Enhancer heterogeneity in acute lymphoblastic leukemia drives differential gene expression in patients

Abstract

Genetic alterations alone cannot account for the diverse phenotypes of cancer cells. Even cancers with the same driver mutation show significant transcriptional heterogeneity and varied responses to therapy. However, the mechanisms underpinning this heterogeneity remain underexplored. Here, we find that novel enhancer usage is a common feature in acute lymphoblastic leukemia (ALL). In particular, KMT2A::AFF1 ALL, an aggressive leukemia with a poor prognosis and a low mutational burden, exhibits substantial transcriptional heterogeneity between individuals. Using single-cell multiome analysis and extensive chromatin profiling, we reveal that much transcriptional heterogeneity in KMT2A::AFF1 ALL is driven by novel enhancer usage. By generating high-resolution Micro Capture-C data in primary patient samples, we identify patient-specific enhancer activity at key oncogenes such as MEIS1 and RUNX2, driving high levels of expression of both oncogenes in a patient-specific manner. Overall, our data show that enhancer heterogeneity is highly prevalent in KMT2A::AFF1 ALL and may be a mechanism that drives transcriptional heterogeneity in cancer more generally.

Figure 1. Patients with B-ALL display enhancer heterogeneity between individuals.

(A) Correlation of accessibility at promoter regions (<2.5 kb from a TSS) measured by ATAC-seq signal between patient with DUX4/ERG, ETV6-RUNX1 (ETV-RUNX), and hyperdiploid subtypes. Data obtained from GSE161501. (B) Correlation of accessibility at putative enhancers (≥2.5 kb from a TSS) as measured by ATAC-seq signal between DUX4/ERG, ETV-RUNX, and hyperdiploid subtypes. (C) Principal component analysis of chromatin accessibility at promoters (left) and enhancers (right) for all 3 B-ALL subgroups. (D) ATAC-seq at the INTS9 locus for hyperdiploid samples. Putative enhancer regions with a high degree of intersample variability are highlighted in blue. (E) ATAC-seq at the SAMD12 locus for ETV6-RUNX1 samples. Putative enhancer regions with a high degree of intersample variability are highlighted in blue. (F) ATAC-seq at the MLLT3 locus for DUX4/ERG samples. Putative enhancer regions with a high degree of intersample variability are highlighted in blue.

Figure 2. Differential enhancer regions in KMT2A::AFF1 cell lines are functional enhancers that drive differential gene expression.

(A) Volcano plot of differentially expressed genes between RS4;11 (1883; red) and SEM cells (2468; blue) or no significant change (gray) from 3 biological replicates; false discover rate (FDR) of <0.05. (B) Volcano plot of enhancers with significantly increased accessibility in RS4;11 (1273; red) or SEM cells (1357; blue), or enhancers with unaltered accessibility (gray) from 8 biological replicates, FDR < 0.05. (C) Chromatin immunoprecipitation sequencing (ChIP-seq) tracks at the GNAQ locus for KMT2A, AFF1, H3K27ac, H3K4me1, H3K79me2, and H3K4me3 together with ATAC-seq and Capture-C in SEM cells using the GNAQ promoter as a viewpoint. The SEM-specific GNAQ enhancer (S1) is highlighted in blue. (D) ChIP-seq tracks at the ARID1B locus for KMT2A, AFF1, H3K27ac, H3K4me1, H3K79me2, and H3K4me3 together with ATAC-seq and Capture-C in SEM cells using the ARID1B promoter as a viewpoint. The RS4;11 specific intergenic enhancer (R1) is highlighted in red, and the SEM-specific intragenic enhancer (S2) is highlighted in blue. (E) Reverse transcription-qPCR comparing the expression of enhancer deletion mutants (light shading) with wild type (dark shading) in RS4;11 (red) or SEM cells (blue) when deleting either the intragenic GNAQ enhancer (S1; left) or ARID1B intergenic enhancer (R1; right). Significance of alterations in relative copy number were determined by a 2-sided t test with correction for multiple testing (Benjamini-Hochberg), n = 6 biological replicates. *Adjusted P value < .05; **P < .01. ns, not significant.

