A Pegylated Growth Hormone Receptor Antagonist, Pegvisomant, Does Not Enter the Brain in Humans

Abstract

Background: GH receptors exist in the hippocampus, cerebral cortex, and hypothalamus, possibly influencing mood, cortical blood flow, and neuronal growth and mediating negative feedback.

Rationale: Pegvisomant is a recombinant mutated GH molecule with high affinity, but little or no activating capability, for the GH receptor. It is used clinically as a GH antagonist.

Hypothesis: Systemic pegvisomant enters brain interstitial fluid via putative choroid-plexus GH receptors, thereby allowing for antagonism of central actions of GH.

Subjects and Location: Six adults requiring a cerebrospinal fluid (CSF) examination for nonneoplastic and noninflammatory syndromes participated at a tertiary medical center.

Methods: Direct assays were conducted of serum and CSF pegvisomant concentrations 18–24 h after sc injection of pegvisomant (20 mg).

Outcomes: Median (range) concentrations of pegvisomant in serum were 215 (74–539) μg/liter and in CSF 0.035 (0.010–0.28) μg/liter (P = 0.016). CSF drug levels were indistinguishable from assay threshold. Corresponding GH values were 0.29 (0.010–1.3) in serum and 0.075 μg/liter (0.01–0.13) in CSF. The geometric mean ratios of serum/CSF pegvisomant and GH concentrations were 5116:1 and 3.5:1, respectively, thus defining a more than 1400-fold difference between mutated and natural GH.

Conclusions: Based upon CSF measurements, a pegylated GH-receptor antagonist does not cross the human blood-brain barrier, thereby sparing inhibition of central nervous system GH actions. Thus, the capability of this antagonist to stimulate GH secretion predominantly reflects its actions outside the blood-brain barrier, such as via the median eminence and/or via suppression of IGF-I concentrations.

Based upon direct cerebrospinal-fluid measurements, a pegylated GH-receptor antagonist does not detectably cross the human blood-brain barrier, thereby sparing inhibition of central nervous system GH actions.

GH secretion is influenced by age, visceral fat mass, sex steroids, and negative feedback by GH and IGF-I (1). Pegvisomant is a selective GH receptor antagonist (2,3,4), which decreases plasma IGF-I levels and elevates blood GH concentrations (5) without reducing systemic GH clearance (6). In principle, augmented GH secretion could reflect diminished negative feedback by lower IGF-I concentrations (7) or direct antagonism of central nervous system (CNS) GH receptors by the drug (1). If the latter is the case, then pegvisomant should be detectable in cerebrospinal fluid (CSF), much as GH, insulin, IGF-I, and a GHRH receptor antagonist are (8,9,10,11). The present study addresses this possibility. Central neural actions of pegvisomant, if documented, would have major implications to interpreting clinical outcomes, because GH is purported to enhance cognition, memory, mood, motivation, and vigor via CNS effects (10,12,13,14,15).

Patients and Methods

Overview

The design entailed giving a single sc injection of pegvisomant (20 mg) (Pfizer Inc., New York, NY) the day before a scheduled elective spinal tap. Eighteen to 24 h later, 5-ml samples of blood and CSF were obtained when the spinal tap was performed. Pegvisomant and GH were measured in both serum and CSF. Peak pegvisomant concentrations are attained in blood about 36 h after single-dose injection (16). The subsequent drug half-life in plasma is even longer (16). For molecules of similar molecular weight like insulin, CSF uptake and washout occur with nominal half-times of 0.5–2 and 3–8 h, respectively (17). Thus, the present study allowed a minimum of 18 h and a maximum of 24 h between pegvisomant injection and the CSF examination to allow for ingress into brain. Although very high doses of pegvisomant (e.g. 1 mg/kg sc) have been used in feedback experiments (5), a dose of 20 mg per subject was chosen here for clinical relevance. Indeed in one study of 112 patients with acromegaly, 62.5% of patients received this daily dose (18).

