Discovery and optimization of a pH-responsive ultra-long-acting VHH-based growth hormone mimetic

ABSTRACT

Weekly formulations of long-acting growth hormone (LAGH) effectively treat growth hormone (GH) deficiency in both children and adults while offering improved convenience compared to daily GH. To further enhance patient convenience and compliance, longer-acting GH preparations have become a key research focus. However, due to the inherent short half-life of protein-based drugs, no ultra-long-acting growth hormone or analog products (with dosing intervals longer than once weekly) are currently available worldwide. In this study, we developed a potentially ultra-long-acting growth hormone receptor (GHR) agonist pH.VHH03 with GH-mimetic activity by engineering hinge region and variable region of a VHH antibody. Through optimization of the hinge region, the in vitro cell proliferation activity of this molecule was significantly enhanced. As a result, it demonstrated LAGH-like effects in promoting body weight gain and stimulating IGF-1 secretion in rats. Furthermore, pH-dependent binding was introduced into the variable region. The optimized molecule exhibited a 227-fold difference in dissociation rates between neutral and acidic conditions. This modification prolonged its in vivo efficacy in rats from 6 days to over 15 days, far exceeding the 3-day duration observed with PEGylated GH. Subsequent in vivo experiments in rats confirmed that the final optimized molecule dose-dependently promoted body weight gain, insulin-like growth factor 1 (IGF-1) secretion, tibial growth, and significantly increased growth plate thickness. The in vivo profile fully replicated the physiological activities of GH. Moreover, at medium to high doses, it induced a notably flatter and sustained IGF-1 response compared to PEGylated GH, suggesting a potentially longer duration of pharmacological activities in humans.

Introduction

Growth hormone is secreted by the pituitary gland and regulates multiple endocrine functions. Composed of 191 amino acids, it binds to pre-existing homodimeric growth hormone receptors (GHR) on the cell membrane. Binding of GH to its receptor induces a conformational change of the receptor, transmitting the signal intracellularly via the transmembrane domain and activating multiple intracellular signaling pathways.¹ GH primarily binds to receptors in the liver, stimulating the production of insulin-like growth factor-1 (IGF-1), which regulates protein synthesis, fat, and mineral metabolism, and promotes the growth of the skeletal system and other internal organs throughout the body.² Furthermore, GH activates GHR in the growth plates, promoting longitudinal bone growth during postnatal and adolescent development and contributing to the attainment of peak bone mineral density (BMD). Consequently, children with growth hormone deficiency (GHD) exhibit short stature and delayed bone maturation.³ GH also binds to GHR in adipose tissue and muscle, modulating fat and muscle metabolism. Therefore, adults with GHD may present with premature aging, decreased muscle mass, and increased adiposity (particularly abdominal obesity).⁴

GH replacement therapy has become the standard treatment for inherent GHD. However, due to its short half-life in humans, early generations of GH, e.g., Nutropin, and Norditropin, require daily injections. The associated pain and inconvenience brought by frequent dosing, particularly for pediatric patients, led to significant compliance issues. The resulting high incidences of missing doses can largely compromise treatment efficacy.^5,6

To address the challenges of compliance of daily GH, several long-acting GH (LAGH) with weekly dosing regimen have been developed. Currently marketed LAGH products include Somapacitan (Sogroya®), Somatrogon (NGENLA®), Lonapegsomatropin (SKYTROFA®), and PEGylated recombinant human growth hormone (Jintrolong®). Their mechanisms of action vary: Somapacitan applies fatty acid chain conjugation to GH, enabling reversible binding to serum albumin; this increases molecular size and extends half-life via neonatal Fc receptor (FcRn)-mediated recycling⁷ Somatrogon is a fusion protein where one or two copies of the C-terminal peptide (CTP) from the β-chain of human chorionic gonadotropin are attached to the N- and C-termini of GH; glycosylation of the CTP increases molecular weight, thereby reducing renal clearance⁸ Lonapegsomatropin is a prodrug where somatropin is conjugated to a carrier via a proprietary TransCon linker, designed to release unmodified, endogenous-like GH in the body.^9,10 Jintrolong is a PEGylated GH (PEG-GH), where conjugation with polyethylene glycol increases the molecular weight of recombinant human GH and extends its elimination half-life.^11–13

These weekly dosing LAGH products significantly reduced injection frequency compared to daily GH and markedly improved patient adherence.^14,15 It can be postulated that products with even longer-acting formulations, such as monthly injection, could further reduce frequency and potentially enhance compliance.

While current clinical studies indicated that the adverse effect (AE) profiles of weekly LAGH were not significantly different from those of daily GH,^16,17 several limitations and theoretical concerns remain.¹⁸ These include pronounced peak-trough fluctuations in GH and IGF-1 levels, the potential development of neutralizing anti-GH antibodies, difficulties in cross-comparing the safety and efficacy of LAGH products due to differing technological platforms.¹⁸ The pharmacokinetic (PK) profiles of weekly LAGH formulations differ substantially from those of daily GH, often accompanied by greater IGF-1 variability.¹⁹ The extent of these fluctuations also varies among different LAGH technologies, making it challenging to maintain flat therapeutic concentrations. Although human studies have not yet confirmed significant clinical metabolic consequences from this variability, it remains a concern based on physiological role of IGF-1.¹⁸ Sustained supraphysiological IGF-1 levels shortly after dosing raise safety concerns, as persistently elevated IGF-1 is associated with potential tumor risk.²⁰ Conversely, low GH levels toward the end of the dosing interval may pose an unnecessary risk for infants with severe GHD who are prone to hypoglycemia.^18,21

Therefore, a next-generation ultra-long-acting GH that offers both an extended dosing interval and a flatter IGF-1 response profile would provide significant clinical value by potentially improving safety, efficacy, and patient adherence.

GHR agonistic antibodies

Endogenous GH has a very short plasma half-life of approximately 20 min.²² and is susceptible to degradation by extracellular proteases in the bloodstream and tissues.^23,24 This vulnerability to extracellular degradation also applies to LAGHs like Somapacitan.^25,26 Theoretically, the prolonged systemic and tissue exposure achieved through long-acting modifications may increase the likelihood of protease-mediated degradation. Furthermore, the sequence modifications required for extending half-life carry a potential risk of immunogenicity. Indeed, the development of anti-drug antibodies (ADAs), including neutralizing antibodies, has been observed in human clinical trials of current LAGHs.^16,18

Therefore, further extension of the in vivo duration of action and reducing the dosing frequency beyond that of current weekly LAGHs through direct modification of GH presents substantial challenges. One alternative and promising strategy is to leverage the excellent drug-like properties of antibodies to develop GHR agonistic antibodies that mimic GH function. Multiple antibody-based cytokine mimetics are being successfully engineered, which may establish them as a highly promising class of therapeutic tools.²⁷

Since GHR exists as a homodimer, early studies identified a subset of monoclonal antibodies against GHR possessing agonistic activities. These antibodies were capable of mimicking GH functions in vitro and in vivo.²⁸ Mab263, a murine hybridoma-derived monoclonal antibody,²⁹ is one of the most extensively studied GHR agonist. It can activate intracellular signaling downstream of GHR and promote cell proliferation in vitro.²⁸ However, it remained less potent and less efficacious.^28,29 While Mab263 promotes growth in hypophysectomized rats, it does not possess the insulin-like effects of GH, hence it does not regulate blood glucose metabolism.³⁰ Subsequent discoveries of other GHR agonistic antibodies also revealed weaker in vitro or in vivo activities compared to the GH.^30–32 Additionally, these antibodies were raised against animal GHRs and did not necessarily exhibit cross-reactivities with the human GHR.

Therefore, to obtain an agonistic antibody targeting human GHR with activities comparable to that of endogenous GH, novel discovery and optimization strategies are required.

