Purpose:

The aim of this study is to determine immune-related biomarkers to predict effective antitumor immunity in myelodysplastic syndrome (MDS) during immunotherapy (IMT, αCTLA-4, and/or αPD-1 antibodies) and/or hypomethylating agent (HMA).

Experimental Design:

Peripheral blood samples from 55 patients with MDS were assessed for immune subsets, T-cell receptor (TCR) repertoire, mutations in 295 acute myeloid leukemia (AML)/MDS-related genes, and immune-related gene expression profiling before and after the first treatment.

Results:

Clinical responders treated with IMT ± HMA but not HMA alone showed a significant expansion of central memory (CM) CD8+ T cells, diverse TCRβ repertoire pretreatment with increased clonality and emergence of novel clones after the initial treatment, and a higher mutation burden pretreatment with subsequent reduction posttreatment. Autophagy, TGFβ, and Th1 differentiation pathways were the most downregulated in nonresponders after treatment, while upregulated in responders. Finally, CTLA-4 but not PD-1 blockade attributed to favorable changes in immune landscape.

Conclusions:

Analysis of tumor–immune landscape in MDS during immunotherapy provides clinical response biomarkers.

Translational Relevance

Although currently, immune checkpoint inhibitors are under active investigation for myelodysplastic syndrome (MDS), there are limited data on immune biomarkers and novel mechanisms of response or resistance to immune checkpoint blockade in MDS. For these purposes, we performed a comprehensive immune monitoring strategy in 55 patients with MDS who received immunotherapy (IMT, αCTLA-4, and/or αPD-1 antibodies) and/or hypomethylating agent (HMA). Longitudinal analysis of tumor–immune landscape in MDS during checkpoint blockade supports an essential role of T cells in antitumor immunity. CTLA-4, but not PD-1 blockade, attributes to shaping a favorable immune landscape in MDS. Autophagy inducers and T-cell costimulators, such as ICOS agonists, may play a role in overcoming immune resistance.

Myelodysplastic syndrome (MDS) is a myeloid clonal disorder that mostly occurs in the elderly population, and is characterized by cytopenia, dysplasia, and a propensity to transform into acute myeloid leukemia (AML; ref. 1). Although hypomethylating agents (HMA) are the standard treatment for patients with high-risk MDS, especially elderly patients (2–4), resistance to HMAs occurs in most MDS patients leading to a dismal outcome (5). Therefore, novel therapeutics for high-risk MDS patients are needed to improve clinical outcome.

The aberrant expression of immune checkpoint proteins in the tumor and tumor microenvironment is thought to be the mechanism of immune evasion (6, 7), as blockage of the immune checkpoint pathway has been shown to effectively eradicate certain solid tumors (8–10) and hematologic malignancies (11, 12). As the overexpression of immune checkpoint proteins was observed in patients with MDS during HMA treatment (6), immune checkpoint blockade can not only resensitize the tumor–immune microenvironment (TIME) in patients with MDS who failed HMAs, but also prolong the duration of response to HMAs in patients with MDS. However, there are limited data on immune biomarkers and novel mechanisms of response or resistance to immune checkpoint blockade in MDS.

In this study, we aimed to determine immune-related biomarkers to predict effective antitumor immunity and uncover the mechanisms of immune resistance in 55 patients with MDS who received immune checkpoint blockade (αCTLA-4 and/or αPD-1 antibody) ± 5-Azacitidine or guadecitabine alone. As we have previously demonstrated that the immune and genetic landscape of peripheral blood (PB) mirrored those of bone marrow in patients with MDS (13), we utilized PB for a comprehensive immune monitoring strategy, and investigated the evolution of the immune landscape through comprehensive immunophenotypic analysis of T cells, next-generation sequencing (NGS)-based T-cell receptor (TCR) beta repertoire analysis, and immune-related gene expression during therapy. In addition, we evaluated the coevolution of tumor landscape through NGS-based targeted sequencing of 295 genes that are commonly mutated in AML/MDS.

Patient cohort

A total of 55 patients with MDS who receive immunotherapy (IMT, αCTLA-4, and/or αPD-1 blockade) ± HMA, and HMA alone in clinical trials (NCT02530463, NCT03094637, NCT02131597, and NCT02599649) were included in this study (Supplementary Table S1; Supplementary Fig. S1). Patients with chronic myelomonocytic leukemia (CMML) were permitted to enroll clinical trials, and 6 patients with CMML were included in the analysis. Clinical responses were determined according to the International Working Group responses criteria for MDS (14). Written informed consent was obtained from each patient or each patient's guardian before participation in this study. All research was conducted in accordance with the Declaration of Helsinki and The University of Texas MD Anderson Cancer Center Institutional Review Board guidelines.

Blood sample collection and isolation of PB mononuclear cells

A total of 10 cc of PB samples were collected prior to the initiation of treatment, each cycle of treatments up to cycle 8, and at the time of relapse (Supplementary Fig. S2). PB samples were processed within 24 hours of collection. Briefly, white blood cells (WBC) were obtained by lysing red blood cells (RBC) from 1 cc of peripheral blood, snap-frozen, and stored at −80°C until genomic DNA was isolated for TCRβ repertoire and mutation analysis. PBMCs were isolated from the remaining blood by standard density centrifugation using Ficoll-Paque (Sigma), portion of PBMCs was used for upfront flow cytometry immune-subset analysis. Remaining PBMCs were cryopreserved in freezing media containing 10% DMSO (Sigma) and 90% FBS (Corning) in N2 for future use including NanoString gene expression analysis.

Multiparameter flow cytometric analysis

Briefly, freshly prepared PBMCs were stained with the following antibodies: CD56-BUV395, CD16-BUV395, CD3-FITC, CD4-APC, CD8α-PerCP, HLA-DR-PE, CD45RA-PB, CD62L-BV786, CD161-BV510, CD25-APC-Cy7, CD127-PE-Cy7, and fixable viability stain 620 to exclude dead cells. After fixation with 2% paraformaldehyde, cells were acquired using an LSR Fortessa Cell Analyzer (BD Biosciences), and subsequent analysis was performed using FlowJo version 10.3 (Supplementary Fig. S3). All the antibodies were purchased from either BioLegend or BD Pharmingen. Previously frozen aliquots of PBMCs from healthy donors were included in each MFC analysis as controls for antibody staining, instrumental acquisition, and analysis. Clinical samples containing more than 500 live CD3+ T cells were included in the final analysis, and an average of 18,349 T cells (minimum 544, maximum 184,362) per sample were subjected to further phenotypic analysis.

TCRβ sequencing and analysis

Genomic DNA was isolated from previously snap-frozen WBC pellets using the QIAmp DNA Blood Mini Kit (Q), and used for amplification of the CDR3 regions of human TCRβ chains and subsequent library preparation using an immunoSEQ hsTCRB Kit (Adaptive Biotechnologies) according to the manufacturer's instructions. The library was sequenced using the NextSeq 500/550 Mid Output V2 (150 cycle) kit (Illumina), and processed by Adaptive Biotechnologies. Subsequent downstream analysis was performed using an Immunoseq Analyzer (Adaptive Biotechnologies).

Mutation analysis of 295 AML/MDS–associated genes

NGS-based targeted sequencing of 295 AML/MDS–associated genes using the SureSelect custom panel (Agilent Technologies) was performed at an average depth of 443 on 50 ng of genomic DNAs extracted from WBC pellets, and high-confidence mutations were selected according to previously reported bioinformatic methodology (15). Subsequent mutation analysis was performed as previously described (16, 17). Briefly, a total number of detectable mutations and sum of variable allelic frequency (VAF) of all detectable mutations per subject or VAFs of individual mutation was compared between two time points–pretreatment and at the time of best response within responders or nonresponders from each treatment groups. Distribution of mutated genes were tabulated according to responders versus nonresponders from each treatment cohort, and changes in VAFs of mutations in gene were represented as “increase (orange)”, “decrease (sky blue)”, “clearance (blue)”, or “new appearance (red)”, “mixed responses (green)”.

NanoString analysis

Total RNAs were isolated from previously cryopreserved PBMCs using the RNeasy Plus Micro Kit (Qiagen), and further amplified using a NanoString Low Input Amplification Kit (NanoString). Up to 100 ng amplified RNAs were hybridized with target-specific, fluorescence-labeled 594 CodeSet and reporter probes from the nCounter Immunology panel (Human v2 – NanoString) for 20 hours, and added to the robotic nCounter Prep Station for automated sample processing for the remaining steps. The cartridge was transferred to an nCounter Digital Analyzer MAX for scanning and data collection. The nSolver software version 4 (NanoString) was used for preprocessing of raw counts, subsequent normalization, and downstream analysis including cell type profiling analysis, gene expression profiling (GEP), and gene set enrichment analysis (GSEA).

