Skip to main content
NCBI home page
Blood logoLink to Blood
. 2025 Apr 17;146(4):471–481. doi: 10.1182/blood.2024028215

Differentiation-dependent EBF1 activity determines CD22 transcription and leukemia sensitivity to inotuzumab ozogamicin∗

Carolin S Escherich 1,2, Zhenhua Li 1, Kelly R Barnett 1, Yizhen Li 1,3, Megan Walker 1, Satoshi Yoshimura 1, Wenjian Yang 1, Xin Huang 4, Jiyang Yu 4, Wendy Stock 5, Elisabeth Paietta 6, Marina Y Konopleva 7, Steven M Kornblau 8, Elias Jabbour 8, Mark R Litzow 9, Hiroto Inaba 10, Ching-Hon Pui 10, Mignon L Loh 11, William E Evans 1, Daniel Savic 1,12, Jun J Yang 1,12,∗
PMCID: PMC13675886  PMID: 40435412

Key Points

  • •

    Early B-cell differentiation arrest is associated with resistance to InO in B-cell ALL.

  • •

    B-lineage transcription factor EBF1 directly regulates the expression of CD22 and modulates ALL sensitivity to InO.

Visual Abstract

graphic file with name BLOOD_BLD-2024-028215-ga1.webp

Open in a new tab

Abstract

Inotuzumab ozogamicin (InO) is an antibody-calicheamicin conjugate with high efficacy in lymphoid malignancies. It targets the B-cell surface protein CD22, which is expressed in most B-cell acute lymphoblastic leukemia (B-ALL) cases, albeit with variable intensity. However, factors governing CD22 expression and thus leukemia sensitivity to InO remain incompletely understood. Using multiomic characterization of 196 human B-ALL samples, coupled with ex vivo InO sensitivity profiling, we showed that early leukemia differentiation arrest at the pre–pro-B stage is associated with resistance to InO. Screening of 1639 transcription factor genes identified early B-cell factor 1 (EBF1) as a key regulator of CD22 expression (false discovery rate of 7.1 × 10−4). When comparing the assay for transposase-accessible chromatin with sequencing profiling results of the most InO-sensitive and -resistant cases (50% lethal concentration <10th vs >90th percentile, n = 18), the binding motif for EBF1 was strikingly enriched in regions with differential open chromatin status (P = 8 × 10−174). CRISPR interference targeting EBF1 binding sites at the CD22 locus led to an ∼50-fold reduction in cell surface CD22 expression and, consequently, an ∼22-fold increase in InO resistance in ALL cell lines. Interestingly, within BCR::ABL1 ALL, we observed intrasubtype heterogeneity linked to EBF1 transcriptional downregulation (P = 1.1 × 10−15) and/or somatic alteration (P = .004), which led to reduced CD22 expression (P = 8.3 × 10−11) and ex vivo and in vivo resistance to InO. Collectively, these findings point to the direct impact of EBF1 on CD22 expression during B-cell development, which, in turn, contributes to interpatient variability in InO response, even within the same subtype of B-ALL.

The efficacy of the anti-CD22 antibody calicheamicin drug conjugate inotuzumab ozogamicin (InO) in B-cell acute lymphoblastic leukemia (B-ALL) is variable for unknown reasons. Escherich and colleagues used patient samples and preclinical models to demonstrate that InO sensitivity in B-ALL depends on the stage of B-cell developmental arrest and identified CD22 as a direct transcriptional target of early B-cell factor 1 (EBF1). The authors showed that B-ALL samples exhibiting gene-regulatory profiles resembling those of normal pre-pro-B cells prior to EBF1 activation have reduced responsiveness to InO, providing strong clues to the mechanisms underpinning interpatient heterogeneity in response.

Introduction

Inotuzumab ozogamicin (InO) is an antibody-cytotoxin conjugate approved for the treatment of CD22+ hematologic malignancies, especially B-cell acute lymphoblastic leukemia (B-ALL).1,2 In adults with relapsed or refractory CD22+ B-ALL, InO has been shown to be superior to conventional chemotherapy.3,4 In addition, experience in pediatric patients have shown similarly promising results with overall response rates of 81.5% (ITCC-059)5 and 58.3% (COG AALL1621, ClinicalTrials.gov identifier: NCT02981628)6 among children with relapsed or refractory disease.

InO targets the B-cell–specific glycoprotein CD22, thereby delivering the chemotherapeutic agent, calicheamicin, directly to leukemic cells.7 Consequently, InO cytotoxicity is dependent on CD22 expression on the leukemic cell surface. Downregulation of CD22 expression and/or loss-of-function mutations in this gene are the main reasons for acquired InO resistance and resistance to other CD22-directed therapies.8, 9, 10, 11, 12 Variation in clinical response to this drug also exists in newly diagnosed ALL, possibly linked to interpatient and intraleukemia heterogeneity of CD22 expression.12, 13, 14 However, reports on the extent to which CD22 expression is attributed to InO sensitivity vary across clinical trials.3,5,14,15

CD22 functions as a negative regulator of B-cell receptor signaling and exhibits high expression across multiple developmental stages of the B-cell lineage.16,17 In B-ALL, about 90% of the cases express CD22, albeit with great variability in expression levels.18,19 There is emerging evidence that this variation may be associated with the degree of maturation arrest of leukemia blasts.18 However, the exact relationship between CD22 expression and leukemia sensitivity to InO, particularly in the context of ALL developmental state, remains unclear. Furthermore, the molecular mechanisms governing CD22 transcription in B-ALL are not well characterized.

In this study, we performed a multiomic characterization of 196 pediatric and adult patients with B-ALL using transcriptomic sequencing, assay for transposase-accessible chromatin with sequencing (ATAC-seq), and ex vivo profiling of InO sensitivity (Figure 1A). Based on these profiling data, we sought to comprehensively investigate the genomic basis of interpatient variability in CD22 expression and thus ALL sensitivity to InO. Our study revealed EBF1 as the main regulator of CD22 expression, and its loss or downregulation in B-ALL was associated with resistance to InO.

Figure 1.

Figure 1.

