CD44 activation in mature B-cell malignancies by a novel recurrent IGH translocation

Mature B-cell non-Hodgkin lymphomas (NHLs) are often associated with chromosomal translocations involving IGH at chromosome band 14q32.^1–5  V(D)J rearrangement and class switch recombination (CSR) are IG rearrangement processes that occur during B-cell development.^6–9  In CSR, DNA breakage is mediated by activation-induced cytidine deaminase in the germinal centers (GCs) of secondary follicles.^10,11  Any aberrant rearrangement during the DNA breakage and repair processes composed of V(D)J rearrangement and CSR may consequently generate a chromosomal translocation.^12,13 IGH translocations contribute to the pathogenesis of B-cell lymphomas via deregulated expression of the genes located at the partner chromosome locus partly because of the presence of potent B cell–specific transcriptional enhancers within the IGH gene locus.^1–5 

Gastric B-cell lymphomas usually exhibit chromosomal translocations involving IGH.^14–16  Molecular cloning of IGH translocation breakpoints has been successfully used to identify novel cancer-related genes in B-cell lymphomas.^17,18  In this study, we used inverse polymerase chain reaction (PCR)¹⁸  to identify a novel translocation involving the 5′ Sμ region of the IGH gene (IGHSμ) and CD44 (located at chromosome 11p13) in gastric as well as other mature B-cell NHLs. Functional studies suggest a possible pathogenic role of this IGHSμ/CD44 translocation in mature B-cell malignancies.

Genomic DNA extracted from frozen sections of tumor specimens from gastric¹⁹  and other mature B-cell NHL cases (supplemental Table 1, available on the Blood website; see the Supplemental Materials link at the top of the online article) was digested with HindIII, and long-distance inverse PCR was performed to detect translocations involving the IGHSμ region using the primers SAE/JXE, followed by SAI/JXI¹⁸  (supplemental Table 2).

CD44ΔEx1 (CD44 variant mRNA lacking exon 1) and CD44s (standard form; wild-type) cDNAs were amplified by reverse-transcribed (RT)–PCR (supplemental Table 2) from total RNA extracted from a case of gastric lymphoma with the IGHSμ/CD44 translocation (GL47) and from the peripheral blood lymphocytes of a healthy volunteer; they were then cloned in-frame with green fluorescent protein (GFP) at the C-terminus of the pmaxFP-Green-N vector (Amaxa) and transfected into the CD44⁻ GC B cell–like (GCB)–diffuse large B-cell lymphoma (DLBCL) cell line BJAB by nucleofection (Amaxa). The subcellular localization of the GFP-tagged proteins was assessed by confocal laser microscopy. The effect of overexpressing CD44ΔEx1 on the growth rate of stably transfected BJAB cells in vitro was determined using the CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega); clonogenic capacity was assessed with colony formation assays in methylcellulose.²⁰ 

Using an inverse PCR approach, we identified a novel translocation, t(11;14)(p13;q32), involving the 5′Sμ region of IGH and CD44 at 11p13 in 4 of 49 gastric lymphomas, 2 of 7 nongastric extranodal lymphomas, 2 of 30 nodal follicular lymphomas (FLs), and 1 of 26 nodal DLBCLs (Figure 1A; supplemental Table 1). This IGHSμ/CD44 translocation juxtaposes the IGHSμ enhancer to the 5′ regulatory region of CD44 in a tail-to-head orientation, leading to the removal of exon 1 of CD44 (Figure 1B). Functional switch recombination was also detected in the untranslocated allele in all translocation-positive cases. Further studies of all lymphoma cases by interphase fluorescence in situ hybridization (FISH) with CD44 break-apart probes (supplemental data) confirmed breakage at the CD44 locus in the 9 IGHSμ/CD44 translocation-positive cases identified by inverse PCR. No additional cases were identified with breakage at the CD44 locus among the inverse PCR⁻ cases (Figure 1C).

The 9 different IGHSμ/CD44 translocation breakpoints were distributed within small regions of IGHSμ and at the 5′ end of intron 1 of CD44 (Figure 1D). Microhomology sequences found at the junctional breakpoints of all the IGHSμ/CD44 translocations suggest that these translocations were facilitated by homologous sequences present on both chromosomes (supplemental Figure 1).^12,13  The CD44 gene at 11p13 is the first cell adhesion molecule²²  known to be involved in an IGH translocation in B-cell lymphomas (Figure 1E). CSR occurs in antigen-stimulated B cells in the GC of secondary follicles.¹¹  Notably, 2 FL cases were detected with IGHSμ/CD44 translocations, and all 4 translocation-positive DLBCL cases were GCB-type DLBCLs (supplemental Table 1), suggesting that these translocations resulted from illegitimate CSR occurring in GC B cells. This correlation is plausible, given that a high proportion of GCB-DLBCLs have undergone immunoglobulin class switching and that these tumors are trapped in the differentiation stage of the GC.^23–25 

