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 Kopchick, J. J. Articles by Trainer, P. J. Search for Related Content PubMed PubMed Citation Articles by Kopchick, J. J. Articles by Trainer, P. J.

Growth Hormone Receptor Antagonists: Discovery, Development, and Use in Patients with Acromegaly

Edison Biotechnology Institute (J.J.K., E.C.S.), Department of Biomedical Sciences, College of Osteopathic Medicine, Ohio University, Athens, Ohio, 45701; and Christie Hospital (C.P., P.J.T.), Department of Endocrinology, Manchester, M20 4BX, United Kingdom

Correspondence: Address all correspondence and requests for reprints to: Dr. P. J. Trainer, Department of Endocrinology, Christie Hospital, Manchester, M20 4BX, United Kingdom. E-mail: Peter.Trainer{at}man.ac.uk

	Abstract

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

 
An understanding of the events that occur during GH receptor (GHR) signaling has facilitated the development of a GHR antagonist (pegvisomant) for use in humans. This molecule has been designed to compete with native GH for the GHR and to prevent its proper or functional dimerization—a process that is critical for GH signal transduction and IGF-I synthesis and secretion. Clinical trials in patients with acromegaly show GHR blockade to be an exciting new mode of therapy for this condition, and pegvisomant may have a therapeutic role in diseases, such as diabetes and malignancy, in which abnormalities of the GH/IGF-I axis have been observed. This review charts the discovery and development of GHR antagonists and details the experience gained in patients with acromegaly.

I. Introduction

II. GH

A. The GH gene family

B. GH gene structure and regulation

C. GH tertiary structure

III. GHR

A. Class I cytokine receptor superfamily

B. GHR gene and primary structure

C. GHR tissue distribution

D. GHR signal transduction

IV. Biological Activities of GH

V. Development of a GH Antagonist

A. Early GH structure-function studies

B. The importance of GH’s helix 3

C. The GH/GHR interaction

VI. Engineering of a Long-Acting GH Antagonist

VII. In Vivo Experience with Pegvisomant

A. Rodent

B. Primate

C. Human

VIII. Acromegaly

A. Mortality and acromegaly

B. Criteria for "remission"

C. Current treatment of acromegaly

IX. Pegvisomant in Acromegaly

A. Monitoring disease activity during pegvisomant therapy

B. Phase II studies

C. Phase III studies

D. Long-term experience with pegvisomant in acromegaly

E. Outstanding questions for pegvisomant in the treatment of acromegaly

X. Summary

	I. Introduction

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

 
WITHIN 12 yr the GH receptor (GHR) antagonist has developed from a scientific concept to a commercially available product (pegvisomant) of proven efficacy.

GH is a protein that contains 191 amino acids with 2 disulfide bonds and 4 -helices (Fig. 1). Its molecular mass is approximately 22,000 Da. By combining site-specific mutagenesis studies of the GH gene with the in vivo assay of the ability of GH analogs to regulate growth of transgenic mice, a GH antagonist was discovered (1, 2, 3, 4). Glycine in the third -helix of GH was found to be particularly important for GH’s biological activity. If it is replaced with arginine, lysine, or with a variety of amino acids, GH is converted from a growth enhancer to a growth suppressor or a GH antagonist (4).

View larger version (36K): [in this window] [in a new window]   Figure 1. Computer representation of porcine GH crystal structure. The four antiparallel -helical bundles are represented as cylindrical rods and labeled with Roman numerals I–IV. The amino and carboxyl termini are indicated with an N and C, respectively. The numbers indicate amino acid positions at the N and C termini of each independent -helix. [Adapted from S. S. Abdel-Meguid et al.: Proc Natl Acad Sci USA 84:6434–6437, 1987 (30 ).]

Pegvisomant is a GH analog that includes a single-amino-acid substitution at position 120 that generates the antagonist. Additional changes include amino acid substitutions within binding site 1 and a further modification by the addition of polyethylene glycol moieties that increase the half-life and reduce the immunogenicity of the molecule.

Pegvisomant’s development has been paralleled by increased awareness of the need for tight control of the GH/IGF-I axis in patients with acromegaly. Despite the best endeavors of dedicated pituitary surgeons combined with dopamine agonist and somatostatin (SMS) analog therapy, a significant proportion of patients remain inadequately controlled and in need of additional treatment.

The evidence from early clinical studies indicates pegvisomant to be the most potent medical treatment for acromegaly, with serum IGF-I being reduced into the age-related reference range in excess of 90% of patients. A great deal more knowledge will be gained in terms of long-term effects of pegvisomant on both the metabolic consequences of acromegaly and on pituitary adenoma tumor growth as these clinical studies continue. Epidemiological and preclinical studies suggest pegvisomant may also have a role in the management of microvascular complications of diabetes mellitus and certain human malignancies, specifically those in colon, breast, and prostate. This review will not discuss these issues but will review and discuss the discovery, development, and clinical experience with pegvisomant.

	II. GH

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

 
A. The GH gene family
GH is a member of a hormone family that includes PRL and the placental lactogens. The genes that encode these hormones are believed to have evolved over the last 350 million years via the duplication of a common ancestral gene (8). In this regard, the primary structure of GH shares a high degree of identity with both PRL and the placental lactogens (8, 9, 10). Further characterization of these three hormones has demonstrated that they are structurally (10, 11), genetically (8, 9), and functionally (10, 11) similar and clearly represent a distinct family of proteins. Based on structural studies, mouse proliferin, mouse proliferin-related protein, rat decidual PRL-like protein, and somatolactin have been identified as additional members of the GH gene family (12, 13, 14, 15, 16). These four helical bundle hormones have structural homology with other members of the class 1 cytokine family such as leptin and erythropoietin (11).

B. GH gene structure and regulation
The hGH gene is located on the long arm of chromosome 17 as part of a gene cluster composed of five closely related genes (17). This cluster spans 66.5 kb and is composed of the genes for GH, choriosomatomammotrope hormone L, choriosomatomammotrope hormone A, GH variant, and choriosomatomammotrope hormone B (17). The 5' flanking area of the cluster contains a transcriptional control region consisting of the following cis elements (from 5' to 3'): enhancer region, glucocorticoid responsive element, two copies of the pituitary-specific GH transcription factor (GHF-1 or pit-1), a cAMP responsive element, and a TATA box (18).