Statistical analysis

Nonparametric Wilcoxon matched paired signed rank or Mann–Whitney tests were used to assess statistical differences between two paired values within response groups or unpaired values across response groups, respectively. The Kruskal–Wallis test was used to compare differences among the three values within the group. All statistical analyses were performed using GraphPad Prism 8.0. P value less than 0.05 was considered as statistically significant. Only the statistically significant or notable P values are shown.

Data availability statement

NanoString data were deposited to the Gene Expression Omnibus (GEO), a public functional genomics data repository at GSE219085.

Patient characteristics

A total of 55 patients with MDS who had paired samples with adequate quality and quantity peripheral blood lymphocytes, collected at before and after the initial cycle of treatment were included in this study. Patients’ characteristics are described in Supplementary Table S1. Thirty-seven patients who received IMT ± HMA consisted of 21 clinical responders [complete response (CR) or partial response (PR)] and 16 nonresponders [stable disease (SD) or progression of disease (PD)]. Eighteen patients who received HMA alone included 12 clinical responders and 6 nonresponders. Detailed information on the treatment regimen and available testing results for the analysis are summarized in Supplementary Figs. S1 and S2.

In vivo expansion of central memory CD8+ T cells and Tregs are associated with clinical responses to IMT

The therapeutic efficacy of immune checkpoint blockade may depend on the presence of potential neoantigen-specific, naïve, or memory T cells in the periphery, and their ability to be activated and expanded during treatment. Therefore, we first evaluated the composition of various T-cell subsets at baseline and the subsequent changes after the initial treatment and at the time of best responses.

First, there was no significant difference in the major T-cell subsets such as CD4+, CD8+, or CD4CD8 T cells between responders versus nonresponders before and average 30.9 days (range: 11–61 days) after the first cycle of treatment (Fig. 1) or at the time of the best responses (Supplementary Fig. S4). There were increased frequencies of DR+CD4+ or DR+CD8+ T cells after the first cycle of treatment in both responders and nonresponders after IMT ± HMA, but not after HMA alone, suggesting that the degree of activated T cells after treatment does not predict clinical responses but rather implicate the effects of immune checkpoint blockade (Fig. 1A). Interestingly, significant in vivo expansion of Tregs and CM CD8+ T cells, and a trend toward increased CM CD4+ T cells was observed in clinical responders (P = 0.0290, 0.0319, and 0.0822, respectively) after the first cycle of IMT ± HMA, but not in nonresponders, or after HMA alone. This change in Tregs and CM CD8+ T cells was best observed after the initial treatment, rather than at the time of best responses (Supplementary Fig. S4). While increased Tregs are thought to have a negative impact on prognosis of MDS (18–20), a significant increase in circulating Treg post treatment in clinical responder may reflect a compensatory response during immunotherapy. Our results indicate that in vivo expansion of CM CD8+ T cells or Treg after the initial immunotherapy can predict the clinical response in MDS during immunotherapy.

Figure 1.

CM CD8+ T cells and Treg expansion are associated with clinical responses in IMT. Multiparameter flow cytometry was used to assess the composition of various immune subsets such as activated T cells (HLA-DR+), Treg (CD4+CD25+CD127low), and naïve (CD45RA+CD62L+), CM (CD45RACD62L+), effector memory (CD45RACD62L), and effector (CD45RA+CD62L) T cells from PB of patients with MDS before and after the initial treatments – IMT ± HMA or HMA only. Clinical responders (R) or nonresponders (NR) were patients who achieved either complete response (CR)/partial response (PR), or stable disease (SD)/partial disease (PD), respectively. In vivo expansion of Treg (A) and central memory CD8+ T cells (B and C) after the initial treatment were associated with clinical responses. A symbol represents a value from one subject and mean ± SD was represented. Wilcoxon match paired signed rank or unpaired Mann–Whitney tests were used to assess statistical differences within or across response groups, respectively. P value less than 0.5 was considered as “significance”, and only significant P value or notable P values were presented.

Figure 1.

CM CD8+ T cells and Treg expansion are associated with clinical responses in IMT. Multiparameter flow cytometry was used to assess the composition of various immune subsets such as activated T cells (HLA-DR+), Treg (CD4+CD25+CD127low), and naïve (CD45RA+CD62L+), CM (CD45RACD62L+), effector memory (CD45RACD62L), and effector (CD45RA+CD62L) T cells from PB of patients with MDS before and after the initial treatments – IMT ± HMA or HMA only. Clinical responders (R) or nonresponders (NR) were patients who achieved either complete response (CR)/partial response (PR), or stable disease (SD)/partial disease (PD), respectively. In vivo expansion of Treg (A) and central memory CD8+ T cells (B and C) after the initial treatment were associated with clinical responses. A symbol represents a value from one subject and mean ± SD was represented. Wilcoxon match paired signed rank or unpaired Mann–Whitney tests were used to assess statistical differences within or across response groups, respectively. P value less than 0.5 was considered as “significance”, and only significant P value or notable P values were presented.

Close modal

Clonal expansion and emergence of novel T-cell clones is associated with clinical responses

While the CTLA-4:B7 interaction works during the initial activation of naïve T cells, PD-1:PD-L1 interaction modulates the homeostasis of activated T cells (21). Therefore, there might be a qualitative difference in the TCR repertoire between clinical responders and nonresponders during immunotherapy. Here, we performed TCRβ repertoire analysis of peripheral blood T cells at before and after the initial treatment to evaluate whether clinical responses correlate with unique changes in TCRβ repertoire during immunotherapy.

First, we compared the clonality and richness of T cells between responders and nonresponders pre- and post-treatment, as well as the changes from pre- to post-treatment with IMT ± HMA or HMA only (Fig. 2A). The responders had a significantly lower clonality than that of nonresponders (P = 0.0415) prior to the treatment, followed by a trend toward increased clonality after treatment with IMT ± HMA (P = 0.0537), indicating that clinical responses during immunotherapy were associated with oligoclonal expansion of rare T cells selected from diverse T-cell clonotypes. There was no significant difference in TCRβ richness between responders and nonresponders or after treatment with HMA alone.

Figure 2.

Clonal expansion and emergence of novel T-cell clones is associated with clinical responses. TCRβ repertoires were assessed from PB T cells prior and post first cycle of treatment in patients with MDS. A, TCRβ clonality and richness at pre- and post-treatment were shown in patients with MDS who received IMT ± HMA (left) or HMA only (right). The presence of decreased TCR clonality at pretreatment and subsequent increase was observed in clinical responders who received IMT ± HMA. Clinical responders (R) or nonresponders (NR) were patients who achieved either complete response (CR)/partial response (PR), or stable disease (SD)/partial disease (PD), respectively. A symbol represents a value from one individual, mean ± SD was presented. Wilcoxon match paired signed rank or unpaired Mann–Whitney tests were used to assess statistical differences within or across response groups, respectively. P value less than 0.5 was considered as “significance”, and only significant P value or notable P values were presented. B, Individual clonotype frequencies at pre- and post-treatment from responders and nonresponders who received IMT ± HMA (top) were shown. In responders, we identified 1,462 significantly changed clonotypes that were either novel (800/1,462; 54.7%), expanded (42/1,462; 2.9%) or contracted (620/1,462; 42.4%). In nonresponders, there were 1,577 significantly changed clonotypes consisting of novel (509/1,577; 32.3%), expanded (228/1,577; 14.5%) or contracted (840/1,577; 53.3%) clonotypes (top left). HMA only (bottom) showed no specific pattern for changes that occurred in the response group. Significantly changed clonotypes –contracted, expanded, and novel clonotypes after the treatment were marked in green, red, or blue, respectively. Average number of significantly changed clonotypes per patients from responders and nonresponders were presented. The emergence of novel T-cell clones after IMT ± HMA was associated with clinical responses. Kruskal–Wallis test was used to compare differences among contracted, expanded, or novel clonotypes within the group, and P value less than 0.05 was considered “significant”. Only significant P value was presented.

Figure 2.

