Development of a Polypeptide-Based Platform to Construct Antibody-Drug Conjugates with High Potency

Abstract

Antibody-drug conjugates (ADCs) represent an emerging pharmaceutical category that leverages antibodies to facilitate drug delivery to target cells. A key challenge for ADCs is increasing the drug-antibody ratio (DAR) to improve efficacy without compromising their properties, such as pharmacokinetics. Here, we present a novel polypeptide-based platform that substantially increases the DAR of antibody-payload ADCs while minimally compromising the ADCs’ inherent characteristics. A polypeptide-based antibody-drug conjugate lead (polypeptide ADC 6c), capable of achieving a DAR of up to 61, was synthesized using the versatile click reaction method based on our polypeptide-based platform. 6c could efficiently recognize antigens, be subsequently endocytosed, and achieve subnanomolar activity at the cellular level. Moreover, 6c exhibited excellent pharmacokinetic parameters and potent antitumor efficacy in mice. Collectively, this study provides and characterizes a highly variable polypeptide backbone platform that can be used to efficiently develop innovative ADCs with high payloads for the treatment of various cancers

Introduction

Cancer has become the second greatest global health threat, accounting for approximately 10 million deaths from cancer that occurred in 2020. There is an urgent need to develop efficient therapies for various kinds of cancers. For the past decades, surgical resection and cytotoxic agent-based chemotherapy have been the main approaches for the treatment of a wide range of cancers. However, these therapies have their disadvantages: it is hard to get rid of cancerous tissue, and off-target exposure causes severe side effects. Subsequently, antibody-drug conjugate (ADC) technology was developed to solve those clinical problems. As a combination of potent chemotherapy and targeted antibody therapy, ADCs have demonstrated powerful clinical efficacy in recent clinical practice across multiple cancers. Currently, 15 ADCs have been approved by the FDA. The success of pioneers inspired active research on the design of more effective and well-tolerated ADCs, which continuously provide better understanding and options for the choice of antibodies, chemistries of linkers, methods of conjugation, and choice of drug payload.

One crucial property that affects ADC potency is the drug-antibody ratio (DAR), which indicates the number of drug molecules per antibody. Although a higher DAR is potently associated with increased potency in vitro, it also has an impact on the properties of ADCs. Due to the increased hydrophobicity of attached payloads, ADCs with higher DAR are more prone to spontaneous aggregation, which negatively affects their physicochemical and pharmacokinetic properties. Moreover, because the payloads of ADCs are conjugated to exposed amino acid side chains of antibodies that have a limited number. Currently, the DAR is limited to ≤8, which may yield greater antitumor efficacy with the traditional approach to ADC construction (Figure A).

1.

Schematic representation of click-based polypeptide ADCs. (A) In traditional ADCs, the ratio of modification sites to payloads is 1:1. (B) Polymer-based ADCs, the ratio extended to 1:N via introduction of polymers. (C) Polypeptide ADC linker described in this work.

Limited DAR narrows down the payload scope of ADCs to highly toxic agents with few mechanisms of action. This leads to the toxicity and drug resistance of ADCs. , To overcome this, ADCs with higher DAR while maintaining the overall properties are needed. Many hydrophilic motifs are incorporated into new linker designs to compensate for the hydrophobicity of payloads, such as PEGs or polysarcosines. , One of the most efficient designs was the highly hydrophilic GGFG linker used in Enhertu, which could deliver up to 8 payloads to target cells and maintain potent efficacy against tumor cells with low HER2 expression in animal models. The limited reactive sites on the antibodies result in low payloads through the typical 1:1 direct drug-antibody conjugation. Subsequently, polymers were applied to alter the linking mechanism to discover a potentially effective means (Figure B). − However, many of them modified the polymer using nonquantitative reactions, which always lead to additional heterogeneity and negatively affect the in vivo properties of ADCs. Others applied long PEG chains, which may cause extra anti-PEG antibodies. Therefore, new polymers are needed to improve the performance of this strategy.

