The TGF-beta type III receptor CD105 is an important activator of endothelial cells in healthy and malignant tissues [13, 32]. Furthermore, it is involved in self-renewal of hematopoietic precursor cells [23, 24]. CD105 has also been described to be expressed on malignant cells in MDS and ALL [26, 27]. In AML, CD105 expression reportedly correlates with poor outcome, which held true even after hematopoietic stem cell transplantation [26, 28, 33]. Whereas distinct mRNA expression of the CD105 gene ENG was found in CLL cells , surface expression of CD105 in CLL has so far never been studied, and to our knowledge, we thus here provide the first report on CD105 surface expression in CLL.
Using flow cytometric analysis of CLL cells, we found CD105 positivity in about 70% of all patients. Similar prevalence was reported for AML, where 37–92% of cases were described as positive [26, 28] as well as for ALL with about 68% reportedly positive cases . Regarding expression levels (SFI), however, AML blasts in general exhibited higher amounts of protein (maximum SFI 65.2) compared to CLL cells (maximum SFI 28.9) as assessed by us using the same antibody for detection . Interestingly, healthy B cells, in contrast to leukemic B cells in CLL, do not express CD105 , which indicates that CD105 is acquired during and may play a role in malignant transformation.
Flow cytometric analysis of peripheral blood leukemic cells is a method that is easy to perform and increasingly used for diagnostic purposes. In CLL, identification of CD19+CD5+ cells within the peripheral blood constitutes the cornerstone of diagnostics. Flow cytometry is broadly applicable and established, and results are rapidly available at relatively low costs. Immunophenotyping of CLL cells not only reliably facilitates diagnosis, but also provides prognostic information by analysis of established markers such as CD38 and CD49d for risk stratification. The presently used panels can easily be adjusted by inclusion of additional prognostic surface markers.
Clinical evaluation according to Rai and Binet stages is long established to be prognostically relevant. However, clinical assessment is often vague and does not fully account for the broad range of disease courses in CLL. While some patients never exhibit symptoms and die eventually from causes unrelated to CLL, others display a steadily progressive disease. Therefore, it is of utmost importance to avoid both insufficient and excessive treatments.
CD105 mRNA expression in CLL reportedly correlates with shorter time to first treatment and a more aggressive clinical course of the disease. We found a significant correlation of CD105hi expression with both shorter OS and TTFT. This is in line with the weak association of other risk markers and high CD105 mRNA levels described in the literature . The strong correlation between CD105 expression and worse outcome in CLL observed in our study suggests that CD105 accelerates disease progression. However, the functional role of CD105 protein in CLL is yet not understood and should be subject to further investigation. Since the prognostic marker CD49d plays an important role in homing of CLL cells to secondary lymphoid organs , CD105 might be linked to interaction with endothelial cells and homing of CLL cells. Furthermore, CLL cells reportedly are capable of exerting immunosuppressive activity. For instance, B regulatory cell-like CLL cells secrete TGF-β and thereby transform naïve T helper cells into T regulatory cells . It is tempting to speculate that the TGF-ß coreceptor CD105 may also be involved in immunosuppression in CLL.
TGF-ß inhibits proliferation in healthy B cells and CLL cells. The latter can overcome this disadvantage by lowering the expression of TGF-ß receptors, thus reducing its antiproliferative effect [35, 36]. It seems possible that CD105 is involved in the process of evading TGF-ß effects, as it was described to modulate TGF-ß signaling, e.g., in T cells [37, 38]. This could at least partially explain the correlation of high CD105 expression on CLL cells and worse outcome.
The high prevalence and expression levels of CD105 as observed in adverse cases indicate that CD105 may constitute an attractive therapeutic target in CLL. Since an anti-CD105 antibody termed TRC105 is already evaluated in clinical trials in other disease entities and so far exhibits a promising safety profile [39,40,41], this antibody could readily be tested in CLL. Therapeutic effects could rely on both the blockade of CD105 signaling and recruitment of immune effector cells alike and ultimately improve and complement established regimens.
However, our study exhibits certain limitations. These include the low number of patients, a long interval between first diagnosis and CD105 testing, and missing data on prognostic factors such as IGHV mutational status and cytogenetics, thus precluding multivariate analysis. The association of CD105 expression and worse outcome in CLL should be further analyzed in cohorts eligible for multivariate analyses. Furthermore, indirect staining of CLL cells with our proprietary antibody clone might not be widely available, thus limiting the reproducibility of our study. The prognostic relevance of CD105 in CLL should be confirmed by flow cytometry data obtained with direct staining methods.
In summary, we first described CD105 expression on the surface of CLL cells and showed a strong correlation with worse outcome in CLL. There was an association of high CD105 expression with shorter TTFT and OS, possibly because of higher proportion of cases with unfavorable biological prognosis. CD105 might serve as a novel prognostic marker in CLL but confirmatory studies in larger patient cohorts and including multivariable analysis are necessary.