A Snake Toxin Derivative for Treatment of Hyponatremia and Polycystic Kidney Diseases

Visual Abstract

Abstract

Key Points

• MQ232, a disulfide-bond reticulated peptide derived from a natural snake toxin, was optimized as a new aquaretic drug candidate.

• MQ232 showed very low acute and chronic toxicity in rat and a biodistribution in mice strongly in favor of the kidney organs.

• MQ232 induced a sole aquaretic effect and demonstrated high in vivo activities on hyponatremia and polycystic kidney disease models.

Background

Vaptans were developed at the end of the previous century as vasopressin type 2 receptor antagonists. Tolvaptan is the most prescribed vaptan for hyponatremia and autosomal dominant polycystic kidney disease (ADPKD). However, its use is not as widespread as it should be due to price issues, a narrow therapeutic window, and some side effects. With the aim of discovering new efficient and safer vasopressin type 2 receptor antagonists, we screened animal venoms and identified several peptide toxins. Among them, mambaquaretin 1 (MQ1) displayed unique biological properties in that regard that it was the starting point for the development of a potential drug candidate.

Methods

Human T-cell assays and bioinformatics were used to mitigate MQ1 immunogenicity risk. MQ232 biodistribution in mice was performed by positron emission tomography. Pharmacodynamics, pharmacokinetics, and acute and chronic toxicity tests were performed on control rats. A rat experimental model of desmopressin-induced hyponatremia, ex vivo mice model of kidney cysts, and mice orthologous model of ADPKD were used to validate MQ232 efficacy in these pathologies.

Results

Three mutations were introduced in MQ1 to mitigate its immunogenicity risk. A fourth gain-of-function mutation was added to generate MQ232. MQ232's safety was demonstrated by a first toxic dose as high as 3000 nmol/kg and a strong kidney organ selectivity by positron emission tomography imaging, while showing almost no interaction with the liver. MQ232's efficacy was first demonstrated with an effective dose of 3 nmol/kg in a hyponatremic model and then in polycystic kidney models, on which MQ232 significantly reduced cyst growth.

Conclusions

We demonstrated, using diverse translational techniques and minimizing animal use, MQ232's safety and efficacy in several rodent models of hyponatremia and ADPKD.

Introduction

The vasopressin type 2 receptor (V2R) is expressed in the distal convoluted tubule and collecting ducts of the kidneys and regulates body fluid homeostasis.^1–3 Its blockade by vaptans is an efficient and validated therapeutic strategy for hyponatremia^4–8 and autosomal dominant polycystic kidney disease (ADPKD).^9–13 Hyponatremia is a prevalent electrolyte disorder characterized by a plasma sodium concentration below 135 mmol/L. It can be induced by an inappropriate secretion of antidiuretic hormone^14–18; congestive heart failure,^19–21 especially when patients are resistant to diuretics^22,23; and also cirrhosis,^24–27 mainly associated with refractory ascites.^28–32 Finally, a large proportion of hyponatremic patients are not treated correctly¹⁴ while hyponatremia is associated with a higher mortality risk^33–36 and greater hospital costs and readmission rates.^37–39 The market-available vaptans, conivaptan (Vaprisol, Cumberland Pharmaceuticals), mozavaptan (Physuline, Otsuka, Japan), and tolvaptan (Samsca, Otsuka Pharma), this last being by far the most used, have limited usage primarily because of their hepatotoxicity, drug–drug interactions, and the challenge for physicians to establish an appropriate, patient-tailored dosage.⁴⁰

Polycystic kidney diseases (PKDs) are a group of ciliopathies causing significant kidney failure in children and adults with no curative treatments.⁴¹ ADPKD is induced by mutations in either the PKD1 or PKD2 gene encoding multimembrane-spanning polycystin-1 and polycystin-2, respectively.^42–45 The recessive form of PKD, autosomal recessive PKD, is due to mutations in the PKHD1 gene encoding for fibrocystin.^46,47 It is suggested that increased cAMP production has a critical role in the formation and expansion of cysts.^48–50 PKDs initiate a ceaseless formation and growth of fluid-filled cysts, leading to the loss of kidney function. Tolvaptan is the only approved drug for adult patients with ADPKD with fast deterioration of their kidney functions.^51–55 Patients are often treated for hypertension by liver-metabolized molecules, making drug–drug interactions difficult to deal with. Emerging studies reported debates among nephrologists with concerns regarding tolvaptan's narrow therapeutic window.^56,57 Autosomal recessive PKD on its own is still an orphan disease.

We screened animal toxins for peptides active on human V2R to address these V2R-related pathologies by vaptan-unrelated molecules with the hope of generating a safe and efficient drug candidate. Animal toxins have evolved for millions of years, allowing them to acquire potent biological activities. Toxins are generally active at very low doses and designed to quickly reach their target in vivo.^58–62 To broaden efficacy while minimizing side effects, we identified nine next-generation V2R antagonists from African mamba snake venom.^63,64 Among them, mambaquaretin 1 (MQ1) emerged as one of the most potent and selective V2R antagonists known to date.⁶³ To harness the interesting properties of this natural peptide and develop it into a promising drug candidate, we undertook an engineering process resulting in the evolution from MQ1 to the enhanced form MQ232 being fully investigated here.

