Epigenetic Reprogramming Potentiates ICAM1 Antibody Drug Conjugates in Preclinical Models of Melanoma

Abstract

Therapeutic benefits and underlying biomechanism(s) of antibody drug conjugates (ADC) in combination with other targeted therapeutics are largely unknown. Here, the synergy between ADC and epigenetic drug decitabine (DAC), a clinically approved DNA methylation inhibitor, in multiple preclinical models of melanoma specifically investigated. Mechanistically, the underlying biomechanisms of how DAC cooperatively worked with ICAM1 antibody conjugated DNA topoisomerase I inhibitor DXd (I1-DXd) is elucidated. DAC treatment significantly enhanced anti-tumor efficacy of I1-DXd by upregulating antigen expression, enhancing antibody internalization and potentiating tumor sensitivity by epigenetically reprogramming of melanoma. Meanwhile, I1-DXd/DAC combination also exerted regulatory effects on tumor microenvironment (TME) by enhancing tumor infiltration of innate and adaptive immune cells and improving penetration of ADCs with a boosted antitumor immunity. This study provides a rational ADC combination strategy for solid tumor treatment.

Keywords: ICAM1; antibody drug conjugate; decitabine; melanoma; targeted therapy.

Figure 1

ICAM1‐ADCs show potent anti‐tumor efficacy in melanoma as monotherapy. (A) Box plots compare ICAM1 mRNA levels via melanoma to normal tissue; (B) Human melanoma cell and normal 293T surface expression of ICAM1 by flow cytometry (PE‐labeled antibody); (C) Representative images of IF staining of ICAM1 in human melanoma cells and normal 293T cell. Scale bar 10 µm; (D) Internalization curve of ICAM1 in melanoma cells (A375, C32 and SK‐MEL‐1) quantified by flow cytometry; (E) Schematic illustration and chemical structures of two ADC linkers and payloads; (F) In vitro cell growth inhibitory activity in A375, C32, SK‐MEL‐1 and 293T cell. Each point represents the mean and SD (n = 3). The red horizontal dotted line indicates the half maximal inhibitory; (G) Representative images of HE and IHC staining for ICAM1 in melanoma PDX model. Scale bar 50 µm; (H) Image of excised subcutaneous PDX tumors from mice treated with PBS (sham), I1 mab, I1‐MMAE, or I1‐DXd group (n = 5 per group); (I) Δ tumor volume in PDX model by tumor volume measurement by caliper. *p < 0.05; **p < 0.01; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (J) Tumor mass at the endpoint of PDX tumors quantified by weight. *p < 0.05; **p < 0.01; Bonferroni‐adjusted p value < 0.05 was considered statistically significant.

Figure 2

Combination with DAC significantly enhances the ICAM1‐ADC anti‐tumor efficacy in vitro and in vivo. (A) Chemical screen of 29 epigenetic probes identifies DNMT1 inhibitor as potential therapeutic target for melanoma. The color scale represents the z‐score of IC50. The lower IC 50 values indicated stronger anti‐proliferative activities effect in melanoma cells; (B) Combined inhibitory effects of various combination matrix of ICAM1‐ADCs (0‐66.7 nM) and DAC (A375: 0–100 µM; C32 and SK‐MEL‐1:0‐50 µM) in melanoma cells; (C) Image of excised subcutaneous PDX tumor from mice treated with PBS (sham), DAC, I1‐DXd, or Combo group (n = 5 per group); (D) Δ tumor volume curve in PDX model by tumor volume measurement by caliper. *p < 0.05; ***p < 0.001; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (E) Tumor mass weight at end point of subcutaneous PDX tumors. **p < 0.01; ***p < 0.001; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (F) Representative HE image of tumor in different groups. Scale bar 20µm. Yellow arrow indicated nuclear condensation; red arrow indicated nuclear fragmentation; (G) Representative Ki67 IHC staining image of tumor in different groups. Scale bar 20µm; (H) Quantitation of Ki67 positive cell proportion in different groups. *p < 0.05; ***p < 0.001; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (I) Blood chemistry parameters (ALT, AST, TB, BUN and Cre) measured when treated mice were sacrificed in different groups.

Figure 3

I1‐DXd monotherapy or in combination with DAC inhibits melanoma metastasis and improves survival in mice bearing A375‐luc xenograft tumors. (A) Representative bioluminescent images of mice A375‐luc xenograft tumors in different groups at different time points after injection; (B) Bioluminescence signal intensity at different time points in different groups (n = 5 per group); (C) Quantitative analysis of natural logarithm of tumor bioluminescence signal intensity in different groups at day 10. **p < 0.01; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (D) Quantitative analysis of natural logarithm of tumor bioluminescence signal intensity in different groups at day 20. **p < 0.01; ***p < 0.001; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (E) Quantitative analysis of natural logarithm of tumor bioluminescence signal intensity in different groups at day 30. **p < 0.01. (F) Kaplan–Meier survival curve of mice in different groups; (G) Bioluminescence signal intensity of major organs (heart, liver, lung, kidney and spleen) when mice were sacrificed in different groups; (H) Representative bioluminescence images of liver in different groups; (I) Quantitative analysis of liver metastasis burden as depicted from natural logarithm of bioluminescence signal intensity. **p < 0.01; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (J) Representative bioluminescence images of lung in different groups; (K) Quantitative analysis of lung metastasis burden as depicted from natural logarithm of bioluminescence signal intensity. *p < 0.05; Bonferroni‐adjusted p value < 0.05 was considered statistically significant.