Discovery and optimization of GH mimetics

The GHR agonistic antibodies discovered to date were traditional heavy chain-light chain antibody structures.^29–32 By contrast, heavy chain-only antibodies (heavy chain variable domains, VHHs, also known as nanobodies) derived from camelids possessed significant structural differences from conventional antibodies, which could lead to differences in their agonistic activities. More importantly, VHH antibodies offered distinct advantages for protein engineering. Their single-domain architecture provided greater conformational flexibility for adjusting the distance and relative orientation between two paratopes,³³ making them an ideal scaffold for constructing cytokine mimetics. Indeed, many cytokine mimetics have been developed using the VHH format.²⁷ In recent years, through engineered design, a variety of VHH-based mimetics have been successfully constructed to mimic the functions of corresponding cytokines.^27,34–38 Previous studies using diabodies and designed ankyrin repeat proteins have demonstrated that the geometry of the antibody (or other scaffolds), as well as the orientation, valency, and linker length of the paratopes, significantly influenced the bias, potency, and efficacy of receptor agonism.^34,39,40 This has also been validated by other teams in their VHH-based cytokine mimetics,^37,38 pointing to possible directions for the optimization of cytokine mimetics.

In addition to renal clearance, the widespread tissue distribution of the GHR renders target-mediated drug disposition (TMDD) a major pathway for the rapid systemic elimination of GH.^41–43 For antibody-based GH mimetics, however, the PK landscape shifts. Their increased molecular size substantially reduces renal clearance, and their half-life is extended via FcRn-mediated recycling. Consequently, with these dominant clearance mechanisms diminished, TMDD may emerge as the predominant elimination route for these engineered molecules. An effective strategy to prolong their in vivo half-life is to confer antibodies with pH-dependent antigen binding, which enables high affinity at neutral pH (e.g., in the bloodstream) to be maintained while affinity under acidic conditions (e.g., within endosomes) is significantly reduced.⁴⁴ Antibodies engineered with this property are termed acid-switched antibodies. This technology was first developed to extend the half-life of Tocilizumab (an anti-interleukin-6 receptor (IL-6R) antibody).⁴⁴ The modified recycling antibody demonstrated significantly improved in vivo PK.⁴⁴ To date, several therapeutic antibodies utilizing pH-dependent binding technology have been approved for marketing, such as Ravulizumab (anti-complement C5),^45,46 Satralizumab (anti-IL-6 R),^44,47 and Crovalimab (anti-complement C5),^48,49 which fully validated the feasibility and effectiveness of this strategy in extending the circulatory half-life of antibodies in vivo.

Current study

Here, we describe the development of a GHR agonistic VHH antibody pH.VHH03. Through dual modifications involving hinge region engineering and pH-dependent optimization of the variable domain, the resulting molecule demonstrated superior in vitro and in vivo functional activities. Furthermore, it exhibited a significantly extended in vivo duration of action in GHD animal models compared to weekly GH products. The results suggest that pH.VHH03 is potentially worth being further developed as an ultra-long-acting GH mimetic.

Results

Enhanced in vitro and in vivo efficacy through hinge region modulation

We obtained a VHH antibody targeting the human GHR through alpaca immunization and phage display library screening. A full-length VHH antibody, VHH-G2 (Figure 1(A)), was constructed by fusing this VHH to the Fc region of human IgG1, with its upper hinge region (UHR) replaced by the corresponding region from IgG2 (Supplemental Figure S1). Although alpaca VHH antibodies naturally possess diverse hinge regions that may potentially influence agonistic activity,⁵⁰ we selected the human IgG2 UHR for clinical development due to its lower immunogenicity risk. In addition, the human IgG2 hinge is known to be more rigid than other isotypes, which often confers stronger agonistic activity for agonist antibodies⁵¹ Following recombinant expression, VHH-G2 demonstrated clear activities in promoting the proliferation of Baf3 cells expressing human GHR (Baf3-hGHR). Its potency was significantly greater than that of the widely referenced GHR agonistic antibody Mab263, yet remained notably lower than the activity of recombinant GH (Figure 1(B)).

Figure 1.

Enhanced in vitro and in vivo efficacy through hinge region modulation.

(A) The schematic illustration of the anti-GHR VHH antibody construct, composed of the VHH variable domain, a hinge region, and a human IgG1 Fc domain (L234A/L235A). The upper hinge region (UHR) is indicated.

(B) Relative cell proliferation activity induced by the parental VHH antibody VHH-G2. A stable Baf3 cell line overexpressing human GHR (Baf3-hGHR) was used for the assay. Numerical values were shown inside the bars. GH (blue) represented positive control and maximal proliferation was designated as 100%. Mab263 (green) was a well-known GHR agonistic antibody. Neg. Ctrl. (purple) was an isotype control antibody targeting an irrelevant antigen.

(C) Cell proliferation of VHH antibodies with modified hinge regions. The left side of the panel listed the specific amino acid sequences of the UHR and the corresponding molecule names. The right side showed the E_max(as a percentage of GH activity) and normalized potency (100/Nor.EC₅₀) values. Baf3-hGHR cells were used for the assay.

(D) Proliferation activities of humanized, hinge-region-modulated VHH antibodies. Humanized VHH-G1.2 and VHH-G1.1 were designated as hzVHH-G1.2 (purple diamonds) and hzVHH-G1.1 (orange diamonds), respectively. Positive controls included GH (green triangle) and PEG-GH (blue squares), while the parental VHH-G2 (teal green diamond) was also included. Neg. Ctr. (gray triangle) was an isotype control antibody targeting an irrelevant antigen. The results demonstrated that humanization preserved the enhanced proliferative activity achieved by hinge region optimization.

(E) Body weight changes of rGHR^KO/hGHR^KItransgenic rats treated with single injection of the humanized antibodies (450 nmol/kg) for 10 days. Purple diamonds: hzVHH-G1.2; Orange diamonds: hzVHH-G1.1; Blue squares: PEG-GH; Gray triangle: Vehicle (Normal saline). Pairwise comparisons: blue, hzVHH-G1.2 vs. PEG-GH; orange, hzVHH-G1.2 vs. hzVHH-G1.1.

Data were presented as mean ± SEM. Statistical significance was determined by two-way ANOVA. ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

To enhance the proliferative activity of this antibody, we designed a series of variants with modifications to the UHR length and amino acid composition (Figure 1(C)). As the UHR serves as the linker between the two VHH domains, its properties directly influence the antibody’s ability to activate GHR. One set of variants was based on the parent molecule: mutating the two cysteine residues in the IgG2 UHR to serine (VHH-G2SS) or retaining only the “ERK” sequence (VHH-G2dCC). These changes aimed to reduce heterogeneity potentially caused by extra disulfide bonds while adjusting the linker’s length and composition. Another set was based on a truncated IgG1 UHR (C220S), with linker lengths ranging from the longest “EPKSSDKTHT” (VHH-G1.10) to the complete absence of any amino acids (VHH-G1.0). All variants shared identical variable region sequences and exhibited no difference in GHR affinity (Supplemental Figure S2), yet showed marked variations in cell proliferation activity. The trends for changes in maximal efficacy (E_max) and potency (indicated by the 100/Nor.EC₅₀ value) were consistent across molecules (Figure 1(C)).

A clear correlation was observed between proliferation activity and UHR length. For variants with the IgG1 hinge (VHH-G1.10 to VHH-G1.0), the activities initially increased and then decreased as the UHR was shortened (Figure 1(C), Supplemental Figure S3A). The variant with the “HT” UHR showed the highest E_max (80%), while the one with “THT” showed the highest potency (100/Nor.EC₅₀ value of 71), indicating an optimal UHR length for maximizing activity. The influence of hinge length was also validated in the three variants derived from the IgG2 UHR: VHH-G2dCC, with a shorter hinge, exhibited a significantly higher E_max (59%) than VHH-G2SS (44%) with a longer hinge. The parent molecule VHH-G2, whose two VHH domains were connected by a hinge containing two cysteines, demonstrated an intermediate spatial distance and correspondingly intermediate activity (Figure 1(C)).