Open in a new tab

Association between CD22 expression and InO response or B-cell differentiation in primary human B-ALL. (A) Circos plot with patient characteristics (sex, age) and leukemia profiling (molecular subtype, InO ex vivo pharmacotyping, and ATAC-seq) data availability for all patients used in this study. (B) The transcriptome-wide association of gene expression with InO LC50 in 196 patients was performed by differential expression analysis using limma-voom by fitting a linear model between gene expression and InO LC50 and using molecular subtypes as covariates. (C) Dynamic change of CD22 expression across B-cell differentiation stages derived from single-cell RNA-seq profiles of healthy human bone marrow (https://scminer.stjude.org).20 (D) CD22 expression across molecular subtypes of B-ALL.21 Only subtypes represented by >10 patients are depicted. CLP, common lymphoid progenitor; HSC, hematopoietic stem cell.

Materials and methods

Patients and samples

In total, we collected 196 primary B-ALL samples for this study (supplemental Table 1, available on the Blood website). Of those, 95 were from pediatric patients from the St. Jude Children’s Research Hospital (Total Therapy XVII, ClinicalTrials.gov identifier: NCT03117751) and 101 were from adult patients from the MD Anderson Cancer Center, the University of Chicago, the ECOG-ACRIN E1910 clinical trial (ClinicalTrials.gov identifier: NCT02003222), and the Alliance A041501 trial (ClinicalTrials.gov identifier: NCT03150693).22 All samples were screened for InO sensitivity ex vivo and subjected to RNA sequencing (RNA-seq; supplemental Table 2); 99 samples were subjected to ATAC-seq. Samples of an additional 91 patients were collected for the analysis of BCR::ABL1 heterogeneity. These patients were enrolled in the previously mentioned clinical trials but were not tested for InO sensitivity. Human subject research was performed with the approval of the respective institutional review boards, and all human samples were obtained with informed consent in accordance with the Declaration of Helsinki.

InO sensitivity screening of patient-derived primary human leukemia blasts

InO was provided by Pfizer Inc under a research collaborative agreement. Drug sensitivity was determined using a mesenchymal stem cell (MSC) coculture system with a fluorescence imaging–based assay, as described previously.23 Briefly, human telomerase reverse transcriptase (hTERT)-immortalized MSCs were seeded in a 384-well plate at 1500 cells per well in 25 μL MSC medium (RPMI-1640, l-glutamine, 10% fetal calf serum, and 1 μM hydrocortisone). After 24 hours, the MSC medium was removed and leukemia cells were added to the stromal cell layer at 25 000 cells per well in 40 μL AIM-VTM medium (Gibco, catalog no. 12055083), along with 6 10-fold serial dilutions of InO. After a 96-hour incubation, drug-induced cytotoxicity was determined by comparing the total number of live leukemia cells at each concentration with vehicle-treated cells, and the 50% lethal concentration (LC50) was calculated using a 4-parameter dose-response model (supplemental Figure 1). For samples in which the lowest drug concentration killed >50% of the leukemia cells, the LC50 was assigned as half of the minimum tested concentration (0.005 ng/mL). Conversely, for cases with >50% viability at the highest drug concentration, the LC50 was assigned as twice that of the highest tested concentration (500 ng/mL). Quality control was performed to remove samples with low viability (<1000 viable blast cells at day 4 in vehicle-treated controls). Details of the methods for other experiments performed and genomic data processing and analysis can be found in the supplemental Methods. Animal studies were conducted in accordance with guidance by the St. Jude Children’s Research Hospital Institutional Animal Care and Use Committee.

Software and statistical analysis

All experiments were independently performed at least 3 times. Student t tests were used to calculate the P values, unless otherwise indicated. Western blot membranes were examined, and protein bands were detected using the LI-COR Odyssey XF Imager (LI-COR). The flow cytometry analysis was performed using the BD LSR Fortessa Flow Cytometer (BD Biosciences), and data analysis was done using FlowJo software (version 10.10, Tree Star). GraphPad Prism (version 10.2.3) and RStudio (version 4.4.0) were used for data visualization.

Results

CD22 expression varies during B-cell development and influences leukemia sensitivity to InO

To comprehensively investigate the biologic basis of interpatient variability in B-ALL response to InO, we performed both genomic and pharmacologic profiling of 196 patients with newly diagnosed B-ALL.22 Representing 21 molecular subtypes, our cohort consisted of 101 adults and 95 children (Figure 1A; supplemental Table 1). We first examined gene expression globally for correlation with leukemia response to InO (LC50) with adjustment for molecular subtypes. As expected, high expression levels of CD22 displayed the strongest association with InO sensitivity across B-ALL subtypes (P = 4.4 × 10–7; Figure 1B), followed by the B-cell transcription factor (TF) genes PAX5 (P = 1.1 × 10–6) and EBF1 (P = 3.0 × 10–6).

Using single-cell RNA-seq data of healthy human bone marrow,20 we next examined the dynamic change in CD22 expression across B-cell differentiation stages. As shown in Figure 1C, most B-cell subsets were CD22+ with a dramatic activation of CD22 transcription at the transition from pre–pro-B to pro-B stage. In B-ALL, CD22 expression also differed between the different molecular subtypes (Figure 1D). In particular, the lowest median CD22 expression was seen in KMT2A-rearranged ALL, a subtype characterized by early B-cell differentiation arrest20 and inferior response to InO.15,24 We therefore hypothesized that maturation arrest of B-ALL leads to differential CD22 expression, which, in turn, leads to varying sensitivity to InO.

Early B-cell differentiation arrest was associated with InO resistance

To further determine the impact of leukemia developmental state on B-ALL sensitivity to InO, we next compared the ex vivo InO sensitivity in 196 diagnostic leukemia samples with the expression of 109 marker genes of each stage of B-cell differentiation (supplemental Table 3).20 Unsupervised hierarchical clustering analysis identified 2 subgroups that were largely in line with the developmental state. Cluster 1 (C1) accounted for 63 of 196 cases (32.1%) and was characterized by high expression levels of common lymphoid progenitor and pre–pro-B marker genes, whereas C2 accounted for 133 of 196 cases (67.9%) and showed prominent expression of pro-B and pre-B genes (Figure 2A). Comparing the InO sensitivity between these 2 groups, we found that cases in C1 were significantly more resistant to InO than those in C2 (P = 1.9 × 10–7; Figure 2B). We then inferred the leukemia developmental state by applying digital cytometry to B-ALL bulk RNA-seq data (Figure 2C; supplemental Table 4).20 As expected, early developmental arrest at the pre–pro-B stage was highly associated with patient cluster C1 (P < .0001), whereas leukemia cells derived from cases in C2 were enriched in pro-B–like cells (P < .0001; Figure 2C; supplemental Figure 2). When looking at the cases that were most sensitive (LC50 <10th percentile) and most resistant (LC50 >90th percentile) to InO, the median percentage of pre–pro-B cells relative to all developmental stages increased from 38% to 54% (P = 1.8 × 10–4), whereas the pro–B-cell percentage decreased from 37% to 19% (P = 1.8 × 10–4; Figure 2D). Taken together, these results suggest that early differentiation arrest at the pre–pro-B stage contributes to leukemia resistance to InO, and the transition to the pro-B stage leads to high CD22 expression and an associated increase in InO sensitivity in B-ALL.

