Beyond Recycling Antibodies: Crovalimab's Molecular Design Enables Four-Weekly Subcutaneous Injections for PNH Treatment

Abstract

The advent of recycling antibodies, leveraging pH-dependent antigen binding and optimized FcRn interaction, has advanced the field of antibody therapies, enabling extended durability and reduced dosages. Eculizumab (Soliris^®) demonstrated the efficacy of C5 inhibitors for paroxysmal nocturnal hemoglobinuria (PNH), while its derivative, ravulizumab (Ultomiris^®), recognized as a recycling antibody, extended the dosing intervals. However, limitations including intravenous administration and inefficacy in patients with the R885H single-nucleotide polymorphism (SNP) in C5 could necessitate alternative solutions. Crovalimab (PiaSky^®), a next-generation recycling antibody, overcomes these challenges with innovative charge engineering, achieving the enhanced cellular uptake of C5-crovalimab complexes and targeting a unique C5 epitope, allowing for efficacy regardless of the R885H SNP. This study highlights crovalimab's distinctive molecular features, showing its eliminated binding to Fcγ receptors and C1q, alongside its optimized antigen binding characteristics. The impact of charge engineering was reconfirmed in mice, demonstrating faster C5 clearance than recycling antibodies. Notably, in the maintenance dosing regimen, crovalimab neutralizes approximately seven C5 molecules per antibody on average. Furthermore, its design also reduces the viscosity to facilitate high-concentration formulations suitable for subcutaneous delivery. Consequently, crovalimab offers a four-weekly subcutaneous injection regimen for PNH, marking a substantial improvement in treatment convenience and potentially transforming patients' quality of life.

Keywords: anti-C5 antibody; charge engineering; crovalimab (PiaSky®); eculizumab; pH-dependent antigen binding; paroxysmal nocturnal hemoglobinuria; ravulizumab; recycling antibody; subcutaneous injection; viscosity.

Figure 1

Schematic overview of the crovalimab creation journey. This diagram illustrates the key steps in the creation of this therapeutic agent, starting with the identification of the lead antibody, followed by the engineering of a recycling antibody and the subsequent enhancements that led to the final antibody, crovalimab. Sites where key antibody engineering technologies were applied are highlighted in the accompanying illustration of crovalimab. FcRn, neonatal fragment crystallizable receptor; FcγR, IgG-Fc receptor.

Figure 2

Amino acid alignment of crovalimab with human germline sequences. The amino acid sequence of the crovalimab Fc region (SG115) is shown together with the human IgG1 and IgG4 sequences. The green and blue regions in the sequence of SG115 indicate, respectively, the human-derived sequences and the artificial mutations introduced to improve the properties of the antibody. Ig, immunoglobulin.

Figure 3

FcγR binding characteristics of crovalimab. SPR analysis of binding between hFcγRs and antibodies using the Biacore T200 system. Crovalimab and trastuzumab (control) were immobilized on a Protein A-coated sensor chip, and various hFcγRs, including polymorphic variants, were injected over them. Binding responses were normalized to 1 RU of antibody. Data represent mean ± SD (n = 3). Unlike trastuzumab, crovalimab showed no detectable binding to any hFcγRs, suggesting a minimal risk of activating effector functions through FcγRs. F, phenylalanine; H, histidine; hFcγR, human Fcγ receptor; NA, neutrophil antigen; R, arginine; RU, resonance units; SD, standard deviation; SPR, surface plasmon resonance; V, valine.

Figure 4

C1q binding characteristics of crovalimab. The binding activity of crovalimab, rituximab (positive control), and natalizumab (negative control) to human C1q protein was evaluated by ELISA. Antibodies were immobilized on 96-well plates, incubated with human C1q protein, and detected using anti-human C1q antibody. Data represent mean ± SD (n = 4). Points represent the average of the measured values for crovalimab (blue), rituximab (red), and natalizumab (green). Crovalimab and natalizumab showed comparably little binding to human C1q, in contrast to rituximab, indicating a minimal risk of complement activation through C1q binding with crovalimab despite its human IgG1 backbone. ELISA, enzyme-linked immunosorbent assay; SD, standard deviation.

Figure 5

Time course profiles of (A) human C5 concentration and (B) anti-C5 antibody concentration in human FcRn transgenic mice after IV administration of human C5 (0.1 mg/kg) with or without an anti-C5 antibody (20 mg/kg). Data are presented as mean ± SD with n = 3. The limit of quantification was 10 ng/mL for C5 and 0.02 µg/mL for anti-C5 antibodies, and values below this threshold are not displayed. Notably, the plasma human C5 concentrations fell below the detection limit by day 7 in both the C5 alone and crovalimab-treated groups, leading to the truncation of the data lines at day 3, as indicated by the arrows. This observation highlights the effectiveness of recycling antibody technology and surface-charge engineering in accelerating C5 clearance. The plasma concentrations of all antibodies remained similar throughout the study period. Reprinted from Sampei Z, et al. [31] and Fukuzawa T, et al. [33]. Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/ accessed on 1 October 2024). FcRn, neonatal Fc receptor; IV, intravenous; SD, standard deviation; SIA, sequence-identical analog.

Figure 6

Relationship between antibody concentration and viscosity within the same formulation: a 20 mM histidine, 150 mM arginine/aspartic acid buffer (pH 6.0). The measured values of the antibody concentration and viscosity are plotted in the graph and presented in the table, which also includes the calculated viscosity at 170 mg/mL and the antibody concentration at which the viscosity reaches 20 mPa·s. Crovalimab demonstrated lower viscosity compared with ravulizumab-SIA at equivalent concentrations, enabling a high-concentration formulation (170 mg/mL) suitable for SC injection. Viscosity measurements were conducted at 25 °C using an EMS-01S viscometer. Each data point represents the mean of three measurements (n = 3), with each measurement being the average of 10 consecutive readings. The buffer viscosity was subtracted from each measurement. An exponential approximation curve was applied to the corrected values for precise modeling, with an R² value approaching 1. SIA, sequence-identical analog.

Figure 7

Cellular uptake of immune complexes and subsequent impact on C5 accumulation across various antibody types, including conventional non-pH-dependent antibodies, recycling antibodies without surface-charge engineering, and recycling antibodies with surface-charge engineering (such as crovalimab).