Although VHH-G2dCC and VHH-G1.3 possessed UHRs of identical length, their proliferation activities differed significantly due to distinct amino acid compositions (Figure 1(C)). This result suggested that, in addition to length, the specific amino acid sequence of the hinge—potentially by influencing the relative binding angles of the two VHH domains—is also a critical factor for modulating the in vitro activities.

To evaluate the in vivo efficacy of the functionally optimized molecules, we selected VHH-G1.2, which had the highest E_max. To concurrently validate the correlation between in vitro and in vivo activities, we also selected VHH-G1.1, which had a notably weaker E_max. These molecules were humanized, designated as hzVHH-G1.2 and hzVHH-G1.1, respectively, for subsequent in vivo studies. Humanization did not alter the variable region sequences, affinity (Supplemental Figure S3B), or cell proliferation activity (Supplemental Figure S3C).

For in vivo efficacy assessment, PEG-GH served as the positive control. Results from in vitro cell proliferation assays indicated that hzVHH-G1.2 and hzVHH-G1.1 had lower E_max but greater potency (higher 100/Nor.EC₅₀) compared to PEG-GH (Figure 1(D)).

We utilized a transgenic rat model—with rat GHR knockout and human GHR knock-in (rGHR^KO/hGHR^KI)—for the evaluation of in vivo efficacy. This model responds specifically to human growth hormone (hGH), exhibiting significant increases in body weight gain rate, body length, and tibial length upon administration. It exhibits a blunted growth phenotype due to the inability (or minimal capacity) of endogenous rat GH to activate the human GHR,⁵² which phenotypically resembles the surgically hypophysectomized rat model. Importantly, this genetic model effectively circumvents the need for the complex and labor-intensive hypophysectomy procedure. This also facilitates the screening of our molecules, which have poor cross-reactivity to rat GHR. Referring to the dose used in the preclinical study of Somapacitan,⁵³ we selected a pharmacologically saturating dose (450 nmol/kg) for testing. Groups included vehicle (saline), PEG-GH, hzVHH-G1.2, and hzVHH-G1.1.

The vehicle group exhibited minimal baseline growth (model background). The PEG-GH group showed pronounced weight gain in the first 3 days, significantly exceeding the vehicle group, followed by a decline in growth rate to near-vehicle levels, marked by a distinct inflection point. The hzVHH-G1.2 group maintained rapid weight gain for the first 6 days, 3 days longer than PEG-GH, while the hzVHH-G1.1 group maintained it for 7 days, 4 days longer than PEG-GH (Figure 1(E)). The order of rapid weight gain rates (slope during the rapid growth phase) was: PEG-GH > hzVHH-G1.2 > hzVHH-G1.1. Conversely, the order of duration of weight gain effect was the opposite (Figure 1(E)). Notably, the ranking of in vitro proliferation E_max aligned perfectly with the ranking of in vivo rapid growth rates, not with EC₅₀. Final body weight gain was determined by both the net growth rate during the rapid phase and the effect duration. Differences in final weight gain among the three molecules were modest, with no significant difference between PEG-GH and hzVHH-G1.2, but hzVHH-G1.2 was significantly higher than hzVHH-G1.1 (Figure 1(E)).

Serum IGF-1 levels served as a biomarker for evaluating the in vivo pharmacodynamic (PD) effects of GH and its analogs in both animals and humans.^54,55 In this experiment, both hzVHH-G1.2 and hzVHH-G1.1 significantly stimulated IGF-1 secretion in rats. The ranking of peak IGF-1 levels induced by the three molecules matched the ranking of their rapid body weight gain rates, and the duration of IGF-1 elevation corresponded closely with the duration of weight gain effect (Supplemental Figure S3D). A particularly striking observation was that the IGF-1 kinetic profiles induced by hzVHH-G1.2 and hzVHH-G1.1 were markedly different from that of PEG-GH. Their IGF-1 levels increased in a sustained and plateau-like manner, lacking the pronounced peak typically observed on day 2 with PEG-GH and other marketed LAGH products.⁵³

pH-dependent binding significantly prolongs in vivo efficacy duration

The in vivo efficacy of the hinge-optimized molecule hzVHH-G1.2 was conclusively demonstrated in rGHR^KO/hGHR^KI rats, where it exhibited a markedly longer duration of growth-promoting action compared to PEG-GH. To mitigate GHR-mediated TMDD and further extend its PK profile, we engineered the variable region of hzVHH-G1.2 to adjust the pH-dependent binding to GHR. The objective was to identify mutants that maintained robust in vitro functional activities while demonstrating a large window for pH-dependent binding to the receptor—specifically, stable GHR binding at neutral pH (7.4) and rapid dissociation under acidic conditions (pH 5.5).

We introduced histidine mutations and performed phage display screening to obtain a series of mutant clones. Four clones expressing the full-length VHH antibody (with the same hinge and Fc as the parent hzVHH-G1.2) exhibited a clear pH-dependent window, with a pH 5.5/pH 7.4 k_off ratio exceeding 40-fold. However, their neutral pH affinity was notably reduced compared to the parent molecule (Figure 2(A)). Given that GHR exists as a preformed dimer on the cell membrane,⁵⁶ the bivalent VHH antibody likely engaged the dimeric receptor in an avidity-driven manner. We therefore reasoned that measuring avidity (with immobilized GHR and the antibody as the analyte) would better reflect the cell surface binding state. The avidity measurements at pH 7.4 for these four mutants and the parent molecule showed K_D values ≤3.43E-10 M, with kinetic parameter rankings similar to their affinity rankings, except for pH.VHH03 and pH.VHH04 where avidity differences were observed (Figure 2(A)).

Figure 2.

Prolonged in vivo efficacy of VHH antibodies with pH-dependent engineering.

(A) Binding affinity and avidity of VHH antibodies with pH-dependent engineering. Kinetic parameters were determined by biolayer interferometry (BLI). Affinity: VHH antibodies were immobilized onto Anti-Human Fc Capture (AHC) biosensors and human growth hormone binding protein with His tag (hGHBP.His) was used as the analyte; Avidity: Biotinylated hGHBP.His was immobilized onto Streptavidin (SA) biosensors, and then VHH antibodies were used as the analytes. Blue curves represent association and dissociation at pH 7.4; Red curves represent association at pH 7.4 followed by dissociation at pH 5.5. pH.VHH03 showed the largest ratio (>200-fold), indicating a strong pH-dependent dissociation switch. Kinetic parameters of affinity and avidity are listed below the binding curves: a full 1:1 fitting mode was used at pH 7.4; a partial fitting mode was used to extract the k_offvalue at pH 5.5.

(B) Proliferation activities of Baf3-hGHR cells induced with pH-dependent VHH antibodies. pH.VHH03 exhibited the highest maximal efficacy (E_max = 95% of GH activity); The potency (indicated by 100/Nor.EC₅₀) of pH-dependent VHH antibodies decreased sequentially from pH.VHH01 to pH.VHH04.

(C) Body weight changes over 15 days of rGHR^KO/hGHR^KIrats following single injection of VHH antibodies. Rats received a single subcutaneous dose (450 nmol/kg) of the indicated antibodies. Parent hzVHH-G1.2 promoted weight gain for approximately 6 days. pH.VHH03 sustained rapid weight gain for >15 days with no clear inflection point till the study endpoint.

(D) Serum IGF-1 levels in rGHR^KO/hGHR^KIrats after single dosing. hzVHH-G1.2 produced a moderate but sustained elevation of serum IGF-1 which lasted approximately 6 days. pH.VHH03 elicited a plateau-like IGF-1 response that remained elevated for at least 15 days.

(E) Serum concentration of VHH antibodies over 15 days after single dosing in transgenic rats. BLOQ (Below the Limit of Quantification) was indicated with a dashed line. Parent hzVHH-G1.2 exhibited rapid clearance consistent with TMDD. pH.VHH03 and pH.VHH04 showed a markedly extended PK and reduced TMDD, with detectable levels at 15 days after dosing.

Symbols: Black square, GH; Green triangle, hzVHH-G1.2; Purple inverted triangle, pH.VHH01; Orange circle, pH.VHH02; Dark red diamond, pH.VHH03; Teal, pH.VHH04; Black circle, Vehicle.