GH is synthesized and secreted by the somatotroph cells in the anterior lobe of the pituitary gland (8, 10). The GH gene is approximately 3 kb long, consists of 5 exons and 4 introns, and encodes a 217-amino-acid precursor protein (8). An amino-terminal signal peptide is subsequently removed by proteolytic cleavage yielding a mature single-chain polypeptide that contains 191 amino acids with a molecular mass of approximately 22 kDa (8, 10, 11, 19, 20, 21). A 20-kDa form of GH has also been found to be secreted by the pituitary and is produced by alternative splicing of the GH precursor mRNA (22).

It has been established that the synthesis and secretion of GH from somatotrophs is a calcium (Ca⁺)-dependent event during which an increase in cytosolic Ca⁺ is required for GH release (23). GH synthesis and secretion are regulated by at least two hypothalamic peptides, GHRH and SMS, via their opposing effects on intracellular Ca⁺ concentration (Fig. 2 and Ref. 24). The GHRH elicits an increase in intracellular Ca⁺ via its G protein-coupled receptor by increasing the level of cellular cAMP (25). In somatotrophs, an increase in cAMP production leads to an increased transcriptional rate of the trans-acting pituitary-specific transcription factor, pit-1. Subsequently, pit-1 (also termed GHF-1) is responsible for enhancing the transcriptional rate of the GH gene (26). Conversely, SMS has an opposing effect on intracellular Ca⁺ concentration compared with GHRH and, therefore, inhibits GH synthesis and secretion (25).

View larger version (36K): [in this window] [in a new window]   Figure 2. Regulation of GH synthesis and secretion. The hypothalamic hormones GHRH and somatostatin elicit an increase and decrease in circulating GH levels, respectively. Circulating GH binds to the GHR on peripheral tissue such as muscle, liver, and bone to induce the secretion of IGF-I. IGF-I itself then has growth-promoting effects on target tissues. IGF-I can also decrease circulating GH levels via feedback inhibition at either the hypothalamic or pituitary level. [Adapted from J. J. Kopchick and J. M. Andry: Mol Genet Metab 71:293–314, 2000 (267 ).]

C. GH tertiary structure
The three-dimensional structures of porcine GH and hGH have been determined by x-ray crystallography (Fig. 1 and Refs. 6 and 30). As expected from their sequence similarities, these molecules share similar topographies. Both molecules contain two disulfide bridges and a four-helical bundle in which the helices are arranged in an "up-up-down-down" topology. Helices 1, 2, 3, and 4 are located between residues 9–34, 72–92, 106–128, and 155–184, respectively, in hGH. The disulfide linkages in hGH occur between residues Cys35-Cys165 and Cys182-Cys189. Three shorter connective helices also exist in hGH, two between helices 1 and 2 and one between helices 2 and 3. GH also contains a hydrophobic central core that is composed of approximately 20 hydrophobic amino acids. Residues in the lower half of helix 3 are hydrophobic and they are buried in this hydrophobic core (see Fig. 3).

View larger version (23K): [in this window] [in a new window]   Figure 3. Edmondson wheel projection of the third -helix (amino acids 109–126) of bGH and hGH (amino acids 110–127). Each amino acid position is given. Amino acids in red are hydrophilic (top half of wheel) and those in blue (bottom half of wheel) are hydrophobic. Glycine-119 (bovine) and glycine-120 (human) are indicated in black.

	III. GHR

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

B. GHR gene and primary structure
The hGHR gene is localized on the short arm of chromosome 5 in the region p13.1-p12 (34). The 5' untranslated region of this gene contains multiple exons that alternatively serve as exon 1 (35). A secretory signal peptide is encoded by exon 2, exons 3–7 encode the extracellular domain, exon 8 encodes the transmembrane domain, and exons 9 and 10 encode the intracellular domain of the receptor (36). The mouse GHR (mGHR) gene was cloned, characterized, and determined to be similar in size and sequence to the hGHR gene but contains two additional exons (37). These include an exon 4B, which is downstream of exon 4, and an exon 8A, which is upstream of exon 8. Exon 4B encodes an eight-amino-acid segment of the extracellular domain of the receptor and is present in all known mGHR transcripts. Exon 8A serves as an alternative splice site for the mGHR and rat GHR gene (37, 38). Alternative splicing via this exon gives rise to the mouse GH binding protein (mGHBP) transcript that contains the first seven exons of the mGHR gene as well as exon 8A but lacks the transmembrane and cytoplasmic domains. Thus, the transcript for the mGHR does not include exon 8A, as splicing occurs directly from exon 7 to exon 8. In humans, GHBP is produced via proteolytic cleavage of membrane-bound GHR; however, in mice, the mGHBP transcript gives rise to the soluble binding protein (35, 37, 38).

The GHRs are a single polypeptide chain that range from 614–626 amino acids in length with a predicted molecular mass of approximately 70 kDa. However, the observed experimental molecular mass of GHR ranges from 100 to 130 kDa due to posttranslational glycosylation and ubiquitination (33, 35, 39). The extracellular region of GHR contains seven cysteine residues and five potential N-linked glycosylation sites that are highly conserved between species. Also, the GHR contains a 24-amino-acid transmembrane domain and the intracellular Box 1 and Box 2 regions that are characteristic of the class I cytokine receptor family (32, 33). As stated above, members of the class I cytokine receptor superfamily contain an extracellular WSXWS motif (33, 40). In the GHR, this region contains the amino acid sequence YXXFS (tyrosine; X is glycine, serine, lysine, or glutamic acid; phenylalanine; and serine; Ref. 40). The importance of the GHR for the growth phenotype has been shown by disruption of the mGHR gene (41). These animals possess a dwarf phenotype and are a model for Laron syndrome (42). One particular and important characteristic of these animals is their extended lifespan (43). If removal of GH action results in lifespan extension, it would certainly be interesting to show a similar effect when using GH antagonists. However, no data exist in support of this notion.

C. GHR tissue distribution
GHRs are present in many biological tissues and cell types. Although first identified in hepatic tissue, GHRs are present in bone, kidney, adipose, muscle, eye, brain, and heart (44, 45, 46, 47, 48). GHRs have also been identified in various immune tissues including cultured human B cells (49), IM-9 lymphocytes (50), spleen, and thymus (51). The fact that GHRs are so ubiquitous suggests the pleotrophic nature of GH’s effects.

