|Home | About | Journals | Submit | Contact Us | Français|
B-Cell CLL/Lymphoma 6 (BCL6) is a proto-oncogene that is highly expressed in acute lymphoblastic leukemia (ALL). BTB and CNC Homology 1 Basic Leucine Zipper Transcription Factor 2 (BACH2) is a suppressor of transcription. The BACH2–BCL6 balance controls selection at the pre-B cell receptor checkpoint by regulating p53 expression. However, the underlying mechanism and the clinical relevance of the BCL6/BACH2 axis are unknown. Here, we found that Ikaros, a tumor suppressor encoded by IKZF1, directly binds to both the BCL6 and BACH2 promoters where it suppresses BCL6 and promotes BACH2 expression in B-cell ALL (B-ALL) cells. Casein kinase 2 (CK2) inhibitors increase Ikaros function thereby inhibiting BCL6 and promoting BACH2 expression in an Ikaros-dependent manner. We also found that the expression of BCL6 is higher while BACH2 expression is lower in patients with B-ALL than normal bone marrow control. High BCL6 and low BACH2 expression is associated with high leukemic cell proliferation, unfavorable clinical and laboratory features, and inferior outcomes. Moreover, IKZF1 deletion is associated with high BCL6 and low BACH2 expression in B-ALL patients. CK2 inhibitors increase Ikaros binding to the promoter of BCL6 and BACH2 and suppress BCL6 while promoting BACH2 expression in the primary B-ALL cells. Our data indicates that Ikaros regulates expression of the BCL6/BACH2 axis in B-ALL. High BCL6 and low BACH2 expression are associated with Ikaros dysregulation and have a potential effect on the development of B-ALL.
Acute lymphoblastic leukemia (ALL) is one of the most common hematologic malignancies. The five-year survival rate has steadily increased over the last fifty years to 85% in the recent diagnosis period. However, up to 20% of patients with ALL relapse and these relapsed patients have poor clinical outcomes in childhood and, more frequently, as adults [1–4]. ALL remains a major cause of death in childhood and causes significant adult mortality. Thus, understanding the mechanisms underlying its leukemogenesis and developing more effective therapeutic approaches could be very helpful for reducing the cancer death toll in both children and adults.
B-Cell CLL/Lymphoma 6 (BCL6) is a proto-oncogene that has been classically described in the setting of its influence on germinal center (GC) B cells. Additionally, BCL6 is a regulator of B cell proliferation, maturation, and resistance to DNA damage [5–17]. More recent work has highlighted the impact of BCL6 on immature and malignant hematopoietic cells . Increased expression of BCL6 in chronic myelogenous leukemia (CML) and ALL has been shown to protect leukemic cells from chemotherapy-induced DNA damage through the repression of p53-induced apoptosis [19–20]. In ALL cells, increased expression of BCL6 results in a tolerance to DNA damage which subsequently increases survival during BCR-ABL1 kinase inhibition .
Another B-lymphoid transcription factor, BTB and CNC Homology 1 Basic Leucine Zipper Transcription Factor 2 (BACH2) also plays critical roles in GC formation after antigen encounter  and class switch recombination in B cells [22–24]. BACH2 is widely characterized as a repressor of transcription although it can activate transcription at selected loci . Deregulated BACH2 expression is associated with lymphoid malignancies. Loss of heterozygosity of BACH2 occurs at a frequency of 20% in human B-cell lymphomas . Deletions of 6q15 that include the BACH2 locus appear in 30% of pre-B ALL cases . In several types of leukemia and lymphoma, disruption of wild type BACH2 expression is attributed to viral integrations [28–31].
Importantly, BCL6 and BACH2 show antagonism during early B cell development, as well as in repertoire selection and counter-selection of premalignant clones for leukemia suppression. BCL6-BACH2 balance regulates selection at the pre-B cell receptor checkpoint by regulating p53 expression . However, the underlying mechanism and the clinical relevance of BCL6/BACH2 axis expression are poorly determined in B-ALL patients.
