprisingly, IGFBP-2 levels were elevated (192), and it seems that the IGFBP-2 is produced by the tumor cells (193). Leukemic T cell lines, but not B cell lines, produce large amounts of IGFBP-2 and express mRNA for IGFBP-2, confirming data from the patients with tumors (193). The biological significance of this discovery is unclear.

Some evidence indicates that the IGFBPs regulate IGF action in lymphoid tissues. Preliminary data suggest that mice that are null for the IGFBP-2 gene show no gross phenotype except for a reduced spleen size, to 50% of normal (194). The above evidence, of IGFBP-2 production by T cells but not B cells, suggests that the distribution of lymphocyte subsets in these null mice may be altered. Overexpression of IGFBP-1 in transgenic mice has led to inconsistent effects on spleen size (195, 196), whereas overexpression of IGFBP-3 causes increased spleen size (197). An intriguing recent paper (198) shows that activation of the tumor suppressor gene p53, which induces apoptosis, stimulates IGFBP-3 expression. It is possible that the IGF system plays a role in the regulation of apoptosis (Fig. 4) via p53 stimulating IGFBP-3, leading to a local inhibition of IGF-I action with the induction of apoptosis. This provides a mechanism by which, in lymphoid tissue, locally produced IGFBPs, by regulating the availability of IGF-I, may help control cell division and survival.

View larger version (13K): [in this window] [in a new window]   Figure 4. Possible mechanism for the antiapoptotic effects of IGF-I and the apoptotic effects of IGFBP-3. Injury can induce the expression of the tumor suppression gene p53, which in turn may increase the expression of IGFBP-3. Because IGF-I is bound by IGFBP-3, this might prevent the activation of the IGF-I receptor and allow apoptosis to proceed. In contrast, exogenous IGF-I can allow cell survival.

	VI. Actions of IGF-I

Top Abstract I. Introduction II. Local GH Axis... III. GH Administration IV. PRL V. Insulin-Like Growth Factors VI. Actions of IGF-I VII. IGF-I in Different... VIII. GH/IGF-I as Antistress... IX. Therapeutic Potential X. Conclusions References

IGF-I has two of these major effects on B cell development (Fig. 5); it acts as a differentiation factor to potentiate pro-B to pre-B cell maturation (200), and it also acts as a B cell proliferation cofactor to synergize with IL-7 (201). The first indication that B cell differentiation factors exist came from clinical studies in infants with cyclic neutropenia in which the production of erythroid and myeloid cells oscillates. In the marrow of these children, pre-B cells also oscillate as does the presence in their urine of a factor that in vitro stimulates normal human marrow cells to generate pre-B cells (202, 203). The factor specifically affects differentiation as this occurs in the absence of proliferation. A bone marrow stromal cell line was found to release a similar activity. This activity was identified as IGF-I based on the use of anti-IGF-I antibodies, antisense to IGF-I, and that recombinant IGF-I could substitute for the activity (200). In the presence of IGF-I, pro-B cells mature to pre-B cells, as judged by their ability to proliferate in response to IL-7 (204). As described above, there is evidence that macrophages are a rich source of IGF-I and that bone marrow stromal cells also produce IGFBPs (147). The treatment of mice with rhIGF-1 confirmed these observations (Fig. 6) as it increased the number of pre-B and mature B cells in bone marrow (205). The mature B cell remains sensitive to IGF-I as immunoglobulin production is also stimulated by IGF-I in vitro and in vivo (206).

View larger version (18K): [in this window] [in a new window]   Figure 5. The differentiative and proliferative actions of IGF-I during the stages of B cell development. IGF-I is unique in that by itself it stimulates the differentiation of Pro-B cells. IGF-I also acts as a proliferative cofactor throughout B cell development.

View larger version (19K): [in this window] [in a new window]   Figure 6. Treatment with rhIGF-1 (4 mg/kg/day, sc infusion for 2 weeks) more than doubles the number of Pre-B cells (B220+, sIgM-) in the bone marrow of normal adult mice. Means and SDs are shown. [Derived from Ref. 205.]