D. GHR signal transduction
Although the mechanism(s) by which GH elicits biological responses are not fully understood, GH-induced GHR dimerization is thought to be the first step in the GHR signal transduction pathway (6, 33, 52). It has been shown that GH-induced GHR dimerization induces the activation of a 121-kDa GHR associated protein (53) that was later identified as Janus-kinase 2 (JAK2). Activation of JAK2s is thought to be the result of transphosphorylation after two JAK2s are brought together by hormone-induced dimerization. JAK2 is a member of the Janus-associated kinase family of cytoplasmic tyrosine kinases (54, 55), consists of 993 amino acid residues, and has a molecular mass of 130 kDa (55). JAK2 noncovalently associates with GHR via the Box-1 region, the area critical for JAK2 binding and phosphorylation (56). Furthermore, once activated, JAK2 is believed to promote self-tyrosine phosphorylation and tyrosine phosphorylation of the GHR at multiple tyrosine residues (57, 58, 59).

Upon GH-induced GHR dimerization and subsequent phosphorylation of JAK2 and GHR, a host of other intracellular proteins is phosphorylated. Among these are several approximately 95-kDa proteins termed the signal transducers and activators of transcription or STATs (60, 61). The STATs are a family of proteins that serve as intracellular signal mediators for cytokine-family members. GH induces the phosphorylation of STAT1, STAT3 (61, 62), and STAT5 (60, 61, 63, 64). Also, the GHR contains several intracellular tyrosine residues that become phosphorylated and provide docking/binding sites for the STAT proteins (57, 64). After binding to the GHR, STATs subsequently become phosphorylated (57, 64), form homo- or heterodimers, enter the nucleus, and regulate GH-specific gene transcription (65). Temporal patterns of GH secretion have been shown to differentially regulate liver STAT5 activation (66).

Recent studies have shed insight into the mechanism by which the GH-induced JAK/STAT signaling pathway is switched off. As well as rapidly activated tyrosine phosphatases [SHP1 and -2 (67)], a novel family of cytokine-inducible inhibitors of signaling, referred to as suppressors of cytokine signaling (SOCS), has been described to down-regulate JAK2 (68, 69, 70). The SOCS family currently consists of eight members [SOCS1–SOCS7 and cytokine inducible Src homology 2-containing protein (CIS)] that participate in a negative feedback loop to regulate cytokine signaling (70). The SOCS genes are highly conserved among species but have limited homology to other family members, suggesting that each member may have a specific role in the feedback signaling (69). The primary target of SOCS activity in the JAK/STAT pathway appears to be at the level of the JAKs and STATs themselves. SOCS-1 has been shown to associate with and reduce the activity of JAK2 (67). SOCS-3 and CIS also have been shown to inhibit the JAK/STAT pathway via competitive binding to STAT5 docking sites on cytokine receptors. Additionally, CIS has been shown to increase proteosome-mediated destruction of the GHR complex (69, 70). Therefore, the method by which SOCS disrupt the JAK/STAT pathway seems to be variable.

	IV. Biological Activities of GH

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

 
In the early part of the last century, it was demonstrated that a pituitary factor, now known to be GH, regulated growth (71, 72). Although a significant physiological effect of GH is the promotion of postnatal longitudinal growth (73), GH also has been associated with alterations in lipid, carbohydrate, nitrogen, and mineral metabolism (74, 75, 76, 77, 78, 79, 80, 81). Other activities of GH include its ability to regulate differentiation of preadipocytes to adipocytes (82), to aid in the maintenance and development of the immune system (83), and to have various effects on the brain and cardiac function (84, 85, 86).

In 1957, Salmon and Daughaday (87) proposed the somatomedin hypothesis, which stated that the observed effects of GH are mediated by the activation of IGF-I. Green et al. (88), in 1985, proposed the dual effector theory of GH action, which states that GH induces differentiation of target cells directly and also induces the secretion of IGF-I. IGF-I was then postulated to promote clonal expansion of the differentiated cells. More recently, IGF-I has been shown to be produced and secreted in response to GH stimulation in many tissue types including liver and bone (Fig. 2 and Refs. 89 and 90). Although all actions of GH are mediated through the GHR, it still is not certain which effects are IGF-I directed and which effects are directly caused by GH.

The IGF-I gene has recently been disrupted conditionally in mouse liver (91, 92), resulting in significantly depressed serum levels of IGF-I, but no corresponding diminution of growth was seen. These data suggest that circulating IGF-I may not be the important agent in growth promotion, but paracrine- and/or autocrine-derived IGF-I may be the salient molecule.

We must point out that the molecular mechanisms resulting in autocrine, paracrine, and endocrine actions of IGF-I are still to be determined. However, independent of the mechanism of IGF-I action, GH stimulation of IGF-I synthesis and secretion is presumed to occur via interaction of GH with the GHR. Therefore, a GH antagonist would be expected to inhibit endocrine, autocrine, and paracrine actions of IGF-I. The correction by pegvisomant of the major metabolic complications of acromegaly, such as insulin resistance, cortisol metabolism, and lipoprotein abnormalities (see Section IX.E.5), indicates that this therapy does more than just lower circulating IGF-I.

Abnormal secretion of GH is observed in multiple disease processes. In children, hyposecretion of GH results in short stature, whereas hypersecretion before epiphyseal fusion yields gigantism. Although GH is clearly not important for longitudinal bone growth in adults, deficiency is associated with reduced well-being, central obesity, increased fat mass, dyslipidemia, and increased mortality. Hypersecretion in adults is usually caused by a pituitary somatotroph tumor and results in acromegaly. In animals, transgenic mice that express bovine (b)GH are significantly larger than nontransgenic controls and develop severe kidney disease (93). A link between GH and the occurrence and progression of malignancies has been shown (94, 95). Elevated levels of GH have also been correlated with diabetic complications (96, 97, 98, 99, 100). In addition, overexpression of GH has been associated with increased renal and glomerular growth (101). Thus, scenarios in which GH levels are elevated have been implicated in several physiological disorders.