Figure 2.

Figure 2.

Open in a new tab

Early B-cell differentiation arrest conferred resistance to InO in primary human B-ALL. (A) Unsupervised hierarchical clustering of 196 B-ALL samples with RNA-seq data, and focusing on 109 protein coding genes specifically upregulated during different stages of B-cell differentiation, identified 2 groups of cases, namely C1 and C2. Each row represents 1 gene, each column represents 1 case, and expression levels were scaled to a mean value of 0 and standard deviation (SD) of 1. The Pearson correlation coefficients of gene expression (RNA-seq) and InO sensitivity (LC50 ex vivo) in 196 cases are shown as bars. Blue indicates negative correlation (high expression in InO-sensitive cases); red indicates positive correlation (high expression in InO-resistant cases). (B) The ex vivo InO LC50 distribution is presented in box plots for the comparison between cases in C1 and C2. Box plots span the first to third quartiles and whiskers show the 10th and 90th percentile. The P value for the comparison between the LC50 values was determined using a 2-sided unpaired Student t test. (C) CIBERSORTx-inferred relative abundance of each B-cell stage cell type within the leukemia sample. Each column of stacked bars represents an individual patient, colored by B-cell differentiation stage populations within the leukemia sample. (D) Box plot showing the proportion of cells in pre–pro-B-like (left) and pro-B–like state (right). B-ALL cases were grouped based on their ex vivo InO sensitivity into highly sensitive (<10th percentile [P10], light green), sensitive (10th-25th percentile [P25], green), moderate (25th-75th percentile [P25-75], gray), resistant (75th-90th percentile [P75], red), and highly resistant (>90th percentile [P90], light red). Box plots span the first to third quartiles and whiskers show the 10th and 90th percentile. P values for the comparison between cell type abundance and InO response were estimated using linear regression. CLP, common lymphoid progenitor.

TF CRISPR screen identified EBF1 as a key regulator of CD22 expression

With the variable expression of CD22 during B-cell development (Figure 1C), we next sought to comprehensively identify TFs that were responsible for regulating CD22 expression in B-ALL. A pooled CRISPR/Cas9 single-guide RNA (sgRNA) library that targeted 1639 human TFs was used with 7 sgRNAs for each TF.25 First, Nalm6-Cas9 cells were infected with the lentiviral sgRNA library before fluorescence-activated cell sorting was performed to collect the top 10% and bottom 10% of cells with the highest and lowest cell surface CD22 expression, respectively. The TF genes that regulated CD22 expression were identified by comparing the sgRNA enrichment in the 2 groups (Figure 3A). We discovered 11 TFs for which CRISPR knockout led to drastic downregulation of CD22 expression (log2 fold change [FC] less than −1; false discovery rate, <0.05; supplemental Table 5) with EBF1 being the top hit (log2FC = −2.28; false discovery rate, 7.1 × 10–4; Figure 3B). To validate the EBF1 results, we generated new sgRNAs independent of those used in the original CRISPR screen (supplemental Table 6) that targeted all transcripts of the EBF1 gene. EBF1 pooled knockout cells (EBF1–/–) were established for 2 B-ALL cell lines, namely Nalm6 (with DUX4 rearrangement) and REH (with ETV6::RUNX1 rearrangement). As shown in Figure 3C-E, EBF1 loss significantly downregulated cell surface CD22 expression (P < .0001 in both cell lines) and reduced B-ALL sensitivity to InO when compared with parental wild-type Nalm6 or REH cells. In contrast, CD22 expression and sensitivity to InO was largely restored when EBF1 was reintroduced in Nalm6 knockout cells through ectopic expression, confirming that drug resistance was specifically driven by the EBF1 loss (supplemental Figure 3). In addition, calicheamicin sensitivity was not affected by the EBF1 genotype (supplemental Figure 4), suggesting that InO resistance in EBF1–/– cells was caused by the loss of CD22, which impairs InO binding and uptake into the leukemia cell. We further evaluated leukemia EBF1 expression in 196 B-ALL cases in our cohort using RNA-seq quantification. Although EBF1 was expressed in most cases (Figure 3F), 8 cases showed expression that was >1.5-fold interquartile range below the first quartile, and these were classified as EBF1 low (Figure 3F). Similar to the results in B-ALL cell lines, EBF1-low cases displayed a 19.4-fold decrease in the median CD22 expression (P < .0001; Figure 3G) and a 123-fold increase in the median InO LC50 (P < .0001; Figure 3H) when compared with all other B-ALL cases.

Figure 3.

Figure 3.