The effects of IGF-I on T cell development are not as well characterized, although thymic T cell progenitors proliferate in response to IGF-I before they respond to any other known cytokine (210). It is also clear that thymic epithelial cells produce IGF-I, functional IGF-I receptors are present on thymocytes (181), and the administration of IGF-I to animals affects the number of T cells in the thymus (92). There is as yet no information on the thymic role of IGF-I in processes of positive or negative selection of thymocytes. It has been claimed that rhGH can induce significant migration of resting and activated human T cells (211). These authors speculate that, by directly altering their adhesive and migratory capacities, GH may play a role in normal lymphocyte recirculation. This finding is as yet unconfirmed and raises the question of the activity of IGF-I in these assays.

B. Effects on lymphoid organ size
An involvement of the somatomedins (IGFs) in the regulation of lymphoid organ growth was suggested when IGF-I was administered to hypophysectomized rats because it caused preferential thymic and splenic growth to a greater degree than did GH (31). Such studies were then extended to the mutant dw/dw rat, which showed a similar disproportionate growth of lymphoid tissue (30) whereas in normal aged 18-month-old rats, IGF-I stimulated thymic growth (Fig. 7) and increased lymphocyte numbers (212). Figure 8 shows the effect of 28 days treatment with rhIGF-1, rhGH, or rhIGF-1 plus rhGH on thymic architecture. It is clear (Fig. 7) that rhIGF-1, and to a lesser extent rhGH, stimulates thymic growth. Because the rat is not the preferred species for immunological studies, the mouse was selected as an experimental animal. At the time, there was very little information on effective doses or dosing regimens of IGF-I in the mouse apart from some early anabolic studies using somatomedin preparations in mutant dwarf mice (213) and isolated more recent studies with rhIGF-1, also in mutant mice (214). Instead of using GH-deficient animals, the effects of IGF-I were studied in normal, 9-month-old, retired breeder "middle-aged" male mice (92).

View larger version (156K): [in this window] [in a new window]   Figure 7. Thymic histology in 18-month-old male rats treated for 4 weeks with either excipient (panel A), rhIGF-1 (panel B, 1.1 mg/rat/day, sc minipump infusion), rhGH (panel C, 1 mg/rat/day, daily sc injection), or rhIGF-1 plus rhGH (panel D). Formalin-fixed hematoxylin-eosin-stained sections. The hormone treatments, particularly rhIGF-1, caused a dramatic expansion and rejuvenation of the involuted thymus of aged rats [R. Clark, unpublished data].

View larger version (23K): [in this window] [in a new window]   Figure 8. Antibody production in mice immunized with 0.1 µg DNP2OA at week 0, then boosted at week 5. In addition, the mice were treated with either excipient or rhIGF-1 (4 mg/kg/day) for 2 weeks after each immunization. Treatment with rhIGF-1 enhanced antibody production, especially after the secondary immunization. Means and SDs are shown. [Derived from Ref. 206.]

The administration of IGF-I has been shown to increase the size of lymphoid organs in several species. In rats and mice numerous studies report increases in lymphoid tissue mass with IGF-I administration (36). In 1-yr-old sheep, an 8-week regimen of three daily injections of rhIGF-1 (50 µg/kg) increased spleen weight by 40% (215). In the rabbit, cat, and dog similar effects of IGF-I have been observed (R. Clark, unpublished observations). In the rhesus monkey, IGF-I also expands lymphocyte numbers (216). This finding is discussed in more detail in Section IX. Lymphoid organ expansion has been reported in children with GH insensitivity who have been treated long-term with rhIGF-1 (110). This is the first direct confirmation that the immunological effects of administering IGF-I to animals are also present in humans.

Mice transgenic for GH or IGF-I have enlarged lymphoid organs (217). An effect of endogenous IGF-I on lymphoid tissue growth was presumed in a large study (218) where lines of mice were selected over many generations on the basis of high or low serum IGF-I levels. The high IGF-I line had spleens 30% heavier than the low IGF-I line. Thymus weights were also greater in high-line than in the low-line mice, and developmental patterns of thymus weight closely paralleled those of circulating IGF-I (218).