As the principal molecular consequences of an IGH translocation is the deregulated expression of the partner gene,^1–5 CD44 mRNA expression was then studied by semiquantitative RT-PCR. RT-PCR primers across exons 2 and 18 (supplemental Table 2) designed to amplify all CD44 mRNA variants²⁶  were used to monitor CD44 expression in microdissected CD10⁺ GC B cells from reactive tonsils²⁷  and the 9 translocation-positive cases (supplemental data). CD44 mRNA was either minimally or not expressed in CD10⁺ microdissected reactive GC B cells (Figure 2A), consistent with the results of Alizadeh et al who showed by gene expression profiling that CD44 mRNA expression is very low in reactive GC B cells.²⁸  Our results suggest that this translocation caused overexpression of a new CD44 variant (CD44ΔEx1) activated from derivative 11 in all 6 IGHSμ/CD44 translocation-positive GC-derived lymphomas (Figure 2A).

5′-Rapid amplification of cDNA ends (RACE) analysis (supplemental data) identified Iμ-CD44ΔEx1 hybrid mRNA transcripts in all 9 lymphomas with IGHSμ/CD44 translocation, resulting from splicing of the Iμ exon upstream of Sμ to exon 2 of CD44 (Figure 2B). Similar Iμ-BCL6 and Iμ-MMSET hybrid transcripts associated with t(3;14)²³  and t(4;14)^30,31  translocations have been previously described. Because CD44 exon 2 has an in-frame stop codon upstream of the predicted translational initiation codon ATG at nucleotide 254, the Iμ transcripts that splice to CD44 exon 2 would not result in a chimeric fusion protein, and the Iμ-CD44ΔEx1 ORF would encode for the CD44ΔEx1 protein starting from the ATG at nucleotide 254 with a strong Kozak sequence²⁹  (Figure 2C). The new CD44ΔEx1 variant showed some similarity to CD44v5³² ; but it lacked the leader peptide, and the 3′ end of exon 2 of CD44ΔEx1 was spliced into a new internal splice acceptor site present within exon 17b, altering the reading frame (Figure 2C).

Notably, the IGHSμ/CD44 translocation was detected in 4 patients with advanced (stages II-IV) gastric, follicular, and nodal DLBCL. CD44 is a cell adhesion molecule and signaling regulator.²²  When overexpressed in vitro in the CD44⁻ GCB-DLBCL cell line BJAB, CD44s-GFP was localized to the plasma membrane, whereas CD44ΔEx1-GFP was present in the cytoplasm and nucleus (Figure 2D). The ectopic expression of CD44ΔEx1 resulted in a significant increase in BJAB cell growth (Figure 2E) and clonogenicity in methylcellulose (Figure 2F).

Because of the unavailability of antibodies for the new CD44ΔEx1 isoform, immunohistochemical analysis to confirm the localization of CD44ΔEx1 in vivo could not be performed on the lymphoma biopsy specimens of IGHSμ/CD44 translocation-positive cases expressing the Iμ-CD44ΔEx1 hybrid mRNA. However, 7 of 9 translocation-positive lymphomas showed MIB1 (Ki-67 antibody; a proliferation marker) nuclear staining in the tumor cells (supplemental Figure 2), supporting the in vitro results showing that CD44ΔEx1 overexpression gives cells a proliferation advantage, and suggesting a possible pathogenic role of the IGHSμ/CD44 translocation.

This work was supported by the Research Grants Council of Hong Kong Special Administrative Region, China (grant HKU 7448/04M; G.S. and R.H.S.L.).

Contribution: X.-T.H. and Y.-W.C. designed and performed experiments, interpreted data, and drafted the manuscript; A.C.T.L., K.-Y.W., T.S.K.W., M.L.Y.W., K.-K.C., and T.G. performed experiments; W.-Y.A., C.-S.C., Y.-L.K., and R.H.S.L. selected lymphoma cases; K.-M.C. provided gastric lymphoma surgical specimens; F.L., L.S., and W.W.L.C. selected tumor blocks and analyzed the immunohistochemistry slides; Q.T. and L.L. advised on the overall design of the study; C.-C.S. and L.C.C. analyzed interphase FISH data; and G.S. designed and supervised the study and finalized the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Gopesh Srivastava, Department of Pathology, University of Hong Kong, Pokfulam, Hong Kong; e-mail: gopesh@pathology.hku.hk.