Open in a new tab

CRISPR screen of 1639 TF genes identified EBF1 as a main driver of CD22 expression. (A) Nalm6/Cas9 cells were lentiviral transduced with the human TF knockout library (Addgene catalog no. 162275) that targeted 1638 TF genes. Transduced cells were first selected by puromycin for 7 days and then stained with anti–CD22-PE antibody. Based on fluorescence intensity, the top 10% and bottom 10% of the cells were separated by FACS. sgRNAs were amplified from the genomic DNA of each population and sequenced using NovaSeq to analyze sgRNA enrichment. (B) Volcano plot depicting the TF genes that positively or negatively regulate CD22 expression. EBF1 was found to be the top activator of CD22 expression (log2FC = –2.28; P = 2.86 × 10–6). (C-E) CD22 expression (C), InO cytotoxicity in the B-ALL cell lines Nalm6 (D), and REH (E) with WT (gray) or EBF1 pooled knockout (EBF1–/–; red) genotype. CD22 expression was quantified by flow cytometry using anti–CD22-PE fluorescent antibody. Each dot represents the mean ± SD of 3 independent experiments. Dose-response curves were fitted using a 4-parameter nonlinear regression model. Validation of EBF1 expression was done by western blotting using anti-human EBF1 antibody. Vinculin was used as the internal control. (F) Distribution of the EBF1 expression levels in 196 primary B-ALL cases as quantified by RNA-seq. Box plots span the first to third quartiles and whiskers show 1.5× the interquartile range (IQR). Outliers with EBF1 expression below the first quartile −1.5× IQR were considered as EBF1-low cases (red dots). CD22 expression (G) and ex vivo InO sensitivity (H) were compared between the EBF1-low cases (n = 8, red) and all other cases (n = 188, gray). Violin plots span the minimum to maximum value with the first to third quartiles indicated by the horizontal lines. Each dot represents a case. Statistical significance was assessed using a 2-sided unpaired Student t test. FACS, fluorescence-activated cell sorting; FI, fluorescence intensity; PE, phycoerythrin; WT, wild-type.

EBF1-mediated enhancer activation at the CD22 locus dictated leukemia sensitivity to InO

To delineate the epigenetic mechanisms that linked EBF1 to CD22 expression in ALL, we comprehensively profiled genome-wide open chromatin status using ATAC-seq in a subset of 99 cases of our original cohort (Figure 1A). Global TF activity was estimated using the TOBIAS (Transcription factor Occupancy prediction by Investigation of ATAC-seq Signal) framework26 and compared between cases with extreme sensitivity and those with resistance to InO (LC50 <10th vs >90th percentile). Strikingly, EBF1 (signified by binding motif MA0154.2, MA0154.3, and MA0154.4) was the most differentially active TF between these 2 groups (P = 8 × 10–174; Figure 4A; supplemental Figure 5; supplemental Tables 7 and 8) among the 810 motifs tested. In fact, we identified 4 regions with prominent open chromatin status within ±10 kb of the CD22 transcription start site (Figure 4B). Of those regions, 3 showed strong correlation with CD22 expression (P < 1 × 10−3; supplemental Table 9) and demonstrated EBF1 binding using chromatin immunoprecipitation sequencing (Figure 4B). We therefore prioritized these regions for further functional validation using a CRISPR interference assay in Nalm6 and REH cells. We designed 3 sgRNAs for each region with 1 sgRNA directly targeting the EBF1 motif (sgRNA RX.2) and 2 sgRNAs within 100 bp upstream and downstream of the EBF1 motif (sgRNA RX.1 and RX.3; Figure 4B). As shown in Figure 4C-E, interference with EBF1 binding in all 3 target regions strongly repressed CD22 expression and led to InO resistance. The strongest effect was observed when the EBF1 motif was blocked directly in target region 2 (Figure 4D) with the sgRNA2-transfected Nalm6 and REH cells showing a 50-fold and 49-fold decrease, respectively, in cell surface CD22 expression, and a 22-fold and 25.7-fold increase, respectively, in the InO LC50. Across the different EBF1-targeting sgRNA combinations, there was a significant correlation between the degree of disruption in EBF1-mediated CD22 expression and InO LC50 in vitro (Pearson correlation coefficient, −0.95; P < .0001; supplemental Figure 6), further emphasizing the relevance of CD22 in InO-induced leukemia cell death. In addition, we also directly modified these 3 putative EBF1 binding sites at the CD22 locus using CRISPR/Cas9 editing. As shown in supplemental Figure 7, disrupting these motifs led to a similar pattern of CD22 loss and InO resistance with the strongest effect observed with the guide targeting the EBF1 motif in region 2. As expected, the loss of EBF1 binding did not affect calicheamicin sensitivity (supplemental Figures 7C and 8).

Figure 4.

Figure 4.

Open in a new tab

EBF1 transcriptional activity regulated CD22 expression and InO sensitivity in vitro. (A) Differential TF footprinting between patients with InO-sensitive (n = 10) and InO-resistant (n = 8) B-ALL. A positive score indicates higher predicted binding in sensitive cases, and a negative score indicates higher predicted binding in resistant cases. EBF1 binding was signified by the motif MA0154. (B) Integrative Genomics Viewer (IGV) image of the CD22 intronic enhancer and promotor region. The combined ATAC-seq trace (green) of B-ALL samples (n = 99) show the regions of open chromatin status. The correlation between the ATAC-seq signal and CD22 expression is reflected in the top bar. EBF1 ChIPseq (GSE126300, blue trace) was used to identify the EBF1 binding sites. (C-E) The impact of EBF1 binding to its motifs in the intronic enhancer region on CD22 expression was validated by CRISPRi assay for target region 1 (C), target region 2 (D), and target region 3 (E). In total, 9 sgRNAs with 3 for each target region were designed and transfected into Nalm6 or REH cells that constitutively expressed dCas9-KRAB. Cell surface CD22 expression (left) was measured by flow cytometry using CD22-PE antibody in sgRNA-transfected (sgRNA RX.1, RX.1+2, RX.2+3, and RX.1+2+3) or untransfected cells. InO-induced cytotoxicity in Nalm6 (middle) and REH (right) B-ALL cells were measured after 48 hours of drug treatment. Each dot represents the mean ± SD of 3 independent experiments. Dose-response curves were fitted using a 4-parameter nonlinear regression model. Contr, control; CRISPRi, CRISPR interference; EBF1 ChIPseq, EBF1 chromatin immunoprecipitation sequencing; FI, fluorescence intensity.

