Combined TNF-α and OX40L targeting as a new treatment option for hidradenitis suppurativa

Abstract

Background: Chronic inflammatory conditions are among the leading causes of disability and mortality. Although therapies have been significantly improved with the introduction of target-specific biologics, many chronic inflammatory conditions can be only moderately controlled by inhibition of individual cytokines.

Objective: We sought to compare individual versus simultaneous blockade of TNF-α and OX40L in controlling inflammation.

Methods: Analysis was conducted of the murine xenograft graft-versus-host disease model, a novel cynomolgus monkey model of simultaneous T cell-dependent antibody response (TDAR) and delayed-type hypersensitivity (DTH) skin reaction, and samples of patients with hidradenitis suppurativa (HS).

Results: Compared with individual inhibition, combined targeting of TNF-α and OX40L using mAbs more potently suppressed TDAR and DTH in cynomolgus monkey. We therefore created SAR442970, a bispecific pentavalent Nanobody containing 2 domains each binding to OX40L and TNF-α and 1 domain binding serum albumin to extend half-life. In xenograft graft-versus-host disease, disease control by SAR442970 was superior compared with treatment with monospecific anti-TNF-α or anti-OX40L Nanobodies. In cynomolgus monkey, SAR442970 potently suppressed TDAR and DTH. The transcriptional signature of inflamed monkey skin showed similarities to that of lesional skin of patients with atopic dermatitis and, even more so, HS and was inhibited by SAR442970 treatment. Increased numbers of cells expressing OX40 were observed in HS lesions, and bioinformatics analyses of single-cell and bulk RNA sequencing data identified cells expressing OX40 as T cells with a highly mobile phenotype.

Conclusion: OX40-OX40L interaction is a key pathogenetic feature in HS, and patients with HS might benefit from combined TNF-α/OX40L blockade by SAR442970, which is currently investigated in a clinical phase 2 trial.

Keywords: Nanobody; OX40; OX40L; T cell–dependent antibody response; TNF-α; delayed-type hypersensitivity skin reaction; hidradenitis suppurativa; xenograft graft-versus-host disease.

Fig 1

Impact of individual versus combined treatment with anti-TNF-α (adalimumab) and anti-OX40L (oxelumab) mAbs in cynomolgus monkey. (A) Graphical scheme of treatments in cmTDAR-DTH model. (B) Specific anti-KLH IgM levels and (C) anti-KLH IgG levels were measured using a specific ELISA method. Graphical representations for the time course of anti-KLH IgM and IgG concentrations (μg/mL) for all groups (mean ± SEM; n = 4). (D) Immunohistochemistry assessment of CD68⁺, CD3⁺, CD4⁺, CD8⁺, CD20⁺, and Ki67⁺ cells in intradermal DTH challenge sites in skin biopsy specimens collected on day 34 and at necropsy on day 59, 3 days after intradermal immunization with KLH on days 31 and 56. Heatmap based on mean scores obtained from microscopic analyses in 3 (day 34) or 2 (day 59) biopsy specimens each for all groups (n = 4). (E) PBMCs isolated from blood samples of cynomolgus monkeys taken on day 34 and day 59 were evaluated for IFN-γ secretion capacity in response to ex vivo KLH stimulation using ELISPOT analysis (n = 4 per treatment group). Frequencies of cells secreting IFN-γ are shown as paired (day 34 and day 59) individual mean values. Ctrl, Control; IV, intravenous; sfc, spot-forming cells.

Fig 2

Design of SAR442970 and in vitro characterization. (A) Schematic representation of the pentavalent, bispecific, trifunctional Nanobody compound consisting of 5 VHH domains derived from heavy-chain llama antibodies. (B) Binding sensorgrams with binding level (RU) of different sample injections flowed over SAR442970 captured on a surface immobilized with human serum albumin. (C) Functional inhibition of human or cynomolgus OX40L mediated T-cell costimulation by SAR442970 and oxelumab. Representative experiment graph for human OX40L and mean IC₅₀ values of independent experiments (n = 3) with corresponding 95% CI. (D) Functional inhibition of human and cynomolgus TNF-α by SAR442970 and adalimumab in the nuclear factor-κB dependent reporter assay. Representative experiment graph for human TNF-α and mean IC₅₀ values of independent experiments (n = 3) with corresponding 95% CI. cps, Counts per second; HSA, human serum albumin.

Fig 3

Impact of individual and combined TNF-α and OX40L blockade in NOG mice treated with human PBMCs from 2 donors to undergo xGVHD. (A) NOG mice, irradiated with 1 Gy, were injected with 2 × 10⁷ human PBMCs on day 0. Mice were then treated 3 times weekly with PBS, isotype control neg-ctrl-nb-cmpd, monospecific anti-TNF-α-nb-cmpd, monospecific anti-OX40L-nb-cmpd, or bispecific SAR442970 each at 3 nmol/animal, respectively, until the study was terminated at day 127. (B) Schematic representation of the different Nanobody VHH constructs used. (C) Kaplan-Meier survival curves for pooled NOG mice from 2 donors undergoing xGVHD. (D) IFN-γ levels (pg/mL) in plasma at day 14 (mean ± SEM; n = 14). ANOVA statistical test followed by Tukey’s multiple comparisons test. IV, Intravenous.