Herein, we describe a novel polypeptide backbone ADC platform formed by the ring-opening polymerization (ROP) of α-amino acid N-carboxyanhydrides (NCA) (Figure B, C). This polyvalent biomimetic material may overcome the limitations of traditional conjugation methods. By easily tuning the functional groups on the side chains, high drug loading can be achieved without a significant decrease in pharmacokinetic parameters or ADC stability. In this work, we construct a polypeptide-based ADC lead (polypeptide ADC 6c) with a DAR of nearly 61 using the aforementioned polypeptide backbone platform, with Herceptin as the targeting moiety and monomethyl auristatin E (MMAE) as the cytotoxic payload, linked to the backbone via vc-linkers. Polypeptide ADC 6c retains potent affinity for the antigen and is subsequently internalized by cells, demonstrating subnanomolar potency across multiple breast cancer cell lines. Importantly, polypeptide ADC 6c is not only observed to be highly stable in plasma from multiple species, but also maintains decent pharmacokinetic parameters in mice. Unsurprisingly, polypeptide ADC 6c reveals potent antitumor activity in tumor-bearing mice. Taken together, this work provided and characterized a novel polypeptide backbone platform to develop ADCs with high DAR and excellent metabolic parameters for tumor treatment.

Results and Discussion

The Construction of a Novel Polypeptide ADC 6c

The configuration and properties of polypeptides obtained by ROP of NCAs derived from (L)-glutamic acid have already been studied. We modified the homopolymer of GluEG₃Me-NCA 1 to a copolymer of GluEG₃Me-NCA 1 and GluEG₃yne-NCA 2. We expected that this copolymer could retain the copper­(I)-catalyzed azide–alkyne cycloaddition (CuAAC) reactivity while maintaining the homopolymer’s properties. Two NCA monomers 1, 2, and the initiator 3 were synthesized following previously reported methodologies (Figure S1). − To study the impact of the ratio of MMAE-containing segments in the entire polymer chain, we first synthesized three polypeptides, 4a, 4b, and 4c, with varying amounts of alkyne segments (Figure A). In our design, the MMAE payloads would be connected to the polypeptide backbone via a CuAAC reaction. We prepared two kinds of payloads, MMAE conjugated to alkyne groups with cleavable or uncleavable linkers (S10 and S14), as described in the Supporting Information (Figure S2).

2.

Synthesis process of Herceptin-polypeptide-MMAE. (A) Synthesis of polypeptide 4 and MMAE-conjugated polypeptide 5. (B) Structures of polypeptide ADC 6 and traditional ADC 7.

We then obtained two MMAE-conjugated polypeptides, 5a and 5b. During the reaction, polypeptide 5b precipitated from the solution, indicating that the ratio of MMAE-containing segments should be less in later designs. As a replenishment, we prepared polypeptide 5c with a portion of alkyne groups conjugated with carboxylic acids to “dilute” the hydrophobicity of MMAE. To our delight, this new polypeptide retained its solubility in water. To test the impact of linkers, we constructed polypeptide 5d with cleavable Val-Cit linkers. We were also interested in the electrical properties of the polypeptides; thus, one more positively charged polypeptide, 5e, was included.

With access to MMAE-conjugated polypeptides, we learned how to conjugate them to antibodies. We infused a protected mercaptan at one end of the polypeptide. Reduction with TCEP, followed by mixing with succinimide hexamethylene carbonate (SMCC)-modified Herceptin, could yield the desired polypeptide ADC. In our investigation, 8 equiv of reduced polypeptide completely converted SMCC-modified Herceptin (Figure S3A). We conducted the optimized conjugation procedure to obtain the polypeptide ADCs 6a–d (Figure B, Figure S3B, C). A traditional ADC 7 (Figure B) with MMAE conjugated by a cleavable linker was synthesized as the positive reference.

Using polypeptide ADCs with varying MMAE per polypeptide, we investigated their performance against SK-BR-3 cells. The polypeptide ADC with fewer drug molecules per polypeptide chain (6a) showed no significant increase in cytotoxicity compared with the traditional ADC (7) (Figure S4A). When the number of drug molecules increased to 13, the corresponding polypeptide ADC (6b) exhibited a 10-fold increase in cytotoxicity relative to the traditional ADC. Next, compared with the uncleavable linker, the cleavable linkers proved to be better for the release of payloads, as 6c showed better performance against SK-BR-3 cells than its uncleavable counterpart 6b, although its DAR is less than the latter one’s (Figure S4B). Finally, we studied the influence of the electrical properties of the polypeptide chains. Two types of polypeptide ADCs, negatively charged 6c and positively charged 6d, were tested. The results showed that the two ADCs were quite similar in cytotoxicity against SK-BR-3 cells. However, the positively charged polypeptide chains 5e exhibited unexpected activity against SK-BR-3 cells (Figure S4C). We presume that these positively charged α-helical polypeptide chains share certain properties of cell-penetrating peptides, potentially leading to off-target effects. Overall, 6c was confirmed as the preferred choice for further investigation.