Clonal expansion and emergence of novel T-cell clones is associated with clinical responses. TCRβ repertoires were assessed from PB T cells prior and post first cycle of treatment in patients with MDS. A, TCRβ clonality and richness at pre- and post-treatment were shown in patients with MDS who received IMT ± HMA (left) or HMA only (right). The presence of decreased TCR clonality at pretreatment and subsequent increase was observed in clinical responders who received IMT ± HMA. Clinical responders (R) or nonresponders (NR) were patients who achieved either complete response (CR)/partial response (PR), or stable disease (SD)/partial disease (PD), respectively. A symbol represents a value from one individual, mean ± SD was presented. Wilcoxon match paired signed rank or unpaired Mann–Whitney tests were used to assess statistical differences within or across response groups, respectively. P value less than 0.5 was considered as “significance”, and only significant P value or notable P values were presented. B, Individual clonotype frequencies at pre- and post-treatment from responders and nonresponders who received IMT ± HMA (top) were shown. In responders, we identified 1,462 significantly changed clonotypes that were either novel (800/1,462; 54.7%), expanded (42/1,462; 2.9%) or contracted (620/1,462; 42.4%). In nonresponders, there were 1,577 significantly changed clonotypes consisting of novel (509/1,577; 32.3%), expanded (228/1,577; 14.5%) or contracted (840/1,577; 53.3%) clonotypes (top left). HMA only (bottom) showed no specific pattern for changes that occurred in the response group. Significantly changed clonotypes –contracted, expanded, and novel clonotypes after the treatment were marked in green, red, or blue, respectively. Average number of significantly changed clonotypes per patients from responders and nonresponders were presented. The emergence of novel T-cell clones after IMT ± HMA was associated with clinical responses. Kruskal–Wallis test was used to compare differences among contracted, expanded, or novel clonotypes within the group, and P value less than 0.05 was considered “significant”. Only significant P value was presented.

Close modal

Next, we assessed qualitative changes in the TCRβ repertoire during treatment by comparing the number of contracted, expanded, and novel clonotypes that occurred posttreatment in responders and nonresponders after IMT ± HMA or HMA only (Fig. 2B). There was a significantly higher abundance of novel clonotypes compared with expanded or contracted clonotypes in responders (P = 0.01), whereas contracted clonotypes were more prevalent in nonresponders (P = 0.0458). These findings suggest that the emergence of new clonotypes was associated with responders, while nonresponders had a contracted clonotype repertoire following IMT ± HMA. In contrast, there was no specific pattern for changes that occurred in the response group following HMA only. Our results indicate that clonal expansion of T cells with novel clonotypes selected from diverse clonotypic pools is associated with clinical responses during immunotherapy in MDS.

Evolution of mutation landscape predicts clinical responses during immunotherapy

Considering the therapeutic efficacy of immunotherapy using checkpoint inhibitors relies on the presence of targetable tumor antigens (22), we performed targeted gene sequencing of 295 recurrently mutated genes in AML/MDS from pre- and matching post-treatment samples after the first cycle, at the best responses and relapse if applicable, to evaluate whether the presence or change of nonsynonymous mutations correlated with the clinical response (Fig. 3).

Figure 3.

Higher individual mutation burden prior to the treatment is associated with clinical responses in patients with MDS during checkpoint blockade ± HMA. NGS-based targeted sequencing of 295 recurring mutated genes in AML/MDS were performed on pre- and matching post-treatment peripheral blood of patients with MDS who received IMT ± HMA or HMA alone, and nonsynonymous somatic mutations were identified. Clinical responders (R) or nonresponders (NR) were patients who achieved either complete response (CR)/partial response (PR), or stable disease (SD)/partial disease (PD), respectively. A, The number of somatic mutations per patient (A, left), the sum of variant allelic frequencies (VAF) of all detectable mutations per patient (A, middle), and VAF of individual mutation of all patients within group – responders versus nonresponders (A, right) were shown in left, middle, and right panel, respectively. A symbol represents a value from one individual (A, left and middle) or one mutation (A, right), and mean ± SD was represented. Wilcoxon match paired signed rank or unpaired Mann–Whitney tests were used to assess statistical differences between two values within or across response groups, respectively. P value less than 0.05 was considered “significant”, and only significant or notable P values were presented. B, Changes of VAFs of mutations in individual genes from pretreatment to best-response time-point were shown in patients with MDS: new appearance (red), increased (orange), decreased (sky blue), clearance (blue), mixed response (green, some of subclones decreased, the others increased). C, Frequencies of mutation in genes detected prior to the treatment were shown in clinical responders (blue) and nonresponders (orange) who received IMT ± HMA (top) and HMA alone groups (bottom).

Figure 3.

Higher individual mutation burden prior to the treatment is associated with clinical responses in patients with MDS during checkpoint blockade ± HMA. NGS-based targeted sequencing of 295 recurring mutated genes in AML/MDS were performed on pre- and matching post-treatment peripheral blood of patients with MDS who received IMT ± HMA or HMA alone, and nonsynonymous somatic mutations were identified. Clinical responders (R) or nonresponders (NR) were patients who achieved either complete response (CR)/partial response (PR), or stable disease (SD)/partial disease (PD), respectively. A, The number of somatic mutations per patient (A, left), the sum of variant allelic frequencies (VAF) of all detectable mutations per patient (A, middle), and VAF of individual mutation of all patients within group – responders versus nonresponders (A, right) were shown in left, middle, and right panel, respectively. A symbol represents a value from one individual (A, left and middle) or one mutation (A, right), and mean ± SD was represented. Wilcoxon match paired signed rank or unpaired Mann–Whitney tests were used to assess statistical differences between two values within or across response groups, respectively. P value less than 0.05 was considered “significant”, and only significant or notable P values were presented. B, Changes of VAFs of mutations in individual genes from pretreatment to best-response time-point were shown in patients with MDS: new appearance (red), increased (orange), decreased (sky blue), clearance (blue), mixed response (green, some of subclones decreased, the others increased). C, Frequencies of mutation in genes detected prior to the treatment were shown in clinical responders (blue) and nonresponders (orange) who received IMT ± HMA (top) and HMA alone groups (bottom).

Close modal

Although there was not a significant difference in the total number of somatic mutations and the sum of variant allelic frequencies (VAF) of all detectable mutations at pretreatment between responders and nonresponders, higher VAFs of individual mutations prior to the treatment and subsequent reduction were observed in responders among patients with IMT ± HMA than nonresponders (P = 0.0026; Fig. 3A). Moreover, 6 of 16 nonresponders (37.5%) showed new mutations while only 1 of 19 responders (5.3%) had the emergence of a new mutation. As expected, more clinical responders (13/19, 68.4%) contained mutations whose VAFs decreased or disappeared after treatment than nonresponders (8/16, 50%; Fig. 3B). Almost all patients who received HMA only had reductions in individual mutation burdens regardless of their responses, suggestive of cytoreductive action by HMA (Fig. 3A and B). However, the emergence of new mutations or worsened VAF in nonresponders appeared in 8 of 8 nonresponders in IMT only but also in 6 of 8 nonresponders in IMT+HMA group. Therefore, we cannot make a definitive conclusion that mutation clearance/reduction seen in clinical responders are solely attributed to the addition of HMA, but rather synergistic action between immunotherapy and HMA treatment. Finally, higher frequencies of TP53 mutations were present prior to treatment in responders (6/of 19, 31.6%) than in nonresponders (2/of 16, 12.5%) with IMT ± HMA, suggesting that TP53 mutations may be associated with clinical responders (Fig. 3C).

Thus, the longitudinal mutation analysis of tumor–immune microenvironment demonstrated that a higher mutation burden of individual mutations and subsequent reduction/clearance during immunotherapy and the presence of p53 mutations were associated with clinical responses in patients who received IMT ± HMA but not HMA alone.

Distinct immune-related gene expression signature is associated with clinical response during IMT

The presence of various mutations in epigenetic regulators may contribute to the immune dysregulation in MDS, and influence the therapeutic efficacy of immunotherapy (6). Here, we investigated the immune-related gene expression signature in the tumor–immune microenvironment (TIME) associated with response or resistance during IMT, by performing targeted GEP and GSEA using a custom NanoString panel composed of 579 immune-related genes (Fig. 4).

Figure 4.

