Chronic lymphocytic leukemia (CLL) is characterized by a CD20 expression lower than other lymphoproliferative disorders, hampering treatment efficacy with anti-CD20 antibodies. NOTCH1 mutations determine a decrease of CD20 expression via a NOTCH1-driven repression of CD20 transcription by an epigenetic HDAC-mediated mechanism (Pozzo et al. Leukemia 2016). This may explain the association between NOTCH1 mutations and resistance to anti-CD20 immunotherapy in CLL clinical trials (Stilgenbauer et al. Blood 2014). SF3B1 mutations, which are found in 10% CLL at diagnosis, induce transcriptome-wide alterations of splicing patterns, thus affecting several pathways, including the NOTCH1 pathway (Wang et al. Cancer Cell 2016). In this context, one of the most aberrantly-spliced genes is DVL2, a key component of the Wnt pathway and a known negative regulator of the NOTCH1 pathway (Collu et al. Development 2012).
To define the association between CD20 expression and SF3B1-mutation in CLL, by investigating NOTCH1-mutation-independent activation mechanisms of the NOTCH1 pathway through loss-of-function splicing alterations of the NOTCH1 negative regulator DVL2.
Mutational status of NOTCH1 and SF3B1 was evaluated by NGS. CD20 expression was evaluated by flow cytometry by computing mean fluorescence intensities (MFI) and/or the percentages of the CD20dim population. Transcript levels of spliced DVL2 were evaluated by qRT-PCR and NGS. Gene expression profile (GEP) was performed by a one-color labeling strategy using the Agilent 8x60K platform.
In a cohort of 537 unselected CLL, 89 cases carried NOTCH1 mutations (NOTCH1-mut); 48 cases carried SF3B1 mutations (SF3B1-mut), the K700 hotspot being the most frequent (50%) followed by the G742 (30%) hotspot; 6 cases were both NOTCH1/SF3B1- mut. Confirming previous findings, NOTCH1-mut cases were characterized by lower CD20 MFI compared to wild-type (WT) cases (p = 0.0154). CD20 expression was found diminished also in SF3B1-mut cases (p = 0.0059) and, consistently, NOTCH1 or SF3B1 mutations were also characterized by a greater proportion of CD20dim CLL cells (Fig.1A). Alternatively spliced DVL2 (altDVL2) was detected only in SF3B1-mut CLL and its expression correlated with the SF3B1 mutational burden (Fig.1B). As DVL2 can act as negative regulator of NOTCH1, we speculated that loss of wild-type DVL2 due to alternative splicing may result in a more active NOTCH1 pathway. The NOTCH1 target gene DTX1, more expressed in NOTCH1-mut cases compared to WT, was also increased SF3B1-mut cases (Fig.1C), where it positively correlated with altDVL2 levels and with a greater proportion of CD20dim CLL cells. In a GEP carried out on selected NOTCH1-mut versus WT cases (6 vs 7), 760 probes were differentially expressed (272 probes were up-regulated, including the NOTCH1-related genes DTX1 and HNRNPH3 and the NF-kB-related gene RELA; 488 probes were down-regulated, including the CD20 gene, MS4A1, and DUSP22). When applying this NOTCH1-related signature to SF3B1-mut cases (5 cases), the latter segregated with NOTCH1-mut cases (Fig.1D), suggesting a GEP relationship between NOTCH1-mut and SF3B1-mut CLL cells. Finally, loss of wild-type DVL2 in the CLL-like MEC1 cell line by siRNA transfection resulted, after 24 hours, in a downregulation of CD20 at both protein (p = 0.0062) and transcript (p = 0.0129) level (Fig.1E).
CLL cases bearing SF3B1 mutations are characterized by a lower CD20 expression, allegedly through a more active NOTCH1 pathway, potentially resulting in clinical resistance to anti-CD20 monoclonal antibodies.