Exclusion criteria

Patients with known or suspected meningitis, encephalitis, intracranial surgery, epidural or dural disease, multiple sclerosis, and any condition associated with neoplasia or systemic inflammation were excluded. Other exclusion criteria included a history of liver disease or abnormal liver function tests within the last 6 months, chronic renal disease, malignancy, systemic infection, latex hypersensitivity, and/or unwillingness to provide written informed consent.

Inclusion criteria

Volunteers comprised patients prescheduled for a neurologically indicated spinal tap. Subjects had a clinically intact blood-brain barrier and provided Institutional Review Board-approved voluntary witnessed informed consent. Indications for spinal taps are summarized in Table 1.

Table 1.

Clinical features of the study cohort

Age (yr)	Gender	Height (cm)	Weight (kg)	Body mass index (kg/m2)	Diagnosis
45	Female	184	66	22	Syncope
60	Female	161	89	23	Neuropathy
61	Male	192	130	52	Neuropathy
69	Male	168	93	35	Ataxia
49	Female	165	83	21	Neuropathy
52	Male	173	70	22	Neuropathy

Biospecimens

CSF and serum samples were frozen at −70 C and stored before being assayed as a batch. Pegvisomant and GH assays were performed after shipment of the samples on dry ice.

Immunofluorometric assay of pegvisomant and GH

Pegvisomant concentrations were determined by a two-site competitive immunofluorometric assay, exactly as described (19,20). Serum samples were first diluted 100-fold, rendering any interference by GH negligible. For analysis of CSF samples, calibrators were diluted in a buffer mimicking protein content of CSF (PBS/1% BSA). CSF samples were also analyzed without predilution. Microtiter plates were precoated with monoclonal antibody 10A7 before adding pegvisomant standards and clinical samples. Incubation proceeded for 3 h at room temperature, followed by washing and addition of indicator antibody (biotinylated monoclonal antibody 6F1) for 12 h. Activity was quantified after incubation with Europium-labeled streptavidin for 15 min, washing, and time-resolved fluorometry (Wallac, Inc., Turku, Finland). Within-assay and between-assay variability averaged 6.2 and 7.2%, respectively, at a pegvisomant concentration of 1000 μg/liter. Assay detection thresholds were 0.5 μg/liter for serum (corresponding to 50 μg/liter for the undiluted serum sample) and 0.01 μg/liter for PBS and 1% BSA spiked with pegvisomant standard (used to analyze CSF samples).

GH was assayed using a pegvisomant-insensitive two-site competitive monoclonal immunofluorometric assay, exactly as described earlier (5,6). In particular, serum GH concentrations were determined in each sample in duplicate by an assay shown not to cross-react with pegvisomant up to 50 mg/ml (20). The standard was 22-kDa recombinant human GH International Reference Preparation 88/624. All samples were analyzed in one run. Median intra- and interassay coefficients of variation were 6.5 and 8.7%, respectively, at the GH concentrations measured.

Statistical analysis

A paired Student’s t test with unequal variance was used to determine whether there was a statistical difference between the logarithms of serum and CSF concentrations of pegvisomant or GH. Data are given as the median (range). Significance at P < 0.05 was confirmed by the nonparametric signed-ranks test.

Results

Two subjects described mild pain at the injection site. There were no other adverse events. Data from the six individuals studied are given in Fig. 1 (left) as vertical dot plots. Pegvisomant concentrations in serum ranged from 74–539 μg/liter and in CSF from 0.010–0.19 μg/liter. Median values were 215 (serum) and 0.04 (CSF) μg/liter, respectively. Pegvisomant (drug) levels in serum and CSF differed at P = 0.016. Drug concentrations in CSF samples were analytically indistinguishable from assay blank (buffer assayed without pegvisomant addition). Individual values were 0.02, 0.05, 0.01, 0.19, 0.03, and 0.09 μg/liter. GH concentrations are shown in Fig. 1 (right) at a 400-fold scale difference. Serum and CSF GH levels (median, range) were 0.29 (0.01–1.3) and 0.025 (0.01–0.50) μg/liter, respectively (P value not significant). Individual CSF GH values were 0.03, 0.01, 0.01, 0.04, 0.02, and 0.05 μg/liter. When pegvisomant standards were diluted in serum instead, CSF drug concentrations ranged from 0.01–0.28 (median 0.035) μg/liter, and CSF GH levels from 0.1–0.13 (median 0.08 μg/liter). The results did not differ from those obtained using PBS with 1% BSA as the diluent.