	V. Development of a GH Antagonist

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

 
A. Early GH structure-function studies
Early structure-function studies attempted to determine specific regions of the GH molecule that were important for the binding of GH to GHR. Via homolog-scanning mutagenesis, three such epitopes on the hGH molecule were identified (102). These GHR binding determinants include the loop between amino acids 54 and 74, the amino-terminal region of helix 1, and the central portion of helix 4 to the carboxyl terminus. Subsequently, alanine-scanning mutagenesis was performed on these GH determinants to identify specific side chains required for the hGH-hGHR interaction (103). These studies identified a total of 12 amino acids, located in helix 1, the loop connecting helix 1 and 2, and helix 4, that are involved in the ligand-receptor interaction. Alanine-scanning mutagenesis of hGHR revealed that tryptophan-104 and -169 were key amino acids that made contact between GH and the GHR (6, 33, 33, 102, 103).

B. The importance of GH’s helix 3
Independently, after the elucidation of the three-dimensional structure of GH (30), Chen et al. (1) observed that the amino acids composing the third -helix, residues 109–126 in bGH and 110–127 in hGH, are arranged in an amphiphilic orientation (Fig. 3). Furthermore, it was noted that three amino acids, 117, 119, and 122 of bGH, are positioned so that this helix 3 does not form an idealized amphiphilic helix. Because it was previously shown that peptides containing the third -helix of GH possessed low but significant growth-promoting activity (104), it was hypothesized that the generation of a perfect amphiphilic helix 3 would result in a GH analog with enhanced biological activity. Thus, a mutant bGH gene was engineered, bGH-M8, which encoded the following amino acid substitutions: glutamate-117 to leucine (E117L), glycine-119 to arginine (G119R), and alanine-122 to aspartate (A122D) (1). Surprisingly, transgenic mice expressing this bGH analog had decreased circulating IGF-I concentrations and exhibited a dwarf phenotype (1, 3). The growth ratio between the transgenic mice and nontransgenic littermates ranges from 0.58 to 1.00 and was directly proportional to the circulating concentration of bGH-M8 (1). The observed dwarf mouse phenotype, and the fact that bGH-M8 inhibited the binding of ¹²⁵I-bGH to liver membranes, revealed that bGH-M8 was the first reported example of a GHR antagonist (1). In vitro, this GH antagonist was also shown to antagonize other actions of GH including adipocyte differentiation and the insulin-like effect of stimulating glucose oxidation and lipolysis by adipose tissue (2, 105). Similar studies with the hGH analog, hGH G120K, were performed and yielded similar results (106).

Additional bGH analogs, with single amino acid substitutions at positions 117, 119, and 122 were generated to investigate further which residues are critical for the growth-promoting activity of GH. These studies revealed that the glycine at position 119 of bGH and glycine-120 of hGH are critical for the growth-promoting activity of the respective molecule. Expression of these mutated GH genes in vivo resulted in dwarf mice (Fig. 4 and Ref. 4). Thus, alteration of one residue (i.e., glycine-119 of bGH or glycine-120 of hGH) changed the molecule from a growth enhancer to a growth suppressor or GH antagonist. It was further demonstrated that substituting the glycine residue in the third helix with any amino acid other than alanine converted the GH molecule from an agonist to an antagonist (4). Subsequently, these results were confirmed by a second group (6) who demonstrated that substitution of glycine-120 of hGH with arginine resulted in a GH antagonist.

View larger version (176K): [in this window] [in a new window]   Figure 4. Transgenic mouse expressing bGH G119K. The mouse with a dwarf phenotype (right) expresses bGH G119K and is pictured beside a sex-matched nontransgenetic littermate (left).

View larger version (68K): [in this window] [in a new window]   Figure 5. Partial space-filling model of the third -helix of bGH. A, The "cleft" due to the glycine residue in the third -helix of GH is indicated. B, Partial space-filling model of GH. When GH is altered such that the glycine residue is substituted, in this case with arginine, the cleft is filled preventing proper interaction of the GH molecule with a second GHR.

View larger version (80K): [in this window] [in a new window]   Figure 6. "Stick model" of the co-crystal structure of one molecule of hGH (blue) and two molecules of the extracellular domain of the GHR (red and green). The two GHR molecules utilize almost identical binding determinants to interact with different determinants on hGH. Glycine-120 hGH is indicated in yellow interacting with tryptophan-104 of hGHR, in white.

	VI. Engineering of a Long-Acting GH Antagonist

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

Independently, and at the time the GH antagonist was reported by Chen et al. (1, 2, 3, 4), Cunningham and Wells (108) were attempting to engineer more potent GH analogs via identifying critical binding determinants in site 1 of the GH molecule. Eight separate amino acids were identified that, when altered, increased the binding affinity for site 1. Furthermore, it was noted that these mutations had an additive effect and, when combined, greatly increased the affinity of GH analog for binding site 1 (7, 108). Thus, a GH analog was produced that contained these eight amino acid substitutions (which increased the binding affinity of GH to the GHR) plus the original G120K substitution (the amino acid substitution that generates a GH antagonist) and would be expected to yield a GH antagonist with high potency. Furthermore, when these nine mutations were combined with the aforementioned addition of PEG-5000, the resulting molecule, known as PEG-hGH G120K (B2036 peg), was shown to maintain both its binding and antagonistic properties. PEG-hGH G120K (pegvisomant) is now in the final stages of clinical trials, under the trade name Somavert (Pharmacia, Peapack, NJ) for sc injection in the treatment of acromegaly.

After GH binding to the GHR, the complex is internalized. The precise mechanisms responsible for the GH/GHR internalization is outside the scope of this review; however, Refs. 109, 110, 111, 112, 113, 114, 115 describe these events. In this context, the nonpegylated GH antagonist binds to the GHR with approximately the same affinity as GH, forms dimers, and is internalized (52) but cannot transduce an intracellular GH-specific signal. Recently, pegvisomant, like the nonpegylated form of the GH antagonists, has been shown to form dimers with the GHR and to be internalized (52, 116, 117). Thus, the antagonists do not inhibit dimer formation, but presumably prevent "proper" or functional dimerization of the GHR.

