Increased Incidence of IKZF1 deletions and IGH-CRLF2 translocations in B-ALL of Hispanic/Latino children—a novel health disparity

Hispanic/Latino (H/L) children and adolescents are 1.2–1.75 times more likely to develop acute lymphoblastic leukemia (ALL) than Non-Hispanic Whites (NHW) [1]. Once they develop ALL, H/L children have a 40% higher death-rate than NHW, after correcting for socioeconomic factors [2]. Although H/L children with B-ALL have a worse prognosis than non-H/L children, the biological basis for this health disparity is largely unknown. Single nucleotide polymorphisms (SNPs) in ARID5B and GATA3 that are associated with predisposition to B-ALL and/or poor prognosis are more frequent among H/Ls [3, 4]. However, the major drivers of B-ALL through which these SNPs might contribute to health disparities have not been defined. A previous study of children with high-risk B-ALL showed increased incidence of CRLF2 gene rearrangement in H/L children as compared to the non-H/L population. CRLF2 gene rearrangement was also associated with deletion of the IKZF1 tumor suppressor [5]. A study of adult H/L patients with B-ALL showed increased incidence of Ph-like B-ALL that was associated with CRLF2 rearrangement and IKZF1 deletion [6]. Both studies were limited to subsets of B-ALL patients with high-risk features. Thus, the question of whether CRLF2 gene rearrangement and/or IKZF1 deletion provide a biological basis for the overall health disparity in pediatric B-ALL for H/L children remains unanswered. Here, we address this question by performing a single-center, unbiased analysis to determine and compare the incidence of CRLF2 rearrangement and IKZF1 deletion in H/L vs. non-H/L children with B-ALL. We analyzed clinical and molecular data [7] from 239 pediatric B-ALL patients treated at Childrens Hospital Los Angeles between 3/2016 and 7/2019 (See Supplemental Materials). Of 239 patients diagnosed with B-ALL, 164 self-reported as H/L and 75 were classified as non-H/L (Table 1). CRLF2 rearrangements include two types of genetic alterations: IGH-CRLF2 translocation, where the immunoglobulin heavy chain locus (IGH) is translocated to CRLF2 [8]; and P2RY8-CRLF2 fusion, where the PAR1 deletion juxtaposes the noncoding exon of P2RY8 to CRLF2 [9]. Analysis of each of these genetic alterations separately, showed significantly increased incidence of IGH-CRLF2 translocation in the H/L vs. non-H/L groups, 19/164 (12%) vs. 2/75 (2.7%), p= 0.026. However, the incidence of P2RY8-CRLF2 fusion was not significantly different between the two populations. B-ALL in the H/L population showed a significantly higher incidence of IKZF1 deletion as compared to non-H/Ls, 48/164 (29%) vs. 11/75 (15%), p= 0.016. These results suggest that IKZF1 deletion is a novel biological determinant of the health disparity in pediatric B-ALL for H/L children. Since the previous data suggested an association between CRLF2 translocations and IKZF1 deletion, we analyzed the incidence of patients with concomitant IKZF1 deletion and * Gordana Raca graca@chla.usc.edu


To the Editor:
Hispanic/Latino (H/L) children and adolescents are 1.2-1.75 times more likely to develop acute lymphoblastic leukemia (ALL) than Non-Hispanic Whites (NHW) [1]. Once they develop ALL, H/L children have a 40% higher death-rate than NHW, after correcting for socioeconomic factors [2]. Although H/L children with B-ALL have a worse prognosis than non-H/L children, the biological basis for this health disparity is largely unknown. Single nucleotide polymorphisms (SNPs) in ARID5B and GATA3 that are associated with predisposition to B-ALL and/or poor prognosis are more frequent among H/Ls [3,4]. However, the major drivers of B-ALL through which these SNPs might contribute to health disparities have not been defined.
A previous study of children with high-risk B-ALL showed increased incidence of CRLF2 gene rearrangement in H/L children as compared to the non-H/L population.
