Surface IgM Levels Independently Influence Clinical Behavior and Associate with Altered Phenotype and Genetics in Chronic Lymphocytic Leukemia

The tumor B-cell receptor (BCR) is the key to survival and proliferation in chronic lymphocytic leukemia (CLL) and is now a therapeutic target of very effective BCR-associated kinase inhibitors. CLL cells are typically characterized by a variable state of anergy, defined by low surface IgM (sIgM) levels and signaling ability with consequent relatively low proliferation rate. The subset with unmutated IGHV (U-CLL) is generally less anergic with a poorer prognosis than that with mutated IGHV (M-CLL), and appears more sensitive to BCR-inhibitors (Byrd et al., NEJM, 2013). However it remains unclear if the variable anergy reflects the nature of the cell of origin and/or the functional state of the BCR, thereby determining clinical behavior.

In this study we investigated if variations of BCR sIgM levels and signaling correlate with clinical behavior of CLL and assessed the phenotypic and genetic associations with the variations.

The study included samples at diagnosis or prior to treatment from 222 patients with sIgM/D CLL diagnosed according to the iWCLL2008 criteria (99 months median follow up). sIgM levels on the CD19+/CD5+ CLL cells and signaling [% intracellular Ca2+ mobilization (iCa2+)] were determined using soluble F(ab’)2 polyclonal antibodies (Mockridge at al, Blood, 2007). sIgM levels were analyzed for their association with i) survival, ii) phenotype (CD38, ZAP70 and BCR regulators CD5, CD19, CD20, CD22) and iii) genetic lesions (Del13q, Trisomy12, Del11q/ATM, Del17p/TP53, or NOTCH1 and SF3B1 mutants). Time from diagnosis to progression requiring first treatment (TTFT) was used as primary endpoint, while overall survival (OS) was used as a secondary endpoint to avoid the influences of chemotherapy. Best cut-offs for progression requiring treatment were determined with ROC and Youden’s T-tests (treatment as a state variable).

Levels of sIgM (5^th-95^th percentile 12-378, median 52) varied markedly between patients. Levels were higher at more advanced stage of disease and sIgM expression (MFI>56) and signaling (iCa2+>6%) associated with significantly more rapid TTFT. Although sIgD levels and signaling also associated with shorter TTFT, multivariate Cox regression adjusted for IGHV status, sIgM levels, sIgM signaling, sIgD levels and sIgD signaling revealed that only U-IGHV (p<0.001, 95% CI 1.64-3.97) and high sIgM levels (p=0.004, 95% CI 1.24-3.16) were independent predictors of shorter TTFT. Also, high sIgM levels and not high sIgD associated with shorter OS, despite chemotherapy. These data confirmed the specific influence of anergy which drives down sIgM and not sIgD, and is associated with less aggressive tumor behavior in CLL.

By focusing on sIgM, a multivariate analysis adjusted for IGHV status, sIgM levels, CD38 and ZAP70 confirmed U-IGHV (p=0.002, 95% CI 1.35-3.56) and high sIgM (p=0.003, 95% CI 1.25-3.04) as the only independent predictors of shorter TTFT. Consistently, high sIgM levels associated with a more aggressive CLL even when U-CLL or M-CLL were investigated separately. Phenotypic analyses showed that high sIgM CLL had higher CD20, CD38, ZAP70, and CD22, than low sIgM CLL. Genetic analysis revealed that CLL subsets harboring Trisomy12 or Del17p/TP53 had significantly higher sIgM levels and signaling capacity than Del13q, even when U-CLL or M-CLL were analyzed separately. Trisomy12 U-CLL cases were enriched for NOTCH1 mutations. Consistently, U-CLL with NOTCH1 mutations (with or without Trisomy12) expressed higher sIgM than Del13q U-CLL.

This study has several biological and clinical implications.

First, it supports the critical role of sIgM in CLL-associated anergy. Our previous studies had shown dynamic variations of sIgM on circulating CLL cells following (auto)antigen engagement in tissue (Mockridge et al., Blood, 2007; Coelho et al. Blood, 2013). The subpopulation with higher sIgM is potentially the most dangerous and the most sensitive to kinase-inhibitors.

Second, it documents the dominant role of sIgM levels on cell signaling and on tumor progression in CLL, with relevance even within U-CLL and M-CLL.

Third, this study suggests influences of altered genetics on sIgM levels. However BCR-inhibitors overcome the influence of genetic lesions like Del17p/TP53. This points to a dominant role of reversed sIgM anergy in the behavior of CLL in vivo and claims the need of studies to verify sensitivity to kinase-inhibitors in CLL patients with different sIgM levels.