IKZF1 encodes a kruppel-like zinc finger protein, Ikaros, that is essential for normal hematopoiesis and acts as a tumor suppressor in ALL. The impairment of Ikaros function, as a result of deletion and/or an inactivating mutation of a single IKZF1 allele, is linked to the development of ALL that is characterized by a high rate of relapse and poor outcome. Recently, we first reported that CK2 inhibition could restore Ikaros function in B-ALL cells [15, 16]. CK2 inhibitors function as Ikaros activators [33–36]. We identified Ikaros’ binding profile in B-ALL cells  and demonstrated that Ikaros exerts its antitumor effect by regulating the expression of its target genes . We also reported that CK2 inhibitors restore Ikaros function by increasing Ikaros binding to gene targets and regulating the expression of Ikaros targets in B-ALL cells [33, 34].
We reported the global Ikaros binding profile in ALL , and found the apparent binding peaks in promoter regions of BCL6 and BACH2 in B-ALL patients using ChIP-seq data. Here, we further observed how expression of BCL6 and BACH2 correlates with clinical features and with Ikaros dysfunctions in adult B-ALL. We found high BCL6 expression and/or low BACH2 expression is associated with leukemic cell proliferation, poor overall survival (OS), and poor event-free survival (EFS). We also found that Ikaros directly suppresses BCL6 and activates BACH2 expression, and that IKZF1 deletion is associated with significantly higher BCL6 and lower BACH2 expression in the patients. Our results indicate that Ikaros directly suppresses BCL6 but promotes BACH2 expression in B-ALL patients, and that patients with BCL6highBACH2low expression also have dysfunctional Ikaros and represent a subset of high-risk B-ALL.
We assessed BCL6 and BACH2 mRNA expression in 79 newly diagnosed adult B-ALL patients. We found that, compared to the normal bone marrow controls, expression of BCL6 is significantly higher (Figure (Figure1A)1A) and BACH2 is significantly lower (Figure (Figure1B)1B) in B-ALL patients. We also observed the expression of high BCL6 and low BACH2 through a reported microarray expression cohort of ALL patients (Supplementary Figure 1 and 2). These data suggest that the patients with both BCL6 high and BACH2 low expression (BCL6highBACH2low) characterize a novel subset of B-ALL.
This cohort of patients was divided into high and low BCL6 or BACH2 expression groups (Quartiles 3-4 vs Quartiles 1-2, respectively) and their expression correlated with clinical features in adult B-ALL (Supplementary Tables 1 and 2). The percentage of the patients with age ≥35 years old, an important prognostic factor of poor outcome in ALL, was significantly higher in the BCL6high compared to the BCL6low subgroup (73.0% vs. 38.1%, P=0.002). This association was confirmed by multivariate analyses (HR 2.949, 95% confidence interval [CI], 1.042, 8.344; P=0.042; Supplementary Table 1). The BCL6high cohort also had a higher frequency of IKZF1 deletion and BCR/ABL1 fusion compared to BCL6low cohort (45.8% vs. 13.9%, P=0.006; 67.6% vs. 39.0%, P=0.012) as determined using chi-squared tests. The association with positive BCR/ABL1 fusion is confirmed in multivariate analyses (HR 2.826, [1.003, 7.966]; P=0.049; Supplementary Table 1).
BACH2Low mRNA levels were associated with both higher WBC (median) and WBC ≥30×10E+9/L (%) compared to BACH2high expression (61.0 vs. 24.0, P<0.0001; 72.7% vs. 38.6%; P=0.003). These associations were confirmed in multivariate analyses (HR 0.994, [0.988, 1.000], P=0.035; HR 0.320, [0.105, 0.972], P=0.045, Supplementary Table 2). The median number of peripheral blood blasts is significantly higher in patients with expression of BACH2low than that of BACH2high (75.5% vs. 56.0%, P=0.025) as determined using chi-squared tests. The BACH2low cohort also had a higher frequency of IKZF1 deletion and BCR/ABL1 fusion compared to the BACH2high cohort (43.3% vs. 13.3%, P=0.003; 69.7% vs. 40.0%, P=0.009) by chi-squared tests. The association with IKZF1 deletion is further confirmed in multivariate analyses (HR 0.256, [0.077, 0.850], P=0.026; Supplementary Table 2).