As stated above, it was believed that the GH antagonist with eight amino acid residue changes in site 1 would bind to the GHR with increased affinity and that the G120K substitution would prevent GHR dimerization. Although the site 1 amino acid substitutions increase affinity for antagonist binding to GHBP, recent evidence has suggested that the eight amino acid substitutions within site 1 do not increase binding of the nonpegylated GH antagonist for the cell-surface GHR (116). However, of the eight amino acid changes made within binding site 1, two (namely lysine to alanine and lysine to arginine at positions 168 and 172, respectively) are nevertheless critical with regard to site 1 binding of the pegylated antagonist. Pegylation at these lysine residues would block or sterically hinder binding of the antagonist to the first GHR. Substitution of these residues removes potential pegylation sites within binding site 1 and, thus, ensures that site 1 of pegvisomant remains accessible to the GHR (116). Nevertheless, pegylation of the antagonist does reduce site 1 binding affinity and, thus, large doses of pegvisomant are required to effectively antagonize GH action in patients with acromegaly.

Although not specifically pertinent to the efficacy of the GHR antagonist, recent evidence has altered our concept of GHR dimerization, as data suggest that rather than inducing sequential binding of two GHRs, ligand binding may produce a conformational change within preformed GHR dimers. Therefore, GH antagonists containing the G120K mutation may not inhibit the recruitment of a second GHR after site 1 binding, but instead prevent functional dimerization by sterically inhibiting conformational changes within these GHR dimers (116).

	VII. In Vivo Experience with Pegvisomant

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

 
A. Rodent
As discussed previously, glycine at positions 119 and 120 of bGH and hGH, respectively, is critical for their respective growth-promoting activity. Unlike bGHG119R, the G120R molecule is relatively ineffective in rodents (118) and, as a result, pharmacological studies of pegvisomant have concentrated on primates and humans. The treatment of hypophysectomized and GH-deficient dwarf rats with G120R during GH infusion failed to antagonize GH action (118). In these animals, G120R paradoxically stimulated skeletal growth, and co-infusion of G120R with hGH produced an additive effect. In normal mice, pegvisomant lowers serum IGF-I to 70% of baseline values when doses of 4 mg/kg (over 10 times the standard daily dose used for the treatment of acromegaly in humans) are administered. Lower efficacy of pegvisomant in rodents is presumed to be due to the amino acid substitutions within the antagonist having been designed to increase binding to the hGHR. This is supported by in vitro studies in which pegvisomant binds to mouse hepatocyte microsomal membrane preparations with 20- to 50-fold less affinity than similar preparations of human liver. Rat hepatocytes have a 10-fold reduced affinity for hGH than mouse cells. Rodents are therefore not an ideal model of GHR blockade by pegvisomant, but at high doses some suppression of IGF-I is observed.

Thus, mice and rats are not good models in which to determine the efficacy and activity of pegvisomant. However, nonpegylated molecules with a single amino acid change, namely glycine to lysine at position 119 within the third -helix, are potent GH antagonists in mice. Several studies have shown that these GH antagonists inhibit the progression of diabetes-induced glomerulosclerosis (96, 98, 99, 100, 101), inhibit hypoxia-induced retinal neovascularization (119), and inhibit the size and progression of dimethyl butyric acid-induced breast tumors in rodents (120).

B. Primate
Two separate studies in which a 1-mg/kg sc injection of pegvisomant was administered to ovariectomized female rhesus monkeys (Macaca mulatta) demonstrated significant reductions in serum IGF-I and other GH-dependent factors. In the first study (121), serum IGF-I fell by 61% within 3 d of administration after a single dose, was maximally suppressed by d 10, and returned to normal levels within 14 d of drug withdrawal. The reduction in serum IGF-I was accompanied by a significant increase in serum GH. Mean serum GH was 86 ± 11 µg/liter in pegvisomant-treated monkeys compared with 15 ± 6 µg/liter in controls.

In a second study (122), five monkeys were treated with the same 1.0-mg/kg sc dose of pegvisomant at weekly intervals. Serum IGF-I was observed to fall within 24 h of administration and continued to decline until d 5, with a fall of 78 ± 4% from baseline. Serum IGF binding protein (IGFBP)-3 and acid labile subunit (ALS) were also significantly suppressed by d 5 to 30 ± 5% and 36 ± 10% below baseline, respectively. Administration of recombinant IGF-I during GHR antagonism restored serum IGF-I and IGFBP-3 values to normal, but the ALS response was inconsistent (122).

C. Human
A phase I, double-blind, randomized, placebo-controlled, single rising-dose study of pegvisomant in healthy male volunteers has recently been reported (123). Thirty-six subjects (mean age of 25 yr) were randomized into four groups (n = 6) and received either placebo or pegvisomant at a dose of 0.03, 0.1, 0.3, or 1 mg/kg as a single sc injection. Serum pegvisomant concentrations showed a dose-dependent increase with peak concentrations occurring approximately 36 h after administration and values greater than 5000 µg/liter from 24–144 h after administration in the 1.0-mg/kg-dose group. Mean serum IGF-I fell in a dose-dependent manner and reached statistical significance on d 3 with doses of 0.3 mg/kg or more. In the 0.3-mg/kg group, there was 28% suppression of serum IGF-I compared with baseline at 72 h, and 49% on d 5 in the 1.0-mg/kg group. There was no effect of pegvisomant on other hormone systems studied.

	VIII. Acromegaly

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

 
A. Mortality and acromegaly
Acromegaly is the result of GH hypersecretion, in the vast majority of cases, from a pituitary somatotroph tumor. GH-secreting pituitary adenomas may occur as part of the multiple-endocrine neoplasia syndrome type I, and, when GH hypersecretion develops before epiphyseal fusion, pituitary gigantism results.

The first clinical description of acromegaly is generally attributed to Marie and de Souza-Leite (1886; Ref. 124), who noted hypertrophy of the extremities including the face, hands, and feet. In 1887, Minkowski (125) noted that such hypertrophy was accompanied by enlargement of the pituitary gland, and later, in 1894, Tamburini (126) documented the presence of a pituitary adenoma. In 1909, Cushing (127) postulated that the production of a growth-promoting hormone by the hyperfunctioning pituitary resulted in the clinical condition of acromegaly.

Acromegaly occurs with equal frequency in both sexes with a prevalence of between 55 and 59 patients per million and an incidence of 3–4 per million (128, 129). The diagnosis is often made incidentally, as less than 20% of patients actively seek medical attention for a change in appearance or enlargement of the extremities (130). The condition develops insidiously over many years and, in most cases, the diagnosis is established after 4–10 yr of active disease. Clinical features reflect excess secretion of GH, local effects of an expanding pituitary mass, and hypopituitarism and include bony proliferation, coarsening of the facial features, soft tissue swelling, hyperhidrosis, macroglossia, headache, visual disturbance, amenorrhea, impotence, diabetes mellitus, or, more commonly, impaired glucose tolerance, hypertension, carpal tunnel syndrome, and sleep apnea (130). Increased serum IGF-I and the failure of GH suppression to below 1 µg/liter after oral glucose (75 g) administration are diagnostic.