Methods

Information on commercially available kits, buffers, and chemistry used in the study and other methodical details, including methods relevant to supplemental results, are provided in Supplemental Material.

Prediction and Identification of CD4 T-Cell Epitopes in Humans

Putative immunogenic sequences (CD4 T-cell epitopes) were predicted using the netMHCIIpan3.2 algorithm.⁶⁵ T-cell epitopes were identified by generation of human T-cell lines and assessment of their peptide specificity by IFN-γ enzyme-linked immunospot using overlapping peptides.⁶⁶

Chemistry

MQs were produced by solid-phase peptide synthesis, deprotected, purified, and folded as described.⁶⁴ Deferoxamine (DFO) MQ232 was radiolabeled with ⁸⁹Zr (90 MBq) at 37°C for 60 minutes. Radiochemical yield and purity were determined using instant thin-layer chromatography and further confirmed by radio-high performance liquid chromatography. The final specific activity was determined to be 28.6±6.2 MBq/ng.

In Vitro Pharmacology

Binding experiments were performed using 1 nM [³H]assays under vasopressin (AVP) in a 100-μl reaction mixture.⁶⁴ Data were fitted to a one-site inhibition mass action curve using GraphPad Prism (San Diego), and IC₅₀ values were converted to K_i using 1.1 nM as [³H]AVP Kd.

cAMP cell-based assay: 5×10³ human V2R-transfected chinese hamster ovary cells/well were incubated for 30 minutes at 37°C with AVP DMEM solutions in the presence of MQs. The reaction was stopped with cAMP Gs dynamic kit homogeneous time-resoluted fluorescence from Cisbio.⁶³

V2-β-arrestin interaction bioluminescence resonance energy transfer (BRET) assay: 75×10³ electroporated murine adenocarcinoma cells/well with human V2R-RLuc and β-arrestin-1-yellow fluorescent protein were incubated with Krebs/Coelenterazine H before addition of AVP/Coelenterazine H solutions in the presence of MQs. EC₅₀ values were calculated from Kinetics BRET.⁶³

V2R internalization assay by diffusion-enhanced resonance energy transfer: 9×10⁴ electroporated human embryonic kidney Magik cells/well with Flag-ST-human V2R were labeled with soluble NSF attachment protein (SNAP)-Lumi4-Tb and incubated with AVP in Krebs/BSA 0.2%/fluorescein in the presence of MQs. Human V2R surface expression level was determined by measuring SNAP-Lumi4-Tb fluorescence at indicated times and area under kinetics curve to obtain related EC₅₀ values from dose-response curves.

Mean arterial pressure kinase phosphorylation: 5×10⁴ Myc-hV2R human embryonic kidney cells/well were starved and stimulated with AVP in the presence of MQs. AVP-stimulated mitogen-activated protein kinases phosphorylation was evaluated using the Cellul'erk kit from Cisbio Bioassays (Codolet, France).⁶³

Mini G protein recruitment assays under vasopressin stimulation were performed on 10⁶ HEK293T cells transiently coexpressing the human V1aR or human V1bR or human V2R, plus the cognate mini Gq protein fused to NanoLuc (mini-Gq for human V1a/bR and mini-Gs for human V2R) plus the membrane anchored venus-KRas protein. Cells were stimulated in the presence of AVP and MQ232 or tolvaptan at various concentrations. BRET recordings were performed in the presence of the NanoLuc substrate furimazine. Values at plateaus of the kinetic curves were used to obtain the dose-response curves and IC₅₀ values.

Pharmacodynamics, Pharmacokinetics, and Safety Analysis

For pharmacodynamic experiments, rats were injected with MQ232 before urine collection in metabolic cages.

For the pharmacokinetic experiments, blood samples were obtained at various times and analyzed by mass spectrometry. Plasma plus the internal standard (MQ232-A39K) was treated by solid-phase extraction, reduced and alkylated and analyzed with the liquid chromatography–mass spectrometry/mass spectrometry Nexera (Shimadzu, France) coupled to mass spectrometry on Quantum Ultra (ThermoFisherScientific, France) for the quantification of MQ232.

For the safety experiments, MQ232 was subcutaneously (s.c.) administered acutely or chronically for 21 days before urine, plasma, and organ analysis.

Positron Emission Tomography Imaging

[⁸⁹Zr]Zr-DFO-MQ232 was intravenously (i.v.) administered through tail vein on anesthetized mice and imaged with a Siemens Inveon (Siemens) by small-animal–adapted positron emission tomography (PET)/computed tomography (CT) sessions. Images were corrected for attenuation using a 20-minute CT scan performed consecutively to the PET, reconstructed using a 3D ordered subset expectation maximization iterative algorithm, and data extraction performed using an image processing, analysis and quatification software (PMOD, Bruker) for signal intensity for pharmacokinetic parameter determination.