Distinct immune-related gene expression signature and immune cell types are associated with clinical response after IMT ± HMA. The gene expression profiling (GEP) with NanoString nCounter Immunology panel was performed on PBMCs collected from a patient with MDS pre- and post-treatment with IMT ± HMA (36 pairs from 21 responders and 15 nonresponders) and HMA alone (13 pairs from 9 responders and 4 nonresponders). Volcano plot was used to demonstrate the fold change in gene expression between two groups [log2 (fold change)] against its statistical significance [log10(unadjusted P value)]. Gene with the FDR-adjusted P value of fold change less than 0.5 were in purple, and top 20 differentially expressed genes between two groups were labeled. Following thresholds represented FDR-adjusted P value less than 0.50 (), 0.10 (), 0.05 (), and 0.01 (). Gene Set Enrichment Analysis was performed to investigate significantly upregulated or downregulated immune pathways between groups. The changes in gene expression and immune-pathways after IMT ± HMA compared to pretreatment (reference) were shown for responders in A, and nonresponders in B. Next, significantly regulated immune-pathways and immune cell types using NanoString gene sets according to the clinical response after IMT ± HMA were evaluated. The normalized expression of significantly regulated genes from top upregulated or downregulated immune-pathways after HMA ± HMA in responders or nonresponders, respectively, are shown in C. Top upregulated genes in responders after treatment included CLTA-4, CD3D, CD8B, IFNγ, and TBX21 from TCR signaling pathways, and FOXP3 from Treg differentiation, while the majority of these genes except CTLA-4 were either downregulated or showed no significant changes in nonresponders. While responders did not have major changes in gene expression in TGFβ singling pathways except IFNγ, nonresponders downregulated almost all genes involved in TGFβ singling pathways. Major autophagy-related genes such as MAPK1, ATG5, ATG7, and ATG12 were significantly downregulated posttreatment in nonresponders while responders did not have significant changes in these genes except BCL2. Relative abundances of immune cell subpopulations estimated by CellType Analysis were shown in D. A symbol represents a value from one individual and mean ± SD was represented. Wilcoxon match paired signed rank or unpaired Mann–Whitney tests were used to assess statistical differences between two values within or across response groups, respectively. P value less than 0.05 was considered “significant”, and only significant or notable P values were presented.

Figure 4.

Distinct immune-related gene expression signature and immune cell types are associated with clinical response after IMT ± HMA. The gene expression profiling (GEP) with NanoString nCounter Immunology panel was performed on PBMCs collected from a patient with MDS pre- and post-treatment with IMT ± HMA (36 pairs from 21 responders and 15 nonresponders) and HMA alone (13 pairs from 9 responders and 4 nonresponders). Volcano plot was used to demonstrate the fold change in gene expression between two groups [log2 (fold change)] against its statistical significance [log10(unadjusted P value)]. Gene with the FDR-adjusted P value of fold change less than 0.5 were in purple, and top 20 differentially expressed genes between two groups were labeled. Following thresholds represented FDR-adjusted P value less than 0.50 (), 0.10 (), 0.05 (), and 0.01 (). Gene Set Enrichment Analysis was performed to investigate significantly upregulated or downregulated immune pathways between groups. The changes in gene expression and immune-pathways after IMT ± HMA compared to pretreatment (reference) were shown for responders in A, and nonresponders in B. Next, significantly regulated immune-pathways and immune cell types using NanoString gene sets according to the clinical response after IMT ± HMA were evaluated. The normalized expression of significantly regulated genes from top upregulated or downregulated immune-pathways after HMA ± HMA in responders or nonresponders, respectively, are shown in C. Top upregulated genes in responders after treatment included CLTA-4, CD3D, CD8B, IFNγ, and TBX21 from TCR signaling pathways, and FOXP3 from Treg differentiation, while the majority of these genes except CTLA-4 were either downregulated or showed no significant changes in nonresponders. While responders did not have major changes in gene expression in TGFβ singling pathways except IFNγ, nonresponders downregulated almost all genes involved in TGFβ singling pathways. Major autophagy-related genes such as MAPK1, ATG5, ATG7, and ATG12 were significantly downregulated posttreatment in nonresponders while responders did not have significant changes in these genes except BCL2. Relative abundances of immune cell subpopulations estimated by CellType Analysis were shown in D. A symbol represents a value from one individual and mean ± SD was represented. Wilcoxon match paired signed rank or unpaired Mann–Whitney tests were used to assess statistical differences between two values within or across response groups, respectively. P value less than 0.05 was considered “significant”, and only significant or notable P values were presented.

Close modal

First, we compared GEP and immune-related pathways between pre- and post-treatment in both responders and nonresponders to explore the dynamic changes in individual TIME after IMT ± HMA treatment. Again, the expression of genes related to processes such as cell adhesion, T-cell differentiation (Th1, Th2, Treg), TGFβ, and autophagy pathways was significantly upregulated after treatment in responders (Fig. 4A; Supplementary Table S2), while these pathways were reciprocally downregulated in nonresponders (Fig. 4B; Supplementary Table S3). Significantly regulated genes were shown in Fig. 4C and Supplementary Tables S2 and S3. In summary, our immune-related gene expression profiling analysis confirmed the activation and differentiation of T cells in responders, and suggests that downregulation of Th1 differentiation, TGFβ, and autophagy pathways may be novel mechanisms of immune resistance that can be targeted to enhance the therapeutic efficacy of immune checkpoint blockade.

Finally, cell type profiling analysis showed that T cells, Th1 cells, CD8 cells, Treg, exhausted CD8 T cells, and cytotoxic cells were increased after treatment in responders, while assessed cell types except exhausted CD8 T cells and NK cells did not show significant changes posttreatment in nonresponders (Fig. 4D). Interestingly, nonresponders had a higher prevalence of T, Th1, and CD8 T cells, exhausted T cells, cytotoxic cells, and NK cells compared with responders in prior treatment, but most of the key immune cells were either downregulated or without significant change after treatment, suggesting that immune cell dysfunction beyond CTLA-4:B7 or PD-1:PD-L1/2 pathways may be present at baseline in nonresponders. In conclusion, T cell–mediated immunity in responders may be suppressed compared with responders prior to the treatment, and subsequently rejuvenated via immune checkpoint blockade.

αCTLA-4 blockade derives a favorable immune landscape in MDS during immunotherapy

As the CTLA-4/B7 or PD-1/PD-L1/2 pathways are involved in the activation of naïve or memory T cells, respectively, the blockade of these immune checkpoint interaction will likely elicit antitumor immunity via distinct mechanisms of action. Our study included cohorts of patients who received a series of immune checkpoint inhibitors, αCTLA-4 versus αPD-1 versus αCTLA-4 plus αPD-1 ± HMA, which may serve as confounding factor to determine predictive biomarkers for responses versus immune resistance. However, this heterogeneity of treatment in turn provides insights into which immune checkpoint blockade is attributed to favorable changes in the immune landscape. Moreover, we added HMA only group to interpret biologic changes for IMT only and in combination with HMA. First, immune subset analysis revealed that patients who received αCTLA-4 containing treatments showed increased frequencies of DR+CD4+ T cells, Treg, and central memory (CM) T cells (Fig. 5A). These findings suggest that CTLA-4 but not PD-1 blockade likely drove the activation of CD4+ T cells, expansion of Treg, and CM T cells observed in clinical responders.

Figure 5.

