Immune profiles in acute myeloid leukemia bone marrow associate with patient age, T-cell receptor clonality, and survival

Abstract

The immunologic microenvironment in various solid tumors is aberrant and correlates with clinical survival. Here, we present a comprehensive analysis of the immune environment of acute myeloid leukemia (AML) bone marrow (BM) at diagnosis. We compared the immunologic landscape of formalin-fixed paraffin-embedded BM trephine samples from AML (n = 69), chronic myeloid leukemia (CML; n = 56), and B-cell acute lymphoblastic leukemia (B-ALL) patients (n = 52) at diagnosis to controls (n = 12) with 30 immunophenotype markers using multiplex immunohistochemistry and computerized image analysis. We identified distinct immunologic profiles specific for leukemia subtypes and controls enabling accurate classification of AML (area under the curve [AUC] = 1.0), CML (AUC = 0.99), B-ALL (AUC = 0.96), and control subjects (AUC = 1.0). Interestingly, 2 major immunologic AML clusters differing in age, T-cell receptor clonality, and survival were discovered. A low proportion of regulatory T cells and pSTAT1+cMAF- monocytes were identified as novel biomarkers of superior event-free survival in intensively treated AML patients. Moreover, we demonstrated that AML BM and peripheral blood samples are dissimilar in terms of immune cell phenotypes. To conclude, our study shows that the immunologic landscape considerably varies by leukemia subtype suggesting disease-specific immunoregulation. Furthermore, the association of the AML immune microenvironment with clinical parameters suggests a rationale for including immunologic parameters to improve disease classification or even patient risk stratification.

Graphical abstract

Figure 1.

Immunocharacterization of the AML, B-ALL, CML, and control BM. (A) Visualization of the quantitative BM immunocharacterization pipeline. FFPE BM tissue blocks of AML patients (n = 69) and age- and sex-matched controls (n = 12) were retrieved from the Helsinki Biobank. TMAs were constructed from duplicate punches (1 mm in diameter) from each subject. TMAs were cut onto tissue slides and stained with mIHC consisting of ≤4 primary antibodies detected with fluorescence dyes and 4′,6-diamidino-2-phenylindole (DAPI) counterstain as well as 2 primary antibodies detected with chromogenic probes and hematoxylin counterstain. Tissue slides were scanned after both staining procedure and corresponding images registered to ensure that the location of individual cells is matched in parallel stainings. Following cell segmentation, marker colocalization and intensity were quantified in identified cells. (B) Immune cells (as a proportion of all cells in a TMA spot) and their immunophenotypes (as a proportion of their parent immune cell) derived from mIHC and computerized image analysis are plotted on a heatmap and organized by hierarchical clustering using Spearman correlation distance and the Ward linkage (ward.D2) method. Immunologic parameters are arranged in rows and patients in columns. Red denotes higher and blue lower proportions. (C) Using 10-fold crossvalidated elastic net–regularized logistic regression analysis, 4 subtype-specific classifiers were developed to identify AML, B-ALL, and CML patients and controls. Classifiers were developed with a training group (n = 94) and assessed with a test group (n = 95). The accuracy of the classifiers was evaluated on the test group with receiver-operator curves (AUC).

Figure 2.

Comparison of the AML and control BM. (A) Immune cells (as proportion to all cells in a TMA spot) and their immunophenotypes (as proportion to their parent immune cell) derived from mIHC and computerized image analysis are plotted on a heatmap and organized by hierarchical clustering using Spearman correlation distance and the Ward linkage (ward.D2) method. Immunologic parameters are arranged in rows and patients in columns. Red denotes higher and blue lower proportions. Horizontal column bars indicate AML etiology, complex karyotype, NPM1 and FLT3-ITD molecular genetics, ELN 2017 risk classification, BM blast proportion (%), PB leukocyte count (E9/mL), and TCR clonality. (B) To focus only on significant comparisons (Benjamini-Hochberg corrected q < 0.05), the median AML-to-control ratio of each immunologic parameter was transformed to twofold logarithmic scale and grouped as anticancer (green) or procancer immunologic markers (orange) according to the literature. (C) Spearman correlation of immunologic parameters. Red denotes positive and blue negative correlations. Insignificant correlations (Benjamini-Hochberg corrected q < 0.05) were blanked. NA, values are not defined (gray bars in panel A).

Figure 3.

Clinicoimmunologic analysis of major AML immune profiles. (A) The proportion of immune cells and their single-marker immunophenotypes in AML patients (n = 69) in clusters 1 and 2 were compared (Mann-Whitney U test), P values corrected (Benjamini-Hochberg procedure), and results plotted on a volcano plot. Immunologic features with log2 fold change (FC) >0 are enriched in cluster 1 and features with log2 fold change <0 are more frequent in cluster 2. Age (Mann-Whitney U test) (B), allo-HSCT (χ² test) (C), and induction treatment frequency (Fisher’s exact test) (D) distributions of patients in clusters 1 and 2 were compared. The survival of patients in clusters 1 and 2 who were intensively treated (HD-cytarabine–based induction treatment; n = 59) were compared (Cox regression, log-rank test) using EFS (E), relapse-free survival (RFS) (F), and overall survival (OS) (G) as an end point. TCR clonality was compared between patients in cluster 1 and cluster 2 (Mann-Whitney U test) (G) and correlated with patient age (Spearman correlation) (I).

Figure 4.

Immunologic prognostic biomarkers in intensively treated AML patients (HD-cytarabine–based induction treatment; n = 59). (A) Volcano plot of the HRs for an EFS incident of immune cells and their immunophenotypes and PB and BM laboratory values. Variables increasing the risk for an EFS event have a positive log10 HR and deviate to the right side of the plot. Parameters with a negative log10 HR diverge to the left. Survival plot of the proportion of CD68⁺ monocytes expressing pSTAT1⁺cMAF⁻ (M1-like monocytes) (B) and helper T cells expressing FOXP3⁺ (%, regulatory T cells) (C). Covariates were divided into tertiles, and the highest 2 subgroups were combined. A forest plot displaying the HR and 95% CI of the categorized proportion of M1-like monocytes (D) and Tregs (E) in univariate and combined with another essential clinical biomarker (Cox regression analysis).

Figure 5.

Comparison of immunophenotypes in BM and PB samples. (A) Boxplots and dotplots displaying only significant comparisons (Mann-Whitney U test, P < .05) of CD8⁺ T- and NK-cell immunophenotypes quantified from paired BM and PB samples of AML patients at diagnosis. The center line of boxplots displays the median value and whiskers the interquartile range. The color of the bar reflects whether the immunophenotype is from BM (red) or PB (blue). P values for the comparison are listed on the top (**P < .01, *P < .05). (B) Differences in the median proportion of immunophenotypes from BM and PB samples in AML patients and healthy controls. The length of the bar signifies the amplitude of the difference between BM and PB samples (value in BM subtracted by value in PB). Bars orientated to the right represent immunophenotypes more prevalent in BM than PB samples. All common immunophenotypes analyzed with flow cytometry in both AML and control BM and PB samples are presented. As control samples were unpaired, no P values were computed.