Pathological Rodent Model Experiments

For the secretion of antidiuretic hormone model of hyponatremia, rats were infused with 10 ng/h of desmopressin. Natremia was measured over time.⁶⁷

Mice embryonic kidneys were cultured^68,69 and treated for 6 days with MQ232. The cystic index was quantified by analyzing the ratio of cyst area to the total area of the kidney section using ImageJ (Version 1.53c). MQ232 or vehicle were s.c. administered to the wild-type (Pkd1+/+; Ksp-Cre) and ADPKD (Pkd1^flox/flox; Ksp-Cre)⁷⁰ mice from day 6 to day 12. Kidneys were harvested to calculate the total kidney-to-body weight (BW) ratio. Kidneys were fixed for histological examination.

In Vivo Experiment Authorizations

All animal procedures were performed in accordance with the Care and Use of Laboratory Animals of the Directive 2010/63/EU of the European Parliament. These procedures were approved by the local ethics committee and obtained the authorizations.

Data Processing and Statistical Analyses

Data are shown as means±SEM of the mean. Normality was assessed using the d'Agostino–Pearson test, comparisons between two groups using the unpaired Student's t test, and comparisons between more than two groups using the Kruskal–Wallis test followed by Dunn's test. ANOVA multiple factors followed by Tukey's test were used to compare repeated measurements. Statistical analyses were performed using GraphPad Prism 9.5.1 software.

Results

Engineering MQ232

The immunogenicity of protein therapeutics may be an issue resulting in lower efficacy, anaphylaxis, and occasionally life-threatening autoimmunity.⁷¹ European and American regulatory agencies recognize limited predictive value of animal models for human immunogenicity and recommend alternative approaches for risk evaluation. Natural MQ1⁶³ being a nonhuman peptide, we used human T-cell assays to evaluate its immunogenicity risk and bioinformatics tools to mitigate it. CD4 T cells specific for MQ1 were derived from human blood samples, and their peptide specificity were assessed by stimulating them with 15-mer peptides overlapping with 12 amino acids encompassing the MQ1 sequence (Figure 1A).^72,73 T-cell lines were found to be specific for peptides P6, P7, and P8 and to a lesser extent for P1, P14, and P15 (Figure 1B). The bioinformatics tool NetMHC 3.2 (Jensen⁷⁴) identified major anchoring amino acids for HLA-DR molecules—F20 (P6–P7), Y21 (P6–P8), and F44 (P14–P15). However, we never succeeded in the substitution of these residues because of their critical role in the MQ1 structure. We thus mitigated the immunogenicity of peptides P1, P14, and P15 by substituting the putative HLA anchor F4 with glycine and I48 with glutamic acid, resulting in MQ211 (Figure 1C and Supplemental Figure 1). MQ211 kept the MQ1's affinity for human V2R and showed a lower HLA-binding score. In the P6–P8 zone, T27 was successfully substituted by an aspartic acid (MQ228). MQ228 preserves MQ1's affinity but combines a null HLA-binding score. In addition to these three mutations that reduce the risk of immunogenicity, we added the K39A gain-of-function substitution that was previously identified.⁶⁴ Added to MQ228, the resulting MQ232 (Supplemental Figure 1) showed a subnanomolar affinity and HLA-DR–binding score close to zero (Figure 1C). Through computational simulations and iterative experimental design, natural MQ1 properties were optimized to enhance its safety and efficacy, thanks to extensive efforts in preserving the toxin's structure.

Figure 1.

Human CD4 T-cell response to MQ1 and design of MQ1 variants with mitigation of T-cell epitope and enhanced binding activity for human V2R. (A) Sequences of the overlapping 15-mer peptides. Main anchor residues predicted from NetMHC3.2 were highlighted in red. (B) T-cell epitope mapping was established by testing the recognition of individual peptides by T-cell lines raised against MQ1. (C) Design of MQ1 variants. F4G, T27D, and I48E in blue. K39A in green. K_i for human V2R. The score of HLA class 2 binding of the variants was calculated for high and low affinity by the number of peptides with a percentile below 10% and 20%, respectively, using NetMHC3.2. MQ1, mambaquaretin 1; V2R, vasopressin type 2 receptor.