αCTLA-4 blockade drives the activation of CD4+ T cells, expansion of Treg and central memory T cells, and perturbation of TCRβ repertoire. A, PB T-cell subsets from MDS patients were assessed at pre- and post-treatments consisting of αCTLA-4 ± HMA, αPD-1 ± HMA, αCTLA-4/αPD-1 ± HMA, or HMA. Immune subset analysis revealed that the frequency of HLA-DR+CD4+ (P = 0.0107) and Treg (P = 0.0085 for % T cells) significantly increased in patients who received αCTLA-4 therapy ± HMA. Similarly, patients treated with αCTLA-4 plus αPD-1 ± HMA showed an increased frequency of HLA-DR+CD4+ T cells (P = 0.0040) and Treg (P = 0.0134 for % T cells). In addition, fractions of CM CD4+ and CD8+ T cells increased significantly in patients who received αCTLA-4 ± HMA (P = 0.0050 and P = 0.0254, respectively). A similar pattern was observed in patients treated with αCTLA-4 plus αPD-1± HMA with a significant increase in fraction of CM CD8+ T cells. A symbol represents a value from one subject. Mean ± SD was represented. Wilcoxon match paired signed rank were used to assess statistical differences within the treatment groups. P value less than 0.5 was considered as “significance”, and only significant P value or notable P values were presented. B, Changes in TCRβ clonality in patients after following treatment: αCTLA-4 ± HMA, αPD-1 ± HMA, αCTLA-4/αPD-1 ± HMA, or HMA. A symbol represents a value from one subject. Matching pre and post values were lined. Responders were presented in red symbol and line, and nonresponders were presented in blue symbol and line. Wilcoxon match paired signed rank were used to assess statistical differences within the treatment groups. P value less than 0.5 was considered as “significance”, and only significant P values were presented. C, Individual clonotype frequencies at pre- and post-treatment (αCTLA-4 ± HMA, αPD-1 ± HMA, αPD-1 ± HMA excluding UPN11, αCTLA-4/αPD-1 ± HMA, or HMA) were plotted. Significantly changed clonotypes –contracted, expanded, and novel clonotypes after the treatment were marked in green, red, or blue, respectively. D, Average number of significantly changed clonotypes per patients following treatment (αCTLA-4 ± HMA, αPD-1 ± HMA, αPD-1 ± HMA excluding UPN11, αCTLA-4/αPD-1 ± HMA, or HMA) were shown. Kruskal–Wallis test was used to compare differences among contracted, expanded, or novel clonotypes within the group, and P value less than 0.05 was considered “significant”. Only significant P value was presented. Although there was no significant change in overall TCRβ clonality following all treatments, αCTLA-4 containing treatment elicited a larger number of clonotypes that were significantly changed (novel, contracted) after the treatment.

Figure 5.

αCTLA-4 blockade drives the activation of CD4+ T cells, expansion of Treg and central memory T cells, and perturbation of TCRβ repertoire. A, PB T-cell subsets from MDS patients were assessed at pre- and post-treatments consisting of αCTLA-4 ± HMA, αPD-1 ± HMA, αCTLA-4/αPD-1 ± HMA, or HMA. Immune subset analysis revealed that the frequency of HLA-DR+CD4+ (P = 0.0107) and Treg (P = 0.0085 for % T cells) significantly increased in patients who received αCTLA-4 therapy ± HMA. Similarly, patients treated with αCTLA-4 plus αPD-1 ± HMA showed an increased frequency of HLA-DR+CD4+ T cells (P = 0.0040) and Treg (P = 0.0134 for % T cells). In addition, fractions of CM CD4+ and CD8+ T cells increased significantly in patients who received αCTLA-4 ± HMA (P = 0.0050 and P = 0.0254, respectively). A similar pattern was observed in patients treated with αCTLA-4 plus αPD-1± HMA with a significant increase in fraction of CM CD8+ T cells. A symbol represents a value from one subject. Mean ± SD was represented. Wilcoxon match paired signed rank were used to assess statistical differences within the treatment groups. P value less than 0.5 was considered as “significance”, and only significant P value or notable P values were presented. B, Changes in TCRβ clonality in patients after following treatment: αCTLA-4 ± HMA, αPD-1 ± HMA, αCTLA-4/αPD-1 ± HMA, or HMA. A symbol represents a value from one subject. Matching pre and post values were lined. Responders were presented in red symbol and line, and nonresponders were presented in blue symbol and line. Wilcoxon match paired signed rank were used to assess statistical differences within the treatment groups. P value less than 0.5 was considered as “significance”, and only significant P values were presented. C, Individual clonotype frequencies at pre- and post-treatment (αCTLA-4 ± HMA, αPD-1 ± HMA, αPD-1 ± HMA excluding UPN11, αCTLA-4/αPD-1 ± HMA, or HMA) were plotted. Significantly changed clonotypes –contracted, expanded, and novel clonotypes after the treatment were marked in green, red, or blue, respectively. D, Average number of significantly changed clonotypes per patients following treatment (αCTLA-4 ± HMA, αPD-1 ± HMA, αPD-1 ± HMA excluding UPN11, αCTLA-4/αPD-1 ± HMA, or HMA) were shown. Kruskal–Wallis test was used to compare differences among contracted, expanded, or novel clonotypes within the group, and P value less than 0.05 was considered “significant”. Only significant P value was presented. Although there was no significant change in overall TCRβ clonality following all treatments, αCTLA-4 containing treatment elicited a larger number of clonotypes that were significantly changed (novel, contracted) after the treatment.

Close modal

Although there were no significant changes in TCRβ clonality between pre- and post-treatment within each treatment cohort, αCTLA-4 and αCTLA-4/PD-1 ± HMA treatment resulted in a larger number of significantly changed clonotypes (1547, 1171 respectively) than treatment with αPD1 ± HMA (321) or HMA alone (86; Fig. 5B and C). Within the treatment group, αCTLA-4 ± HMA had a significantly higher abundance of novel (728/1,547; 47.1%), and contracted clonotypes (758/1,547;49.0%) than expanded clonotypes (61/1,547; 3.9%; P < 0.0001), while there were no significant differences in αPD-1 and αCTLA-4 plus αPD-1 groups (Fig. 5D). Considering that one responder (UPN11) in αPD1 ± HMA attributed 305 of 321 significantly changed clonotypes, αCTLA-4 blockade likely contributed to the alteration of TCRβ repertoire in responders.

Finally, we investigated differential immune-related gene expression signatures and pathways associated with each treatment to determine which immunotherapy elicited changes in the immune landscape favorable for clinical responses (Fig. 6; Supplementary Tables S4–S7). Both αCTLA-4 ± HMA and αCTLA-4 plus αPD1 ± HMA upregulated genes related to pathways of TCR signaling, T-cell differentiation, and lymphocytes activation seen in responders while αPD-1 ± HMA failed to elicit such changes. There were synergistic effects of αPD-1 and αCTLA-4 treatment as αPD-1 plus αCTLA4 ± HMA showed more upregulated genes and pathways compared to αCTLA-4 ± HMA. Interestingly, HMA alone drastically upregulated multiple immune–inflammation signaling pathways and genes supporting potential immune dysregulation in patients with MDS (Fig. 6D; Supplementary Table S7). Interestingly, TGFβ signaling and autophagy pathways were downregulated in all treatment groups but were significantly upregulated in responders (Fig. 4). Thus, GEP analysis confirmed that αCTLA-4 blockade may contribute to shaping a favorable immune landscape to eliminate leukemia in patients with MDS.

Figure 6.

αCTLA-4 blockade shapes a favorable immune-landscape in MDS during immunotherapy. A–D, Differential gene expression between pre- and post-treatment [αCTLA-4 ± HMA (14 pairs), αPD-1 ± HMA (8 pairs), αCTLA-4/αPD-1 ± HMA (14 pairs), or HMA (13 pairs)] were assessed. A volcano plot was used to demonstrate the fold change in gene expression after treatment [log2(fold change)] against its statistical significance [log10(unadjusted P value)]. Genes with a FDR-adjusted P value of fold change less than 0.5 were in purple, and the top 40 differentially expressed genes between two groups were labeled. The following thresholds represented FDR-adjusted P value less than 0.50 (), 0.10 (), 0.05 (), and 0.01 (). GSEA was performed to investigate significantly upregulated or downregulated immune pathways between pre- and post-treatment.

Figure 6.

αCTLA-4 blockade shapes a favorable immune-landscape in MDS during immunotherapy. A–D, Differential gene expression between pre- and post-treatment [αCTLA-4 ± HMA (14 pairs), αPD-1 ± HMA (8 pairs), αCTLA-4/αPD-1 ± HMA (14 pairs), or HMA (13 pairs)] were assessed. A volcano plot was used to demonstrate the fold change in gene expression after treatment [log2(fold change)] against its statistical significance [log10(unadjusted P value)]. Genes with a FDR-adjusted P value of fold change less than 0.5 were in purple, and the top 40 differentially expressed genes between two groups were labeled. The following thresholds represented FDR-adjusted P value less than 0.50 (), 0.10 (), 0.05 (), and 0.01 (). GSEA was performed to investigate significantly upregulated or downregulated immune pathways between pre- and post-treatment.