Acromegaly is associated with considerable morbidity and mortality, primarily from cardiovascular causes (128, 129, 131, 132, 133, 134, 135). In one case-note review of 79 individuals with acromegaly treated between 1964 and 1989 in Stoke-on-Trent, United Kingdom, the ratio of observed to expected deaths was 2.68 (132). This review also documented that normalization of serum GH was the best determinant of therapeutic outcome, as the mortality rate for subjects in whom serum GH was suppressed below 2.5 ng/ml (5 mU/liter) was not statistically different than that of the normal population (132). The importance of tight control of the GH/IGF-I axis has also been demonstrated in two long-term outcome studies after surgery, where mortality rates achieved by patients with a biochemically defined postsurgical "cure" were equivalent to those of matched normal controls (136, 137). In contrast a 2.4- to 4.8-fold increased rate of death was observed in those patients with persistent disease activity after treatment (136, 137). Rajasoorya et al. (134), in 1994, reported a series of 145 individuals with acromegaly over the period of 1964 to 1989 in which the last known serum GH value was significantly higher in patients who had died compared with those alive (33 ± 17.5 vs. 4 ± 0.5 ng/ml respectively; P < 0.0002). Other determinants predicting outcome were age and serum GH level at the time of diagnosis and the presence of hypertension or cardiac disease. In the cohort described by Rajasoorya et al. (134), survival, irrespective of treatment, was reduced by an average of 10 yr. Although the most common cause of death is cardiovascular disease, there are significant increases in death rates due to lung infections and possibly malignancies (131, 138, 139). Up to 46% of patients with acromegaly harbor colonic polyps, which are frequently precursors of colonic carcinoma, and a 2.45-fold increased incidence of malignant tumors was reported in one study (139, 140). In a further multicenter retrospective cohort study of 1362 subjects with acromegaly (133), overall cancer incidence was not elevated, but mortality from colonic malignancy was increased. This study also confirmed that serum GH values below 2.5 ng/ml after treatment were associated with a reduction in mortality, and, along with earlier studies, clearly demonstrated the crucial importance of achieving biochemical control in acromegaly and, to that end, the need for aggressive management. Subsequently, a retrospective analysis has alluded to a reduction in mortality after normalization of serum IGF-I by transsphenoidal surgery (137).

B. Criteria for "remission"
Therapeutic goals in acromegaly include protection of the optic chiasm, amelioration of symptoms, preservation of pituitary function, and treatment of the pituitary tumor, but the primary goal of treatment is the restoration of normal serum GH and IGF-I levels, as this is associated with normalization of life expectancy (132, 133, 134). An epidemiologically safe mean serum GH value after treatment (2.5 ng/ml) should, ideally, be accompanied by serum IGF-I reduction into the normal range for age and sex (137). Thus, the recent consensus statement defined acceptable biochemical "cure" as the achievement of an age- and sex-matched serum IGF-I value, in addition to a serum GH level of less than 1 µg/liter after 75 g of oral glucose (141).

C. Current treatment of acromegaly
1. Surgery.
Existing treatment modalities for acromegaly modify disease activity by reducing pituitary GH release (Table 1). Consequently, their efficacy is defined by tumor-dependent characteristics. Surgery is the first-line treatment for acromegaly, as it offers the prospect of cure in a proportion of subjects with rapid reduction of serum GH and improved well-being in the remainder. Ideally, pituitary surgery should completely remove the GH-secreting adenoma while preserving or restoring normal anterior pituitary function and protecting the optic apparatus where appropriate. Numerous factors determine surgical outcome with respect to biochemical cure, but of paramount importance is the choice of pituitary surgeon. In a study conducted in Manchester, United Kingdom, 73 patients with acromegaly were operated on by any one of nine surgeons between 1974 and 1997, and in fewer than 20% of subjects did serum GH levels fall below 2.5 ng/ml (142). In a similar study conducted in Birmingham, United Kingdom, 8 surgeons operated on 89 patients with acromegaly between 1963 and 1993, 33% of whom achieved a serum GH value less than 2.5 ng/ml. However, remission rates (<2.5 ng/ml) increased to 64% when practice was altered, and all transsphenoidal surgery undertaken by a single dedicated pituitary surgeon (143). Other centers have reported similar findings (137, 144, 145, 146). Factors influencing postsurgical GH concentration include pituitary tumor size, degree of extrasellar extension (particularly into the cavernous sinus), and high presurgery serum GH levels (136, 137, 142, 144, 145, 146, 147). Sheaves et al. (145), in 1996, observed a 65% remission rate when the preoperative serum GH level was below 20 mU/liter (50 ng/ml) but only 18% in subjects with values above 100 mU/liter (200 ng/ml). When strict criteria for surgical remission are adopted (i.e., serum GH < 2.5 ng/ml, or normalization of serum IGF-I), surgical remission is possible in 61–91% of subjects with microadenomas, but for macroadenomas, remission rates may be as low as 23% (137, 145, 146, 148). Such variable surgical results have led some to question the role of pituitary surgery in acromegaly for all patients (149, 150).

2. Radiotherapy.
Conventional, three-field, pituitary radiotherapy (CRT) has a well-established role in the treatment of acromegaly, either to effect remission after the failure of surgery or when surgery is contra-indicated or refused. Standard treatment schedules involve the administration of 4000–5000 cGy delivered in 160- to 180-cGy fractions over 5–6 wk. Such regimes do not damage the optic chiasm but do cause hypopituitarism (157, 158). The possibility of second brain tumors after CRT has been the subject of debate, and this question has yet to be satisfactorily answered, but, if anything, the possibility appears to be low (159, 160). Serum GH falls slowly after CRT with, in general, a 50% decline within 2 yr, improving to 75% within 5 yr (161, 162, 163, 164, 165). However, it may take 20 yr before 90% of patients achieve serum GH values below 5 ng/ml (163, 165). Barkan et al. (166), in 1997, reported the failure of CRT to normalize serum IGF-I values despite the achievement of epidemiologically safe serum GH concentrations, but more recent data suggest that serum IGF-I is often normalized after pituitary radiotherapy (165, 166, 167).