Pharmacological Characterization of MQ232

MQ1^63,64,67 displays binding affinities of 4.56 nM (pK_i 8.37±0.07) for the human V2R using tritiated AVP as the radioligand (Figure 2A). MQ232 displays binding affinities of 0.37 nM (pK_i 9.60±0.19) for the human V2R and 1.89 nM (pK_i 8.73±0.04) for the rat V2R (Figure 2A and Supplemental Table 1). MQ232 consistently outperforms MQ1 in its ability to antagonize the three AVP-induced V2R signaling pathways linked to the cAMP production, β-arrestin-1 recruitment, and mean arterial pressure kinase phosphorylation. In addition, both peptides block the AVP-induced Flag-Snap-hV2R internalization (Figure 2, B–E, Supplemental Figure 2, and Supplemental Table 2). Like MQ1, MQ232 displays a selectivity factor higher than 100 times in favor of human V2R versus human V1aR and human V1bR (Supplemental Figure 3). Control Sprague Dawley rats spontaneously urinated on average 1.79 ml/h per kilogram BW with an osmolality of 1475 mOsm/kg H₂O, corresponding to an electrolyte excretion of 60 Osm/d per kilogram BW, without sex-linked differences. Acute s.c. injections of MQ232 with doses from 1 to 3000 nmol/kg BW dose dependently increased 24-hour aquaresis from 2.6 to 12.0 ml/h per kilogram BW in parallel with a decrease in osmolality (Figure 2F and Supplemental Table 3). No significant loss of electrolytes was noticed, no perturbation of any of the 32 blood or urine measured parameters could be detected (Supplemental Table 3), and no histologic sign of toxicity on the six examined organs could be seen (Supplemental Figure 4 and Supplemental Table 4). The unique toxicity sign was detected on only one rat of the six injected with the highest dose of 3000 nmol/kg. This rat presented an apathetic behavior, 24% weight loss, and enlarged right kidney (Supplemental Figure 5 and Supplemental Information 1). Aquaresis was maximum between the first and second hour with a maximal aquaresis around 50 ml/h per kilogram BW and exhibited a monoexponential decrease with a t_1/2 of 1.33±0.13 hours (0.7–1.7 hours, Figure 2G, and Supplemental Table 5). Pharmacokinetics parameters of MQ232 were determined after i.v. (Supplemental Figure 6) and intraperitoneal (i.p., Supplemental Figure 7) injection of 30 nmol/kg BW (Figure 2H and Supplemental Tables 6 and 7). A monocompartmental distribution was observed with a t_1/2 of 0.96 and 1.1 hours and total areas under the curve of 0.47 and 0.52 µg/ml×hour, respectively, giving a peptide biodistribution of 100% by i.p. versus i.v. route (Figure 2H and Supplemental Table 8). i.p. injection of 300 nmol/kg BW of MQ232 (Supplemental Figure 8) displayed a ten-times higher total area under curve of 5.9 µg/ml×hour and t_1/2 of 2.1 hours, demonstrating a good linearity between the injected doses and circulating concentrations of MQ232. MQ232 distributes with a small apparent volume ranging from 450 to 780 ml/kg, its clearance being between 6.34 and 7.53 ml/min per kilogram. Finally, there is a good linear relationship between aquaresis and MQ232 plasma concentration (R²=0.91, Figure 2I, and Supplemental Table 9). Aquaresis of rats continuously infused for 21 days with 0, 0.5, or 5 nmol/h per kilogram BW of MQ232 raised from 1.12 ml/h per kilogram BW (0 nmol/kg per hour, control) to 4.09 (0.5 nmol/kg per hour) and 14.40 ml/h per kilogram BW (5 nmol/kg per hour, Supplemental Table 10). For the dose of 5 nmol/h per kilogram BW, MQ232 plasma concentrations ranged between 11.3 and 15.2 nM (Supplemental Table 12). Those aquaresis raises were associated with a urine osmolality decrease without electrolyte loss. No modification in the 34 urine and blood biochemical parameters nor in the six observed organs, including the liver, were noted (Supplemental Figures 9–11, Supplemental Table 11, and Supplemental Information 2). The expression of lipocalin 2, a highly sensitive tubule injury marker remained undetectable in the collecting ducts of the rats treated for 21 days with 0 or 5 nmol/h per kilogram of MQ232 (Supplemental Figure 10).

Figure 2.

MQ232 pharmacological characterization. (A) Competitive binding inhibition of ³H-AVP (1 nM) on human (full lines, stable CHO-hV2R cell line) and rat (dashed line, expressed in transfected HEK) V2R by MQ1 (red), MQ232 (blue), and [⁹¹Zr]Zr-DFO-MQ232 (green), n=5. (B–E) Arunlakshana–Schild plots plotted as mean±SEM. mBU corresponding to the competitive inhibition by MQ1 (red) and MQ232 (blue) of AVP-induced (B) cAMP production in a stable CHO-hV2R cell line, n=3, P = 0.052; (C) β-arrestin-1 recruitment by BRET of β-arrestin-1–yellow fluorescent protein and human V2R-Rluc in murine adenocarcinoma transfected cell line, n=4, P = 0.004; (D) internalization of the Flag–Snap human V2R in HEK magic transfected cell line, n=3, P < 0.001; and (E) MAP kinase phosphorylation in stable Myc-hV2R HEK cell line, n=3, P = 0.048 (Supplemental Figure 2). (F) Rat aquaresis (red) and urine osmolality (blue) over a 24-hour period for various doses of MQ232. n=6 rats per dose. (G) Rat aquaresis versus time with increasing doses of MQ232 (0–3000 nmol/kg BW) s.c. injected in control male rats at time zero, n=6 rats per dose. Aqueous diuresis was measured over 60 hours. Half-times (t_1/2, hour) were calculated when possible, following a noncompartmental analysis. (H) Plasma MQ232 concentrations were determined by mass analysis after i.p. injection of 30 and 300 nmol/kg BW and after i.v. injection of 30 nmol/kg BW in control rats (Supplemental Figures 6–8). n=3 rats per group, noncompartmental analysis. (I) Relationship between aquaresis (30 nmol/kg BW s.c.) and plasma MQ232 concentration (30 nmol/kg BW i.p.). n=3 rats per group. AUCtot, total area under the curve; AVP, assays under vasopressin; BRET, bioluminescence resonance energy transfer; BW, body weight; CHO-hV2R, human V2R-transfected chinese hamster ovary; DFO, deferoxamine; HEK, human embryonic kidney; i.p., intraperitoneal; i.v., intravenous; MAP, mean arterial pressure; mBU, milliBRET unit; s.c., subcutaneous; SNAP, soluble NSF attachment protein; Vss, extrapolated total volume of distribution.