Figure 3. Enhancer heterogeneity persists in genetically matched KMT2A::AFF1 leukemias derived from a single donor.

(A) Representative flow cytometry plots of the sorting strategy used for KMT2A::AFF1 ALL samples. (B) Scaled H3K27ac level at the 500 most variable enhancer peaks between the 2 KMT2A::AFF1 models. (C) Tornado plot of H3K27ac CUT&Tag signal in 2 HSPC-derived KMT2A::AFF1 ALL models at the 500 most variable enhancer peaks, k-means clustering separates the regions into enhancers showing increased activity in KMT2A::AFF1 ALL 1 (top) vs KMT2A::AFF1 ALL 2 (bottom). (D) CUT&Tag for H3K27ac at the PLXNA4 locus, enhancer regions with increased activity in KMT2A::AFF1 ALL 1 are highlighted in blue. (E) CUT&Tag for H3K27ac at the ENSG00000287092 locus, enhancer regions with increased activity in KMT2A::AFF1 ALL 2 are highlighted in red. FSC, forward scatter; HSPC, hematopoietic stem and progenitor cell.

Figure 4. Differential enhancer regions in KMT2A::AFF1 patients are readily observed.

(A) UMAP of the single-cell ATAC-seq modality for 4 KMT2A::AFF1 blast samples from the VIVO Biobank, United Kingdom. (B) Regions of accessible chromatin displaying significantly increased accessibility in 1 of 4 patient samples. (C) Genomic distribution of uniquely accessible ATAC-seq peaks relative to the nearest TSS, the dotted gray line indicates 2.5 kb. (D) Annotation of the genomic location of unique ATAC-seq peaks. (E) Schematic of the strategy used to identify unique enhancer peaks using H3K27ac ChIP-seq data sets. (F) Tornado plot of H3K27ac signal in KMT2A::AFF1 samples at enhancers identified as being patient specific. UMAP, uniform manifold approximation and projection.

Figure 5. Differential enhancer activity in KMT2A::AFF1 patients drives oncogene specific expression such as at MEIS1 and RUNX2.

(A) UMAP of single-nucleus gene expression (snGEX) for 4 KMT2A::AFF1 blast samples (chALL1, iALL3-5; VIVO Biobank, United Kingdom) and 3 single-cell GEX (scGEX) samples (EGAS00001003986). (B) Dot plot of marker gene analysis between 7 KMT2A::AFF1 blast samples, showing the top 5 marker genes per sample. (C) Normalized MEIS1 (left) and RUNX2 (right) expression in KM2TA::AFF1 sn/scGEX samples. (D) TOPmentation for H3K27ac and the N terminus of KMT2A (KMT2A-N) in KMT2A::AFF1 blast sample (VIVO Biobank, United Kingdom) at the MEIS1 locus. The chALL1 unique enhancer region downstream of MEIS1 is highlighted in red. (E) TOPmentation for H3K27ac and KMT2A-N in KMT2A::AFF1 blast samples at the RUNX2 locus. The chALL1- and iALL2-specific enhancer region upstream of RUNX2 is highlighted in red. (F) Survival curve comparing high (red) and low (green) MEIS1 expression in B-ALL. Data analyzed from COG P9906 childhood B-ALL clinical trial. (G) Survival curve comparing high (red) and low (green) RUNX2 expression in B-ALL. Data analyzed from the Eastern Cooperative Oncology Group E2993 adult B-ALL clinical trial. UMAP, uniform manifold approximation and projection.

Figure 6. MCC reveals patient-specific enhancer-promoter interactions in primary patient cells.