Intrasubtype heterogeneity of EBF1 and CD22 drives variation in InO response within BCR::ABL1 ALL

Although CD22 expression differs by ALL subtype (Figure 1D),20 there was also significant intrasubtype variability. In particular, BCR::ABL1 ALL exhibited the greatest variation in CD22 expression (Figure 1D). Therefore, we sought to explore possible genomic abnormalities that drive this heterogeneity. Uniform manifold approximation and projection analysis of the gene expression profiles of 132 BCR::ABL1 ALL cases identified 2 distinct clusters (Figure 5A; supplemental Figure 9A), corresponding to cases with multilineage (BCR::ABL1-M) or lymphoid only (BCR::ABL1-L) involvement.27,28 BCR::ABL1-M (39%) was characterized by upregulation of the myeloid markers CD13 and CD33, whereas BCR::ABL1-L (61%) was associated with upregulation of the B-cell markers CD19 and CD20 (supplemental Figure 9B), as well as CD22 and its TF EBF1 (Figure 5B-C). BCR::ABL1-M cases were significantly more resistant to InO ex vivo (P = 1.9 × 10–4; Figure 5D) than BCR::ABL1-L cases, consistent with their differences in EBF1 and CD22 expression. Interestingly, when EBF1, CD22, and the BCR::ABL1 subtype (M or L subcluster) were included in a multivariate regression model, only EBF1 expression was significantly associated with the InO LC50 (P = .019), signifying its role as the main regulator of InO sensitivity in BCR::ABL1 ALL. Using whole-genome sequencing or whole-exome sequencing in combination with single-nucleotide polymorphism microarrays, we further profiled the somatic genomic alterations (Figure 5E). High frequencies of CDKN2A/B deletions (63% vs 5%; P = 1.3 × 10–8) and PAX5 alterations (29% vs 5%; P = .0052) were seen in BCR::ABL1-L cases, whereas EBF1 deletion was more frequently found in BCR::ABL1-M cases (27% vs 6%; P = .004), in line with previous reports.28,29 Taken together, these data imply that the variability in the InO response among BCR::ABL1 ALL cases is, in part, governed by the CD22/EBF1 axis.

Figure 5.

Figure 5.

Open in a new tab

Heterogeneity in BCR::ABL1+ B-ALL was associated with InO sensitivity ex vivo and in vivo. (A) UMAP analysis of leukemia gene expression profiles in BCR::ABL1+ B-ALL identified 2 distinct clusters of patients, namely BCR::ABL1-M (red) or BCR::ABL1-L (blue). (B-D) Comparison of CD22 expression (B), EBF1 expression (C), and ex vivo InO sensitivity (D) of primary leukemia blasts between BCR::ABL1-M (red) and BCR::ABL1-L (blue) B-ALL. Differences between the 2 groups were assessed using Mann-Whitney U tests. (E) Somatic genomic abnormalities in BCR::ABL1+ B-ALL. Each column represents 1 sample, grouped by their subtype (BCR::ABL1-M or -L) identity. The BCR::ABL1 junction type and karyotype anomalies are indicated (top) and somatic genomic alterations of frequently mutated genes are shown (bottom). (F) The schematic plot shows the design of the InO in vivo treatment of patient-derived xerograph (PDX) models. PDX cells derived from SJAALL062194 (BCR::ABL1-L, blue) or SJALL048708 (BCR::ABL1-M, red) were injected into NSG mice by tail vein injection and InO or vehicle treatment was started once the peripheral blast percentage reached >10%. InO was administrated at 16.7 μg/kg body weight by intraperitoneal injection once weekly for 3 weeks. (G) Human leukemia PB blast (hCD45+/hCD19+) was plotted as a function of time after injection (BCR::ABL1-L, left; and BCR::ABL1-M, middle). The red shade indicates the treatment period, and black arrows indicate InO injections, dotted line indicates vehicle treatment, solid line indicates InO treatment. (H) Kaplan-Meier survival curves were plotted for PDX mice. IP, intraperitoneal; UMAP, uniform manifold approximation and projection.

To validate these findings in vivo, we evaluated the InO antileukemic efficacy using 2 BCR::ABL1-rearranged, patient-derived xenograft models. SJALL048708 (model 1) was derived from a male adult patient with BCR::ABL1-M B-ALL. This case was a carrier of a homozygous EBF1 deletion, expressed low levels of cell surface CD22, and was resistant to InO ex vivo (supplemental Figure 10). SJALL062194 (model 2) was derived from a male adult patient with BCR::ABL1-L B-ALL who had high cell surface CD22 expression, no EBF1 alteration, and ex vivo sensitivity to InO. Patient-derived xenograft mice were treated with InO (16.7 μg/kg body weight) or vehicle by intraperitoneal injection once a week for 3 weeks. The treatment was initiated once the human leukemia blast percentage in the peripheral blood (PB; PB blasts) reached >10% and was monitored weekly using flow cytometry (Figure 5F). SJALL048708 (BCR::ABL1-M) was nonresponsive to InO treatment with no impact on the PB blast burden or leukemia-free survival when compared with vehicle-treated mice (Figure 5G-H). In contrast, SJALL062194 (BCR::ABL1-L) showed rapid response to InO with elimination of the PB blasts after 3 InO injections. By the time vehicle-treated mice reached the humane end point (>75% PB blasts), BCR::ABL1-L animals in the InO arm remained leukemia free (<0.1% PB blasts; P < .0001; t test; Figure 5G) with marked increase in survival relative to the vehicle group (P < .0001; Figure 5H). In summary, InO has superior in vivo antileukemia efficacy in BCR::ABL1-L than in BCR::ABL1-M ALL, pointing to the direct impact of EBF1 activity and CD22 expression on the in vivo InO response in BCR::ABL1 B-ALL.

Discussion

The CD22 level on ALL blasts dictates the amount of chemotherapeutic payload delivered to the target leukemia cells and hence response to InO. In fact, CD22 loss caused by somatic mutation or alternative splicing is documented in ALL, leading to acquired resistance to this drug.8,9 However, there is significant interpatient variability in CD22 expression and the degree to which this contributes to variability in ALL response to InO remains unclear.14 Results from the adult INO-VATE trial showed that patients with both high (>90%) and low (<90%) CD22 expression demonstrated similar responses to InO.3 In the European pediatric ITCC-059 trial, CD22 expression levels were similar between patients with positive and those with negative minimum residual disease after InO treatment.5 In contrast, the Children’s Oncology Group trial AALL1621 showed that patients with CD22 expression below 90% consistently had positive post-InO minimum residual disease,6 which is in line with the ex vivo InO sensitivity results described in this study. Several reasons can be hypothesized to explain these differences. First, the in vivo response in patients not only reflects leukemia drug sensitivity but is also strongly influenced by pharmacokinetics. It is, therefore, easier to detect the effects of leukemic intrinsic factors, such as CD22, in the ex vivo setting. Second, the composition of ALL subtypes differs greatly among different studies, for example, relapsed and newly diagnosed diseases or adults and children.3,5,6,30 Because leukemia subtype generally affects drug response,31 the impact of CD22 on InO sensitivity can vary depending on the context of ALL subtype. Nonetheless, key questions regarding the role of CD22 in InO therapy remain unanswered. What is the minimum CD22 expression required for sufficient InO-induced leukemia cell death? What is the impact of CD22 relative to other genetic determinants of InO sensitivity?