MQ232 Pharmacokinetic and Tissue Distribution Derived from PET Imaging in Control Mice

PET is a powerful imaging modality thanks to its high sensitivity and in vivo dynamic imaging capabilities. DFO is the most widely used chelator for ⁸⁹Zr labeling.⁷⁵ A DBCO-derived DFO was cycloadded to the azido-MQ232 to generate the DFO-MQ232 (Figure 3A and Supplemental Figure 1). [⁹¹Zr]Zr-DFO-MQ232 displays affinities of 0.82 (pK_i=9.12±0.08) and 2.47 nM (pK_i=8.64±0.08) for human V2R and rat V2R, respectively (Figure 2A and Supplemental Table 1), demonstrating the minimal influence of this modification. The radiolabeling yield of DFO-MQ232 with ⁸⁹Zr demonstrated high efficiency, exceeding 95% (Figure 3B). Capitalizing on PET's significant sensitivity, we administered [⁸⁹Zr]Zr-DFO-MQ232 at a low dose of 40±5 µg (4.28±0.18 MBq) into control mice. Notably, there were no observed toxic effects after injection, emphasizing the ability to achieve distinct distribution without adverse effects on the biological milieu. The dynamic acquisition, PET reconstruction, and fusion of PET with CT images allowed the determination of MQ232's whole-body behavior over time (Figure 3C). 62%±8% of the injected [⁸⁹Zr]Zr-DFO-MQ232 accumulated favorably in the kidney during the first hour. The gallbladder, colon, and liver accumulated 9%±0.4%, 6%±1%, and 5%±0.5% of the injected dose, respectively (Figure 3D and Supplemental Table 13). It is worth noting that the liver signal is classically overestimated because of the transmetalation occurring between Zr and Fe ions.⁷⁶ No other organs were significantly labeled, except the bladder, as the kidney is also the metabolizing organ for [⁸⁹Zr]Zr-DFO-MQ232. The analysis of MQ232's uptake within the blood pool allows specific pharmacokinetic parameter extraction (Figure 3E and Supplemental Table 4). A factor of almost eight was obtained between the blood and kidney AUC in favor of the latter (Figure 3F), showing a high plasmatic clearance and strong affinity of the radiotracer for the kidneys. We used a one-compartment open model to fit the image-derived blood activity, enabling the determination of MQ232 t_1/2 to be 27±4 (16–40) minutes (Figure 3F and Supplemental Table 14), in accordance with the one determined in rats.⁷⁷

Figure 3.

Biodistribution and pharmacokinetic of MQ232 derived from PET imaging. (A) Model of complexation of ⁸⁹Zr in DFO DBCO.⁷⁵ (B) Instant thin layer chromatography (iTLC) chromatogram of [⁸⁹Zr]Zr (red) and [⁸⁹Zr]Zr-DFO-MQ232 (green). (C) Dual PET/CT imaging, with left: CT scan of a healthy mice showing the bone structure (gray scale). Right: PET/CT imaging at 1 and 4 hours after injection of 40±5 µg (4.28±0.18 MBq) of the radiolabeled peptide [⁸⁹Zr]Zr-DFO-MQ232 i.v. injected in the tail of control mice (n=6). A strong signal is observed in the kidneys, whereas the liver exhibits minimal residual signal, and there is a low uptake in the gallbladder attributed to restrained hepatic metabolism. It is worth noting that this signal might be overestimated because of transmetalation. The urinary bladder is prominently visible at 4 hours after injection, highlighting the substantial renal metabolism and excretion of the radiolabeled compound. The presented images showcase a composite of CT (gray scale) and PET (color scale) for enhanced subject visualization. (D) PET-derived biodistribution of [⁸⁹Zr]Zr-DFO-MQ232 in the kidneys, gallbladder, colon, liver, blood (image derived input functions using the heart, a large blood pool, which does not express V2R), bone junction, and brain using specific volumes of interest at 1 and 4 hours after injection. The kidneys (first column in dark red) concentrated more than 60% of the injected radioactivity. (E) [⁸⁹Zr]Zr-DFO-MQ232 blood concentration over time from which the blood half-life of the compound was extracted using a mono exponential model for data fitting. (F) Pharmacokinetic parameters derived from [⁸⁹Zr]Zr-DFO-MQ232 PET kinetics in blood and kidneys. AUC, area under the curve; CT, computed tomography; DBCO, dibenzocyclooctyne; %ID, % of injected dose; PET, positron emission tomography.

