Immune and Epigenetic Landscape of TP53-mutated Acute Myeloid Leukemia (AML) and Higher-Risk Myelodysplastic Syndromes (HR-MDS)

Introduction: Mutations in TP53 occur in 10% of patients (pts) with AML and HR-MDS and have been associated with worse outcomes and an immunosuppressive phenotype. To define the immune and epigenetic landscape in TP53-M advanced myeloid neoplasms (MN), we compared data from 61 pts with HR-MDS or AML with TP53 mutations (TP53-M) to 143 TP53 wildtype (TP53-WT) pts who were followed prospectively with serial samples in a well-annotated clinical trial in which all pts received hypomethylating agent (HMA)-based therapy.

Methods: The FUSION trial (NCT02775903) was a large, randomized phase 2 study comparing azacitidine (AZA) monotherapy with AZA + anti-PD-L1 antibody durvalumab in 2 separate cohorts of previously untreated unfit AML and HR-MDS pts (Zeidan A et al, ASH 2019). Responses were classified by IWG 2003 and 2006 criteria for AML and MDS, respectively. Survival was estimated using Kaplan-Meier techniques. Samples from peripheral blood (PB) and bone marrow (BM) were obtained at baseline and serially on trial. A 38-targeted mutation analysis was performed at Munich Leukemia Laboratory. Only level 1 pathogenic TP53 mutations with variant allele frequency (VAF) ≥2% were included. DNA methylation was assessed using Illumina's Infinium Human Methylation EPIC methylation array. Immunophenotyping and immune checkpoint expression was performed using flow cytometry. Gene expression profiles were studied by RNA-sequencing.

Results: Of 129 AML and 84 HR-MDS pts enrolled in FUSION trial, 37 had TP53-M AML, 88 TP53-WT AML, 24 TP53-M HR-MDS, and 55 TP53-WT HR-MDS pts. The average VAF for TP53 mutations were 37%, and 90% had ≥10% VAF. TP53-M AML pts were more likely to have poor-risk cytogenetics, therapy-related disease, and lower BM blast percentage compared to TP53-WT AML pts. TP53-M HR-MDS were more likely to have secondary MDS, very poor risk cytogenetics by IPSS-R, and very high risk IPSS-R score. There were no statistically significant differences in overall response rate (ORR) between TP53-M and TP53-WT pts (AML cohort: ORR: 35.1% [95% CI: 21%-53%] vs. 34.1% [CI: 26%-45%]; HR-MDS cohort: ORR: 41.7% [CI: 23%-63%] vs. 60% [CI: 46%-73%]). Median OS was 8.1 [95% CI: 5.4 - 13] months (mos) among TP53-M AML pts and 16.6 [95% CI: 13 - 21] mos for TP53-WT AML pts [Figure 1A]. Median OS was 9.8 [95% CI: 9.3 - 20+] mos for TP53-M HR-MDS pts and 23.5 [95% CI: 12 - 25+] mos for TP53-WT HR-MDS pts [Figure 1B]. Global DNA methylation was independent of TP53 mutation status in both AML (global methylation score x10^5: TP53-M: 4.9 [SD: 0.23] vs TP53-WT: 4.8 [SD: 0.33]; p=0.35) and HR-MDS pts (global methylation score: TP53-M: 4.8 [SD: 0.20] vs TP53-WT: 4.7 [SD: 0.25]; p=0.24) at baseline. DNA methylation changes after one cycle of AZA treatment were similar in both cohorts (AML: TP53-M:4.4 [SD: 0.38] vs TP53-WT: 4.5 [SD: 0.41, p=0.33]; HR-MDS: TP53-M: 4.3 [SD: 0.35] vs TP53-WT: 4.3 [SD: 0.36]; p=0.52). In RNA sequencing (Figure 2), TP53-M pts had higher expression of T-cell genes (e.g. IL7R) in both AML and HR-MDS compared to TP53-WT pts. Compared to TP53-WT pts, IFN alpha signature genes were reduced only in TP53-M AML pts but were increased in TP53-M HR-MDS pts. PD-L1 (CD274) expression was correlated with the T-cell gene signature and had a higher expression in TP53-M samples. TP53-M pts showed lower expression of tumor associated genes (e.g. CD34) consistent with the tumor cell percentages seen by BM flow cytometry. However, in gene set enrichment analysis, MYC target genes, MTORC1, and E2F were enriched in TP53-M samples consistent with the higher expression of proliferation genes (e.g., MKI67). In the bone marrow flow cytometry of AML pts, more T-cells were seen in TP53-M pts (Figure 3A), and PDL1 positive tumor cells were trending higher in TP53-M pts while the total abundance of tumor cells was slighter higher in TP53-WT (Figure 3B).

Discussion: We confirm here that achieving a response to AZA therapy in AML or HR-MDS pts is not impacted by presence of TP53 mutations, however as expected median OS was substantially shorter among TP53-M pts for both AML and HR-MDS. In analyzing the epigenetic landscape, there were no differences in baseline global DNA methylation by TP53 status. RNA sequencing showed enrichment of T-cell genes and PD-L1, and an increase in gene expression of proliferation genes in TP53-M pts. Taken together, these findings support the presence of immunosuppressive microenvironment among TP53-M pts with advanced MN.

Figure 1

View largeDownload slide

Figure 1

View largeDownload slide

Close modal