CD22 expression is low in the early B-cell developmental stages, rises dramatically at the pro-B stage, and continues to increase as B cells mature.17,18 EBF1 is a critical driver of early B-cell differentiation and has previously been linked to the dynamic change in CD22 expression, consistent with the results presented here.32 Because CD22 expression is closely related to B-cell differentiation, we hypothesized that CD22 is also regulated by TFs other than EBF1. To investigate this possibility, we analyzed the correlation between the expression of B-cell–related TF genes and CD22 expression in B-ALL (supplemental Table 10). In addition to EBF1, PAX5, LEF1, and TCF3 all exhibited strong positive correlations with CD22 expression. However, the knockout of PAX5 had only moderate effects on CD22 surface expression in the TF CRISPR screen (log2FC = −0.33; P = .03, supplemental Tables 5 and 10), and the inferred PAX5 binding activity (using ATAC-seq) was only marginally different between the InO-sensitive and -resistant cases (log2FC = 0.04; P = 7.9 × 10–69; supplemental Tables 7 and 10). For other B-cell TFs, such as IKZF1, no clear evidence was found for its role in CD22 expression regulation and InO sensitivity (supplemental Table 10). That said, it is most likely that CD22 modulation is dependent on a multitude of factors, although EBF1 emerged as the strongest modulator in our laboratory model systems and in our primary ALL cohorts (supplemental Table 10).

Recently, we and others reported that DNA nucleotidylexotransferase (DNTT) is a new determinant of InO sensitivity in B-ALL.9,22 Similar to EBF1 and CD22, DNTT expression also varies during B-cell differentiation.20 As shown in supplemental Figure 11, patients with low expression of both DNTT and EBF1 were most resistant to InO, suggesting that the 2 nonoverlapping mechanisms had additive effects; EBF1 influences InO sensitivity via CD22, whereas DNTT causes resistance by altering the InO-induced DNA damage response.

B-ALL subtypes are closely linked with the B-cell differentiation state.20 In line with the association between CD22 expression and B-cell differentiation, B-ALL molecular subtypes exhibit distinct sensitivity to InO.22 For example, KMT2A ALL, characterized by leukemic arrest at the pre–pro-B stage,20 showed relatively inferior response to InO in clinical trials.15,24 However, molecular subtype is an imprecise predictor of InO response because there is substantial heterogeneity in this phenotype for nearly every single ALL subtype. As shown in Figure 5, B-cell differentiation state and EBF1/CD22 activity can lead to distinct InO response within BCR::ABL1 ALL, suggesting that molecular subtype–based stratification is insufficient for directing InO therapy. In contrast, somatic focal deletion is recurrent in B-ALL33 and can be a genomic predictor of InO resistance. Exploring this in the BCR::ABL1 ALL subset of our pharmacotyping cohort, we observed a trend that cases with an EBF1 deletion (all hemizygous) had a higher LC50 than those with wild-type EBF1 (supplemental Figure 12), but this did not reach statistical significance, probably because of the small sample size. It is also possible that ALL with homozygous loss of EBF1 would exhibit a greater degree of InO resistance, which warrant further evaluation in future studies.

In summary, our study demonstrated the direct impact of EBF1 on CD22 expression during B-cell development. We highlight the direct impact of EBF1-mediated transcriptional regulation of CD22 on InO response ex vivo and in vivo. Finally, the variability in the EBF1/CD22 axis activity explains interleukemia variability in InO sensitivity across and within ALL molecular subtypes, potentially informing the individualization of InO therapy for this cancer.

Conflict-of-interest disclosure: M.Y.K. reports receiving grants from AbbVie, Allogene, AstraZeneca, Cellectis, Daiichi Sankyo, Forty Seven, Genentech, Gilead, MEI Pharma, Precision Bio, Rafael Pharmaceutical, Sanofi, and Stemline-Menarini; receiving personal fees from AbbVie, AstraZeneca, Auxenion, Genentech, Gilead, F. Hoffman-La Roche, Janssen, MEI Pharma, Sellas, and Stemline-Menarini; reports having patent US 7,795,305 B2 “CDDO-compounds and combination therapies” with royalties paid to Reata Pharmaceuticals, the patent “Combination Therapy with a mutant IDH1 Inhibitor and a BCL-2” licensed to Eli Lilly, and patent 62/993,166 “Combination of a MCL-1 inhibitor and Midostaurin, uses and pharmaceutical compositions thereof” to Novartis. H.I. reports receiving grants from Servier, Incyte, and Amgen and personal fees from Servier and Jazz. J.J.Y. reports receiving research funding from Takeda Pharmaceutical and AstraZeneca Plc. None of these relationships were related to this work. The remaining authors declare no competing financial interests.

Acknowledgments

The authors thank the patients and families who participated in the clinical trials included in this study. The authors also thank the Pharmacotyping Resource in the Department of Pharmacy and Pharmaceutical Sciences for performing the drug sensitivity profiling, the Flow Cytometry and Cell Sorting Shared Resource, and the Animal Resources Center at the St. Jude Children’s Research Hospital for their assistance in animal studies. The authors thank C. Li at the St. Jude Children’s Research Hospital for sharing the transcription factor CRISPR library.

This study was supported by the National Institutes of Health (NIH), National Cancer Institute (NCI) (grants R01 CA264837 and U01 CA264610 [J.J.Y.], R01 CA234490 [D.S.]), the NIH, National Institute of General Medical Sciences (NIGMS) (grant R01GM134382 [J.J.Y.]), the NIH, NCI (grants U10CA180820, UG1CA189859, and UG1CA232760 [ECOG-ACRIN clinical trial]), the St. Jude Children’s Research Hospital Cancer Center Support Grant (P30 CA021765 [J.J.Y.]), the NIH, NIGMS (grant P50 GM115279 [J.J.Y., W.E.E.]), and the German Research Foundation grant (519605261 [C.S.E.]). J.J.Y. is the Endowed Chair of Pharmacogenomics at St. Jude Children's Reseach Hospital. C.-H.P. is the Fahad Nassar Al-Rashid Chair of Leukemia Research. M.L.L. is the Aldarra Foundation, June and Bill Boeing Founders, Endowed Chair of Pediatric Oncology.

