High frequency of p53 dysfunction and low level of V_H mutation in chronic lymphocytic leukemia patients using the V_H3-21 gene segment

The recent report by Tobin et al¹  that chronic lymphocytic leukemia (CLL) patients using the V_H3-21 gene segment have a poor prognosis despite mostly having more than 2% V_H mutation has generated a considerable amount of interest.^2-4  This is partly because such cases constitute an exception to the rule that “mutated” CLL necessarily has a good prognosis, and partly because their poor outcome remains unexplained.

We have previously reported that p53 dysfunction can arise in CLL through inactivating mutation of the genes encoding either p53 or ATM (a kinase that regulates p53) and have devised a simple screening test for these defects.⁵  The test measures the accumulation of p53 and one of its transcriptional targets (p21) in response to ionizing radiation (IR). Both defects are associated with impaired p21 accumulation. In the type A defect due to p53 mutation, baseline levels of p53 are increased, whereas in the type B defect, due to ATM mutation, baseline p53 is not increased, but IR-induced accumulation is impaired. We have further shown that both defects are, like CD38 expression, largely confined to patients with less than 5% V_H mutation and associated with a short survival.⁶  On the basis of these findings we proposed a prognostic model of CLL in which a 5% V_H mutation cut-off separates a homogeneous group of good-risk “mutated” patients from a heterogeneous “unmutated” group. The latter can then be risk-stratified according to the presence or absence of other adverse prognostic factors (eg, p53 dysfunction, CD38 expression) that are confined to this group.

It was of obvious interest to ascertain how CLL patients using V_H3-21 fitted into this prognostic model and whether the poor outcome of such patients could be explained by an association with p53 dysfunction or CD38 expression. To do this, we have examined the relationship between V_H3-21 usage and V_H mutation, p53 dysfunction, and CD38 expression in an updated series of 83 patients with CLL attending the Royal Liverpool University Hospital. The median age at diagnosis was 64 years and the male-to-female ratio was 1.48. The median interval between diagnosis and blood sampling was 26 months. The overall median follow-up was 60 months, and survival data were obtained on all but one patient.

V_H3-21 was used in 8 patients (9.6%). In keeping with the findings of Tobin et al,¹  the survival of such patients was significantly shorter than that of patients using other V_H gene segments (Figure 1A). Two (25%) of the 8 patients using V_H3-21 had less than 2% V_H mutation, 6 (75%) had 2% to 5% mutation, and none had more than 5% mutation (Figure 1B). In contrast, among the 75 patients with V_H segments other than V_H3-21, 33 (44%) had less than 2% V_H mutation, 12 (16%) had 2% to 5% mutation, and 30 (40%) had more than 5% mutation (Figure 1B). The association between V_H3-21 usage and less than 5% V_H mutation was found to be statistically significant (P < .05; Fisher exact test). Furthermore, among cases with 5% V_H mutation or less, there was a significant association between V_H3-21 use and 2% to 5% V_H mutation (P < .05). These findings are very much in accordance with the point made by Nollet et al²  that, despite being “mutated” by conventional criteria, most CLL patients using V_H3-21 would fall into the “unmutated” category if a higher V_H mutation threshold were employed.

In addition to confirming that V_H3-21 usage is associated with short survival and a relatively low level of V_H mutation, we also observed a strong association between V_H3-21 usage and p53 dysfunction (P < .05; Fisher exact test). Specifically, among the 8 patients using V_H3-21, 5 (62%) had a functional defect of the pathway (2 type A, 3 type B) compared with 20 (27%) of the 75 patients using other V_H gene segments (10 type A, 10 type B). The higher frequency of p53 dysfunction observed among patients using V_H3-21 was not associated with delayed blood sampling relative to patients using other V_H gene segments (P = .64; Mann-Whitney U test).

A statistically significant association was not observed between V_H3-21 usage and CD38 expression (defined as positivity on more than 20% of CLL cells using the HB7 monoclonal antibody). Thus, among the 8 patients using V_H3-21, 4 (50%) expressed CD38, compared with 20 (26.7%) of the 75 patients using other V_H gene segments (P = .22; Fisher exact test). However, the high odds ratio (2.75) meant that a β error due to small sample size could not be excluded.

We conclude that CLL patients using V_H3-21 fit into our prognostic model extremely well and that their adverse clinical outcome might, at least in part, be due to associated p53 dysfunction.