ETx-22, a Novel Nectin-4-Directed Antibody-Drug Conjugate, Demonstrates Safety and Potent Antitumor Activity in Low-Nectin-4-Expressing Tumors

Abstract

ETx-22, a novel ADC combining a tumor nectin-4-specific antibody and an innovative linker to exatecan, demonstrates significant and durable responses in low-target-expressing tumor models that are resistant to MMAE-based EV and has a better toxicity profile. This new ADC has the potential to benefit additional patient populations beyond its current indication.

Figure 1

Epitope characterization of the 15A7.5 antibody. A, The apparent EC₅₀ ratio of chimeric 15A7.5 and 5A12.2 and human HA22 anti–nectin-4 mAbs on normal differentiated human keratinocytes vs. the SUM190 cell line. B, Specificity of chimeric mAb 15A7.5 for human nectin-4. Flow cytometry normalized MFI signal from MDA-MB-231 cells (open triangles) and MDA-MB-231 cells transfected with human nectin-4–expressing plasmid (plain triangles) after incubation with a dose range of chimeric 15A7.5 mAb and staining with a PE goat anti-human antibody. Inlay: MFI signal recovered from MDA-MB-231 cells stained with anti-hNectin-1 (R1.302), anti-hNectin-2 (R2.477), or anti-hNectin-3 (N3.12) mAbs. C, Specificity of humanized mAb 15A7.5 for human nectin-4. The binding profile of humanized 15A7.5 was assessed with a human plasma membrane protein cell array. Shown are confirmation screen images of 293HEK live cells expressing membrane nectin-4 (green) or secreted tethered (blue) that demonstrate the specificity for nectin-4 of humanized mAbs 15A7.5 and 5A12.2. D, Chimeric 15A7.5 mAb binds to the IgV domain of nectin-4. Binding of chimeric 15A7.5 mAb was evaluated by ELISA on wells coated with the complete extra cellular domain of nectin-1 (Nec1-VCC), the complete extra cellular domain of nectin-4 (Nec4-VCC), or the IgV domain of nectin-4 only (Nec4-V). E, Cross-reactivity of chimeric 15A7.5 for cynomolgus nectin-4. Cos cells were transiently transfected with expression plasmids coding for human, cynomolgus, rat, or mouse nectin-4. Cells were incubated with a dose range of chimeric 15A7.5 and stained with PE goat anti-human antibody. F, Epitope of humanized 15A7.5, as determined by deep mutational scanning and represented on the reference structure of human nectin-4. MFI, mean fluorescence intensity.

Figure 2

Differential recognition and activity of the 15A7.5 antibody. A, Quick scoring results for the staining of each tissue with murine 15A7.5 and 5A12.2 antibodies. Frozen tissue sections of high- and low-nectin-4–expressing CDX breast cancers (SUM190 and SUM149, respectively), as well human skin, were processed to determine the level of nectin-4 expression by IHC. Right, The quick scoring ratio for both mAbs of SUM190 cells over human skin. B, Internalization of chimeric 15A7.5 and 5A12.2 mAbs. Shown is the ratio of apparent EC₅₀ value for internalization in SUM190PT cells over that in differentiated keratinocytes. C, Chimeric 15A7.5 (blue) and 5A12.2 (green) mAbs were conjugated to α-amanitin. Comparative in vitro cytotoxicity experiments (n = 3) were performed on nectin-4–expressing tumor cell lines (MDA-MB-468 and SUM190PT) and on differentiated keratinocytes. Shown are the apparent mean EC₅₀ values obtained and their calculated ratios for each cell type. Viability was measured via CellTiter-Blue for open symbols and IncuCyte for plain symbols. ns, not significant; QS, Quick Score.

Figure 3

Characterization of ETx-22 ADC. A, Schematic representation of ETx-22. B, Comparative HIC profile between the naked humanized 15A7.5 and ETx-22. C, DAR measure by LC/MS. Data indicate that cysteine from both light and heavy chains of the antibody are fully loaded with the linker payload yielding a DAR of 8. D,Ex vivo stability assessment. ETx-22 (from chimeric 15A7.5 mAb) was incubated in mouse, cynomolgus, or human serum. At the indicated times, ETx-22 was affinity-captured via anti-human LC-κ (mouse) or Fc–nectin-4 fragment (cynomolgus and human), and the DAR was measured by LC/MS. E, ETx-22 in vitro cytotoxic activity. A concentration range (1 pmol/L–6.67 nmol/L) of ETx-22 (from humanized 15A7.5 mAb, blue triangles) or an isotype control coupled to the same linker payload (open diamonds) was incubated with HCT-116-2G10 clone expressing human nectin-4 for 8 days. Viability was monitored through mitochondrial oxidation (cell titer blue reagent) using a BMG Labtech fluorometer. Ab, Antibody; LC, Light Chain; HC, Heavy Chain.

Figure 4

PK/PD characterization of ETx-22 ADC. NSG mice and PDX TNBC400-bearing NSG mice (n = 3 per time point) were injected intravenously when tumors averaged 150 mm³ (T0) with 10 mg/kg of ETx-22 (from humanized 15A7.5 mAb). At the indicated time points, terminal blood sampling was performed, and plasma was prepared from naïve and PDX TNBC400 mice. A, PK analysis of ETx-22 [ADC + total antibody (tmAb)] by Meso Scale Discovery technique. B,In vivo stability measurement by LC/MS after affinity capture of ETx-22. C, determination by LC/MS of circulating free exatecan concentration in plasma.