Although randomized trials of stereotactic and conventional radiotherapy are lacking, stereotactic radiotherapy is attractive, because, by restricting the treatment field to encompass only tumor, it is possible to deliver a larger dose of highly focused radiation exclusively to the pathological lesion. This may lead to more rapid inhibition of tumor growth and GH secretion, with a lower incidence of hypopituitarism and less damage to surrounding tissues. It may also allow the treatment of tumors beyond the reach of the surgeon, particularly in the cavernous sinus.

There are three forms of stereotactic radiotherapy: proton beam therapy, the Gamma Knife (Electa Instruments AB, Stockholm, Sweden), and the linear accelerator (LINAC). Facilities for proton beam therapy are very restricted such that most experience has been gained with photon therapy (Gamma knife and LINAC; Refs. 168, 169, 170).

3. Medical therapy.
The first effective medical treatment of acromegaly was dopamine agonists that stimulate GH secretion in normal individuals but paradoxically suppress GH secretion in a proportion of patients with acromegaly (171). In 549 cases reported in 31 series in which the dose of bromocriptine ranged from 7.5–60 mg/d, GH levels were reduced below the suboptimal value of 5 ng/ml in just 20% of patients. Seven of the 31 reports measured serum IGF-I, which normalized in only 10% of those treated (172, 173, 174, 175, 176, 177, 178, 179). Newman (180) recently reviewed the efficacy of two dopamine agonists, cabergoline and quinagolide, in acromegaly. Two reports including 28 subjects with acromegaly treated with quinagolide (a dose of 0.15–0.6 mg/d) reported serum IGF-I normalization in 43% of patients, and 5 reports including a total of 96 subjects treated with cabergoline (0.3–7.0 mg/wk) suggest IGF-I normalization is achieved in 34% of patients (181, 182, 183, 184, 185, 186). In the largest of the cabergoline series, 64 patients treated with between 1 and 3.5 g/wk (1 patient received 7 g/wk), 39% were observed to have serum IGF-I concentrations below 300 µg/liter (183). A random measurement of serum GH was less than 2 ng/ml in 46% of patients. However, in the 16 PRL-cosecreting tumors, a random serum GH of less than 2 ng/ml was observed in 50% of cases (183). As these data demonstrate, dopamine agonists are more effective in the 33% of GH-secreting tumors that also secrete PRL (185). Overall, dopamine agonist therapy is of limited efficacy in acromegaly and is less effective than the SMS analog octreotide (174, 175).

Analogs of SMS are the most widely used form of medical therapy for acromegaly. Native SMS inhibits GH release and exists in two forms, SMS-14 and an amino-terminal extended SMS-28, with both peptides being encoded by a common preprosomatostatin gene located on chromosome 3q28 (187, 188). The various actions of SMS are mediated through specific cell-surface receptors expressed in the central nervous system, leptomeninges, anterior pituitary, mucosa of the gastrointestinal tract, and both the endocrine and exocrine pancreas (189). Five human subtypes of the SMS receptor have been cloned, all of which possess seven membrane-spanning domains (190, 191, 192, 193) and are functionally linked to adenylate cyclase through coupling mechanisms involving guanine nucleotide-binding (G) protein (194, 195, 196). Octreotide (Sandostatin, Novartis Pharma, Basel, Switzerland), a synthetic octapeptide, was the first analog introduced into clinical practice and is 45 times more effective than SMS-14 at inhibiting GH secretion (Fig. 7 and Ref. 197). Unlike native SMS, which binds with equal affinity to all five SMS-receptor (SSTR) subtypes, octreotide binds with high affinity to SSTR-2 and -5, with moderate affinity for SSTR-3, but does not bind to SSTR-1 and -4 (196, 198, 199). The acute suppression of GH secretion in response to octreotide administration in patients with acromegaly is dependent on SSTR availability (200), which in turn predicts the long-term effect of octreotide therapy on serum GH and IGF-I concentrations in most patients (201, 202, 203). Maximal suppression of serum GH and IGF-I is seen with sc doses of 300–600 µg/d with little additional benefit from further increments (204, 205). Most long-term studies of octreotide therapy in acromegaly report statistically significant suppression of serum GH, with serum IGF-I normalization in approximately 50–60% of patients (204, 205, 206, 207), although the ability of octreotide to normalize serum IGF-I is dependent on preoctreotide serum GH levels and SSTR number (206). In the largest series analyzing the efficacy of octreotide over 6 months, GH levels fell below 2.5 µg/liter (5 mU/liter) in 22–45% of cases, with normalization of serum IGF-I in 45% (206, 207). The combination of octreotide and bromocriptine may prove more effective than either drug alone, although this finding has been questioned (208, 209). In nonrandomized clinical trials, octreotide is reported to reduce somatotroph tumor size in up to 50% of patients (201, 204, 206). In one study, a reduction in tumor size of more than 20% was observed in 15 of 34 (44%) patients receiving octreotide (207).

View larger version (30K): [in this window] [in a new window]   Figure 7. Peptide structure of native SMS 14 and the SMS analogs octreotide and lanreotide. The presence of D-tryptophan, a nonphysiological amino acid, within the ring structure of these analogs stabilizes the molecule and reduces enzymatic degredation. Half-life is thereby prolonged in comparison to native SMS.

Slow-release lanreotide (Somatuline, Ipsen Biotech Lab; Fig. 7) has a fixed injection dose of 30 mg given at variable intervals of 14, 10, or 7 d. During a six-month, open-label, prospective study involving 14 patients with acromegaly that was not controlled after transsphenoidal surgery and (in 10 cases) pituitary radiotherapy, mean serum GH levels fell from 21.4 µg/liter to 7.3 µg/liter and mean serum IGF-I fell to 272 µg/liter (215). In a further open-label study involving 57 patients of whom 50 received 6 months of lanreotide therapy, serum GH suppressed below 2 µg/liter in 66% and IGF-I normalized in 38% of patients (216), with similar results having been observed other groups (214, 217).