Data were presented as mean ± SEM. Statistical significance was determined by two-way ANOVA: ns, not significant; *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Color code for pairwise comparisons: red, pH.VHH03 vs. hzVHH-G1.2 at each time point; purple, orange, and teal indicated comparisons of pH.VHH03 vs. pH.VHH01, pH.VHH02, and pH.VHH04, respectively, at the endpoint.

The five molecules displayed significant differences in their in vitro cell proliferation EC₅₀, a ranking that closely paralleled their avidity, suggesting a correlation between avidity and functional potency. Notably, pH.VHH03 demonstrated the highest E_max in the proliferation assay (95%), significantly surpassing the parent hzVHH-G1.2 (85%) and approaching the activity level of recombinant GH (Figure 2(B)).

All four molecules engineered for pH dependency at a 450 nmol/kg dose demonstrated extended duration of body weight gain in rGHR^KO/hGHR^KI transgenic rats. While pH.VHH01 showed a discernible inflection point in weight gain around day 11, the other three molecules (pH.VHH02, pH.VHH03, and pH.VHH04) exhibited no such inflection point at the study endpoint on day 15, indicating that their growth-promoting effects likely persisted beyond 15 days. This represented a significant extension of growth duration compared to the approximately 6-day duration observed for the parent molecule, hzVHH-G1.2 (Figure 2(C)). Among them, pH.VHH03 achieved the greatest terminal body weight gain, significantly outperforming the parent molecule, pH.VHH01, and pH.VHH04.

The duration of serum IGF-1 elevation correlated with the length of the body weight gain effects. The pH-dependent mutants induced a significantly more prolonged IGF-1 response than the parent molecule. pH.VHH03, in particular, maintained elevated IGF-1 levels even at day 15, significantly higher than pH.VHH01 and pH.VHH02, suggesting that its weight gain effects likely extended beyond 15 days (Figure 2(D)). Consistent with the parent molecule pre-pH engineering, all mutants elicited a sustained and plateau-like IGF-1 profile, devoid of the sharp peak characteristic of PEG-GH (Figure 2(D)).

Subsequent PK analysis confirmed significant improvement for the pH-dependent mutants. Notably, pH.VHH03 and pH.VHH04 displayed a PK profile nearly devoid of the pronounced TMDD observed with the parent molecule (Figure 2(E)). Interestingly, the degree of TMDD observed in the PK curves inversely correlated with the avidity k_off value at pH 5.5: faster dissociation under acidic conditions corresponded to weaker TMDD. Since pH.VHH03 exhibited virtually no TMDD, this result suggested that its acidic dissociation rate may be near-optimal, and further increases (as seen in pH.VHH04) may not confer additional PK benefit.

In summary, pH-dependent engineering led to molecules with: 1) significantly prolonged duration of body weight gain and increased terminal body weight in rGHR^KO/hGHR^KI rats (Figure 2(C)); 2) markedly extended duration of IGF-1 stimulation (Figure 2(D)); and 3) substantially improved PK profiles with attenuated TMDD (Figure 2(E)). pH.VHH03 appeared to be the optimal molecule, demonstrating the highest terminal weight gain and, based on endpoint IGF-1 levels, a growth duration exceeding 15 days. This is substantially longer than the ~6 days for its parent molecule and far exceeded the 3-day effect of PEG-GH at an equivalent dose (Figure 1(E)).

An intriguing finding is that among the five molecules tested, pH.VHH03 was not the best in terms of proliferation 100/Nor.EC₅₀ or the fastest in pH 5.5 dissociation. However, it possessed the highest cell proliferation E_max and the largest pH 5.5/pH 7.4 k_off ratio (Figure 2(A)). These results suggested that the in vivo efficacy of pH-dependent GHR agonistic antibodies correlated with several parameters, including pH 7.4 avidity K_D, pH 5.5 avidity k_off, and in vitro proliferative activity. A more precise, quantitative analysis of these correlations warrants further investigation.

Given that pH.VHH03 exhibited the best overall performance in both body weight gain and IGF-1 stimulation in vivo, we performed additional characterization of this molecule. Multi-concentration avidity measurements under neutral and acidic conditions revealed that pH.VHH03 possesses a pronounced pH-dependent binding profile (Supplemental Figure S4), with a pH 5.5/pH 7.4 k_off ratio that closely matched the value obtained from single-concentration screening. In contrast, the parental molecule (hzVHH-G1.2) before pH-dependent engineering showed no significant difference in avidity between pH 7.4 and pH 5.5. Notably, native GH behaved similarly to hzVHH-G1.2, indicating that the GH–GHR interaction is not intrinsically pH-dependent (Supplemental Figure S4). These results suggest that the pH-dependent binding of pH.VHH03 is primarily attributable to the two histidine residues introduced into complementarity-determining region 3 (CDR3), a finding further supported by in silico structural modeling of the VHH/ GH binding protein (GHBP) complex (Supplemental Figure S5). The predicted low immunogenicity risk (Supplemental Table S1) and favorable developability profile (Supplemental Table S2, S3) together support pH.VHH03 as a lead molecule for further validation and development.

pH.VHH03 promoted growth in rGHR^KO/hGHR^KI transgenic rats in a dose-dependent manner

To systematically evaluate the in vivo PD profile of pH.VHH03 in the rGHR^KO/hGHR^KI transgenic rat model, we conducted a single-dose escalation study (experimental design detailed in Table 1). Throughout the 24-day observation period, pH.VHH03 significantly promoted body weight gain at all seven tested doses, with terminal body weight increases ranging from 30 to 100 g, demonstrating a clear dose–response relationship (Figure 3(A)).

Table 1.

Experimental design of the pharmacodynamic studies.

Test article	Route	Dose (nmol/kg)	Frequency	Number of animals	Readouts	Blood sampling schedule
pH.VHH03	s.c.	1, 3, 10, 50, 150, 450, 900	Single dose	8/group	Daily body weight, IGF-1 concentration, tibia length	Day 0, 1, 3, 5, 7, 9, 11, 13, 15, 17, 19, 21, 23. Blood samples are collected before each administration.
GH	s.c.	4.55	QD * 18

Figure 3.

pH.VHH03 promoted rGHR^KO/hGHR^KI rat growth in a dose-dependent manner.

Vehicle and daily administration of GH (4.55 nmol/kg) were served as controls.

(A) Dose-dependent effects of pH.VHH03 on body weight gain;

(B) Tibia length on day 24;

(C) Serum IGF-1 concentration;

(D) AUC of IGF-1 concentrations after a single subcutaneous administration in transgenic rats.

Data were presented as mean ± SEM (n = 8). Statistical significance: pH.VHH03 vs. vehicle: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; pH.VHH03 vs. GH: #p < 0.05, ##p < 0.01, ###p < 0.001, ####p < 0.0001. Ordinary one-way ANOVA followed by Tukey’s multiple comparisons test.

Notably, all single-dose pH.VHH03 groups exhibited superior body weight gain compared to the daily-administered GH group (4.55 nmol/kg, equivalent to 0.1 mg/kg, the optimal effective dose in this model) during the first 10 days post dosing. Importantly, a single dose of 10 nmol/kg pH.VHH03 achieved non-inferior terminal weight gain compared to the cumulative effect of daily 4.55 nmol/kg GH injections over 24 days, despite the GH group receiving a much higher total molar dose (4.55 nmol/kg × 18 administrations). The 450 nmol/kg pH.VHH03 group maintained rapid weight gain for 17 days, substantially longer than the duration observed with PEG-GH at an equivalent dose (Figure 1(E)), underscoring the exceptional in vivo longevity of pH.VHH03. Furthermore, the 900 nmol/kg group sustained rapid weight gain for up to 20 days, with both terminal body weight and growth rate significantly surpassing those of the daily GH regimen. This prolonged growth-promoting effect suggested that pH.VHH03 has the potential for a longer dosing interval in rats, providing crucial rationale for subsequent design of clinical dosing regimen.