Fig 4

Impact of SAR442970 treatment in cmTDAR-DTH model. (A) Graphical scheme of treatments in cmTDAR-DTH model, including time points for application of subcutaneous and intradermal KLH challenges for TDAR and DTH, respectively. (B) Specific anti-KLH IgM levels and (C) anti-KLH IgG levels were measured using a specific ELISA method. Graphical representations for time course of anti-KLH IgM and IgG concentrations (μg/mL) for all groups (mean ± SEM; n = 4). (D) Exposure-response of the bispecific SAR442970 with standard monotherapy in the pooled analysis from first and second cynomolgus monkey studies. (E) Immunohistochemistry assessment of CD68⁺, CD3⁺, CD4⁺, CD8⁺, CD20⁺, and Ki67⁺ cells in intradermal DTH challenge sites in skin biopsy specimens collected on day 34 and at necropsy on day 59, 3 days after intradermal immunization with KLH on days 31 and 56. Heatmap based on mean scores obtained from microscopic analyses in 3 (day 34) or 2 (day 59) biopsy specimens each for all groups (n = 4). (F) Frequency of cells secreting IFN-γ determined by ELISPOT analyses for the specific immune response against KLH ex vivo in PBMCs isolated from animals on day 59. Distribution of individual animal values with mean shown (n = 4). AUC, Area under the curve; Ctrl, control; sc, subcutaneous.

Fig 5

SAR442970 effect on cynomolgus monkey DTH response and its comparison with HS patient RNA-Seq data. (A) Numbers of significantly regulated genes in cynomolgus monkey DTH skin in different treatment groups compared with a vehicle control at day 34 and day 59. (B) Overlap of significantly regulated genes in different SAR442970 dose groups compared with vehicle control at day 34. (C) Visualization of overlap of differentially expressed genes in cynomolgus DTH skin with differentially expressed genes in human skin diseases HS, atopic dermatitis, and psoriasis (always fold change ≥ 1.5, false discovery rate [FDR] < .05). (D) Clustering of significantly regulated genes present in at least 2 SAR442970 treatment groups based on their expression (transcripts per million) levels at day 34 in monkey DTH skin and of respective human genes in HS nodules or human healthy skin. (E) IPA analysis for the enrichment of pathways in human HS lesional versus healthy skin compared with monkey DTH skin from the 100 mg/kg versus vehicle treatment group at day 34. The y-axis depicts the percentual overlap of significantly upregulated (red), significantly downregulated (blue), not regulated (no change), and genes not overlapping with the corresponding pathway genes (no overlap) in both comparisons. Asterisks depict the Benjamini-Hochberg corrected P value of the enrichment: ∗∗P < .01, ∗∗∗P < .001. (F) Visualization of top 20 significantly regulated upstream regulators in human HS patient skin compared with skin from healthy human donors (HS vs healthy), and monkey DTH skin biopsy specimens taken on day 34 from animals treated with increasing doses of SAR442970 compared with vehicle-treated controls. (G) Expression of IFNG, TNF, IFNAR1, IFNAR2, and IL1B in human healthy skin (n = 8) and HS lesional skin (n = 31). (H) Expression of IFNG, TNF, IFNAR1, IFNAR2, and IL1B in cynomolgus monkey DTH model skin at day 34. Significantly regulated genes are with |fold change ≥ 1.5 (for monkey) or fold change ≥ 2 (for HS), FDR < .05 and ≥1 transcripts per million expression in at least 2 samples per group. n = 4 per group for cynomolgus monkey cohort and n = 31 lesional versus n = 8 healthy samples for entire human HS cohort. Asterisks depict the following FDR for differentially expressed genes: ∗∗FDR < .01, ∗∗∗FDR < .001. AD, Atopic dermatitis; ns, not significant; Pso, psoriasis.

Fig 6

OX40 and OX40L in human HS lesional skin. (A) mRNA expression of OX40 and OX40L in human healthy skin (n = 8) and HS lesional skin (n = 31). Values represent transcripts per million with ∗∗∗∗P < .0001 (Mann-Whitney test). (B) Quantitative density image analyses of immunohistochemical stains of cells expressing OX40 and OX40L in healthy control skin (n = 15) and HS patient skin (n = 23) samples (mean ± SD) with ∗∗∗∗P < .0001 (Mann-Whitney test). (C) Representative images of immunohistochemical staining of OX40 in healthy control skin (I) and HS patient skin (II, III, and IV). (D) Expression of OX40 mRNA in different single-cell populations in HS patient skin from the Gudjonsson et al dataset. Higher color intensity corresponds to higher expression. (E) Molecules in OX40-positive cells whose expression levels correlate with OX40 levels. Molecules with a highly significant (rs > 0.75; adjusted P < 6.5 × 10⁻⁶) or significant (rs > 0.50, adjusted P < .01) positive correlation with OX40 expression are depicted in red or blue, respectively. ab, Abscess; ns, not significant; TPM, transcripts per million; tu, tunnel. Created with BioRender.com.

Supplementary Fig E1

Supplementary Fig E2