Abstract
Background.NOTCH1 mutations are found in about 10% of chronic lymphocytic leukemia (CLL) cases at diagnosis, and with higher frequencies in chemorefractory CLL, CLL in advanced disease phases and in Richter Syndrome transformations (Rossi et al. Blood, 2013). In CLL, NOTCH1 mutations have been recently associated with clinical resistance to anti-CD20 immunotherapy (Stilgenbauer et al. Blood, 2014, Dal Bo et al. AHO, 2014). In lymphoproliferative disorders, susceptibility to rituximab is determined by CD20 levels (Golay et al. Blood, 2001), which are in turn epigenetically modulated via HDAC (Shimizu et al., Leukemia, 2010).
Aim. To investigate whether NOTCH1 mutations could affect CD20 expression in CLL.
Methods. NOTCH1 mutations and NOTCH1 mutational burden were evaluated by ARMS PCR, QRT-PCR, conventional and next generation sequencing. NOTCH1 protein levels were evaluated by western blot. CD20 expression was evaluated by flow cytometry. Transcript levels of MS4A1, encoding for CD20 protein, were evaluated by QRT-PCR. CLL cells and CLL-like cell line MEC-1 cells were treated with the HDAC inhibitor valproic acid (VPA) at a concentration of 3mM. Susceptibility to rituximab was evaluated by complement dependent cytotoxicity (CDC) assay. MEC-1 cells were transfected with a vector containing a NOTCH1 intracellular domain (NICD) carrying either the 7544-7545delCT or a stop codon at the beginning of NICD sequence (c.5304G>A), as control.
Results.i) In a cohort of 452 CLL, 54 cases carried NOTCH1 mutations. NOTCH1-mutated cases (NOTCH1-mut) with high mutational burden showed 3.0- fold higher transmembrane NOTCH1 and 2.1- fold higher NICD protein expression compared to NOTCH1 wild-type cases (NOTCH1-wt). NOTCH1-mut showed a lower Mean Fluorescent Intensity (MFI) of CD20 expression than NOTCH1-wt both in trisomy 12 (tris12, 18 NOTCH1-mut vs. 44 NOTCH1-wt, p<0.001) and in non-tris12 (36 NOTCH1-mut vs. 354 NOTCH1-wt, p=0.005) cases. MS4A1 transcript levels were lower in NOTCH1-mut than in NOTCH1-wt (tris12 CLL, p=0.024; non-tris12 CLL, p=0.002). By performing cell sorting to isolate CD20low and CD20high subpopulations in 4 cases with subclonal NOTCH1 mutations (range 26%-45%), the CD20low subpopulation always showed an enrichment in NOTCH1 mutational burden when compared to the CD20high counterpart, along with lower MS4A1 expression levels. NICD mutated MEC-1 cells, expressing higher NICD transcript (2.2 over the control) and NICD (1.3 over the control) protein, were characterized by lower CD20 (0.1 over the control) and MS4A1 transcript (0.77 over the control) expression. ii) In keeping with CD20 levels, primary NOTCH1-mut CLL cells showed lower relative lysis induced by rituximab than NOTCH1-wt CLL cells (mean relative lysis: NOTCH1-mut, 5% vs. NOTCH1-wt, 30%, p=0.008). This phenomenon was confirmed by using the MEC-1 cell model (mean relative lysis: NICD mutated, 3% vs. control, 30%, p=0.006). iii) NICD mutated MEC-1 cells expressed higher levels of both HDAC1 (2.0 over the control) and HDAC2 (1.4 over the control) transcripts. Moreover, in co-immunoprecipitation experiments, NICD mutated MEC-1 cells were characterized by lower levels of HDAC1/HDAC2 bound to the transcriptional factor CBF-1, due to a competition for the binding to CBF1 with the high NICD levels present in these cells, and suggesting a nuclear interplay between NOTCH1 and HDAC where CBF-1 is the connecting element. The dependency of low CD20 expression by HDAC activity was demonstrated by the capability of VPA to increase CD20 expression in primary NOTCH1-mut CLL cells (1.3 over the control) and NICD mutated MEC-1 transfectants (1.4 over the control).
Conclusions. CLL cases carrying NOTCH1 mutations are characterized by low CD20 expression levels that may confer resistance to anti-CD20 immunotherapy. The low CD20 expression, at least in part due to HDAC-dependent repression mechanism(s), can be reverted by HDAC inhibitor therapy.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.