We have also measured the tibia length at the end of the study for the evaluation of the effects of the molecule on rat bone growth (Figure 3(A)). pH.VHH03 induced a clear, dose-dependent increase in tibial growth. A statistically significant difference compared to the vehicle group emerged at the 3 nmol/kg dose. Furthermore, pH.VHH03 demonstrated superior efficacy over daily GH starting at the 50 nmol/kg dose. These results confirmed that pH.VHH03 retained the dose-dependent, bone-growth-promoting capability characteristic of recombinant GH.

In the multi-dose escalation study, the IGF-1 response exhibited a biphasic, dose-dependent pattern. A clear dose-dependent increase in the peak concentration of serum IGF-1 (day 1) was observed at lower doses of pH.VHH03. However, at and above the 50 nmol/kg dose, the initial peak did not increase further and even showed a slight decline. Subsequently, a new, broader IGF-1 elevation emerged gradually, peaking around Day 7. With further dose escalation (150, 450, and 900 nmol/kg), the duration of IGF-1 elevation was significantly prolonged to at least 23 days, and the stimulation profile transitioned to a sustained, plateau-like state with a corresponding increase in the area under the serum concentration-time curve (AUC) (Figures 3(C,D)). Critically, this evolving IGF-1 profile closely paralleled the dose-dependent trends in body weight gain and tibial growth. This correlated response pattern strongly corroborated the long-acting growth-promoting efficacy of pH.VHH03. In contrast, due to the short half-life of human GH in rats (approximately 0.5 h.⁵⁷), serum IGF-1 levels are expected to return to baseline within 24 h after a single hGH injection. As blood samples in this study were collected either pre-dose or 24 h post-dose, no significant IGF-1 stimulation was detected in the GH control group.

pH.VHH03 induces a dose-dependent increase in growth plate thickness in rGHR^KO/hGHR^KI transgenic rats

The growth plate is a critical site for longitudinal bone growth, and changes in its thickness directly reflect the activity of bone growth.⁵⁸ To further investigate the effect of pH.VHH03 on skeletal development, we quantified the thickness of the proximal tibial growth plate in rats on day 11 post-administration.

The results demonstrated that, compared to the vehicle group, pH.VHH03 significantly increased growth plate thickness across three dose groups (10, 50, and 150 nmol/kg) in a clear dose-dependent manner (Figure 4(A,B)). The increase in growth plate thickness indicates active chondrocyte proliferation, providing further mechanistic support for the bone-growth-promoting action of pH.VHH03. This finding aligns with the trends observed in body weight gain (Figure 3(A)) and tibial length increase (Figure 3(B)), offering histological validation of the in vivo efficacy of pH.VHH03.

Figure 4.

pH.VHH03 induced a dose-dependent increase in growth plate thickness in rGHR^KO/hGHR^KI rats.

(A) Representative safranin O–fast green staining of tibial tissue sections of transgenic rats at day 11 after a single administration of pH.VHH03.

(B) Quantification of growth plate thickness. Growth plate thickness (μm) was calculated as growth plate area (μm²) divided by growth plate length (μm). Data were presented as mean ± SEM (n = 9). Statistical significance: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 versus vehicle control (ordinary one-way ANOVA followed by Tukey’s multiple comparisons test).

pH.VHH03 prolongs its PK exposure and continuously activates IGF-1

To comprehensively characterize the in vivo pharmacological profile of pH.VHH03, we conducted an integrated PK/PD study in 8–9-week-old male rGHR^KO/hGHR^KI transgenic rats, using PEG-GH as a comparator. Both molecules were administered at two dose levels: 50 and 150 nmol/kg. Following a single administration, blood samples were collected at 0, 6, 12, 24, and 48 h, and then every 24 h up to 360 h for systematic analysis of serum IGF-1 levels and drug concentrations.

In the PEG-GH groups (50 and 150 nmol/kg), IGF-1 stimulation peaked on days 1–2 (Cmax ~550 ng/ml and ~750 ng/ml, respectively) and then rapidly declined to baseline levels (~50 ng/ml), indicating a duration of IGF-1 stimulation of approximately 4–5 days. In contrast, pH.VHH03 induced a sustained elevation of IGF-1. Although its peak level was lower than that of PEG-GH, pH.VHH03 maintained IGF-1 at a flat plateau of approximately 400–500 ng/ml, which persisted until the study endpoint (day 15) (Figure 5(A)). This flat and prolonged IGF-1 stimulation profile may help avoid large fluctuations in IGF-1 levels, offering a potential safety advantage for clinical application.⁵⁹

Figure 5.

pH.VHH03 demonstrated sustained and prolonged activity in IGF-1 stimulation and PK profiles in transgenic rats.

(A) Serum IGF-I concentration – time profiles of pH.VHH03 in transgenic rats after a single s.c. dose of 50, 150 nmol/kg;.

(B) Mean plasma concentration – time profiles.

Data are presented as mean ± SEM (n = 8).

A systematic analysis of PK parameters (Figure 5(B), Table 2) revealed that pH.VHH03 exhibited significantly superior PK characteristics compared to PEG-GH.

Table 2.

Comparison of PK parameters of pH.VHH03 and PEG-GH in transgenic rats.

PK Parameter	PEG-GH (nmol/kg)		pH.VHH03 (nmol/kg)
	50	150	50	150
T1/2(hr)	16.37	38.56	42.20	39.54
Tmax(hr)	15.0	21.0	15.0	19.5
Cmax(pmol/ml)	132.02	465.35	184.62	572.78
AUClast(hr*pmol/ml)	3630.56	14,274.93	12,423.69	41,728.50
Vz_F_obs(ml/kg)	322.62	600.45	247.28	224.14
Cl_F_obs(ml/hr/kg)	14.64	10.69	4.04	3.83
MRTlast(hr)	21.51	23.84	60.86	63.30

Half-life (T_1/2): pH.VHH03 demonstrated a T_1/2 of 39.54–42.20 h, compared to 16.37–38.56 h for PEG-GH. At the 50 nmol/kg dose, the T_1/2 of pH.VHH03 was approximately 2.6-fold longer; Exposure (AUC_last): pH.VHH03 showed markedly greater systemic exposure, with an AUC_last approximately 3 times higher than that of PEG-GH at equivalent doses; Clearance (Cl_F_obs): The clearance of pH.VHH03 (3.83–4.04 ml/hr/kg) was substantially slower than that of PEG-GH (10.69–14.64 ml/hr/kg); Volume of Distribution (Vz_F_obs): pH.VHH03 had a smaller volume of distribution (224.14–247.28 ml/kg) versus PEG-GH (322.62–600.45 ml/kg), suggesting more confined distribution to the blood and extracellular fluid, which may contribute to higher bioavailability; Mean Residence Time (MRT_last): The MRT_last of pH.VHH03 was 60.86–63.30 h, about 2.7–2.9 times longer than that of PEG-GH (21.51–23.84 h), confirming its prolonged presence in vivo.

Integrating the IGF-1 and PK data, the long-lasting growth-promoting effect of pH.VHH03 was closely linked to its distinct PK profile. While PEG-GH requires weekly dosing in human patients, the significantly extended exposure duration of pH.VHH03 in rats suggested the potential for a longer dosing interval in humans—a finding that awaits validation in clinical trials.

Discussion

Long-acting GH mimetic

This study described the successful development of a GHR agonistic antibody based on the VHH (nanobody) format, derived from alpaca immunization and phage display library screening. Through VHH engineering strategies, including hinge region optimization and pH-dependent modification, we have generated a novel anti-GHR agonistic antibody with high bioactivity and an extended duration of action. While retaining potency and efficacy of GHR activation, this engineered antibody significantly prolonged the drug’s half-life and the duration of IGF-1 stimulation, achieving growth-promoting efficacy comparable to the weekly PEG-GH in a preclinical model of GHD at a potentially reduced dosing frequency.

