Recent immunogenetic studies of chronic lymphocytic leukemia (CLL) suggest a dichotomy: those developing from naive, pregerminal B lymphocytes exhibiting germline configuration of Ig VHstatus (poor outcome) and those stemming from more mature, postgerminal center memory B cells with mutated Ig VHgenes1-3 (good prognosis). Due to the cost and expertise required for this technique, a simpler, inexpensive substitute has been sought. CD38 antigen expression is a marker that strongly correlates with Ig VH gene mutational status1,4 and predicts outcome in CLL,1,4-9 although not all studies agree.10-12 Another class of markers, chemokines, and their respective receptors, may also correlate with clinical stage and prognosis. CXCR4, a chemokine receptor found on CLL B cells, may play a role in the marrow infiltration observed in CLL.13 We evaluated (1) whether CD38 or CXCR4 expression is associated with immunoglobulin gene usage and (2) whether these markers are prognostic for survival in familial CLL cases.
Thirty-nine individuals from the NCI Familial CLL Registry14 had cells labeled, using a 4-color staining method, with monoclonal antibodies against the lymphoid antigens CD5 phycoerythrin, CD38 allophycocyanin, CD19 peridinin chlorophyll protein (BD Biosciences, San Jose, CA), and CXCR4 flourescein (R&D Systems, Minneapolis, MN). Antibody expression was measured by FACSCaliber (Becton Dickinson, San Jose, CA) flow cytometer. CD38 and CXCR4 expressions were scored as percent positive of the B-CLL cells (CD5+/CD19+).6 DNA of rearranged immunoglobulin heavy chain genes were amplified by polymerase chain reaction and cloned and sequenced, using a previously described method.15 Assays were conducted with laboratory personnel blinded to VH mutation and clinical characteristics of the patients.
Generalized linear models were used to estimate least-squared mean CD38 expression by heavy chain mutation status on 21 patients using SAS Version 8.0 (SAS, Cary, NC) and to evaluate the association with CXCR4 expression data. Survival was estimated on all 39 familial CLL patients by the Kaplan-Meier method, using SPLUS 2000 (Insightful, Seattle, WA). Median levels of CD38 and CXCR4 expression in the subset of cases with VH data were used to group the patients into lower and higher risk; differences in survival were tested by the Wilcoxon test. All tests of statistical significance were two-sided.
In agreement with earlier reports,1 3 unmutated VH cases displayed a higher percentage of CD38+cells than mutated cases (23.62% versus 5.80%, P = .03). We did not observe an association between CD38 expression and survival in the 39 patients (Wilcoxon, P = .45) (Figure1A). Rather, expression of the CXCR4 chemokine receptor was more strongly correlated with VHmutational status (9.53% versus 40.82%, P = .004; adjusted for age) and a better predictor of survival (Wilcoxon,P = .02) (Figure 1B).
CD38 expression has been proposed as a surrogate marker for VH mutation status to predict the clinical course in CLL with mixed findings.4-12 Although we observed an association between CD38 expression and VH mutation status, CD38 expression, in contrast with VH mutation status, did not predict survival in our familial CLL cases. The 2 assays yielded discordant results in 8 of the 21 cases, which may account for these findings. Furthermore, only 2 cases had advanced disease at the time of the specimen collection, resulting in the generally lower CD38 expression observed in this series. Variability in CD38 expression may also be lower in familial CLL than in sporadic cases, making it difficult to observe a difference with the available number of cases. Only 2 of 21 patients with VH mutational data had CD38 expression levels greater than 30%.
The association between underexpression of the CXCR4 chemokine receptor and survival is consistent with its involvement in the trafficking and homing of B-CLL cells to bone marrow, and the strong down-modulation of CXCR4 on CLL B cells that migrate into the marrow stromal cell layer that has been reported.13 Although we were unable to examine CXCR4 expression in relation to the extent of marrow involvement, our data also suggest that CXCR4 down-regulation may play a role in the marrow infiltration observed in CLL. CXCR4 may be one factor that regulates neoplastic B cells' survival by making cells more resistant to apoptosis and allowing for unchecked proliferation of the malignant clone.
In summary, our results suggest that CXCR4 receptor expression levels could potentially distinguish CLL patients who will develop aggressive disease from those who will not. In our familial CLL cases, however, no association with prognosis with CD38 expression was observed. Future studies in a larger study population with extensive treatment and clinical information are clearly warranted.