The patients with both high BCL6 and low BACH2 (BCL6highBACH2low) mRNA levels were associated with having a higher WBC (median) and WBC ≥30×10E+9/L (%) compared to BCL6lowBACH2high expression (115.0 vs. 14.0, P=0.003; 90.9% vs. 27.8%; P=0.002; Supplementary Table 3). The association with WBC ≥30×10E+9/L (%) was also significant with multivariate analyses (HR 26.0, [2.607, 259.29], P=0.005; Supplementary Table 3). The median number of bone marrow blasts is significantly higher in patients with BCL6highBACH2low than BCL6lowBACH2high expression (91.0% vs. 84.0%, P=0.044) as determined using chi-squared tests. The BCL6highBACH2low expression cohort also had a higher frequency of IKZF1 deletion and BCR/ABL1 fusion compared to BCL6lowBACH2high expression cohort (60.0% vs. 15.8%, P=0.032; 81.8% vs. 10.5%, P<0.0001) as determined using chi-squared tests. The association with positive BCR/ABL1 fusion was also confirmed in multivariate analyses (HR 0.042, [0.002, 0.715]; P=0.028; Supplementary Table 3).
Subjects with expression of BCL6high had a shorter median OS and EFS than those with BCL6low (16.0 months [10.735, 21.265] vs. 35.5 months [27.982, 43.018], P=0.046, Figure Figure2A;2A; 24.0 months [10.681, 37.319] vs. 43.0 months [27.674, 58.326], P=0.052, Figure Figure2B).2B). Subjects with BACH2low expression had shorter median OS and EFS than those with BACH2high expression (16.0 months [10.934, 21.066] vs. 35.5 months [22.711, 48.289], P=0.050, Figure Figure2C;2C; 12.0 months [6.615, 17.385] vs. 17.0 months [6.536, 27.464], P=0.191, Figure Figure2D).2D). Although no statistical significance was achieved, we observed the trend of shorter EFS in patients with BCL6high and shorter OS and EFS in those with BACH2low. Furthermore, we did observe the significantly shorter median OS and EFS in patients with BCL6highBACH2low than BCL6lowBACH2high expression cohort (13.0 months [3.952, 22.048] vs. 33.0 months [13.696, 52.304], P<0.0001, Figure Figure2E;2E; 11.1 months [0.394, 21.806] vs. 16.0 months [9.725, 22.275], P=0.015, Figure Figure2F).2F). These data indicate that patients with BCL6highBACH2low expression could be defined as a subgroup with much worse outcomes.
Our recent ChIP-seq data observed the robust binding peaks of Ikaros on the promoter regions of both BCL6 and BACH2 in Nalm6 B-ALL cells (Figure (Figure3A3A and and3B)3B) as well as in primary B-ALL cells (Supplementary Figure 3). The significant binding of Ikaros to the promoter of BCL6 and BACH2 was confirmed by qChIP in both Nalm6 cells (Figure (Figure3C)3C) and primary cells from ALL patients (Figure (Figure3D).3D). To further demonstrate the direct transcriptional regulation by Ikaros on BCL6 and BACH2, we showed that Ikaros suppresses the promoter activity of BCL6 and activates that of BACH2 by luciferase reporter assay (Figure (Figure4A).4A). The effect of Ikaros over-expression or knock-down on BCL6 and BACK2 expression was further tested in Nalm6 cells. The results showed that Ikaros overexpression suppressed BCL6 and increased BACH2 mRNA levels (Figure (Figure4B).4B). Conversely, Ikaros knockdown induced an increase in BCL6 expression and a decrease in BACH2 expression (Figure (Figure4C).4C). Treatment of Nalm6 cells with CX4945, a CK2 inhibitor that restores Ikaros function, was found to suppress BCL6 expression (Figure (Figure5A)5A) and increase BACH2 expression (Figure (Figure5B)5B) in a dose-dependent manner. CK2 knockdown with shRNA suppresses BCL6 but increases BACH2 mRNA levels (Figure (Figure5C).5C). Importantly, Ikaros knockdown could reverse the CX4945-induced decrease in BCL6 expression and increase in BACH2 expression (Figure (Figure5D).5D). These data indicate that both BCL6 and BACH2 are direct Ikaros targets in B-ALL and that Ikaros transcriptionally regulates their expression.
We found that IKZF1 expression is positively correlated with BACH2 expression but negatively correlated with BCL6 expression in the cohort studies of patients with B-ALL (Supplementary Figure 4). Ikaros 6 (Ik6) is the most frequent type of IKZF1 deletion. We tested for Ik6 deletion in our cohort and analyzed the BCL6/BACH2 expression in patients with and without Ik6. Our data indicated significantly higher BCL6 expression and lower BACH2 expression in patients with (Ik6+) compared to those without Ik6 (Ik6-) (Figure (Figure6A6A and and6B).6B). We also observed that the BCL6highBACH2low had a much higher frequency of Ik6+ cases compared to BCL6lowBACH2high subgroup (60.0 vs. 15.8, P=0.032) as determined using chi-squared tests (Supplementary Table 3). These data further support the regulatory effect of Ikaros on both BCL6 and BACH2 in B-ALL patients; they also reveal that IKZF1 deletion may be responsible for high BCL6 and low BACH2 expression in patients.