The GHR agonistic antibody features a unique design involving both hinge region engineering and pH-dependent antigen binding. Hinge region modifications were key to maximizing receptor activation efficiency. In vitro cell proliferation assays confirmed that its activity is comparable to that of native GH. Comprehensive in vivo functional characterization demonstrated that this antibody effectively mimicked the physiological functions of GH, including promoting body weight gain, stimulating IGF-1 secretion, increasing growth plate thickness, and enhancing tibial growth in rats. Crucially, the pH-dependent engineering, aimed at reducing receptor-mediated clearance, fundamentally improved its PK profile. A single dose (450 nmol/kg) sustained growth promotion in transgenic rats for >15 days, far exceeding the 3-day duration observed with PEG-GH, and other LAGHs (Supplemental Figure S6A, S6C) at the same dose, highlighting its ultra-long-acting properties.

This study has demonstrated for the first time that by applying GHR VHH nanobody format with modification of hinge region and pH-dependent mutation, it is possible to create a molecule that can fully mimic GH activity while extending its working duration. Given the established clinical evidence that long-acting GH preparations (e.g., weekly formulations) significantly improve patient adherence compared to daily injections.^14,15,60, the VHH agonist developed in this study—with its comprehensive GH-mimetic activity and exceptional longevity—holds promise for extending the dosing interval longer than a week. This positions it as a potential novel, highly effective ultra-long-acting therapeutic option for GHD, which could dramatically enhance treatment compliance.

Sustained and plateau-like IGF-1 stimulation

In transgenic rats, the hinge-optimized VHH antibody elicited a notably flat and sustained elevation in IGF-1 levels. Importantly, this favorable kinetic profile persisted following pH-dependent engineering. The multi-dose escalation study of pH.VHH03 in rats revealed that: at the 50 nmol/kg dose, the IGF-1 response reached a clear plateau. With further dose escalation, the duration of IGF-1 elevation was significantly prolonged, and the stimulation peak appeared progressively delayed while increasing in magnitude. This overall plateau-like IGF-1 profile at medium-to-high doses contrasts sharply with the significant peak-trough fluctuations induced by other LAGH products, as observed in our experiment (Supplemental Figure S6B, S6D) and reported in non-clinical and clinical studies.^53,61 Such fluctuations complicate IGF-1-guided dose titration, and sustained supraphysiological IGF-1 levels pose a potential physiological concern.¹⁸ In contrast, the VHH antibody developed here displayed unique, sustained IGF-1 kinetics. In transgenic rats, it achieved body weight gain comparable to PEG-GH, yet elicited a significantly flatter IGF-1 response with a lower peak (Cmax). This demonstrated that comparable growth-promoting efficacy can be achieved with a lower Cmax and a more sustained IGF-1 elevation.

Based on these data, we reasonably speculate that this VHH antibody may exhibit a similar profile in humans: achieving growth promotion equivalent to PEG-GH while maintaining IGF-1 levels within a flat, physiological range without significant fluctuations—a pharmacodynamic profile more akin to that of daily GH in clinical practice.⁶² For pediatric patients with growth hormone deficiency, this suggests a novel therapeutic potential—effectively promoting linear growth while potentially mitigating the risks associated with excessively high or widely fluctuating IGF-1 levels, thereby possibly enhancing treatment safety and physiological tolerability.

Furthermore, the property of extending the duration of IGF-1 stimulation with increased dose—without proportionally elevating the peak IGF-1 level—offers the possibility of further prolonging the in vivo effect and reducing the dosing frequency through dose escalation. This implies that the ultimate dosing interval may be determined by the formulation’s concentration rather than by tolerability limits related to IGF-1 levels.

VHH antibodies and hinge modulation

Although numerous early studies investigated GHR agonistic antibodies, these were all based on traditional heavy-light chain formats. They exhibited a low incidence of agonistic activity and generally weak functional potency, failing to fully recapitulate the physiological functions of native growth hormone.^28,30–32 In a previous study, only 2 out of 14 antibodies (mab263 and 1C9) showed weak proliferative activity in BaF-B03 cells expressing human GHR, with maximal efficacy reaching only 5.4% and 6.4% of that induced by GH.²⁸ The low activity of Mab263 was consistent with our observations in the Baf3-hGHR proliferation assay (Figure 1(B)).

In contrast, among the anti-GHR VHH antibodies discovered through alpaca immunization in our study, the frequency of clones possessing agonistic activity was markedly higher (Supplemental Figure S1B) than that reported in a previous study,²⁸ and their proliferative potency significantly exceeded that of Mab263. Beyond potential differences in epitope recognition, this advantage likely stems from the intrinsic structural properties of VHHs. Compared to conventional antibodies, the two paratopes of a bivalent VHH format may be positioned at a shorter distance and with a more favorable relative orientation, thereby inducing a GHR conformational change closer to that triggered by GH. Our initial construct, which incorporated the more rigid IgG2 hinge region, further underscores the advantage of the VHH platform for discovering potent GHR agonists.

The greater advantage of VHH antibodies lies in their exceptional suitability for protein engineering. In subsequent hinge region (functionally serving as the linker) optimization, we found that agonistic activity varied with hinge length. At an optimal length and amino acid composition, the VHH agonist achieved in vitro cell proliferation potency comparable to that of recombinant GH, establishing a critical efficacy foundation for antibody-based GH mimetics. This dependence of activity on hinge properties aligns with similar observations made by other teams in bispecific VHH formats,^34,36 suggesting a universal regulatory role of linker-mediated spatial constraints across different formats of cytokine mimetics.

Furthermore, we demonstrated that altering the amino acid composition of the hinge, independent of length, also significantly modulated biological activities. Combined with a previous report on the differential effects of flexible versus rigid linkers,³⁶ this strongly suggested that the physicochemical properties of the hinge (e.g., flexibility, charge, hydrophobicity) precisely fine-tuned the distance and relative orientation between the two VHH domains. This spatial arrangement ultimately determined the efficiency with which the antibody induced the active conformation of GHR. Our work thereby provided a framework for the rational design of hinge region length and sequence to optimize the activity of future VHH-based therapeutics.

Correlation between avidity and in vitro/in vivo activity

During the pH-dependent engineering of the variable region to mitigate TMDD, we observed that a reduction in neutral pH avidity generally led to decreased in vitro proliferative activity. However, within a certain range (on the order of 10 pM K_D), this decrease was not drastic (less than a 10-fold decrease in 100/Nor.EC₅₀). Notably, while pH.VHH01 showed a significantly reduced affinity compared to the parent molecule, its 100/Nor.EC₅₀ remained unchanged. In contrast, the molecule with the lowest affinity (343 pM K_D) exhibited a pronounced decline in both in vitro and in vivo activity. This suggested a correlation between avidity and functional activity, indicating that maintaining an affinity in the ~10 pM range may be crucial for preserving potent cell proliferation. Previous studies have reported an optimal affinity window for agonistic antibodies rather than a simple “higher is better” relationship.⁶³ As we did not systematically test a broader panel of mutants with varying affinities, we cannot definitively confirm a similar optimal affinity relationship in our system.

While numerous successful cases exist where pH-dependent engineering extends the in vivo half-life of antibodies targeting soluble proteins,^44,49,64,65 the application of this strategy to membrane proteins remains less common. A notable successful example is a pH-dependent antibody targeting Protease-activated receptor 2 (PAR2), which demonstrated significantly prolonged PK and efficacy.⁶⁶ However, this approach is not universally applicable. For instance, an antibody targeting Asialoglycoprotein receptor 1 (ASGR1)—a receptor with extremely high expression levels on target cells—still exhibited rapid clearance despite its engineered fast dissociation in endosomes.⁶⁷ Thus, the pH-dependent strategy for extending efficacy does not succeed for all non-shedding membrane protein targets.

