CDX2/UBTF Acute Lymphoblastic Leukemia: Subtype Discovery and Clinical Implications

For adult patients with B-lineage acute lymphoblastic leukemia (B-ALL), the cytogenetic and/or molecular genetic alterations define the specific subtypes of this leukemia and form the basis of the World Health Organization (WHO) classification system for lymphoid malignancies.¹  Importantly, the specific cytogenetic or molecular genetic alterations inform the biologic and clinical course, provide the basis for risk classification, and in some subtypes (i.e., Philadelphia chromosome [Ph] –positive B-ALL), influence therapeutic decisions. Many cases of B-ALL, particularly in adult patients, however, have neither chromosome nor molecular genetic alterations as detected with conventional cytogenetic and genomic analysis, and are classified in the fifth edition of the WHO classification system as B-cell lymphoblastic leukemia/lymphoma not otherwise specified.¹  This subtype of B-ALL continues to be the focus of rigorous laboratory investigation, specifically with advances in the use of whole-genome sequencing (WGS) and whole-transcriptome sequencing (RNA-seq), in efforts to uncover the “driver” genomic alterations. This can in turn provide insight into the clinical behavior of these undefined B-ALL cases and potentially identify novel targeted therapies.

In the two companion articles, Dr. Marie Passett and colleagues and Dr. Shunsuke Kimura and colleagues used an integrative genome analysis strategy— the use of transcriptome sequencing (RNA-seq) with WGS — to further evaluate cases of B-ALL characterized by the absence of recurrent genetic abnormalities. Dr. Passet and colleagues performed their analysis on 302 primary B-ALL samples obtained at diagnosis from adult patients (age range, 18-84 years) who enrolled on consecutive Group for Research in Adult ALL (GRAALL) clinical trials (GRAALL- 2005, GRAALL-2014), and the European ALL Working Group INO (EWALL-INO) trial. Dr. Kimura and colleagues analyzed 3,398 B-ALL cases obtained at diagnosis or relapse from nine patient cohorts that included adult and pediatric clinical trials and/or academic institutions (age range, 12-70 years). Both groups independently identified a new, rare (n=26 cases in Passett et al; n=22 cases in Kimura et al) subtype of B-ALL with a distinct gene expression profile and discovered two unique genomic alterations that were present in all cases: 1) short (83 and 109 kb) deletions at the 13q12.2 locus resulting in the aberrant expression of the caudal type homeobox 2 (CDX2) gene; and 2) short (10 kb) deletions at the 17q21.31 locus involving the upstream binding transcription factor (UBTF) and ataxin 7 like 3 (ATXN7L3) genes and resulting in the expression of UBTF::ATXN7L3 fusion transcript.

CDX2 is a homeobox transcription factor involved in early stages of embryogenesis with a key role in intestinal development, but also plays a role in hematopoietic development.^2,3  Its expression in adult tissues is restricted to intestinal epithelial cells, however. Aberrant expression of CDX2 has been described recently in B-ALL,^4,5  but the mechanism of its expression has been unknown. Thus, the authors of the current reports further investigated the dysregulation of the CDX2 gene expression and the potential role of the 13q12.2 deletion on CDX2 function. They showed that the upregulation of CDX2 results from a mechanism termed “enhancer release and retargeting,” or “enhancer hijacking,” which in this scientific context refers to the repositioning of an enhancer away from its normal target promotor to an alternate promoter.⁶  In the case of the genomic 13q12.2 deletion in CDX2/UBTF ALL cases, the authors showed evidence of chromatin looping between the PAN3 (poly[A]-specific ribonuclease subunit) enhancer and the CDX2 promoter, providing evidence that the PAN3 enhancer was driving the aberrant expression of the CDX2 gene. Interestingly, 13q12.2 deletions have been identified recently in a cohort of pediatric patients with the high-hyperdiploid B-ALL subtype.⁷  In this subtype, however, the PAN3 enhancer drives aberrant expression of the FLT3 gene, which also resides in the 13q12 locus. Hence, both research teams examined FLT3 expression in the CDX2/UBTF ALL cases, but in these cases, the 13q12.2 deletions removed the promoter region and exon 1 of the FLT3 gene (compared to high-hyperdiploid B-ALL, where the 13q12.2 deletion spared the FLT3 promoter region), and thus FLT3 expression was observed to be low in CDX2/UBTF cases and not thought to contribute significantly to the phenotype. With regard to the deletion identified on chromosome 17 that results in the UBTF::ATXN7L3 fusion transcript, further study is necessary to understand its function and interaction with the CDX2 gene, specifically how do these two distinct genomic alterations cooperate to promote leukemogenesis?

The clinical implications of CDX2/UBTF ALL are distinctive as well. The majority of the cases were identified in young female patients (20 of 26 cases identified by Dr. Passett and colleagues; 17 of 22 cases identified by Kimura and colleagues) with a median age of 31 years (Passett et al cohort) and 35 years (Kimura et al cohort). Of note, in the pediatric cohort of high-hyperdiploid B-ALL that contained 13q12.2 deletions,⁷  the majority of cases also were identified in female patients (14 of 23 cases), despite that the incidence of childhood B-ALL is slightly more common in male patients, and the high-hyperdiploid subtype is equally common in male and female pediatric cases.^8,9  The immunophenotype reveals a blast population negative or with partial expression of CD10, negative expression of CD20, high expression of CD34, CD38, and cytoplasmic IgM. Patients with CDX2/UBTF ALL who were enrolled on GRAALL-2005/2014 protocols demonstrated a slower blast clearance to induction therapy (day 8 blood response of 50% vs. 84% in other B-ALL cases; day 15 marrow response of 13% vs. 60% in other B-ALL cases). Also, patients with CDX2/UBTF ALL were more likely to have detectable minimal residual disease (≥10⁻⁴) following induction therapy (93% vs. 46% of other B-ALL cases), the most important prognostic factor for relapse in ALL. And in fact, the cumulative incidence of relapse (CIR) was higher for patients with CDX2/UBTF ALL (3-year CIR estimate of 75% vs. 32.4% in other B-ALL cases). Overall survival was similar between CDX2/UBTF ALL and other B-ALL cases (3-year overall survival, 68% vs. 66%, respectively), however, and suggests that current salvage therapies are effective in this unique subtype. Patients included in the Kumari et al cohort also showed a young female predominance and revealed a high incidence of detectable minimal residual disease (>1%) at the end of induction therapy (56.3%), and enrichment of the subtype in relapse cases, further supporting the high-risk features of this novel subtype.