Moreover, CX-4945 can increase Ikaros binding to the promoter of BCL6 and BACH2 (Data not shown), where it suppresses BCL6 (Figure (Figure6C)6C) and increases BACH2 expression (Figure (Figure6D)6D) in primary B-ALL. These data not only indicate the effect of CK2 inhibitors on the promotion of Ikaros function, but also suggest that Ikaros-induced changes in expression of the BCL6/BACH2 axis are at least partially responsible for the application of CK2 inhibitors in B-ALL therapy.
The balance between BACH2 and BCL6 controls the pre-B cell receptor checkpoint ; however, its oncogenic effect in leukemia and its clinical relevance are not yet fully clarified. We found that Ikaros directly regulates BCL6/BACH2 expression and their potential oncogenic effect in B-ALL as summarized in Figure Figure77.
Our ChIP-seq data detected robust Ikaros binding peaks on the promoter regions of both BCL6 and BACH2 in B-ALL cell lines and primary B-ALL cells from patients. This encouraged us to test the expression of BCL6 and BACH2 in our cohort of B-ALL patients. Further, BCL6 is highly expressed in ALL and has an oncogenic effect [7, 11, 32]. BACH2 is a tumor suppressor and it is reported to be frequently inactivated in primary pre-B ALL patients via promoter hyper-methylation, missense mutations and deletions, as well as via loss of its upstream regulator paired box 5 (PAX5) [32, 37–39]. However, there are no reports about BCL6/BACH2 expression in leukemia patients. Here, we found that BCL6 is highly expressed but BACH2 is downregulated in patients. Moreover, BCL6high, BACH2low, and particularly BCL6highBACH2low cohorts are associated with poor survival and leukemic cell proliferation. Ikaros deletion is associated with BCL6high and BACH2low expression in the patients. The high-risk markers, Ikaros deletion and BCR/ABL1 fusion, were significantly detected in the patients with BCL6high, BACH2low and BCL6highBACH2low expression. These findings further illuminated the oncogenic effect of the BCL6/BACH2 axis in B-ALL. More importantly, our data revealed a novel subset of high-risk B-ALL characterized by BCL6highBACH2low expression with Ikaros dysfunction and BCL/ABL1 fusion.
It is reported that PAX5 regulates both BCL6 and BACH2 expression in B-cell development [32, 37–39]. BCR/ABL1 is also reported to be involved in the suppression of BACH2 expression in ALL [32, 40]. We found that Ikaros directly binds to the promoter of BCL6/BACH2 and regulates their expression. We demonstrated that BCL6 and BACH2 are direct effectors of Ikaros in B-ALL. CK2 inhibitors suppress BCL6 and promote BACH2 expression in an Ikaros-dependent manner. Our results not only demonstrate a novel mechanism underlying regulation of the BCL6/BACH2 axis, but also indicate that this is one of the mechanisms underlying the therapeutic effect of CK2 inhibitors on high-risk ALL.
Ikaros is a DNA-binding protein with a complex role in transcriptional regulation. Ikaros binds the upstream regulatory regions of its target genes, recruits histone deacetylase complexes, and represses transcription via chromatin remodeling . Recently, genome-wide screening of Ikaros binding sites revealed that Ikaros binding is associated with the global change of histone modification in mouse lymphocyte development . We identified the chromatin state of the Ikaros gene targets in ALL leukemia cells and found that Ikaros suppresses the expression of its targets by recruiting H3K9me3 or H3K27me3 , and up-regulates gene expression by increasing the binding of H3K4me3 via KDM5B in the promoter region of its targets [34, 36, 43]. Therefore, we considered that Ikaros suppresses BCL6 transcription via recruiting H3K9me3 or H3K27me3 to its promoter, and up-regulates BACH2 expression via increasing the H3K4me3 binding on its promoter as we previously reported [34, 36, 43].