Figure 5

ETx-22 mechanism of action. PDX TNBC400-bearing NSG mice (n = 3 per time point) were injected intravenously when tumors averaged 150 mm³ (T0) with 10 mg/kg of ETx-22 (from humanized 15A7.5 mAb). At the indicated time points, animals were euthanized, and their tumors were collected, measured, and weighted. A, IHC analysis of ETx-22 infiltration and PD in PDX TNBC400 tumors. Tumors were fixed and embedded in paraffin. ETx-22 was detected with rabbit anti-human IgG which was revealed with a secondary anti-rabbit IgG coupled to HRP and a ChromoMap DAB kit. Phosphorylated H2A.X, which is a marker of topoisomerase I inhibitor activity, was detected with mouse anti–phospho-Histone H2A.X which was revealed with a secondary rabbit anti-mouse IgG and a tertiary anti-rabbit IgG coupled to HRP and a ChromoMap DAB kit. Scale bar, 50 μm. B, Western blot analysis of tumor lysates. ETx-22 was detected with a goat anti-human IgG conjugated to HRP. C, Quantification of phosphorylated H2A.X-positive cells. Slides in A were numerized using a Hamamatsu scanner, and Tribun Calopix software was used to quantify the percentage of phosphorylated H2A.X-positive cells. D, LC/MS determination of exatecan amount per gram of tumor. E, tumor volume measured using a caliper [V = (L × W × H) × π/6)].

Figure 6

In vivo efficacy of ETx-22 in SUM190 breast cancer cell line and PDX models. A, NSG mice (n = 5/group) were orthotopically xenografted bilaterally with SUM190PT cells embedded in Matrigel. At the indicated time (black arrowhead), three different ADCs were injected intravenously: isotype control (8 mg/kg) conjugated to maleimide–ß-glucuronide–exatecan–PSAR (open diamonds), ETx-22 (from chimeric 15A7.5 mAb) at 4 and 8 mg/kg (open and plain blue arrowheads, respectively), and EV at 4 mg/kg (plain brown arrowhead). B, Same as in A but NSG mice (n = 5/group) were orthotopically xenografted bilaterally with MMAE-resistant SUM190 cells embedded in Matrigel. Statistical significance between groups is indicated at the termination day of control group (P < 0.0001). The number of complete responses in the 8 mg/kg ETx-22 group is indicated. C, NMRI nude mice (n = 8/group) were subcutaneously implanted with tumor fragments from bladder cancer PDX B521. At the indicated time (black arrowhead), three different ADCs were injected intravenously: isotype control (8 mg/kg) conjugated to maleimide–ß-glucuronide–exatecan–PSAR (open diamonds), ETx-22 (from chimeric 15A7.5) at 4 and 10 mg/kg (open and plain blue arrowheads, respectively), and EV at 4 mg/kg (plain brown arrowhead). A chemotherapy control group (red diamonds) was treated with cisplatin (4 mg/kg, Q3W) plus gemcitabine (60 mg/kg, QW × 4). D, Same as in C with tumor fragments from bladder cancer PDX BLCU-003. ADC treatments were administered twice (black arrowheads). ETx-22 was from humanized 15A7.5 mAb. E, NSG mice (n = 5/group) were orthotopically xenografted bilaterally with TNBC317 tumor. At the indicated time (black arrowhead), three different ADCs were injected intravenously: isotype control (4 mg/kg) conjugated to maleimide–ß-glucuronide–exatecan–PSAR (open diamonds), ETx-22 (from chimeric 15A7.5) at 4 mg/kg (plain blue arrowheads), and EV at 4 mg/kg (plain brown arrowhead). F, Same as in E with TNBC348 tumor. Etx-22 and its isotype control were administered at 1 mg/kg and EV at 4 mg/kg. G, NOD/SCID mice (N = 10/ group) subcutaneously implanted with tumor fragments from ovarian cancer PDX OV2018. At the indicated time (black arrowhead), three different ADCs were injected intravenously: isotype control (10 mg/kg) conjugated to maleimide–ß-glucuronide–exatecan–PSAR (open diamonds), ETx-22 (from chimeric 15A7.5) at 4 mg/kg (plain blue arrowhead), and EV at 4 mg/kg (plain brown arrowhead). A chemotherapy control group (red diamonds) was treated with carboplatin (40 mg/kg, QW × 4). H, Same as in G with tumor fragments from ovarian cancer PDX OV2423. ADC treatments were administered twice (black arrowheads). ETx-22 was from humanized 15A7.5 mAb. In each animal model, (i) control group (plain black diamonds) received ADC diluent; (ii) tumor growth was monitored twice a week using a caliper, and tumor volume was calculated in mm³ as L × W² × π/6; and (iii) the H-score for nectin-4 expression as determined by IHC staining is indicated. Statistical significance between groups is indicated at the termination day of the control group (*, P < 0.05; **, P < 0.01; ****, P < 0.0001). On each graph, the number of complete responses in the highest-dosed ETx-22 group is indicated. CR, Complete Remission; d, day; QS, Quick Score.