MQ232's Efficacy in Experimental Rodent Models of Hyponatremia and PKDs

The rat model for the syndrome of inappropriate antidiuretic hormone was obtained by s.c. infusion of 10 ng/h of desmopressin and water gavage from D0 to D4.^78,79 In vehicle-treated rats, plasma sodium concentration ([Na⁺]) lowered from 141.7±0.63 to 124.5±4.17 mM between D0 and D4 (Figure 4A and Supplemental Table 15). MQ232 and tolvaptan were administrated on days 2, 3, and 4. Tolvaptan (22 µmol/kg BW per os) partially restored natremia levels at D4 only (plasma [Na⁺]=137.6 mM) and with a plasma [Na⁺] rise of 4.8 mM/d. MQ232 (3.1 nmol/kg BW s.c.) restored natremia from D3 (plasma [Na⁺]=142.7 mM, Figure 4A, and Supplemental Table 15) and with plasma [Na⁺] rises between 6.9 and 7.5 mM/d. The 9.6 nmol/kg BW s.c. MQ232 dose gave no additional effect. Apart from a moderate weight loss due to the highly stressful model for rats, no toxicity signs were observed (Supplemental Table 16).

Figure 4.

MQ232's efficacy in experimental rodent models of hyponatremia and PKDs. For each experiment, a minimum number of animals were used to obtain statistical differences between control and treated groups. (A) Male rat model of hyponatremia syndrome of inappropriate antidiuretic hormone. Six- to seven-week-old rats were infused through an osmotic pump with 10 ng/h of dDAVP inducing a clear hyponatremia in 2 days. Blood samples were obtained just before MQ232 i.p. administration (3.1 or 9.6 nmol/kg BW) or tolvaptan p.o. administration (22 µmol/kg BW) indicated by red arrows. Ex vivo mouse model of PKD: (B) The isolated kidney harvested at embryonic day 13.5 were treated with 100 μM 8-Br-cAMP to stimulate cyst growth plus saline (vehicle NaCl 0.9%) or 1 μM MQ232 for 6 days. Representative images of embryonic kidneys cultured with vehicle or 1 μM MQ232 on days 0 and 6. (C) Fractional cyst area (i.e., empty white spots) of kidneys on day 6 were counted. Data are presented as mean±SEM (vehicle-treated group n=6, MQ232-treated group n=8). Ortholog male mouse model of ADPKD: (D) The genotype Pkd1^flox/flox; Ksp-Cre was selected on postnatal day 2 and mice administrated s.c. with MQ232 (1 μmol/kg BW per day) or saline vehicle (NaCl 0.9%) from day 6 to day 12. Representative images on day 12 of hematoxylin/eosin staining kidney sections (lower) and kidney photography (upper) from Pkd1^flox/flox; Ksp-Cre mice. (E) KW indexes (KW/BW) of Pkd1^flox/flox; Ksp-Cre mice at day 12. Mean±SEM (n=6). ADPKD, autosomal dominant polycystic kidney disease; dDAVP, desmopressin; KW, kidney weight; PKD, polycystic kidney disease; p.o., per os.

We evaluated MQ232 activities ex vivo on an embryonic kidney developing cyst under 8-Br-cAMP stimulation.^69,70 The cystogenesis effect (Figure 4B and Supplemental Table 17) was significantly inhibited by 1 µM of MQ232 as the ratio of cyst area to total kidney area was reduced from 37% to 26% (Figure 4C). Next, we used a kidney-selective Pkd1 knockout mice (Pkd1^flox/flox; Ksp-Cre) orthologous model of ADPKD.^69,70 Compared with the vehicle-treated group and after 12 days of treatment, MQ232 concentration dependently slowed kidney enlargement (Figure 4D and Supplemental Table 18). Indeed, kidney index (kidney weight/BW, Figure 4E) diminished from 20% to 18% and 13% with 0.1 and 1 µmol/d per kilogram BW of MQ232, respectively.

Discussion

In this work, we use a large array of techniques to investigate the efficacy and safety of a new class of aquaretic agents, led by MQ232. This peptide derived from a natural snake toxin called MQ1 displays much lower risk of immunogenicity by in silico evaluation associated with a 12-fold higher affinity for the human V2R. MQ232's safety is favorable with a maximum tolerate dose in rats of 1000 nmol/kg combined with an in vivo selectivity for the kidney organ. It seems to be predominantly eliminated through the kidneys, this being supported by its creatinine-like high clearance with low reabsorbtion^80,81 and peptide nature. Finally, its minimal uptake by the liver allows the anticipation of a reduced drug–drug interaction probability, particularly appreciated in polymedicated patients.