We demonstrated that the significant therapeutic efficacy of CK2 inhibitor CX4945 on high-risk leukemia by restoring Ikaros function . Here we identified a subset of high-risk leukemia characterized with high BCL6 and low BACH2 expression and associated with Ikaros dysfunction. In addition, CK2 inhibitor CX-4945 suppresses BCL6 and promotes BACH2 expression in an Ikaros-dependent manner. These data indicate the therapeutic efficacy of CX-4945 on this subset of high-risk ALL. Moreover, targeting BCL6 high expression with BCL6-peptide inhibitor has an anti-proliferative effect in leukemia cells and the BCL6-peptide inhibitor strongly enhances the effect of PKI imatinib on apoptosis of CML cells [44, 45]. Also, specific BCL6 antagonists, including small molecule inhibitors, have been developed, and the BCL6 antagonists are active against primary DLBCL [46, 47]. These data suggest the possibility of combining CX-4945 with BCL6 antagonists in the therapy of this subset of high-risk leukemia.
In summary, Ikaros directly regulates BCL6/BACH2 axis expression. BCL6high and/or BACH2low expression is associated with IKZF1 deletion, BCR/ABL1 fusion, leukemic cell proliferation, and inferior outcomes in B-ALL. All of which could help distinguish a novel subgroup of high-risk B-ALL.
Between June 2008 and December 2015, 79 consecutive subjects with newly-diagnosed B-ALL (age 14-77 years old) were studied at the First Affiliated Hospital of Nanjing Medical University and Zhongda Hospital Southeast University. Diagnoses were based on the WHO Diagnosis and Classification of ALL (2008). The study was approved by the Ethics Committee of the First Affiliated Hospital of Nanjing Medical University, Jiangsu Province Hospital, the Ethics Committee of the Zhongda Hospital Southeast University, Nanjing, China, with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. There were no studies with animals.
Therapy details are published protocol (CALLG2008) . Induction was with VDCLP (vincristine, daunorubicin, cyclophosphamide, L-asparaginase, prednisone). Early consolidation used CAT (cyclophosphamide, cytarabine, thioguanine), high-dose methotrexate/L-asparaginase, and mitoxantrone. Late consolidation used VDLP (vincristine, daunorubicin, L-asparaginase, prednisone), COATD (cyclophosphamide, vincristine, cytarabine, epipodophyllotoxin and dexamethasone), high-dose methotrexate/L-asparaginase, epipodophyllotoxinandcytarabine. Maintenance therapy used 6-mercaptopurine and methotrexate. Subjects with BCR/ABL1-positive ALL received Imatinib from day 15 of introduction therapy.
Cytogenetics and detection of the most common IKZF1 deletion (Ik6) were analyzed as described [35, 43, 48]. qPCR was performed on StepOnePlus Real-time PCR system (Applied Biosystem-Thermofisher, Foster, CA, USA). Gene expression values of genes of interest (GOI) were achieved in each sample by a formula derived from a scatter graph of Ct values from serial dilutions of a template standard as described [35, 43, 48]. Expression levels of GOIs were normalized to housekeeping genes expressed as gene expression value of GOI/18s rRNA. Subjects were allocated in a high or low BCL6/BACH2 expression cohort (3rd-4th quartiles vs. 1st-2nd quartiles) with a cut-off value (0.0118226976019986/0.00749525) was determined by SPSS 20.0 [35, 43, 48].
qPCR for BCL6 and BACH2 expression was analyzed as above in Nalm6 and primary B-ALL cells. Results were normalized to those obtained with 18s rRNA and presented as fold-induction over vector controls. Primers: 18s rRNA, Sense: 5’-GTAACCCGTTGAACCCCATT-3’, Anti-sense: 5’-CCATCCAATCGGTAGTAGCG-3’; BCL6 Sense: 5’-CTGTGATGGCCACGGCTAT-3’, Anti-sense: 5’-TCCGGCAAGTGTCCACAA-3’; BACH2 Sense: 5’-GC GGCCCCAAATTAAATGT-3’, Anti-sense: 5’-AACGAT CCGGATTCGTCACT-3’.