Compared with the ample data on the effects of IGF-I, the effects of IGF-II on lymphopoiesis are not as well studied (219, 220). IGF-II binds less well to the Type I IGF-receptor; therefore, higher concentrations are probably needed (although much less than insulin) to mimic the effects of IGF-I. However, there are reports, based on IGF-II transgenic mice, that IGF-II also has immune modulator properties (219, 220). In some situations, in some species, IGF-II may play a role in lymphoid tissue function. Adult humans, unlike rodents, have blood concentrations of IGF-II equal to or greater than those of IGF-I. The regulation of IGF-II production is an active area of research and has produced surprising findings. For example, IGF-II production may be controlled by a rapamycin-sensitive pathway (221).

C. Effect of in vivo treatment on lymphocyte number and in vitro function
In mice treated with rhIGF-1 for 7 or 14 days, analysis of the lymphocyte subsets showed that a large part of the increased spleen weight was due to a doubling in the number of both T- and B lymphocytes (92). The increased thymic mass was also due to a doubling in Thy 1-positive cells (T lymphocytes) and an increase in peanut agglutinin receptor binding (a marker for immature thymocytes), but there were no changes in Thy-1, CD4, or CD8 expression on single or double positive thymocytes. In the spleen, there was a preferential increase in the number or sIg+ cells (B-lymphocytes) compared with T lymphocytes (92). For the splenic T lymphocytes, the numbers of both CD4- and CD8-positive cells were increased. Peripheral lymph nodes were also increased in size and lymphocyte number but in peripheral blood lymphocyte number decreased by 20%, whereas neutrophil number increased. There were no changes in other blood cell numbers (92).

To test lymphocyte function, cells from spleen and lymph nodes were incubated in vitro with mitogens (92). After 14 days of treatment with rhIGF-1, the responses to ConA (T cells), lipopolysaccharide (LPS, B cells), and pokeweed mitogen (both T- and B cells) were increased 4-fold in spleen and doubled in lymph nodes. Cells from rhIGF-1-treated and control mice showed identical responses in a mixed lymphocyte response to allogenic splenocytes, suggesting unchanged antigen-specific T cell responses. To test B cell function, mice were immunized with dinitrophenyl-ovalbumin (DNP₂OA) and given a boost 35 days later and treated with rhIGF-1. When tested in vitro, splenocytes from IGF-treated mice had a doubling of their basal and antigen-stimulated immunoglobulin production. Therefore, there was clear evidence that treatment with rhIGF-1 increased both the number of lymphocytes and their function (92). In comparison, injections of hGH were much less potent and effective (143).

D. Functional effects of IGF-I in vivo: antibody responses
It was clear from the literature (222) and our own studies that B cells are preferentially responsive to rhIGF-1 and show an enhanced immunoglobulin production in vitro (92). To discover whether IGF-I enhanced immune function in vivo, immunization experiments (Fig. 8) were perfomed in retired breeder mice using DNP₂OA as the antigen (206). In the first experiments, mice were treated with rhIGF-1 for 14 days, commencing at the time of an antigen challenge, and showed a dramatically enhanced primary antibody response, as measured by the serum anti-DNP IgG concentration assayed by ELISA. This protocol was then repeated, but a secondary immunization was given after 21 days to test the effect of rhIGF-1 on the memory response to DNP₂OA (206). The anti-DNP IgG concentration at the peak of the secondary response was 4-fold higher in mice treated with rhIGF-1. A second 14-day course of rhIGF-1 treatment (begun 8 weeks after the first course), initiated when lymphoid organ cell numbers had returned to baseline, also increased lymphoid organ cell number. Lastly (Fig. 8), the ability of rhIGF-1 to enhance the antibody response to a suboptimal dose of antigen was tested (206). In this study the mice were treated with rhIGF-1 twice, at the primary immunization and then after 5 weeks at the secondary immunization. Primary and secondary responses to a suboptimal dose of antigen were greatly enhanced, reaching levels similar to those induced by an optimal antigen dose in excipient treated animals (206). In diabetic rats, IGF-I treatment did not improve the primary antibody response to an antigen challenge (223). It is unclear why this study in rats failed to show the effects on antibody generation seen in mice. However, the older literature (126) shows that in hypophysectomized rats, which have very suppressed antibody responses, there is a dramatic enhancement of antibody generation after treatment with GH.