Close modal

As immune checkpoint inhibitors are becoming an emerging therapy for hematologic malignancies including AML and MDS (11, 23–28), the investigation of the tumor–immune landscape in TIME in patients with MDS during immunotherapy can provide clinical response determinants and mechanistic insight into immune resistance in MDS during checkpoint blockade and help identify potential targets to overcome therapeutic resistance or synergize to immune checkpoint blockade. Daver and colleagues published the results of a phase II study of nivolumab in combination with 5-Azacitidine in relapsed/refractory AML (29), and showed that clinical responders had higher effector CD4+ T and CD8+ T cells in the bone marrow than nonresponders prior to treatment, and that the increased frequencies of CTLA-4+ CD4+ or CD8+ T cells after treatment were associated with nonresponders. Furthermore, paired single-cell RNA analysis and TCR repertoire profiling showed that the emergence of novel CD8+ T-cell clonotypes and CD8+ T-cell phenotypes with stem-like properties expressing granzyme K were associated with responders (30). We observed significant increases in activated T cells, CM T cells, and Treg within the T-cell compartment after IMT ± HMA, and did not observe significant differences in major T-cell subsets between responders and nonresponders either in pre- or post-treatment (Fig. 1). These dynamic changes were observed only in patients who received αCTLA-4 containing treatment (αCTLA-4 ± αPD-1 ± HMA) but not αPD-1 ± HMA or HMA alone, suggesting that αCLTA-4 blockade may play a critical role in eliciting favorable T-cell responses. These findings reflected on clinical trial results recently reported by Zeidan and colleagues where a phase II study of 5-Azacitidine with or without durvalumab for higher-risk MDS did not show significant improvement in clinical outcomes by combination therapy (31). Moreover, we found higher levels of T, Th1, CD8 T cells, cytotoxic T and NK cells in nonresponders prior to treatment, which failed to expand after treatment (Fig. 4D). Finally, we observed an increased expression of soluble and transmembrane CTLA-4 transcripts after treatment, which was significantly higher in responders than in nonresponders (Fig. 4C). The differences in findings between the work by Daver and colleagues and our study are likely due to the different types of immune checkpoint blockade patients received – αPD-1 versus CTLA-4 ± αPD-1, location of TIME analyzed – peripheral blood versus bone marrow, and nature of diseases – robust nature of AML compared with MDS. As we investigated the coevolution of the immune and genetic landscapes in PBMCs as a surrogate TIME to the bone marrow of patients with MDS during immunotherapy, further investigation using bone marrow samples warrant.

T cells are key players in the antitumor response mediated by immunotherapies. Upon recognizing cognate antigens, activated T cells can differentiate and proliferate, a process called “clonal expansion” (32). Therefore, determining T-cell clonality even in peripheral blood can provide information on the degree of tumor antigen–driven T-cell expansion and mechanistic insights into the immune checkpoint blockade (28, 33, 34). In our study, clinical responders had a diverse TCRβ repertoire prior to treatment compared with nonresponders, followed by an increase in the TCRβ clonality with the appearance of novel T-cell clones after treatment (Fig. 3). Interestingly, CTLA-4 rather than PD-1 blockade appears to have driven a favorable perturbation in the TCRβ repertoire, which coincides with the expansion of CM T cells, Treg, and activated T cells observed in patients with CTLA-4 blockade (Fig. 5). Our findings support the association of diverse TCRβ repertoire pretreatment and the appearance of novel TCRβ clones post treatment with a clinical response after immune checkpoint blockade (30, 34). Although HMA treatment may have contributed to the expansion of rare TCR clones in clinical responders (Fig. 2; ref. 35), a larger degree of oligoclonal expansion of novel T-cell clones was observed in responders after IMT ± HMA, especially αCTLA-4 blockade ± HMA, compared with HMA alone (Fig. 2). The reason that we did not observe emergence of novel TCR clones in clinical responders from HMA alone, differently from previous report (35) may be due to the potential enrichment of tumor specific T cells in bone marrow (35) versus peripheral blood (our study). As clinical responders had a higher tumor mutation burden with subsequent reduction after IMT ± HMA, our findings support the potential antitumor activity of novel T-cell clones in MDS during immunotherapy.

The IFNγ receptor signaling pathway and “T cell inflamed” TIME has been shown to play an essential role in antitumor T-cell responses elicited by immune checkpoint blockade (36–41). More specifically, the increased frequencies of ICOS+ CD4+ T cells in peripheral blood or TIME after αCTLA-4 or αCTLA-4/αPD-1 therapy were correlated with clinical response in metastatic melanoma and urothelial carcinoma, respectively (42, 43). Our results are in line with previous findings in that type I and II IFN signaling was upregulated in pretreatment peripheral blood of clinical responders, but not supportive of “T-cell inflamed” TIME as several pathways involved in T-cell differentiation, TCR signaling, and T-cell trafficking were downregulated in TIME of clinical responders compared with nonresponders. Nonetheless, these pathways were upregulated in clinical responders after αCTLA-4–containing therapy supporting the role of T cell–mediated antitumor immunity in MDS during checkpoint blockade ± hypomethylation. Interestingly, we observed a significant upregulation in the MHC II antigen presentation pathway in the pretreatment TIME of clinical responders, which was likely driven by concurrent treatment with a hypomethylator. Antigen processing and presentation, one of the major pathways modulated by IFNγ, is a key step to initiate T cell–mediated immune responses and its defects were attributed to acquired immune resistance to immunotherapy (44, 45). Therefore, our findings highlight a critical role of the antigen presentation machinery in eliciting effective antitumor immunity in MDS during checkpoint blockade.

Tumors evade immune surveillance and checkpoint blockade through multiple mechanisms such as defective IFNγ signaling or antigen presenting pathway in tumor impairing the effective T-cell activation (40, 41, 44, 45), and aberrations in Wnt/b-catenin signaling or loss of PTEN or LKB1 in tumor avoiding “T-cell inflamed” TIME by the reduction of intratumoral trafficking of T cells (46–49). In our work, gene expression profiling of peripheral blood TIME revealed that autophagy, Th1 differentiation, and TGFβ singling pathways were the top three pathways significantly downregulated in nonresponders after checkpoint blockade ± hypomethylation. In addition, the majority, if not all, of the pathways related to T-cell differentiation, signaling, and activation were downregulated in nonresponders. Our findings echo the importance of T cell–mediated antitumor immunity and support the ICOS–ICOSL interaction of the TCR signaling pathway as a feasible target to synergize the efficacy of checkpoint blockade in MDS (Fig. 4C; ref. 50). The differential regulation of the TGFβ signaling pathway along with Th2 and Treg pathways seen in clinical responders (upregulated) and nonresponders (downregulated) may reflect compensatory immune suppression toward T cell–mediated inflammation during immunotherapy, rather than the mechanism of tumor–immune evasion given the suppressive role of TGFβ signaling during inflammation to maintain immune homeostasis (51). Finally, we observed an opposite regulation of autophagy pathways between clinical responders (upregulation) and nonresponders (downregulation) after checkpoint blockade ± hypomethylation. Autophagy is a cellular process that adapts to metabolic stress through the degradation and recycling of intracellular components to produce energy. While aberrant autophagy pathways are thought to participate in the pathogenesis and prognosis of MDS, autophagy regulation plays a key role in the homeostasis and function of various immune cells (52–54). In T cells, autophagy pathways are activated through TCR-mediated signaling or common gamma chain containing cytokine receptor signaling, and control essential programs in the activation, proliferation, differentiation, metabolism and function of T cells. As such, the downregulation of the autophagy pathway after checkpoint blockade suggests the presence of defective T cell–mediated immunity in nonresponders and supports the concurrent use of autophagy inducers such as mTOR inhibitors with checkpoint blockade to improve T-cell function during immunotherapy in MDS.

K. Takahashi reports personal fees from Symbio Pharmaceuticals and personal fees from Mission BIo outside the submitted work. N.G. Daver reports grants and other support from Daiichi-Sankyo, Bristol-Meyers Squibb, Pfizer, Gilead, Servier, Genentech, Astellas, AbbVie, ImmunoGen, Amgen, and Trillium; grants from Hanmi, Trovagene, FATE Therapeutics, Novimmune, and Glycomimetic; other support from Arog, Novartis, Jazz, Celgene, Syndax, Shattuck Labs; and other support from Agios outside the submitted work. No disclosures were reported by the other authors.

S. Lee: Conceptualization, investigation, writing–original draft. F. Wang: interpreted results. M. Grefe: Writing–review and editing. A. Trujillo-Ocampo: Investigation. W. Ruiz-Vasquez: Investigation. K. Takahashi: Interpreted results. H.A. Abbas: Interpreted results. P. Borges: interpreted results. D.A. Antunes: interpreted results. G. Al-Atrash: Writing–review and editing. N. Daver: Writing–review and editing. J.J. Molldrem: Writing–review and editing. A. Futreal: Writing–review and editing. G. Garcia-Manero: Resources, writing–review and editing. J.S. Im: Conceptualization, writing–original draft.