The Nalm6 cell line has been previously described [33, 34]. Cells were cultured in RPMI-1640 medium (Cellgro, Tewksbury, MA, USA) supplemented with 10% fetal bovine serum (Hyclone, Logon, Utah, USA). HEK 293T cells were cultured in DMEM (Cellgro) supplemented with 10% fetal calf serum and 1% L-glutamine (Cellgro). Cells were incubated at 37°C in a humidified incubator with 5% CO2. Primary human B-ALL cells were cultured in RPMI-1640 medium (Cellgro) supplemented with 10% fetal bovine serum (GE-Hyclone, Logon, Utah, USA). CX4945 was purchased from Sigma (St. Louis, MO, USA). Cells were cultured with or without CX-4945 and collected for total RNA isolation. Human IKZF1 retroviral construct and retroviral production was described [33–35].
Promoters of BCL6 (-2000bp to +100bp) and BACH2 (-1000bp to +200bp) were cloned into pGL4.15 vector (Promega, Madison, WI, USA). The transient luciferase assay was performed in HEK293T cells using the Promega luciferase assay reagents and measured with a luminometer according to the manufacturer instructions [33, 35, 48]. The firefly luciferase activities were calculated as fold-change relative to values obtained from pGL4.15 vector only control cells, and expressed as a percent of pcDNA3.1-Ikaros transfection-induced luciferase activity vs. the pcDNA3.1 vector. All transfection and reporter assays were performed independently with ≥3 replicates.
qChIP assays were performed by incubating chromatin with antibodies against Ikaros [5, 8, 11] or normal rabbit IgG (Abcam) as a control [33, 35, 43, 44]. Enrichment of the ChIP sample over input was evaluated by qPCR with ≥3 replicates using specific primers in the promoter region of BCL6 (forward: 5’-TGCCGGCCAGTGAAAAA-3’, reverse: 5’-GCCCCTGGCCAACCAA-3’) and BACH2 (forward: 5’-AGCTGGCAGCCTCATTTCC-3’, reverse: 5’-GAGTG GAGTTGGGCACATCA-3’). Relative concentration of the qPCR product is presented as the fold change of the level of DNA-Ikaros compared with controls.
Nalm6 cells were transiently transfected with human IKZF1 or CK2α (CSNK2A1) shRNA constructs in the GFP vector (pGFP-v-RS) (OriGene) using the Neon Transfection System (Invitrogen, Carlsbad, CA, USA). We used scrambled 29-mer shRNA cassette in the pGFP-v-RS vector as a control. Knockdown of IKZF1 was confirmed by IKZF1 mRNA levels [33, 35, 43]. Primers used for qPCR are 5’-GGCGCGGTGCTCCTCCT-3’ (IKZF1-F) and 5’-TCCGACACGCCCTACGACA-3’(IKZF1-R).
Median differences between the cohorts were evaluated using a Mann–Whitney U-test. Frequency differences were analyzed using uni- and multivariate Cox model. Event-free survival (EFS) and overall survival (OS) were estimated by the Kaplan-Meier method and compared using a log-rank test. The starting point for the observation time for EFS and OS was the date of diagnosis. Death in induction, resistance, relapse, and death in continuous complete remission or new cancer were considered events in EFS calculations. Living subjects were censored for survival at last follow-up. Statistical analyses used SPSS version 20.0. Data were represented as mean values with bars representing the standard error of the mean (SEM). Determinations of statistical significance were performed using a Student's t-test for comparisons of two groups or using analysis of variance (ANOVA) for comparing multiple groups.
This work is supported in part by The National Natural Science Foundation of China (81270613, 30973376); Jiangsu Province Key Medical Talents (RC2011077); The Scientific Research Foundation for the Returned Overseas Chinese Scholars; State Education Ministry (39th); China Postdoctoral Science Foundation (20090461134); Special grade of the financial support from China Postdoctoral Science Foundation (201003598); The Six Great Talent Peak Plan of Jiangsu (2010-WS-024); The Scientific Research Foundation for the Returned Overseas Chinese Scholars; Nanjing Municipal Bureau of Personnel (2009); Southeast University Basic Research Fund (2242016k40143) (ZG). This work is supported in part by NIH/NCI R01 CA209829, a Hyundai Hope on Wheels Scholar Grant Award, Alex's Lemonade Stand Grant, Bear Necessities Pediatric Cancer Foundation, the Four Diamonds Fund of the Pennsylvania State University, College of Medicine, and the John Wawrynovic Leukemia Research Scholar Endowment (SD), and Alex's Lemonade Stand Grant (SD and CS); and by St. Baldrick's Foundation Summer Fellows Award (ED and JLP).
CONFLICTS OF INTEREST
All the authors declare no conflicts of interest.