The DAR Characterization of Polypeptide ADC 6c

Due to the properties of the mentioned polypeptide ADC, simple chromatography methods cannot resolve distinct peaks for different DARs, thereby hindering the determination of total DAR (Figure A). To overcome this problem, we applied a slightly modified version of a previously reported protocol to verify the DAR of 6c. First, all MMAE molecules on the conjugation were released via papain-catalyzed cleavage of the Val-Cit linker, and the MMAE content of the processed samples was determined by LC-MS/MS (Figure A). The protein content of Herceptin in the samples was measured using a BCA assay according to the manufacturer’s instructions (Figure B). Finally, the DAR was calculated as the ratio of MMAE to Herceptin content. The DAR of traditional ADC (7) is conventionally about 3.7, and that of the polypeptide ADC 6c is encouragingly up to 61 (Figure C).

3.

DAR characterization of polypeptide ADC 6c and traditional ADC 7. (A–B) The workflow to determine the DAR value of ADC. LC-MS/MS was used to quantify the payload in samples (A), and a BCA assay was used to determine the antibody content of the samples according to the manufacturer’s instructions (B). The DAR can then be obtained by following the definition: DAR = drug/antibody. (C) Proposed structure of polypeptide ADC 6c and traditional ADC 7 based on the calculated DAR.

Antigen Binding, Endocytosis, and Cytotoxicity of Polypeptide ADC 6c

Antigen binding and subsequent endocytosis are two key processes by which an ADC exerts cytotoxicity. Next, we sought to ascertain whether the conjugation of a large MMAE-conjugated polypeptide to random lysine residues had any impact on these key ADC events. We prepared FITC-tagged 6c (6c-FITC) using preprepared FITC-tagged Herceptin. SK-BR-3 (HER2-positive) and MDA-MB-231 (HER2-negative) cells were incubated with 6c-FITC or FITC-tagged Herceptin (Her-FITC) at multiple time points, and intracellular transport of fluorescent signaling was visualized by confocal microscopy to verify target binding at the cell membrane and subsequent endocytosis of ADCs. Indeed, a strong FITC signal was observed at the cell membrane of SK-BR-3, but not MDA-MB-231, confirming the high HER2-binding affinity of Herceptin (Figure A). Encouragingly, marked FITC signaling was also detected in the cytoplasm of SK-BR-3 cells after 2–4 h of incubation, suggesting that 6c retained endocytic activity after binding to HER2 on target cells (Figure A). After endocytosis into the cell, the ADC payload will be released to exert potent cytotoxicity. Traditional ADC 7 was included in parallel as a positive control. We then tested the cytotoxicity of 6c in SK-BR-3 (HER2-positive), MDA-MB-453 (HER2-positive), and MDA-MB-231 (HER2-negative) cells. 7 was active in inducing cytotoxicity in SK-BR-3 cells with an IC₅₀ of 68 pM and MDA-MB-453 cells with an IC₅₀ of 191 pM, meanwhile, the IC₅₀ of 6c in SK-BR-3 cells was 11 pM, and in MDA-MB-453 cells was 101 pM, indicating stronger cytotoxicity as an ADC with higher DAR in HER2-positive target cells (Figure B, C). A fold-change discrepancy was observed between the cytotoxic activity and DAR of 6c and 7. We hypothesize that this is because 6c was produced via lysine conjugation, which exhibits greater heterogeneity than the cysteine conjugation strategy adopted for 7, and this additional heterogeneity may partly offset the advantage conferred by the higher DAR. Besides, 7 was essentially noncytotoxic, but a high concentration of 6c exhibited certain cytotoxic activity against HER2-negative MDA-MB-231 cells (Figure D). It was speculated that the main reason for the difference in MDA-MB-231 cytotoxicity between the two groups is the trace level of HER2 expression in MDA-MB-231 cells, leading to greater cytotoxicity of ADCs with high DAR than those with low DAR. Together, the polypeptide ADC 6c exhibits potent cytotoxicity, owing to its high DAR.