(A) A schematic for the MCC protocol. (B) MCC at the MEIS1 locus using the promoter (blue highlight) as the viewpoint (triangle) for SEM cells (blue) or chALL1 cells (red), or from the chALL1-unique enhancer region (enhancer 27; red highlight), together with TOPmentation for H3K27ac and KMT2A-N. (C) MCC at the RUNX2 locus in chALL1 cells using either the promoter (blue highlight) or open chromatin regions (enhancers 36-47) within the identified enhancer region (red highlight) as the viewpoint (triangle) together with TOPmentation for H3K27ac and KMT2A-N in SEM and chALL1 cells. (D) MCC at the CD69 locus in chALL1 cells using either the promoter (blue highlight) or open chromatin regions (enhancers 14-19) within the identified enhancer regions (red highlight) as the viewpoint (triangle) together with TOPmentation for H3K27ac and KMT2A-N in SEM and chALL1 cells. (E) Expression of CD69 in patients-derived KMT2A::AFF1 blast samples. (F) MCC for the ARID1B viewpoint (triangle) together with ChIP-seq for H3K27ac and KMT2A-N at the ARID1B locus. The SEM-specific intragenic enhancer is highlighted in blue. (G) Expression of ARID1B in patient-derived KMT2A::AFF1 blast samples. (H) Expression of genes linked to chALL1-unique enhancer regions between KMT2A::AFF1 single-cell blast samples.

Figure 7. An unbiased machine-learning model identifies KMT2A::AFF1 complex binding as a driver of differential enhancer usage.

(A) Proportion of enhancers (common = no change in activity, RS4;11 = increased activity in RS4;11 cells, SEM = increased activity in SEM cells) containing either an SNV (blue), an indel (yellow), or both (green) in RS4;11 cells (left) or SEM cells (right). (B) Proportion of enhancers of each enhancer type (common, RS4;11, SEM) containing no heterozygous SNVs (dark gray), SNVs removed due to intrinsic bias (eg, problematic genomic regions or mapping bias; gray), SNVs without allele-specific bias in accessibility (light gray), or those exhibiting allele specific bias in accessibility as measured by ATAC-seq (red). (C) Schematic of the strategy used to determine key predictive features of differential enhancer activity. ChIP-seq signal for 56 factors was extracted over enhancers with increased activity in RS4;11 cells (red; 1522), or SEM cells (blue; 1677) or those common to both (gray; 4232). A gradient boosted decision tree was trained from these data, and predictive features were extracted using SHAP. (D) The relative feature importance for each enhancer category (increased activity in SEM cells [blue], RS4;11 cells [red], or common enhancers [gray]) of the top 20 most important features for differential enhancer prediction. Features that correspond to binding of the KMT2A::AFF1 complex are highlighted, and a schematic of the complex is provided for reference. (E) Tornado plot of AFF1-C and KMT2A-N ChIP-seq signal at enhancers displaying increased activity in RS4;11 cells (top) or SEM cells (bottom) in RS4;11 (left) or SEM (right) cells. (F) Pearson correlation between H3K27ac and KMT2A signal at the 290 blast-specific enhancers identified, for each patient sample. (G) H3K27ac ChIP-seq signal at enhancers with increased activity in SEM cells (blue), RS4;11 cells (red), or common enhancers (gray) upon KMT2A::AFF1 knockdown by small-interfering RNA (siRNA; dashed line). (H) Enhancer-promoter interaction frequency at enhancer regions with increased activity in SEM cells (i-iv) or RS4;11 cells (v) upon treatment of SEM (i) or RS4;11 (v) cells with 2 μM EPZ5676 for 1 week or SEM PAF1-FKBP12^F36V (iii)/SEM SSRP1-FKBP12^F36V (iv) cells treated with dTag13 for 24 hours, together with SEM cells treated with an siRNA against KMT2A::AFF1 (i). Interaction frequency for each enhancer-promoter pair is shown relative to the mean interaction frequency of the control; n = 3 biological replicates per condition. *P < .05; **P < .01; ***P < .001. (I) Example of a loss of enhancer-promoter interactions at the ARID1B locus in SEM cells as assessed by Capture-C in control (gray) or KMT2A::AFF1 knockdown conditions (red) in 3 biological replicates. Enhancers with increased activity in SEM cells are highlighted in blue. ChIP-seq for H3K27ac in control (gray) or KMT2A::AFF1 knockdown conditions (red) in addition to the N terminus of KMT2A and the C terminus of AFF1 are provided for reference. (J) Model for the role of the KMT2A::AFF1 complex in promoting transcription heterogeneity between patients. bp, base pair; dTag, dTAG-13; EPZ, EPZ5676; indel, insertion-deletion; KD, knockdown; SHAP, SHapley Additive exPlanations.