In healthy rats, MQ232 increases aquaresis starting from the dose of 3 nmol/kg. This dose was sufficient to restore natremia in our experimental model of hyponatremia. Aquaretic drugs' administration stands out as the most effective strategy for normalizing natremia in patients with euvolemic and hypervolemic hyponatremia, particularly when it is caused by excessive AVP production. Interestingly, MQ232 quickly restores natremia at a patient safety–compatible rate. It is indeed crucial for hyponatremic patients to (1) receive prompt medical attention to minimize the detrimental effects associated with low plasma sodium concentrations; (2) control the natremia's increase rate, aiming for <12 mM/d to prevent osmotic demyelination syndrome^82–84; and (3) avoid hypernatremia.¹⁴ MQ232 completely respected these conditions. The parenteral administration of MQ232 coupled with its rapid action and swift elimination from the rat body could confer a substantial advantage for emergency physicians involved in natremia rescue. In addition, MQ232 could offer additional safety benefits with a therapeutic window in rats being over 300 times higher than the effective dose and associated with a strong correlation between urine output and plasma MQ232 concentration. These qualities anticipate a wide safety margin and reduced likelihood of adverse effects in the case of accurate dosage adjustments and monitoring.

PKDs, whether dominant or recessive, are still incurable, and only V2R blockade has been shown to be truly effective in humans. MQ232 showed a dose-dependent efficacy on an ex vivo experimental model and orthologous mice model of ADPKD. We have thus validated the use of MQ232 for these pathologies even if these initial findings warrant validation through additional toxicity studies, encompassing both male and female patients and using models that exhibit a gradual cyst progression, mirroring the natural history of ADPKD. The parenteral administration of MQ232 may be an issue for patients with ADPKD. The use of a pump like for insulin or a slow-release formulation like for glutathion-like protein 1 analogs may provide a more convenient solution.

Nature is an incredible source of drug candidates,⁸⁵ and animal venom is a reservoir of millions of peptides selected by evolution for high efficiencies at very low doses. When a venomous animal bites a prey or a predator, it cannot control neither the real quantity injected nor the exact administration route of its venom.⁸⁶ Thus, toxins must act quickly regardless of the injection site. This swift effect is crucial to prevent the prey from escaping or the predator from killing the venomous animal. MQ232 possesses all of these interesting qualities. It acts very quickly on aqueous diuresis after injection, its efficacy at triggering is equal between i.v. and i.p. injection routes, and the first active dose is as low as 3 nmol/kg BW.

MQ232 showed compatible efficacy for in vivo V2R imaging by PET. This tool holds the potential to introduce a novel approach for categorizing patients with ADPKD. Kidney size is currently evaluated by echography and/or magnetic resonance imaging technics,⁵¹ but it gives no information on the density of treatment-accessible V2R. The response to tolvaptan varies among individuals with ADPKD: Some patients experience a beneficial response, whereas others may show no significant improvement. Individual factors, such as genetic variations and disease severity, contribute to the variability of responses to tolvaptan. V2R expression level in patients with ADPKD may also potentially be related to their response variability to this drug. PET-compatible, MQ232-based imaging agents may offer, after validation, a promising pathway to investigate the role of V2R-binding sites' density in treatment outcomes. By studying the relationship between V2R expression and treatment response, physicians may gain valuable insights on the mechanisms underlying interpatient variations in drug efficacy.

In conclusion, MQ232, derived from a natural peptide, is a new class of biological V2R antagonist and a promising drug candidate for hyponatremia and ADPKD.

Supplementary Material

jasn-36-181-s001.pdf ^{(1.4MB, pdf)}

jasn-36-181-s002.pdf ^{(2.7MB, pdf)}

Disclosures

Disclosure forms, as provided by each author, are available with the online version of the article at http://links.lww.com/JSN/E915.

Funding

N. Gilles: Université Paris-Saclay (2023 Technics Project from the Health and Drug Science Graduate School and AAP IDEX “INNOVATION and ENTREPRENEURSHIP” POC IN LABS 2021). G. Stanajic-Petrovic: Commissariat à l'Énergie Atomique et aux Énergies Alternatives (CFR PhD Scholarship).

Author Contributions

Conceptualization: Frank Bienaimé, Nicolas Gilles, Goran Stanajic-Petrovic, Charles Truillet, Wenchao Zhao.

Data curation: Khawla Chmeis, Nicolas Gilles, Mathilde Keck, Christiane Mendre, Goran Stanajic-Petrovic, Apolline Urman.

Formal analysis: Khawla Chmeis, Evelyne Correia, Jean-Paul Duong Van Huyen, Nicolas Gilles, Hervé Nozach, Goran Stanajic-Petrovic, Frederic Theodoro, Apolline Urman.

Funding acquisition: Nicolas Gilles, Goran Stanajic-Petrovic, Charles Truillet.

Investigation: Peggy Barbe, Frank Bienaimé, Florence Castelli, Evelyne Correia, Jean-Paul Duong Van Huyen, Nicolas Gilles, Dong Guo, Pierre Isnard, Mathilde Keck, Pascal Kessler, Christiane Mendre, Bernard Mouillac, Anvi-Laëtitia Nguyen, Hervé Nozach, Stefano Palea, Alain Pruvost, Philippe Robin, Goran Stanajic-Petrovic, Frederic Theodoro, Apolline Urman, Wenchao Zhao.

