Hypomethylation and up-regulation of PD-1 in T cells by azacytidine in MDS/AML patients: A rationale for combined targeting of PD-1 and DNA methylation

Abstract

The hypomethylating agents (HMAs) are standard therapy for patients with higher-risk myelodysplastic syndrome (MDS); however, the majority of the patients will lose their response to HMAs over time due to unknown mechanisms. It has recently been shown that T cell expression of the immunoinhibitory receptor PD-1 is regulated by DNA methylation. In 12 of 27 patients (44%) PD-1 promoter demethylation was observed in sorted peripheral blood T cells isolated over consecutive cycles of treatment with 5-azacytidine (5-aza). The PD-1 promoter demethylation correlated with an increase in PD-1 expression. Moreover, demethylation of the PD-1 promoter correlated with a significantly worse overall response rate (8% vs. 60%, p = 0.014), and a trend towards a shorter overall survival (p = 0.11) was observed. A significantly higher baseline methylation level of the PD-1 promoter was observed in T cells of non-responding patients compared to healthy controls (p = 0.023). Accordingly, in addition to their beneficial function, HMAs induce PD-1 expression on T cells in the MDS microenvironment, thereby likely hampering the immune response against the MDS blasts. Thus, we suggest that activation of the PD-1 checkpoint during HMA treatment can be a possible resistance mechanism, which may be overcome by combination therapy with a PD-1 pathway inhibitor.

Keywords: DNA methylation; T cells; hypomethylating agents; myelodysplastic syndromes; programmed death-1.

Figure 1. Mean PD-1 promoter methylation in six distinct cell populations from 5 healthy donors

All cells are from peripheral blood. The graph shows the means. PBMNC = peripheral blood mononuclear cells. CD3 = CD3⁺ T cells. CD4 = CD4⁺ T cells. CD8 = CD8⁺ T cells. Gran = granulocytes. B cells = CD19⁺ B cells.

Figure 2. Dynamics of PD-1 promoter methylation in peripheral blood mononuclear cells of 15 patients during treatment with 5-azacytidine

(A) PD-1 promoter methylation in the 9 patients in whom we observed a classifiable demethylation. (B) PD-1 promoter methylation in the 6 patients in whom we did not observe a classifiable demethylation. C = course of 5-aza treatment. D = day in treatment course. PBMNC = peripheral blood mononuclear cells.

Figure 3. Dynamics of PD-1 promoter methylation in peripheral blood CD4⁺ and CD8⁺ T cells from the patients with PD-1 promoter demethylation during treatment with 5-azacytidine

(A) PD-1 promoter methylation in CD4⁺ T cells of eight patients. (B) PD-1 promoter methylation in CD8⁺ T cells of eight patients. In patient no. 123 we only observed demethylation in the CD4⁺ T cells (CD8⁺ T cells from patient no. 123 are included in Figure 4B) and in patient no. 60 we only observed demethylation in the CD8⁺ T cells (CD4⁺ T cells from patient no. 60 are included in Figure 4A). C = course of 5-aza treatment. D = day in treatment course.

Figure 4. Dynamics of PD-1 promoter methylation in peripheral blood CD4⁺ and CD8⁺ T cells from patients without PD-1 promoter demethylation during treatment with 5-azacytidine

(A) PD-1 promoter methylation in CD4⁺ T cells of 14 patients. (B) PD-1 promoter methylation in CD8⁺ T cells of 14 patients. In patient no. 123 we observed demethylation in the CD4⁺ T cells (CD4⁺ T cells from patient no. 123 are included in Figure 3A) and in patient no. 60 we observed demethylation in the CD8⁺ T cells (CD8⁺ T cells from patient no. 60 are included in Figure 3A). C = course of 5-aza treatment. D = day in treatment course.

Figure 5. Baseline PD-1 promoter methylation in peripheral blood CD4⁺ and CD8⁺ T cells from 5-azacytidine treated patients and five healthy donors

The patients are grouped according to whether the PD-1 promoter demethylates or not during treatment. The values for CD4⁺ and CD8⁻ T cells are pooled.

Figure 6. PD-1 protein surface expression on CD8⁺ T cells from 5-aza treated patients

PD-1 protein surface expression, as expressed by the median fluorescence intensity (MFI), on CD8⁺/CD45RO⁺ T cells from PBMNCs of four patients at baseline before 5-aza treatment (1) and following the first treatment cycle (2) (the first available sample for every patient after the first treatment cycle. Patient no. 94, 123, 141 and 146: C4D5, C3D1, C2D1 and C3D1, respectively (C = course of 5-aza treatment. D = day in treatment course)). In patient no. 94 and 141 an increase in PD-1 protein expression is observed (MFI: 770 to 853 (MFIR: 1.11), and MFI: 220 to 358 (MFIR: 1.63), respectively), whereas in patient no. 123 and 146 no increase is observed (MFI: 336 to 291 (MFIR: 0.89), and MFI: 426 to 373 (MFIR: 0.88), respectively). Correspondingly, we observed a significant demethylation of the PD-1 promoter during treatment in CD8⁺ T cells in patient no. 94 and 141 (Figure 3B) and no demethylation in patient no. 123 and 146 (Figure 4B). Moreover, we saw a higher baseline MFI of PD-1 in patient 94 compared to patient 141, which is in line with the variation in baseline methylation level.

Figure 7. Correlation between PD-1 promoter methylation and PD-1 gene expression in peripheral blood CD4⁺ and CD8⁺ T cells from 5-azacytidine treated patients

Ten patients and their corresponding values of methylation and relative expression before and during 5-aza treatment (altogether 117 pairs of observations). Methylation level is plotted as co-variate and the relative expression as outcome on a log-log scale with the base 2. The linear regression reveals a statistically significant inverse relationship between the two log-transformed variables (p < .0001) with the regression coefficient −2.0936 (SE 0.2521).

Figure 8. Baseline PD-1 promoter methylation in peripheral blood CD4⁺ and CD8⁺ T cells from 5-azacytidine treated patients and five healthy donors

The patients are grouped according to whether they responded or not responded to the treatment. The values for CD4⁺ and CD8⁻ T cells are pooled.