4.

Antigen binding, endocytosis, and subsequent cytotoxicity of 6c. (A) Fluorescence microscopy to illustrate the binding affinity between Herceptin-FITC or 6c-FITC and cell surface antigen, as well as the subsequent endocytosis after the target recognition. Blue, cell nuclei stained by DAPI; green, antibodies tagged by FITC. Scale bar, 60 μm. (B-D) The cytotoxicity of 6c and 7 against HER2-positive SK-BR-3 (B), HER2-positive MDA-MB-453 (C), and HER2-negative MDA-MB-231 (D). Values are the means ± SEM of three independent experiments.

The Plasma Stability and Pharmacokinetic Parameters of Polypeptide ADC 6c

Plasma stability and ADC exposure are key factors in evaluating eligibility for subsequent animal efficacy testing. First, the stability of the linker region of 6c was verified by monitoring the amount of MMAE released after incubation in human or mouse plasma. Similar to traditional ADC 7, polypeptide ADC 6c showed minimal decomposition in human plasma (Figure A). Consistent with the previous report, traditional ADC could be partly cleaved in mouse plasma. Although 6c was also partly cleaved, it was markedly weaker than 7 (Figure A). Besides linker stability, the whole structure stability of 6c was also detected through the FITC-tagged antibody (Figure B). The results showed no aggregation, indicating that the polypeptide design hardly alters the overall ADC structural stability in both human (Figure C) and mouse (Figure D) plasma. Next, we tested the pharmacokinetic properties of 6c in animals. SCID mice were intravenously (i.v.) treated with a single dose of 6c (10 mg/kg), and serum was collected at the indicated time points postadministration. Then, the antibody content in the serum was monitored using an IgG enzyme-linked immunosorbent assay (ELISA) according to the manufacturer’s instructions; meanwhile, the concentration of conjugated MMAE was determined as the difference between total MMAE and free MMAE by LC-MS/MS (Figure E). Polypeptide ADC 6c showed acceptable pharmacokinetic parameters in general (Figure F). The plasma concentrations of total Herceptin and conjugated MMAE gradually decreased after the first 5 min sampling time point with terminal elimination half-lives of approximately 4.6 and 5.1 days, respectively (Figure G). The plasma exposure AUC_0‑inf for Herceptin and conjugated MMAE was 2161 nmol·d/L and 21782 nmol·d/L, respectively (Figure G). The concentration of free MMAE in plasma was lower than the detection limit of the assay (10 ng/mL) except for the first 5 min time point, indicating that the free MMAE concentration was at least 1,000-fold lower than that of conjugated MMAE in plasma, and MMAE-conjugated polypeptide displayed rapid plasma elimination and was not detectable beyond the 24-h time point (Figure G).

5.

Plasma stability and pharmacokinetic parameters of 6c. (A) The stability of the linker of 6c in human and mouse plasma. (B–D) The whole structure stability test of 6c. The flow diagram of measurement (B), the result of human stability (C) or mouse (D) plasma. (E–G) Pharmacokinetic parameters test of 6c after a single dose (10 mg/kg, i.v.) in mice. Schematic representation (E), pharmacokinetics curves (F), and the calculational parameters (G). T _1/2, half-life; C _max, maximum plasma concentration; AUC, area under the curve. Values are the means ± SD of three independent experiments.