Authorship

Contributions: J.J.Y., D.S., and W.E.E. conceptualized the study; C.S.E., Z.L., X.H., and W.E.E. developed the methodology; C.S.E., Z.L., K.R.B., W.Y., and X.H. were responsible for the formal analysis; C.S.E., M.W., and S.Y. conducted the investigation; Y.L., J.Y., W.S., E.P., M.Y.K., S.M.K., E.J., M.R.L., H.I., C.-H.P., M.L.L., W.E.E., and D.S. provided resources; C.S.E., Z.L., and J.J.Y. wrote the original draft; C.S.E. and Z.L. were responsible for visualization; J.J.Y. supervised the study; and all authors reviewed and edited the manuscript.

Footnotes

∗

C.S.E. and Z.L. contributed equally to this study.

Presented in abstract form at the 66th annual meeting of the American Society of Hematology, San Diego, CA, 9 December 2024.

RNA-sequencing data are available at the European Genome-phenome Archive (https://ega-archive.org/; accession numbers EGAS00001003266, EGAS00001004810, and EGAS00001007473) or at St. Jude Cloud Genomics Platform under the Real-time Clinical Genomics data set (https://platform.stjude.cloud/data/cohorts; accession number, SJC-DS-1007). Assay for transposase-accessible chromatin with sequencing data of primary human B-cell acute lymphoblastic leukemia samples are publicly available in the Gene Expression Omnibus (GEO) database (accession number GSE211631). Early B-cell factor 1 binding using chromatin immunoprecipitation sequencing data used in this study are publicly available in the GEO database (accession number GSE126300).

The online version of this article contains a data supplement.

There is a Blood Commentary on this article in this issue.

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

Supplementary Material

Supplemental Methods, References, Table Legends, and Figures
BLOOD_BLD-2024-028215-mmc1.pdf (7.5MB, pdf)
Supplemental Tables 1-10
BLOOD_BLD-2024-028215-mmc2.zip (17.6MB, zip)