In addition to these general limitations, achieving significant PK extension for a GHR agonistic antibody presented two further, specific challenges distinct from antibodies against soluble antigens, primarily due to its requirement for avidity binding to the dimeric GHR:

Mechanistic hurdle with membrane proteins

For pH-dependent antibodies against soluble antigens, achieving a sufficient affinity differential between neutral and acidic pH can improve PK, even for high-affinity blocking antibodies.⁶⁸ This is because the antibody-antigen complex can be recycled via FcRn; within the endosome, the antibody releases the antigen, which is then degraded. This mechanism is not feasible for membrane-bound targets like GHR. The antibody must dissociate from its membrane-bound receptor within minutes in the early/sorting endosome to be available for FcRn recycling; failure to do so results in lysosomal degradation.⁶⁹

The avidity dilemma for an agonist

GHR exists as a pre-dimerized receptor on the cell membrane,⁵⁶ and its activation by our antibody requires bivalent avidity binding.⁷⁰ To maintain high in vitro activity, a very high neutral pH avidity (K_D ~10 pM) is necessary. However, upon internalization, the same bivalent antibody must rapidly dissociate from GHR in the acidic endosome to avoid degradation. Among our tested molecules, only pH.VHH03 and pH.VHH04 showed minimal TMDD, which in our experimental system required an acidic pH (5.5) dissociation rate (k_off) approaching 0.01 s⁻¹. This creates a significant engineering challenge: achieving extremely high avidity at neutral pH for activity and extremely low avidity at acidic pH to circumvent TMDD. Nevertheless, we identified pH.VHH03, which largely retained the parent molecule’s activity while nearly eliminating receptor-mediated clearance, resulting in the optimal pharmacodynamic outcome—the greatest terminal body weight gain.

Other potential benefits of the GH mimetic

Existing LAGH products were developed by modifying the GH molecule itself. This chemical modification can increase the risk of ADA formation, including neutralizing antibodies,⁷¹ which could theoretically diminish the already insufficient activity of endogenous GH in patients with GHD. Although current clinical data indicate that these ADAs have not led to significant efficacy loss or safety issues, they remain a concern which requires long-term monitoring.^16,18 In contrast, the antibody-based GH mimetic is theoretically free from this issue. Due to its entirely distinct amino acid sequence from GH, any ADAs generated would not target endogenous GH, thereby eliminating this specific risk.

From another perspective, while the emergence of ADAs was common in children receiving daily GH therapy and was often not clinically significant,⁷² switching to a GH mimetic could eliminate any lingering concerns regarding efficacy and safety for these patients. This could be particularly advantageous for the subset of patients with severe GHD, especially those with Type 1A isolated GHD (IGHD), where high-titer anti-GH antibodies have been shown to significantly impact clinical outcomes.^72,73 For such patients, a GH mimetic may represent a superior alternative to other GH analogs.

Furthermore, for some patients with growth hormone insensitivity,⁷⁴ the GH mimetic may offer distinct clinical benefits. Given its potentially different binding mode and epitope on the GHR compared to native GH, it might remain effective in cases where GHR point mutations specifically weaken GH binding, but do not affect mimetic binding. In such scenarios, the GH mimetic could theoretically demonstrate superior clinical efficacy compared to native GH or its analogs.

Limitations and future directions

Here, we have presented the discovery and optimization of a GHR agonistic antibody that possesses the potential to have future dosing interval longer than once weekly while maintaining its efficacy in human. However, several limitations still exist and caution should be taken when interpreting the results.

Firstly, it should be noted that this work primarily focused on the engineering and PD optimization of the GHR agonistic antibody. The detailed biological mechanisms and downstream signaling pathways were not explored herein, and a thorough assessment will be the focus of our next-stage investigation and is not within the scope of this study.

Secondly, although our preclinical data in transgenic rats indicate that pH.VHH03 may induce a longer duration of body weight gain than PEG-GH, Somapacitan (Supplemental Figure S6A) and GX-H9.⁷⁵ (a GH-Fc fusion protein, also known as TJ101 or Eftansomatropin alfa, Supplemental Figure S6C), we acknowledge the limitation of not having included other internationally approved LAGH products, such as lonapegsomatropin (SKYTROFA^Ⓡ) and somatrogon (NGENLA^Ⓡ). We also recognize that the transgenic rat model lacks physiological GH regulation, has fundamentally altered endocrine feedback loops, and does not address species-specific FcRn behavior or TMDD, all of which could potentially exaggerate the apparent PD duration. Nevertheless, we have demonstrated in hypophysectomized wild-type rats—a classical GH-deficient model—that pH.VHH03 significantly outlasts PEG-GH in terms of both body weight gain duration and IGF-1 elevation (Supplemental Figure S7A-C). Finally, formal translational PK/PD modeling to predict human dosing intervals was not performed. Therefore, the actual dosing interval in humans—whether biweekly, monthly, or otherwise—can only be established through clinical PK/PD studies, which were beyond the scope of this study.

Thirdly, compared with existing LAGHs, the plateau-like IGF-1 profile induced by pH.VHH03 more closely resembled that of daily GH, yet it still differed from the pulsatile nature of physiological IGF-1 secretion. The mechanisms underlying this kinetic pattern warrant further investigation, and whether this flat IGF-1 kinetics can be replicated in humans, as well as its long-term safety, must be addressed in future clinical studies.

Fourthly, although the lead molecule showed a low predicted immunogenicity risk by in silico analyses, immunogenicity prediction cannot fully reflect the real-world behavior of the molecule regarding ADA. Such studies and other safety aspects must be systematically validated in clinical settings.

Conclusion

In this study, we discovered and optimized a VHH-based GH mimetic that effectively recapitulated the key physiological functions of GH in vitro and in vivo. The engineered molecule demonstrated a significantly extended duration of action and a unique, sustained IGF-1 stimulation profile in rats, markedly differentiating it from currently marketed LAGH products. This successful development of an ultra-long-acting GHR agonistic antibody not only offers a promising therapeutic candidate but also validates a novel engineering paradigm for creating potent and tunable cytokine mimetics. This agonistic antibody, named GenSci134, has completed comprehensive preclinical safety and efficacy assessments and has now entered clinical development (NCT07016802). To our knowledge, it represents the first antibody-based cytokine mimetic to advance into clinical trials. We anticipate that its clinical performance will address significant unmet needs in the current management of growth hormone deficiency.

Materials and methods

Protein preparation

Recombinant human GH was expressed in mammalian cells. PEG-GH was prepared according to the method described in patent CN101385858B.⁷⁶ Mab263 was purchased from Santa Cruz Biotechnology (sc-57161). Two forms of the human hGHBP were used: one with a C-terminal 6x His tag (hGHBP.His) and another fused to the N-terminus of mouse IgG1 Fc (hGHBP-mIgG1Fc). All other recombinant proteins were produced by Biointron Biological Inc. Briefly, the amino acid sequences were reverse-translated into nucleotide sequences and cloned into eukaryotic expression vectors, followed by transient transfection of CHO or HEK293 cells. Target proteins were purified by affinity chromatography using protein A, protein G, or nickel-affinity resin as appropriate. The purified proteins met quality control specifications, as confirmed by SDS-PAGE, size-exclusion HPLC (SEC-HPLC), and endotoxin testing, before being used in subsequent experiments.

Engineering pH-dependent binding into the VHH variable region

Library design and phage display selection

To engineer pH-dependent antigen binding, histidine mutations were introduced into the CDRs of the parental VHH. Three separate phage display libraries were constructed, each targeting a single CDR (CDR1, CDR2, or CDR3). Libraries were generated using a “soft” mutagenesis strategy, wherein each codon within the targeted CDR was synthesized as a mixture of 80% wild-type nucleotide and 20% nucleotides encoding histidine.

Panning strategy

Selections were performed using biotinylated hGHBP.His as the antigen. For each panning round, the phage library was first allowed to bind the antigen at pH 7.4. After washing with PBS (pH 7.4), specifically bound phage particles were eluted by incubation in PBS at pH 5.5 for 10 min. Eluted phage was used to infect E. coli TG1 cells for amplification, followed by subsequent rounds of panning. After three rounds, individual clones from the output pools were picked for further analysis.