This work was funded by the Catholic Medical Center Research Foundation made in 2020 (S.L.), MD Anderson Cancer Center Leukemia SPORE 2P50CA100632 (J.I.), Khalifa Bin Zayed Al Nahyan Foundation (J.I.), Emerson Collective Cancer Research Foundation (J.I.), MDACC Institutional Start-up Fund (J.I.), Welch Foundation (A.F.), Lydia Hill foundation (A.F.), Anna Darko Foundation (A.F.), MDACC MDS/AML Moonshot (G.G.M), and NCI Cancer Center Support Grant P30CA16672 (Flow Cytometry and Cellular Imaging Facility).

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.

Note: Supplementary data for this article are available at Clinical Cancer Research Online (http://clincancerres.aacrjournals.org/).

1.
Cazzola
M
.
Myelodysplastic syndromes
.
N Engl J Med
2020
;
383
:
1358
74
.
2.
Fenaux
P
,
Mufti
GJ
,
Hellstrom-Lindberg
E
,
Santini
V
,
Finelli
C
,
Giagounidis
A
, et al
.
Efficacy of azacitidine compared with that of conventional care regimens in the treatment of higher-risk myelodysplastic syndromes: a randomised, open-label, phase III study
.
Lancet Oncol
2009
;
10
:
223
32
.
3.
Silverman
LR
,
Demakos
EP
,
Peterson
BL
,
Kornblith
AB
,
Holland
JC
,
Odchimar-Reissig
R
, et al
.
Randomized controlled trial of azacitidine in patients with the myelodysplastic syndrome: a study of the cancer and leukemia group B
.
J Clin Oncol
2002
;
20
:
2429
40
.
4.
Atallah
E
,
Kantarjian
H
,
Garcia-Manero
G
.
The role of decitabine in the treatment of myelodysplastic syndromes
.
Expert Opin Pharmacother
2007
;
8
:
65
73
.
5.
Jabbour
E
,
Garcia-Manero
G
,
Batty
N
,
Shan
J
,
O'Brien
S
,
Cortes
J
, et al
.
Outcome of patients with myelodysplastic syndrome after failure of decitabine therapy
.
Cancer
2010
;
116
:
3830
4
.
6.
Yang
H
,
Bueso-Ramos
C
,
DiNardo
C
,
Estecio
MR
,
Davanlou
M
,
Geng
QR
, et al
.
Expression of PD-L1, PD-L2, PD-1 and CTLA4 in myelodysplastic syndromes is enhanced by treatment with hypomethylating agents
.
Leukemia
2014
;
28
:
1280
8
.
7.
He
X
,
Xu
C
.
Immune checkpoint signaling and cancer immunotherapy
.
Cell Res
2020
;
30
:
660
9
.
8.
Sharma
P
,
Siddiqui
BA
,
Anandhan
S
,
Yadav
SS
,
Subudhi
SK
,
Gao
J
, et al
.
The next decade of immune checkpoint therapy
.
Cancer Discov
2021
;
11
:
838
57
.
9.
Leach
DR
,
Krummel
MF
,
Allison
JP
.
Enhancement of antitumor immunity by CTLA-4 blockade
.
Science
1996
;
271
:
1734
6
.
10.
Twomey
JD
,
Zhang
B
.
Cancer immunotherapy update: FDA-approved checkpoint inhibitors and companion diagnostics
.
AAPS J
2021
;
23
:
39
.
11.
Daver
N
,
Boddu
P
,
Garcia-Manero
G
,
Yadav
SS
,
Sharma
P
,
Allison
J
, et al
.
Hypomethylating agents in combination with immune checkpoint inhibitors in acute myeloid leukemia and myelodysplastic syndromes
.
Leukemia
2018
;
32
:
1094
105
.
12.
Hatic
H
,
Sampat
D
,
Goyal
G
.
Immune checkpoint inhibitors in lymphoma: challenges and opportunities
.
Ann Transl Med
2021
;
9
:
1037
.
13.
Lee
S-E
,
Wang
F
,
Trujillo-Ocampo
A
,
Ruiz-Vasquez
W
,
Cho
H-W
,
Takahashi
K
, et al
.
Fidelity of peripheral blood for monitoring genomics and tumor immune-microenvironment in myelodysplastic syndromes
.
eJHaem
2020
;
1
:
552
7
.
14.
Cheson
BD
,
Greenberg
PL
,
Bennett
JM
,
Lowenberg
B
,
Wijermans
PW
,
Nimer
SD
, et al
.
Clinical application and proposal for modification of the International Working Group (IWG) response criteria in myelodysplasia
.
Blood
2006
;
108
:
419
25
.
15.
Takahashi
K
,
Wang
F
,
Kantarjian
H
,
Doss
D
,
Khanna
K
,
Thompson
E
, et al
.
Preleukaemic clonal haemopoiesis and risk of therapy-related myeloid neoplasms: a case-control study
.
Lancet Oncol
2017
;
18
:
100
11
.
16.
Klco
JM
,
Miller
CA
,
Griffith
M
,
Petti
A
,
Spencer
DH
,
Ketkar-Kulkarni
S
, et al
.
Association between mutation clearance after induction therapy and outcomes in acute myeloid leukemia
.
JAMA
2015
;
314
:
811
22
.
17.
Duncavage
EJ
,
Jacoby
MA
,
Chang
GS
,
Miller
CA
,
Edwin
N
,
Shao
J
, et al
.
Mutation clearance after transplantation for myelodysplastic syndrome
.
N Engl J Med
2018
;
379
:
1028
41
.
18.
Kordasti
SY
,
Ingram
W
,
Hayden
J
,
Darling
D
,
Barber
L
,
Afzali
B
, et al
.
CD4+CD25high Foxp3+ regulatory T cells in myelodysplastic syndrome (MDS)
.
Blood
2007
;
110
:
847
50
.
19.
Giovazzino
A
,
Leone
S
,
Rubino
V
,
Palatucci
AT
,
Cerciello
G
,
Alfinito
F
, et al
.
Reduced regulatory T cells (Treg) in bone marrow preferentially associate with the expansion of cytotoxic T lymphocytes in low risk MDS patients
.
Br J Haematol
2019
;
185
:
357
60
.
20.
Kahn
JD
,
Chamuleau
ME
,
Westers
TM
,
Van de Ven
PM
,
van Dreunen
L
,
van Spronsen
M
, et al
.
Regulatory T cells and progenitor B cells are independent prognostic predictors in lower risk myelodysplastic syndromes
.
Haematologica
2015
;
100
:
e220
2
.
21.
Ott
PA
,
Hodi
FS
,
Robert
C
.
CTLA-4 and PD-1/PD-L1 blockade: new immunotherapeutic modalities with durable clinical benefit in melanoma patients
.
Clin Cancer Res
2013
;
19
:
5300
9
.
22.
Rizvi
NA
,
Hellmann
MD
,
Snyder
A
,
Kvistborg
P
,
Makarov
V
,
Havel
JJ
, et al
.
Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer
.
Science
2015
;
348
:
124
8
.
23.
Ghosh
A
,
Barba
P
,
Perales
MA
.
Checkpoint inhibitors in AML: are we there yet?
Br J Haematol
2020
;
188
:
159
67
.
24.
Boddu
P
,
Kantarjian
H
,
Garcia-Manero
G
,
Allison
J
,
Sharma
P
,
Daver
N
.
The emerging role of immune checkpoint based approaches in AML and MDS
.
Leuk Lymphoma
2018
;
59
:
790
802
.
25.
Berger
R
,
Rotem-Yehudar
R
,
Slama
G
,
Landes
S
,
Kneller
A
,
Leiba
M
, et al
.
Phase I safety and pharmacokinetic study of CT-011, a humanized antibody interacting with PD-1, in patients with advanced hematologic malignancies
.
Clin Cancer Res
2008
;
14
:
3044
51
.
26.
Davids
MS
,
Kim
HT
,
Bachireddy
P
,
Costello
C
,
Liguori
R
,
Savell
A
, et al
.
Ipilimumab for patients with relapse after allogeneic transplantation
.