E. Immune reconstitution
IGF-I treatment increases T and B cell number and improves antibody responses, suggesting that it might have a normal role in B and T cell function. To address the site(s) of action of IGF-I, mice were lethally irradiated and then reconstituted with a transplant of 10 million bone marrow cells from syngeneic donors (205). The catabolic effects of the radiation were attenuated by IGF-I with body weight loss reduced, spleen and thymus weight improved, and splenic T cell number and function improved 23 days after transplantation. Treatment with rhIGF-1 also doubled thymus weight, and thymic cell count tripled. In this model IGF-I increased the rate of peripheral lymphocyte repopulation by acting directly on bone marrow progenitors and by stimulating the entry of mature peripheral splenic lymphocytes into S phase of the cell cycle (205). After chemically destroying lymphoid tissues, similar restorative effects of rhIGF-1 have been shown in rats (224). In mice with severe combined immune deficiency (SCID), GH has also been shown to improve T cell engraftment after the transfer of human or murine cells (225). It is unclear whether IGF-I is involved in these effects of GH, although in our hands GH is much less effective than IGF-I at promoting immune reconstitution after bone marrow transplantation.

F. Mechanism of action: apoptosis
Programmed cell death (apoptosis) is fundamental to many levels of the immune system from the development of precursor cells in bone marrow, selection in the thymus, to deletion of mature cells in the periphery (226). How IGF-I expands B and T cell number is unclear (199); it could act positively to potentiate differentiation or it could act passively to enhance survival, e.g. by reducing apoptosis (Fig. 9). IGF-I has marked anti-apoptotic effects (227) in many tissues and cell types, which may be important in normal growth and differentiation, in tumor growth (228), and for the protection of tissues from damage. IGF-I has been especially impressive at protecting the kidney (229), heart (230), and brain (231) from damage after ischemic injury. This protection may, in part, involve anti-apoptotic mechanisms. The bulk of evidence for the identification of IGF-I and the Type I IGF receptor, as powerful inhibitors of apoptosis and survival factors for cells, comes from the field of tumor biology (228). Apoptosis is regulated by a rapidly growing array of families of signaling molecules; as a consequence, the pathway(s) affected by IGF-I will be described in the future. For example, in several interleukin 3 (IL-3)-dependent cell lines, IGF-I can prevent apoptosis after IL-3 withdrawal (232). In some cell types, even where it is a poor mitogen, IGF-I is the key antiapoptotic growth factor (227); therefore, it is reasonable to assume that IGF-I does signal through apoptotic pathways in lymphoid cells. There is, as yet, no direct published evidence that IGF-I increases lymphoid cell number in animals by inhibiting apoptosis, but a theoretical framework for such activity can be postulated, as illustrated in Fig. 9.

View larger version (19K): [in this window] [in a new window]   Figure 9. A proposed scheme of how the GH/IGF-I axis might modulate apoptosis. An insult or injury induces GH resistance and IGF resistance, which adrenal steroids exacerbate. The inhibition of the effects of GH and IGF-I probably enhances the likelihood of damaged cells dying from apoptosis.

	VII. IGF-I in Different Physiological States

Top Abstract I. Introduction II. Local GH Axis... III. GH Administration IV. PRL V. Insulin-Like Growth Factors VI. Actions of IGF-I VII. IGF-I in Different... VIII. GH/IGF-I as Antistress... IX. Therapeutic Potential X. Conclusions References