CRLF2 gene rearrangement was also associated with deletion of the IKZF1 tumor suppressor [5]. A study of adult H/L patients with B-ALL showed increased incidence of Ph-like B-ALL that was associated with CRLF2 rearrangement and IKZF1 deletion [6]. Both studies were limited to subsets of B-ALL patients with high-risk features. Thus, the question of whether CRLF2 gene rearrangement and/or IKZF1 deletion provide a biological basis for the overall health disparity in pediatric B-ALL for H/L children remains unanswered. Here, we address this question by performing a single-center, unbiased analysis to determine and compare the incidence of CRLF2 rearrangement and IKZF1 deletion in H/L vs. non-H/L children with B-ALL.
We analyzed clinical and molecular data [7] from 239 pediatric B-ALL patients treated at Childrens Hospital Los Angeles between 3/2016 and 7/2019 (See Supplemental Materials). Of 239 patients diagnosed with B-ALL, 164 self-reported as H/L and 75 were classified as non-H/L (Table 1). CRLF2 rearrangements include two types of genetic alterations: IGH-CRLF2 translocation, where the immunoglobulin heavy chain locus (IGH) is translocated to CRLF2 [8]; and P2RY8-CRLF2 fusion, where the PAR1 deletion juxtaposes the noncoding exon of P2RY8 to CRLF2 [9]. Analysis of each of these genetic alterations separately, showed significantly increased incidence of IGH-CRLF2 translocation in the H/L vs. non-H/L groups, 19/164 (12%) vs. 2/75 (2.7%), p = 0.026. However, the incidence of P2RY8-CRLF2 fusion was not significantly different between the two populations.
B-ALL in the H/L population showed a significantly higher incidence of IKZF1 deletion as compared to non-H/Ls, 48/164 (29%) vs. 11/75 (15%), p = 0.016. These results suggest that IKZF1 deletion is a novel biological determinant of the health disparity in pediatric B-ALL for H/L children.
Since the previous data suggested an association between CRLF2 translocations and IKZF1 deletion, we analyzed the incidence of patients with concomitant IKZF1 deletion and Bold values indicate statistical significance p < 0.05. a Statistics presented: median (IQR); n (%). b Statistical tests performed: Wilcoxon rank-sum test; Fisher's exact test. c The P2RY8-CRLF2 translocation is more common in Down Syndrome B-ALL, the H/L cohort included three Down Syndrome cases (one P2RY8-CRLF2 and two unknown genetics); the Other cohort included four Down Syndrome cases (three cases P2RY8-CRLF2, and one hyperploidy).
IGH-CRLF2 translocation are highly increased in B-ALL of children in the H/L population. The incidence of P2RY8-CRLF2 fusion is not different between the two populations. IKZF1 deletion and IGH-CRLF2 translocation are each associated with poor prognosis [10][11][12][13] and both IKAROS and CRLF2 proteins regulate a large number of genes and/or pathways that promote leukemia progression and drug resistance [14,15]. Thus, these data provide evidence of IGH-CRLF2 translocation and IKZF1 deletion as biological determinants of the health disparity in pediatric B-ALL for H/ L patients and suggest a biological rationale for the inferior outcome of H/L children with this disease.
We analyzed whether the age of patients affects the incidence, and/or racial difference of the above genetic alterations in B-ALL. In children ≥10 yrs old (Table 1), the incidence of IKZF1 deletion is 2.8-fold increased in the H/L vs. non-H/L population, 35/59 (59%) vs. 5/24 (21%), p = 0.002, with an odds ratio of 5.4. IKZF1 deletion is highly increased in children ≥10 yrs (59%) vs. <10 yrs (12%) in the H/L population (Table S2), but not in the non-H/L group. In children ≥10 yrs, the incidence of IGH-CRLF2 translocation was strongly increased in the H/L vs. non-H/L population, 18/59 (31%) vs. 0/24 (0%), p = 0.001. The incidence of IGH-CRLF2 translocation is highly increased in children ≥10 yrs, 18/59 (31%) old vs. <10 yrs old, 1/105 (1%) in the H/L population, but not in the non-H/L group. All of the patients ≥10 yrs with IGH-CRLF2 translocations in both the H/L and the non-H/L group also had concomitant IKZF1 deletions. Thus, no patient ≥10 yrs had IGH-CRLF2 translocation without concomitant IKZF1 deletion. In contrast, in patients ≥10 yrs old, the IKZF1 deletion without the presence of the IGH-CRLF2 translocation was detected in 17/59 (29%) of the H/L population.