Figure 1.

Vertical dot plots of individual-subject serum and CSF concentrations of pegvisomant (left) and GH (right). Data were obtained by non-cross-reacting assays. P denotes probability of falsely rejecting the null hypothesis of no serum CSF difference. Individual values near zero are stated in Results, because they are less than the size of the symbol.

Discussion

The present study rejects the null hypothesis that serum and CSF concentrations of pegvisomant do not differ in men and women studied 18–24 h after sc injection of this GH receptor antagonist (P = 0.016). In fact, the geometric mean ratio of serum/CSF concentrations of pegvisomant was 5116:1 and of GH 3.5:1. Thus, mean fractional CSF uptake was 1462-fold lower for pegvisomant than GH.

Pegvisomant reduces IGF-I concentrations by disrupting GH receptor dimer-dependent signaling in peripheral organs, such as the liver (16). The liver accounts for more than 70% of systemic IGF-I production, but IGF-I is also synthesized in the brain, kidney, immune system, muscle, pituitary gland, heart, bone, and skin under GH drive (1). Although pegvisomant is a synthetic GH analog containing nine amino acid substitutions, the clinically administered form is pegylated (pretreated with polyethylene glycol) to reduce immunogenicity and prolong the plasma residence time. Small amounts of free (nonpegylated) analog might exist or be formed in vivo. In principle, free pegvisomant protein could cross the blood-brain barrier analogously to GH via the choroid plexus and other circumventricular organs (see introductory section). The present analysis shows that immunogenic pegvisomant does not enter CSF in amounts exceeding 0.02% of serum concentrations. According to this outcome, pegvisomant would neither 1) block GH’s negative feedback by inhibiting GH receptors internal to the blood-brain barrier nor 2) antagonize putatively beneficial effects of GH on motivation, cognition, mood, vigor, and memory in patients with an intact blood-brain barrier.

Because acromegalic patients have elevated mean and nadir serum GH concentrations (21), CSF levels of GH would be expected to be somewhat higher in this disease than in normal adults (8,9,10). In the present study, the serum/CSF concentration ratio was 1400-fold higher for pegvisomant than GH. Thus, pegvisomant treatment in acromegaly would not be able to block the actions of CSF GH.

An important conceptual implication relates to the mechanisms that mediate approximate doubling of mean GH concentrations during pegvisomant exposure (5). Inasmuch as pegvisomant administration does not alter the half-life of either endogenous or exogenous GH (6), increased GH concentrations during GH receptor blockade reflect a true increase in GH secretion. The present data signify further that elevated GH secretion in this context would not be readily explicable by direct blockade of central neural GH receptors internal to the blood-brain barrier but more likely via depletion of systemic IGF-I concentrations (5). Although no evidence exists for major GH feedback activity on the external median eminence, this region of the CNS can admit larger proteins (1,8,9).

A caveat is that extremely high systemic concentrations of pegvisomant might yield sufficiently large amounts of unpegylated peptide to enter the CNS in small amounts. However, unlike the case of GH, there is no direct evidence that mutated GH molecules are actually transendocytosed intact by choroidal endothelial cells and/or transferred via other pathways into brain interstitial fluids. Nonetheless, the conclusion that pegvisomant does not cross the blood-brain barrier is indirect in that CSF rather than brain interstitial fluid was sampled.

In conclusion, under the conditions of this study, fractional CNS uptake of pegvisomant is about 1500-fold lower than that of GH in individuals with a clinically intact blood-brain barrier. Given that significant receptor antagonism requires an excess of antagonist over agonist, these findings imply that clinical use of pegvisomant in GH-replete patients will not attenuate important brain actions of endogenous GH.