Methodology: Peggy Barbe, Florence Castelli, Khawla Chmeis, Evelyne Correia, Jean-Paul Duong Van Huyen, Nicolas Gilles, Dong Guo, Pierre Isnard, Mathilde Keck, Pascal Kessler, Bernard Maillere, Christiane Mendre, Bernard Mouillac, Anvi-Laëtitia Nguyen, Hervé Nozach, Stefano Palea, Alain Pruvost, Goran Stanajic-Petrovic, Frederic Theodoro, Charles Truillet, Apolline Urman.

Project administration: Nicolas Gilles.

Resources: Florence Castelli, Sékou Siramakan Diarra, Nicolas Gilles, Dong Guo, Mathilde Keck, Christiane Mendre, Anvi-Laëtitia Nguyen, Stefano Palea, Alain Pruvost, Denis Servent, Charles Truillet, Wenchao Zhao.

Supervision: Nicolas Gilles, Denis Servent, Charles Truillet.

Validation: Nicolas Gilles, Charles Truillet.

Writing – original draft: Nicolas Gilles, Goran Stanajic-Petrovic, Charles Truillet.

Writing – review & editing: Khawla Chmeis, Nicolas Gilles, Catherine Llorens-Cortes, Bernard Maillere, Christiane Mendre, Bernard Mouillac, Hervé Nozach, Stefano Palea, Denis Servent, Goran Stanajic-Petrovic, Charles Truillet.

Data Sharing Statement

All data are included in the manuscript and/or supporting information.

Supplemental Material

This article contains the following supplemental material online at http://links.lww.com/JSN/E914.

Supplemental Methods

Supplemental Figure 1. Chemical synthesis of MQs.

Supplemental Figure 2. Functional characterization of MQ1 and MQ232.

Supplemental Figure 3. Functional characterization of MQ232 on hV1aR, hV1bR, and hV2R.

Supplemental Figure 4. Pathological study of rat organs from the acute protocol.

Supplemental Figure 5. Necroscopy of the female rat (n=91,109).

Supplemental Figure 6. MQ232 blood concentration evolution with time after acute i.v. injection of 30 nmol/kg.

Supplemental Figure 7. MQ232 blood concentration evolution with time after acute i.p. injection of 30 nmol/kg.

Supplemental Figure 8. MQ232 blood concentration evolution with time after acute i.p. injection of 300 nmol/kg.

Supplemental Figure 9. Pathological study of rat organs from the chronic protocol.

Supplemental Figure 10. MQ232 does not increase the expression of the injury marker lipocalin 2 (Lcn2) in chief collecting duct cells.

Supplemental Figure 11. Pathological study showing inconspicuous mast cell hyperplasia.

Supplemental Table 1. Binding values of MQ232 and [91Zr]Zr-DFO-MQ232 on human and rat V2R.

Supplemental Table 2. Inhibition of various V2R signaling pathways by MQ232.

Supplemental Table 3. Blood and urine analyses of rats after acute injection with MQ232.

Supplemental Table 4. Rat organ histology after acute s.c. injection of various doses of MQ232.

Supplemental Table 5. Pharmacodynamics (s.c. route) and determination of the half-life of MQ232 on rats.

Supplemental Table 6. Pharmacokinetics parameters after i.v. administration of 30 nmol/kg of MQ232 in rats.

Supplemental Table 7. Pharmacokinetics parameters after i.p. administration of 30 nmol/kg of MQ232 on rats.

Supplemental Table 8. Pharmacokinetics parameters after i.v. administration of 300 nmol/kg of MQ232 on rats.

Supplemental Table 9. Summary of the MQ232 pharmacokinetics parameters.

Supplemental Table 10. Blood and urine analyses of rats infused with MQ232.

Supplemental Table 11. Blood MQ232 concentration during infusion of 5 nmol/kg per hour BW.

Supplemental Table 12. Rat organ histology after chronic infusion of various doses of MQ232.

Supplemental Table 13. Averaged PET-obtained activities, by organ.

Supplemental Table 14. Corrected time–activity curve data for each mouse used for plasma half-life determination.

Supplemental Table 15. Mean and SEM values for natremia (mmol/L) in a hyponatremic rat model treated with MQ232 or tolvaptan.

Supplemental Table 16. Mean and SEM values for body weight (BW, g) in a hyponatremic rat model treated with MQ232 or tolvaptan.

Supplemental Table 17. Kidney index of the ex vivo mouse model of PKD treated with MQ232 (1 μM, n=8) or vehicle (NaCl 0.9%, n=6) on day 6.

Supplemental Table 18. Body and kidney weight of the Pkd1 knockout male mice (Pkd1^flox/flox; Ksp-Cre) i.p. injected for 12 days with MQ232 (1 μmol/d per kilogram BW) or vehicle (NaCl 0.9%).

Supplemental Information 1. Acute safety.

Supplemental Information 2. Chronic safety.