In children <10 yrs, neither IGH-CRLF2 translocation, IKZF1 deletion, nor the combination of these two genetic alterations showed significant difference in incidence between the H/L and non-H/L populations (Table S1).
When analysis included only Ph negative B-ALL (Table S2- The results of our study provide a biological rationale for the worse prognosis of B-ALL in H/L children. Our unbiased, single-institution study identified highly increased incidence of IGH-CRLF2 translocation, IKZF1 deletion and concomitant IGH-CRLF2 translocation with IKZF1 deletion in B-ALL of H/L children. The approximate fourfold increased incidence in H/L children with B-ALL, makes IGH-CRLF2 the single genetic alteration with the highest racial/ethnic pediatric cancer disparity. The very high overall incidence (29%), makes IKZF1 deletion the most frequent genetic alteration that confers adverse prognosis in B-ALL in H/L children. However, the largest difference between H/L and non-H/L children was the presence of the concomitant IKZF1 deletion with IGH-CRLF2 translocation, which was detected in 11% of H/L children, but was not detected in any leukemia of non-H/L children.
The disparity in incidence of IGH-CRLF2 translocation and IKZF1 deletion in H/L children vs. non-H/Ls was very strong in children ≥10 yrs, but not in younger children. The most intriguing finding in our study was that over 94% (18/19) B-ALL in H/L children with IGH-CRLF2 translocation had concomitant IKZF1 deletion. In contrast, 30 H/L children with B-ALL had IKZF1 deletion without concomitant IGH-CRLF2 translocation. This raises the strong possibility that IKZF1 deletion precedes IGH-CRLF2 translocation, and/or that IKZF1 deletion predisposes cells to IGH-CRLF2 translocation in B-ALL of H/L children. IKAROS represses transcription of the RAG1 gene and increased expression of RAG1 due to IKZF1 deletion, might play a role in this process (Fig. 1). The results of our study lead to two main questions: (1) What biological factors cause increased incidence of IKZF1 deletion in H/L children; and (2) Does the presence of IKZF1 deletion in the B-lineage make cells more susceptible to the IGH-CRLF2 translocation, and if so, why is that susceptibility stronger in H/L children than in non-H/L populations. One GATA3 SNP occurs at higher frequency in the H/L population and has been associated with increased susceptibility to CRLF2 rearrangement and IKZF1 deletion [3,4]. However, functional studies to evaluate the potential role of GATA3 or non-H/L biological factors in CRLF2 and IKZF1 alterations have not been performed. Answering these questions will help in understanding the pathogenesis of pediatric B-ALL In summary, the presented data demonstrate that IGH-CRLF2 translocation and IKZF1 deletion provide a biological basis for the health disparity in pediatric B-ALL for H/L children and a strong biological rationale for the higher deathrate they experience due to B-ALL. Our study suggests that, in addition to reducing socioeconomic inequities, the following changes in clinical practice would improve the prognosis of H/L children with B-ALL: (1) due to the high incidence of IGH-CRLF2 translocation and IKZF1 deletion, every child of H/L background with B-ALL should be tested specifically for the presence of both of these genetic alterations; and (2) novel treatment strategies that restore IKAROS function while targeting CRLF2 signaling pathways (e.g., JAK/STAT or PI3K/ AKT/mTOR), should be developed and clinically tested to reduce the health disparity in pediatric B-ALL.
Author contributions SD and KJP analyzed and interpreted data, wrote the paper and designed research: GR and HA collected data, analyzed and interpreted data, and wrote the paper; FY, JB, JLP, CG, DB, JS, DD, ED, and TH analyzed and interpreted data; ASB performed statistical analysis; MER provided critical review and assisted in writing the paper.

Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict of interest.
Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/.