Polypeptide ADC 6c Can Effectively Inhibit Tumor Growth in Animals

We then established a mouse xenograft model to characterize and compare the efficacy of polypeptide ADC 6c and traditional ADC 7 in inhibiting SK-BR-3 tumor growth in animals. Tumor-bearing mice were treated with vehicle (PBS), Herceptin (2 mg/kg), MMAE (0.033 mg/kg), 7 (2 mg/kg, equivalent to 0.033 mg/kg MMAE), and gradient doses of 6c. Polypeptide ADC 6c showed dose-dependent efficacy relative to the vehicle group, confirming its potent antitumor activity in vivo. First, 0.12 mg/kg of 6c (equivalent to 0.033 mg/kg MMAE) was more effective at inhibiting tumor growth than MMAE. Meanwhile, 0.12 mg/kg of 6c exhibited an antitumor effect similar to that of the traditional ADC 7 at 2 mg/kg. But it was worth noting that the amount of antibody used in the 6c group was only 1/16 that in the 7 group. Additionally, at the same antibody dose (2 mg/kg), 6c could carried 16 times more MMAE than 7 and showed greater potency in achieving more potent antitumor efficacy. Furthermore, even at half the antibody dose, polypeptide ADC 6c (1 mg/kg) still showed greater efficacy than traditional ADC 7 (2 mg/kg). Treatment with both MMAE and Herceptin alone was moderately active (Figure A). All treatments were well tolerated, with no significant weight loss (Figure B). Even the polypeptide ADC 6c showed greater efficacy than the traditional ADC 7 but did not achieve complete tumor growth inhibition. We hypothesized that a single 2 mg/kg dose would be difficult to sustain sufficient systemic ADC concentration throughout the late experimental period. To further evaluate the in vivo antitumor activity of 6c, MDA-MB-453 tumor-bearing mice received two weekly injections of 6c at 5 mg/kg. 6c almost completely suppressed tumor growth (Figure C), and body weight monitoring confirmed favorable tolerability of this higher dose (Figure D). These data suggest that polypeptide ADC 6c exhibits potent efficacy against SK-BR-3 and MDA-MB-453 tumor growth in vivo without acute toxicity at its effective dose.

6.

Antitumor efficacy of 6c in vivo. (A–B) Mice were randomly grouped and given a single intravenous injection of vehicle, free MMAE, Herceptin, ADC 7 or 6c at the specified doses for antitumor efficacy assessment in the SK-BR-3 xenograft model, and tumor volumes (A) and body weight (B) were measured periodically (n = 6–7). (C–D) Mice were randomly assigned to groups and received two intravenous injections of vehicle or 5 mg/kg of 6c, administered once weekly, in the MDA-MB-453 xenograft model. Dosing time points are labeled with arrows, and tumor volume (C) and body weight (D) were monitored (n = 5–6). Values are the means ± SEM of animals in each treatment group. *P < 0.05, **P < 0.01, ***P < 0.001.

Conclusion

One of the most challenging aspects of ADC drug development is delivering sufficient payloads to target cancer cells. One approach is to increase the ADC’s DAR. Still, since the payloads tend to be highly hydrophobic, a high DAR often means faster plasma clearance, which is one of the crucial reasons why the DAR of ADC drugs under research is generally low. Herein, we developed a novel polypeptide backbone ADC platform, which can achieve high DAR loading and maintain the good medicinal properties of the resultant ADC. A polypeptide ADC 6c with a DAR up to 61 was produced using the combination of Herceptin and MMAE, based on the aforementioned polypeptide backbone platform. 6c retains the characteristics of antigen recognition, endocytosis, and subsequent cytotoxicity in multiple human cell lines. Meanwhile, 6c shows decent stability, pharmacokinetic parameters, and potent antitumor efficacy in animals. Due to the high DAR, the platform-based polypeptide ADC can be effective with fewer antibodies, achieving the payload of a traditional ADC, thereby reducing immunogenicity and improving pharmacokinetic properties. Besides, with the ease of adjusting the structure and payloads, the platform’s potency and hydrophilicity can be smoothly adjusted. We can envision adapting the novel polypeptide-backbone ADC platform reported herein to generate additional derivatives with various combinations of antibodies and payloads, potentially providing an effective means to develop innovative ADC drug candidates for the treatment of various cancers.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacsau.6c01025.

• Supplementary Figures, materials and methods, experimental arrangement details, synthetic methods, reaction schemes, and spectroscopic data (PDF)

X.L. conceptualized the project. X.L. and F.G. revised and finalized the manuscript. J.B. and R.T. designed and performed the chemical synthesis. F.G. designed and performed the cellular and animal experiments. F.G. and R.T. wrote the manuscript. All authors edited and approved the final manuscript. R.T. and F.G. contributed equally to this work

The authors declare no competing financial interest.