References

  • 1.DiJoseph JF, Popplewell A, Tickle S, et al. Antibody-targeted chemotherapy of B-cell lymphoma using calicheamicin conjugated to murine or humanized antibody against CD22. Cancer Immunol Immunother. 2005;54(1):11–24. doi: 10.1007/s00262-004-0572-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Dijoseph JF, Dougher MM, Armellino DC, Evans DY, Damle NK. Therapeutic potential of CD22-specific antibody-targeted chemotherapy using inotuzumab ozogamicin (CMC-544) for the treatment of acute lymphoblastic leukemia. Leukemia. 2007;21(11):2240–2245. doi: 10.1038/sj.leu.2404866. [DOI] [PubMed] [Google Scholar]
  • 3.Kantarjian HM, DeAngelo DJ, Stelljes M, et al. Inotuzumab ozogamicin versus standard therapy for acute lymphoblastic leukemia. N Engl J Med. 2016;375(8):740–753. doi: 10.1056/NEJMoa1509277. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Jabbour E, Sasaki K, Ravandi F, et al. Chemoimmunotherapy with inotuzumab ozogamicin combined with mini-hyper-CVD, with or without blinatumomab, is highly effective in patients with Philadelphia chromosome-negative acute lymphoblastic leukemia in first salvage. Cancer. 2018;124(20):4044–4055. doi: 10.1002/cncr.31720. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Pennesi E, Michels N, Brivio E, et al. Inotuzumab ozogamicin as single agent in pediatric patients with relapsed and refractory acute lymphoblastic leukemia: results from a phase II trial. Leukemia. 2022;36(6):1516–1524. doi: 10.1038/s41375-022-01576-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.O'Brien MM, Ji L, Shah NN, et al. Phase II trial of inotuzumab ozogamicin in children and adolescents with relapsed or refractory B-cell acute lymphoblastic leukemia: Children's Oncology Group protocol AALL1621. J Clin Oncol. 2022;40(9):956–967. doi: 10.1200/JCO.21.01693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.DiJoseph JF, Armellino DC, Boghaert ER, et al. Antibody-targeted chemotherapy with CMC-544: a CD22-targeted immunoconjugate of calicheamicin for the treatment of B-lymphoid malignancies. Blood. 2004;103(5):1807–1814. doi: 10.1182/blood-2003-07-2466. [DOI] [PubMed] [Google Scholar]
  • 8.Zheng S, Gillespie E, Naqvi AS, et al. Modulation of CD22 protein expression in childhood leukemia by pervasive splicing aberrations: implications for CD22-directed immunotherapies. Blood Cancer Discov. 2022;3(2):103–115. doi: 10.1158/2643-3230.BCD-21-0087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9.Zhao Y, Short NJ, Kantarjian HM, et al. Genomic determinants of response and resistance to inotuzumab ozogamicin in B-cell ALL. Blood. 2024;144(1):61–73. doi: 10.1182/blood.2024023930. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Reinert J, Beitzen-Heineke A, Wethmar K, Stelljes M, Fiedler W, Schwartz S. Loss of CD22 expression and expansion of a CD22dim subpopulation in adults with relapsed/refractory B-lymphoblastic leukaemia after treatment with inotuzumab-ozogamicin. Ann Hematol. 2021;100(11):2727–2732. doi: 10.1007/s00277-021-04601-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Fry TJ, Shah NN, Orentas RJ, et al. CD22-targeted CAR T cells induce remission in B-ALL that is naive or resistant to CD19-targeted CAR immunotherapy. Nat Med. 2018;24(1):20–28. doi: 10.1038/nm.4441. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Wintering A, Ishiyama K, Tamaki S, et al. CD22low/Bcl-2high expression identifies poor response to inotuzumab ozogamicin in relapsed/refractory acute lymphoblastic leukemia. Blood Adv. 2023;7(2):251–255. doi: 10.1182/bloodadvances.2021006810. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Sartor C, Arpinati M, Chirumbolo G, et al. Baseline cluster of differentiation 22 fluorescent intensity correlates with patient outcome after inotuzumab ozogamicin treatment. Hematol Oncol. 2022;40(4):734–742. doi: 10.1002/hon.3029. [DOI] [PubMed] [Google Scholar]
  • 14.Rubinstein JD, O'Brien MM. Inotuzumab ozogamicin in B-cell precursor acute lymphoblastic leukemia: efficacy, toxicity, and practical considerations. Front Immunol. 2023;14 doi: 10.3389/fimmu.2023.1237738. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Brivio E, Locatelli F, Lopez-Yurda M, et al. A phase 1 study of inotuzumab ozogamicin in pediatric relapsed/refractory acute lymphoblastic leukemia (ITCC-059 study) Blood. 2021;137(12):1582–1590. doi: 10.1182/blood.2020007848. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Nitschke L, Carsetti R, Ocker B, Köhler G, Lamers MC. CD22 is a negative regulator of B-cell receptor signalling. Curr Biol. 1997;7(2):133–143. doi: 10.1016/s0960-9822(06)00057-1. [DOI] [PubMed] [Google Scholar]
  • 17.Stoddart A, Ray RJ, Paige CJ. Analysis of murine CD22 during B cell development: CD22 is expressed on B cell progenitors prior to IgM. Int Immunol. 1997;9(10):1571–1579. doi: 10.1093/intimm/9.10.1571. [DOI] [PubMed] [Google Scholar]
  • 18.Raponi S, De Propris MS, Intoppa S, et al. Flow cytometric study of potential target antigens (CD19, CD20, CD22, CD33) for antibody-based immunotherapy in acute lymphoblastic leukemia: analysis of 552 cases. Leuk Lymphoma. 2011;52(6):1098–1107. doi: 10.3109/10428194.2011.559668. [DOI] [PubMed] [Google Scholar]
  • 19.Shah NN, Stevenson MS, Yuan CM, et al. Characterization of CD22 expression in acute lymphoblastic leukemia. Pediatr Blood Cancer. 2015;62(6):964–969. doi: 10.1002/pbc.25410. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Huang X, Li Y, Zhang J, et al. Single-cell systems pharmacology identifies development-driven drug response and combination therapy in B cell acute lymphoblastic leukemia. Cancer Cell. 2024;42(4):552–567.e6. doi: 10.1016/j.ccell.2024.03.003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Gu Z, Churchman ML, Roberts KG, et al. PAX5-driven subtypes of B-progenitor acute lymphoblastic leukemia. Nat Genet. 2019;51(2):296–307. doi: 10.1038/s41588-018-0315-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Escherich CS, Moriyama T, Li Z, et al. DNTT-mediated DNA damage response drives inotuzumab ozogamicin resistance in B-cell acute lymphoblastic leukemia. Blood. 2025;145(11):1182–1194. doi: 10.1182/blood.2024026085. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Rowland L, Smart B, Brown A, et al. Ex vivo drug sensitivity imaging-based platform for primary acute lymphoblastic leukemia cells. Bio Protoc. 2023;13(15) doi: 10.21769/BioProtoc.4731. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Kantarjian HM, Stock W, Cassaday RD, et al. Inotuzumab ozogamicin for relapsed/refractory acute lymphoblastic leukemia in the INO-VATE trial: CD22 pharmacodynamics, efficacy, and safety by baseline CD22. Clin Cancer Res. 2021;27(10):2742–2754. doi: 10.1158/1078-0432.CCR-20-2399. [DOI] [PubMed] [Google Scholar]
  • 25.Zhang H, Zhang Y, Zhou X, et al. Functional interrogation of HOXA9 regulome in MLLr leukemia via reporter-based CRISPR/Cas9 screen. Elife. 2020;9 doi: 10.7554/eLife.57858. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Bentsen M, Goymann P, Schultheis H, et al. ATAC-seq footprinting unravels kinetics of transcription factor binding during zygotic genome activation. Nat Commun. 2020;11(1):4267. doi: 10.1038/s41467-020-18035-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Arber DA, Orazi A, Hasserjian RP, et al. International Consensus Classification of myeloid neoplasms and acute leukemias: integrating morphologic, clinical, and genomic data. Blood. 2022;140(11):1200–1228. doi: 10.1182/blood.2022015850. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Bastian L, Beder T, Barz MJ, et al. Developmental trajectories and cooperating genomic events define molecular subtypes of BCR::ABL1-positive ALL. Blood. 2024;143(14):1391–1398. doi: 10.1182/blood.2023021752. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Kim JC, Chan-Seng-Yue M, Ge S, et al. Transcriptomic classes of BCR-ABL1 lymphoblastic leukemia. Nat Genet. 2023;55(7):1186–1197. doi: 10.1038/s41588-023-01429-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30.Stelljes M, Raffel S, Alakel N, et al. Inotuzumab ozogamicin as induction therapy for patients older than 55 years with Philadelphia chromosome-negative B-precursor ALL. J Clin Oncol. 2024;42(3):273–282. doi: 10.1200/JCO.23.00546. [DOI] [PubMed] [Google Scholar]
  • 31.Lee SHR, Yang W, Gocho Y, et al. Pharmacotypes across the genomic landscape of pediatric acute lymphoblastic leukemia and impact on treatment response. Nat Med. 2023;29(1):170–179. doi: 10.1038/s41591-022-02112-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Ng AP, Coughlan HD, Hediyeh-Zadeh S, et al. An Erg-driven transcriptional program controls B cell lymphopoiesis. Nat Commun. 2020;11(1):3013. doi: 10.1038/s41467-020-16828-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Brady SW, Roberts KG, Gu Z, et al. The genomic landscape of pediatric acute lymphoblastic leukemia. Nat Genet. 2022;54(9):1376–1389. doi: 10.1038/s41588-022-01159-z. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Methods, References, Table Legends, and Figures
BLOOD_BLD-2024-028215-mmc1.pdf (7.5MB, pdf)
Supplemental Tables 1-10
BLOOD_BLD-2024-028215-mmc2.zip (17.6MB, zip)

Articles from Blood are provided here courtesy of The American Society of Hematology

RESOURCES