Primary screening of soluble periplasmic extracts

Selected clones were cultured for soluble expression of the VHH domain in the bacterial periplasm. Periplasmic extracts were prepared and analyzed for pH-dependent binding kinetics using an Octet HTX system. Briefly, biotinylated hGHBP.His was captured on Streptavidin (SA) biosensors. The biosensors were then dipped into the periplasmic extracts to monitor association at pH 7.4, followed by dissociation in buffers at both pH 7.4 and pH 5.5.

Hit confirmation and protein production

Clones exhibiting a pronounced difference in dissociation rate (k_off) between the two pH conditions were identified using Octet Analysis Studio software. Unique sequences were selected, subcloned into a mammalian expression vector to create a VHH-Fc fusion format, and produced by transient transfection in CHO or HEK293 cells for full characterization.

Affinity and avidity measurements

Instrumentation and software. Binding kinetics were measured using an Octet HTX system (Sartorius). All data were processed and fitted using the Octet Analysis Studio software.

Determination of binding constants

For standard affinity measurements, sensorgram data are globally fitted with a 1:1 binding model to obtain the association and dissociation rate constants (k_on and k_off). The equilibrium dissociation constant (K_D) is calculated from the ratio k_on/k_off. To specifically analyze pH-dependent dissociation, a partial fitting mode is used where only the dissociation phase is fitted to extract the k_off value at each pH.

Affinity (monovalent) assay

The monovalent binding affinity of antibodies for the soluble receptor is assessed as follows. Antibodies (2 µg/ml) are first captured onto Anti-Human Fc Capture (AHC) biosensors. The loaded biosensors are then dipped into a solution containing 100 nM of monomeric human GH-binding protein (hGHBP.His) in pH 7.4 buffer for the association phase. Subsequently, dissociation is monitored by transferring the sensors to buffers at either pH 7.4 or pH 5.5.

Avidity (bivalent) assay

The avidity of the bivalent antibody for an immobilized receptor is assessed using a reversed format. Biotinylated hGHBP.His (2 µg/ml) is immobilized on Streptavidin (SA) biosensors. These biosensors are then exposed to a 100 nM solution of the antibody in pH 7.4 buffer to measure association. Dissociation kinetics are similarly recorded in pH 7.4 and pH 5.5 buffers.

Cell proliferation assay

Cell line construction

A lentiviral plasmid overexpressing the human GHR gene (UniProt ID: P10912) was constructed. High-titer lentiviral particles were produced by plasmid transfection, followed by concentration. The resulting lentiviral solution was used to infect Baf3 cells (obtained from CTCC, GDC0656), respectively. After infection, transduced cells were selected under puromycin pressure to establish polyclonal resistant cell pools. GHR protein expression was assessed by flow cytometry. Positive polyclonal cells were then subjected to single-cell cloning via limiting dilution. Following expansion and validation, a stable Baf3 cell line overexpressing human GHR was established and designated as Baf3-hGHR.

Cell culture and plating

The proliferative activity of the stably transfected Baf3-hGHR cell line was assessed using a Cell Proliferation/Cytotoxicity Assay Kit (DojinDo, CK04). Briefly, Baf3-hGHR cells were washed twice with assay medium consisting of RPMI-1640 (Gibco, 11875–093), 3% fetal bovine serum (Gibco, A5669801), and 50 µM 2-mercaptoethanol (Sigma, M3148), followed by a 4-h starvation period in the same medium. Cells were then seeded into 96-well plates at a density of 40,000 cells in 50 µL per well.

Sample treatment and incubation

Test articles were serially diluted in the assay medium using a 4-fold dilution series. A 50 µL volume of each dilution was added to the cell-containing wells (n = 3 technical replicates per concentration), resulting in a final volume of 100 µL per well. The plates were then incubated for 18 h at 37°C.

CCK-8 detection

After the incubation, 50 µL of a 1:5 diluted CCK-8 solution was added to each well. Plates were incubated for an additional 3 h, after which the absorbance at 450 nm was measured using a microplate reader.

Data analysis

Dose–response data were analyzed using GraphPad Prism software. A nonlinear regression curve was fitted using the “log(agonist) vs. response – Variable slope (four parameters)” model to determine the half-maximal effective concentration (EC₅₀) and the plateau value (Span). The E_max of a test sample was calculated as: E_max (%) = [Span(sample) / Span(GH)] × 100. The relative potency is reported as a normalized EC₅₀ value: 100/Nor.EC₅₀ (%) = [EC₅₀ (GH) / EC₅₀ (sample)] × 100.

Rat in vivo studies

Animal model

A transgenic Sprague-Dawley (SD) rat model with rat GHR knockout and human GHR knock-in (rGHR^KO/hGHR^KI) was generated via CRISPR/Cas9 technology by Cyagen Biosciences (Suzhou). These rats exhibit a GH-deficient phenotype, characterized by reduced body size and weight compared to age-matched wild-type SD rats. All animal studies were approved by the Institutional Animal Care and Use Committee (IACUC) of Changchun GeneScience Pharmaceutical Co., Ltd (approval number: CCGS2024052).

Experimental design and dosing

Male and female rGHR^KO/hGHR^KI rats aged 7–9 weeks were randomized by body weight into groups of 6–8 animals. All substances were administered via subcutaneous injection in the neck region. The negative control group received an equivalent volume of vehicle. Positive control (PEG-GH or GH) and test articles were diluted in the same vehicle to achieve the target dose and injected at an equivalent volume.

Body weight monitoring

Body weight was recorded daily post-dosing.

Sample collection and processing

Blood samples (~100–150 µL) were collected from the jugular vein according to the sampling schedule. Blood was placed in clotting-promoter tubes, kept at 2–8°C for 2 h, and then centrifuged twice at 1,500 × g for 10 min at 4°C. The resulting serum was aliquoted and stored at ≤−70°C until analysis.

IGF-1 and PK analysis

Serum IGF-1 concentrations were quantified using a commercial sandwich ELISA kit (R&D Systems, #SMG100) according to the manufacturer’s instructions. Serum concentrations of the test antibody were determined using a validated ELISA method developed in-house. Briefly, microplates were coated with recombinant hGHBP.His protein, and bound antibody was detected using a horseradish peroxidase-conjugated goat anti-human IgG polyclonal antibody (Bethyl, A80-304P). The assay’s quantitative range was 2–1600 ng/ml. Standard curves were fitted using a 4-parameter logistic model with fixed weighting. Each analytical run included one set of standards and two sets of high, medium, and low-quality controls, all assayed in duplicate. PK parameters were derived from the serum concentration-time data by non-compartmental analysis using Phoenix WinNonlin software (Version 8.2, Certara). Reported parameters included, where data permitted, the maximum observed concentration (C_max), time to C_max (T_max), elimination half-life (T_1/2), AUC, and mean residence time (MRT).

Bone and growth plate analysis

At the study endpoint, tibiae were collected for morphometric and histological assessment. The proximal end of one tibia was placed upright on a caliper to measure bone length. The contralateral tibia was fixed in 10% neutral buffered formalin, decalcified, and processed for paraffin embedding. Sections were stained with Safranin O/Fast Green to visualize the epiphyseal plate, allowing for measurement of growth plate thickness and assessment of chondrocyte morphology.

Data analysis

All statistical analyses and curve fitting were performed using GraphPad Prism software (Version 10).

Abbreviations

ADA

Anti-drug antibody

CTP

C-terminal peptide

FcRn

Neonatal Fc Receptor

GH

Growth hormone

GHR

Growth hormone receptor

GHBP

Growth hormone binding protein (Extracellular domain of GHR)

GHD

Growth hormone deficiency

IGF-1

Insulin-like growth factor 1

LAGH

Long-acting growth hormone

PEG-GH

PEGylated recombinant human growth hormone

PK

Pharmacokinetic

PD

Pharmacodynamic

TMDD

Target mediated drug disposition

UHR

Upper hinge region

VHH

Variable domain of Heavy chain of Heavy-chain antibody

100/Nor.EC₅₀

[EC₅₀ (GH) / EC₅₀ (sample)] × 100

Funding Statement

This work was supported by Changchun GeneScience Pharmaceutical Co., Ltd.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19420862.2026.2691350