N Engl J Med
2016
;
375
:
143
53
.
27.
Bashey
A
,
Medina
B
,
Corringham
S
,
Pasek
M
,
Carrier
E
,
Vrooman
L
, et al
.
CTLA4 blockade with ipilimumab to treat relapse of malignancy after allogeneic hematopoietic cell transplantation
.
Blood
2009
;
113
:
1581
8
.
28.
Zeidan
AM
,
Knaus
HA
,
Robinson
TM
,
Towlerton
AMH
,
Warren
EH
,
Zeidner
JF
, et al
.
A multi-center phase I trial of ipilimumab in patients with myelodysplastic syndromes following hypomethylating agent failure
.
Clin Cancer Res
2018
;
24
:
3519
27
.
29.
Daver
N
,
Garcia-Manero
G
,
Basu
S
,
Boddu
PC
,
Alfayez
M
,
Cortes
JE
, et al
.
Efficacy, safety, and biomarkers of response to azacitidine and nivolumab in relapsed/refractory acute myeloid leukemia: a nonrandomized, open-label, phase II study
.
Cancer Discov
2019
;
9
:
370
83
.
30.
Abbas
HA
,
Hao
D
,
Tomczak
K
,
Barrodia
P
,
Im
JS
,
Reville
PK
, et al
.
Single cell T cell landscape and T cell receptor repertoire profiling of AML in context of PD-1 blockade therapy
.
Nat Commun
2021
;
12
:
6071
.
31.
Zeidan
AM
,
Boss
I
,
Beach
CL
,
Copeland
WB
,
Thompson
E
,
Fox
BA
, et al
.
A randomized phase 2 trial of azacitidine with or without durvalumab as first-line therapy for higher-risk myelodysplastic syndromes
.
Blood Adv
2022
;
6
:
2207
18
.
32.
Gudmundsdottir
H
,
Wells
AD
,
Turka
LA
.
Dynamics and requirements of T cell clonal expansion in vivo at the single-cell level: effector function is linked to proliferative capacity
.
J Immunol
1999
;
162
:
5212
23
.
33.
Arakawa
A
,
Vollmer
S
,
Tietze
J
,
Galinski
A
,
Heppt
MV
,
Bürdek
M
, et al
.
Clonality of CD4(+) blood t cells predicts longer survival with CTLA4 or PD-1 checkpoint inhibition in advanced melanoma
.
Front Immunol
2019
;
10
:
1336
.
34.
Postow
MA
,
Manuel
M
,
Wong
P
,
Yuan
J
,
Dong
Z
,
Liu
C
, et al
.
Peripheral T cell receptor diversity is associated with clinical outcomes following ipilimumab treatment in metastatic melanoma
.
J Immunother Cancer
2015
;
3
:
23
.
35.
Abbas
HA
,
Reville
PK
,
Jiang
X
,
Yang
H
,
Reuben
A
,
Im
JS
, et al
.
Response to hypomethylating agents in myelodysplastic syndrome is associated with emergence of novel TCR clonotypes
.
Front Immunol
2021
;
12
:
659625
.
36.
Cristescu
R
,
Mogg
R
,
Ayers
M
,
Albright
A
,
Murphy
E
,
Yearley
J
, et al
.
Pan-tumor genomic biomarkers for PD-1 checkpoint blockade-based immunotherapy
.
Science
2018
;362:eaar3593.
37.
Grasso
CS
,
Tsoi
J
,
Onyshchenko
M
,
Abril-Rodriguez
G
,
Ross-Macdonald
P
,
Wind-Rotolo
M
, et al
.
Conserved interferon-gamma signaling drives clinical response to immune checkpoint blockade therapy in melanoma
.
Cancer Cell
2020
;
38
:
500
15
.
38.
Ji
RR
,
Chasalow
SD
,
Wang
L
,
Hamid
O
,
Schmidt
H
,
Cogswell
J
, et al
.
An immune-active tumor microenvironment favors clinical response to ipilimumab
.
Cancer Immunol Immunother
2012
;
61
:
1019
31
.
39.
Ayers
M
,
Lunceford
J
,
Nebozhyn
M
,
Murphy
E
,
Loboda
A
,
Kaufman
DR
, et al
.
IFN-gamma-related mRNA profile predicts clinical response to PD-1 blockade
.
J Clin Invest
2017
;
127
:
2930
40
.
40.
Gao
J
,
Shi
LZ
,
Zhao
H
,
Chen
J
,
Xiong
L
,
He
Q
, et al
.
Loss of IFN-gamma pathway genes in tumor cells as a mechanism of resistance to anti-CTLA-4 therapy
.
Cell
2016
;
167
:
397
404
.
41.
Zaretsky
JM
,
Garcia-Diaz
A
,
Shin
DS
,
Escuin-Ordinas
H
,
Hugo
W
,
Hu-Lieskovan
S
, et al
.
Mutations associated with acquired resistance to PD-1 blockade in melanoma
.
N Engl J Med
2016
;
375
:
819
29
.
42.
Ng Tang
D
,
Shen
Y
,
Sun
J
,
Wen
S
,
Wolchok
JD
,
Yuan
J
, et al
.
Increased frequency of ICOS+ CD4 T cells as a pharmacodynamic biomarker for anti-CTLA-4 therapy
.
Cancer Immunol Res
2013
;
1
:
229
34
.
43.
Gao
J
,
Navai
N
,
Alhalabi
O
,
Siefker-Radtke
A
,
Campbell
MT
,
Tidwell
RS
, et al
.
Neoadjuvant PD-L1 plus CTLA-4 blockade in patients with cisplatin-ineligible operable high-risk urothelial carcinoma
.
Nat Med
2020
;
26
:
1845
51
.
44.
Sade-Feldman
M
,
Jiao
YJ
,
Chen
JH
,
Rooney
MS
,
Barzily-Rokni
M
,
Eliane
JP
, et al
.
Resistance to checkpoint blockade therapy through inactivation of antigen presentation
.
Nat Commun
2017
;
8
:
1136
.
45.
Gettinger
S
,
Choi
J
,
Hastings
K
,
Truini
A
,
Datar
I
,
Sowell
R
, et al
.
Impaired HLA class I antigen processing and presentation as a mechanism of acquired resistance to immune checkpoint inhibitors in lung cancer
.
Cancer Discov
2017
;
7
:
1420
35
.
46.
Spranger
S
,
Bao
R
,
Gajewski
TF
.
Melanoma-intrinsic beta-catenin signalling prevents anti-tumour immunity
.
Nature
2015
;
523
:
231
5
.
47.
Peng
W
,
Chen
JQ
,
Liu
C
,
Malu
S
,
Creasy
C
,
Tetzlaff
MT
, et al
.
Loss of PTEN promotes resistance to T cell-mediated immunotherapy
.
Cancer Discov
2016
;
6
:
202
16
.
48.
Skoulidis
F
,
Goldberg
ME
,
Greenawalt
DM
,
Hellmann
MD
,
Awad
MM
,
Gainor
JF
, et al
.
STK11/LKB1 mutations and PD-1 inhibitor resistance in KRAS-mutant lung adenocarcinoma
.
Cancer Discov
2018
;
8
:
822
35
.
49.
Kitajima
S
,
Ivanova
E
,
Guo
S
,
Yoshida
R
,
Campisi
M
,
Sundararaman
SK
, et al
.
Suppression of STING associated with LKB1 loss in KRAS-driven lung cancer
.
Cancer Discov
2019
;
9
:
34
45
.
50.
Solinas
C
,
Gu-Trantien
C
,
Willard-Gallo
K
.
The rationale behind targeting the ICOS-ICOS ligand costimulatory pathway in cancer immunotherapy
.
ESMO Open
2020
;
5
:
e000544
.
51.
Batlle
E
,
Massague
J
.
Transforming growth factor-beta signaling in immunity and cancer
.
Immunity
2019
;
50
:
924
40
.
52.
Watson
AS
,
Mortensen
M
,
Simon
AK
.
Autophagy in the pathogenesis of myelodysplastic syndrome and acute myeloid leukemia
.
Cell Cycle
2011
;
10
:
1719
25
.
53.
Macian
F
.
Autophagy in T cell function and aging
.
Front Cell Dev Biol
2019
;
7
:
213
.
54.
Wang
MJ
,
Liu
WY
,
Wang
XY
,
Li
YM
,
Xiao
HY
,
Quan
RC
, et al
.
Autophagy gene panel-based prognostic model in myelodysplastic syndrome
.
Front Oncol
2020
;
10
:
606928
.