Two studies have suggested that Sandostatin-LAR produces improved suppression of serum GH and IGF-I, compared with lanreotide, using established dosing schedules (218, 219). Together, these studies involved a total of 22 patients who initially received lanreotide and subsequently switched to Sandostatin-LAR. Of the 22 patients involved in both studies, GH was suppressed below 2.5 µg/liter (5 mU/liter) in 10 individuals during lanreotide therapy and in 12 patients during Sandostatin-LAR treatment. Serum IGF-I was suppressed into an age-related normal reference range in 9 patients with lanreotide therapy and 12 patients with Sandostatin-LAR therapy. In both studies, Sandostatin-LAR produced a significant further suppression of mean serum GH compared with lanreotide (218, 219). More recently Chanson et al. (220) have described similar findings in group of 125 patients with acromegaly.

The use of a new formulation of lanreotide (Autogel), four weekly deep sc injections, has been reported to be as effective as conventional lanreotide for use in acromegaly (221). There are, however, no comparative studies of Lanreotide Autogel and Sandostatin-LAR.

Both SMS analogs and dopamine agonists have side effects that may limit their clinical usefulness. In particular, SMS analogs are known to inhibit the release of glucagon, insulin, and serotonin, as well as other gut hormones, and are associated with cholesterol gallstone formation (222).

	IX. Pegvisomant in Acromegaly

Top Abstract I. Introduction II. GH III. GHR IV. Biological Activities of... V. Development of a... VI. Engineering of a... VII. In Vivo Experience... VIII. Acromegaly IX. Pegvisomant in Acromegaly X. Summary References

 
With appreciation of the importance of vigorous control of the GH/IGF-I axis, and even more stringent outcome criteria against which to judge the efficacy of therapy, has come the recognition that despite the use of all current modes of treatment, a significant cohort of patients with acromegaly remain inadequately controlled. Pegvisomant is an analog of hGH and represents an exciting new prospect in the medical management of acromegaly. As previously stated, the ability of current treatment modalities to achieve biochemical remission is inextricably linked to pituitary tumor characteristics. Tumor size and position, pretreatment serum GH, and the expression of SSTRs all influence treatment outcome. Whereas SMS analogs bind centrally to specific tumor receptors to mimic physiological inhibition of GH release, pegvisomant acts on peripheral GHRs, blocking "proper" GHR dimerization to prevent GH-induced signal transduction. This approach offers the prospect of higher normalization rates and greater specificity of effect in comparison to current therapies for acromegaly.

A. Monitoring disease activity during pegvisomant therapy
Clearly, serum GH cannot be used as a marker of disease activity in patients with acromegaly who receive pegvisomant, as pegvisomant makes no attempt to lower circulating GH levels. Furthermore, the structural homology between native GH and the GH antagonist, which differ by only 9 amino acids, means pegvisomant is detected in conventional GH assays, resulting in spuriously high serum GH estimation. In the presence of pegvisomant, the direct measurement of serum GH is possible using a customized two-site GH immunoassay (123). An indirect measurement of serum GH using a modified chemiluminescence assay (empirically corrected for pegvisomant cross-reactivity) has been devised and is reported to correlate well with the results of direct measurement (123).

In the absence of serum GH as a marker of disease activity, reduction of serum IGF-I into the age-related reference range has been the primary goal of pegvisomant therapy in acromegaly.

B. Phase II studies
A phase II, randomized, placebo-controlled, multicenter study of pegvisomant provided proof of concept that pegvisomant lowers serum IGF-I in patients with active acromegaly (223). Forty-six patients with a serum IGF-I value at least 50% above the age-related upper limit of normal received placebo, 30-, or 80-mg of pegvisomant once weekly for 6 wk. Mean serum IGF-I levels were unchanged in the placebo group but fell by 16 ± 4.8% and 31 ± 6.7% in the 30-mg (P = 0.04) and 80-mg (P = 0.001) groups, respectively. Free IGF-I fell by 47% in the 80-mg group (226). Although a dose-related decline in serum IGF-I and other markers of the GH/IGF-I axis was seen, serum IGF-I was normalized in only a minority (n = 3) of patients.

The pharmacokinetics of the drug provided a probable explanation for the low serum IGF-I normalization rate. Phase I studies in healthy volunteers had demonstrated peak serum concentration levels at 36 h and maximal suppression of serum IGF-I on d 5 after a single sc dose of pegvisomant (123). As a reversible, competitive receptor antagonist, sustained concentrations of pegvisomant are crucial to efficacy. With a half-life of approximately 70 h in humans and a weekly dosing schedule, it was suspected that adequate serum concentrations of pegvisomant were not maintained throughout the dosing interval. Pharmacokinetic modeling suggested that the trough concentration of pegvisomant would be increased by 20% by delivering the same weekly dose but divided into a daily dose.

C. Phase III studies
A 12-wk, phase III, multicenter, double-blind, randomized, placebo-controlled study of 112 patients provided convincing evidence of the effectiveness of daily pegvisomant (10, 15, or 20 mg) in patients with a serum IGF-I level at least 30% above the upper limit of an age-related reference range. Short-acting SMS analogs and dopamine agonist therapy were discontinued 2 and 5 wk before study entry, respectively. Clinical/laboratory assessments (including completion of a questionnaire designed to evaluate five symptoms and signs of active disease—soft tissue swelling, arthralgia, headache, perspiration, and fatigue) and ring size measurement using standardized European Jewelers’ rings on the left fourth digit were undertaken at baseline, 2, 4, 8, and 12 wk after commencement of therapy or placebo. Pituitary anatomy was assessed at baseline and after 3 months using magnetic resonance imaging.

Dose-dependent reductions in serum IGF-I, free IGF-I, IGFBP-3, and ALS were observed in the three pegvisomant-treated groups. Mean serum IGF-I levels fell significantly in a dose-dependent manner in the 10-, 15-, and 20-mg treatment groups (Fig. 8). The percentage of patients achieving a serum IGF-I value within the age-adjusted reference range was 54%, 81%, and 89%, respectively (Fig. 9 and Ref. 224).

View larger version (21K): [in this window] [in a new window]   Figure 8. Percentage reduction in serum IGF-I in 111 patients with acromegaly receiving pegvisomant for 12 wk. The data show a dose-dependent decline in serum IGF-I. All patients had active disease before starting pegvisomant, based upon a serum IGF-I level at least 30% greater than the upper limit of an age-related reference range. [Adapted from P. J. Trainer et al.: N Engl J Med 342:1171–1177, 2000 (224 ).]