Figure 4

I1‐DXd remodels TME to activate anti‐tumor immunity and synergizes with DAC in immunocompetent mice bearing B16OVA‐hICAM1. (A) Δ tumor volume curve in immunocompetent mice bearing B16OVA‐hICAM1 by tumor volume measurement by caliper. *p < 0.05; **p < 0.01; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (B) Tumor mass at end point of subcutaneous tumors quantified by weight. **p < 0.01; ***p < 0.001; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (C) Heat map of IFN signature, antigen presentation, chemokine, cytokine and collagen formation genes in different groups (n = 3 samples per group); (D) Relative proportion bar chart of 25 infiltrated immune cells in each samples according to CIBERSORT algorithm; (E‐G) Quantitative analysis of CD8+T cells (E), NK cells (F) and M1 macrophage cells (G) in different groups. *p < 0.05; ***p < 0.001; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (H) Representative HE staining images in different groups. Blue arrow indicated lymphoid cells; red arrow indicated tumor cells; green arrow indicated degenerative tumor cells; (I) Representative multiplex immunofluorescence images of tumor samples in different groups for CD8 (pink), CD86 (green) and CD11b (red); (J‐L) Fluorescence intensity analysis of CD8+T cells (J), NK cells (K) and M1 macrophage cells (L) in different groups. *p < 0.05; **p < 0.01; Bonferroni‐adjusted p value < 0.05 was considered statistically significant.

Figure 5

DAC induces ICAM1 expression and increases the internalization of ICAM1 antibody. (A) Heat map showing the down‐regulation of DNMTs and up‐regulation of ICAM1 and dsDNA sensor genes in the DAC group compared with PBS group; (B) GSEA for gene sets associated with cytosolic DNA sensing pathway in DAC group vs. PBS group; (C) Representative images of IF staining of dsDNA in different groups. Scale bar 10 µm; (D‐E) Quantitative fluorescence intensity analysis of dsDNA in A375 cells (D) and SK‐MEL‐1 cells (E). ***p < 0.001; (F) MFI change calculated by flow cytometry after DAC treatment. n = 3 biological replicates; (G) Quantitative analysis of ICAM1 high expression rate after DAC treatment. n = 3 biological replicates. ICAM1 high expression rate was defined by the rate of cells with ICAM1 expression higher than the middle point of the peak of the highest expressing sample in each cell line; (H) Schematic illustration for ICAM1 antibody internalization assay; (I) Quantitative analysis of ICAM1 antibody internalization at 2h and 4 h after DAC treatment in A375 cell (I) and SK‐MEL‐1 cell (J). *p < 0.05; **p <0.01; (K) Representative images of IF staining of ICAM1 antibody at 4 h after DAC treatment. Scale bar 10 µm; (L‐M) Quantitative fluorescence area analysis of tumor cell exposure to ICAM1 antibody at 4 h after DAC treatment in A375 cell (L) and SK‐MEL‐1 cell (M). *p < 0.05; **p <0.01.

Figure 6

DAC enhances the anti‐tumor efficacy of I1‐DXd by synergistically inducing apoptotic cancer cell death. (A) Heat map of apoptosis‐related genes expression in different groups; (B) GSEA for gene sets associated with apoptosis pathway in the Combination group vs. PBS group; (C) Western blots showing the DNA damage and apoptosis response followed by I1‐DXd monotherapy or combination treatment with DAC. β‐Actin is the loading control; (D) Representative images of IF staining of pH2AX in different groups. Scale bar 10 µm; (E‐F) Fluorescence intensity analysis of pH2AX in different groups in A375 cell (E) and SK‐MEL‐1 cell (F) (n = 100 cells per group); *p < 0.05; ***p < 0.001; Bonferroni‐adjusted p value < 0.05 was considered statistically significant. (G‐H) Representative TEM image of A375 (G) and SK‐MEL‐1 (H) after different treatment for 24h.

Figure 7

DAC enhances tumor penetration of I1‐DXd. (A) Schematic design of I1‐DXd biodistribution in PDX tumor; (B) Representative fluorescence images of tumor edge; yellow dotted line indicated the margin of tumor; (C) Representative fluorescence images of tumor core; (D) Quantitative analysis of fluorescence intensity in tumor edge in control or DAC group (n = 5 per group). *p < 0.05; (E) Quantitative analysis of fluorescence intensity in tumor core in control or DAC group (n = 5 per group). *p < 0.05; (F) Representative fluorescence images of the whole tumor in different groups; (G) Euclidean distance map of I1‐DXd‐Cy3 distribution in different groups (n = 5 per group).

Figure 8

Schematic depicting the antitumor mechanisms of I1‐ADC monotherapy (A) and the synergistic antitumor mechanisms of I1‐ADC in combination with DAC (B).