|Home | About | Journals | Submit | Contact Us | Français|
It was hypothesized that contact between chronic lymphocytic leukemia (CLL) B-cells and marrow stromal cells impact both cell types. To test this hypothesis, we utilized a long-term primary culture system from bone biopsies that reliably generates a mesenchymal stem cell (MSC). Co-culture of MSC with CLL B-cells protected the latter from both spontaneous apoptosis and drug-induced apoptosis. The CD38 expression in previously CD38 positive CLL B-cells was up-regulated with MSC co-culture. Up-regulation of CD71, CD25, CD69 and CD70 in CLL B-cells was found in the co-culture. CD71 up-regulation was more significantly associated with high-risk CLL, implicating CD71 regulation in the microenvironment predicting disease progression. In MSC, rapid ERK and AKT phosphorylation (within 30 min) were detected when CLL B-cells and MSC were separated by transwell; indicating that activation of MSC was mediated by soluble factors. These findings support a bi-directional activation between bone marrow stromal cells and CLL B-cells.
Chronic lymphocytic leukemia (CLL) has long been recognized as a result of the accumulation of clonal B-cells that are defective in apoptosis; however, isolated CLL B-cells will undergo relatively rapid apoptosis in vitro. This observation has led to the speculation that the microenvironment is necessary and/or plays a pivotal role in maintaining the enhanced survival of CLL cells in vivo. Human bone marrow stromal cells (BMSC) have been demonstrated to support the survival of CLL cells when both cell types were co-cultured in vitro (Kay, et al 2007, Lagneaux, et al 1998, Panayiotidis, et al 1996). Further investigation has suggested that CLL B-cells need to have intimate contact with BMSC in bone marrow (Lagneaux, et al 1998), T cells in lymph nodes (Kater, et al 2004), and nurse-like cells (NLC) in lymphatic tissues (Burger, et al 2000, Tsukada, et al 2002) to maintain survival. Direct contact between CLL cells and the stromal tissue or the NLC is probably necessary for CLL survival (Lagneaux, et al 1998, Lagneaux, et al 1999) and could play a pivotal role in disease progression.
Among the adverse prognostic factors for CLL patients, CD38 (Deaglio, et al 2003, Deaglio, et al 2007), ZAP-70 expression (Chen, et al 2002), and the immunoglobulin variable heavy chain region (IGHV) mutational status (Lanham, et al 2003) were shown to modulate signal transduction pathways in CLL B-cells, suggesting the activation status of CLL B-cells is important for disease prognosis. We now know that CLL is not a static disease that results simply from accumulation of long-lived lymphocytes but is rather likely to be a dynamic integrative process composed of cells that proliferate and die albeit at reduced levels compared to normal cells (Chiorazzi and Ferrarini 2006). It has been estimated that the CLL B- cell proliferation rate can vary from 0.1% to greater than 1% per day (Messmer, et al 2005). Damle et al (2007) recently demonstrated that CD38 expression detects an activated subset of CLL B-cells that is enriched in proliferating cells. However, the actual sites of in vivo proliferation for CLL B-cells remain to be clearly defined. Proliferation centres are present in lymph node, spleen and bone marrow, and they are likely candidates for sites of CLL proliferation. These proliferation centres contain prolymphocytes and para-immunoblasts, a higher proportion of which express the proliferation marker Ki67 in comparison to surrounding small lymphocytes (Caligaris-Cappio 2003, Lampert, et al 1999, Schmid and Isaacson 1994). Given that proliferating leukemic cells with enhanced survival are probably critical to CLL disease progression, we wished to test the hypothesis that bone marrow stromal cells can regulate or modify CLL B-cell activation and if there is bidirectional activation between stromal cells and leukemic CLL B-cells.
To do this, we developed a system to isolate and generate long-term cultures of bone marrow stromal cells from CLL patients (Kay, et al 2007). This current study further characterized the cultured bone marrow stromal cells as mesenchymal stem cells (MSCs), based on immunophenotype and differentiation capacity. We then tested the effect of CLL MSC on CLL B-cell activation. Important findings include that there is not only significant nurturing of both a lightly and tightly adherent subpopulation of CLL B-cells by MSCs but that there is “cross-talk” between MSCs and CLL B-cells, which results in activation features for both cell populations.
Blood and bone marrow biopsies were obtained from CLL patients who had provided written informed consent under a protocol approved by the Mayo Clinic Institutional Review Board according to the regulations of the Declaration of Helsinki. All CLL patients had a confirmed diagnosis using the National Cancer Intitute Working Group definition (Cheson, et al 1996). The 31 patients used in this study were from all Rai stages and had not been treated for at least 3 months prior to blood processing (Table 1). CLL cells were isolated from heparinized venous blood by density gradient centrifugation. Purified lymphocytes from CLL patients were cultured in AIM-V medium (Invitrogen, Carlsbad, CA) for the laboratory studies described below. This medium was used as we found that there was much less spontaneous apoptosis of CLL B-cells with AIM-V over one week of culture compared to other medium including RPMI with serum (data not shown).
Fresh bone marrow core biopsy samples were obtained from CLL patients (n = 14). Control bone biopsies were obtained from waste bone material of otherwise healthy patients (similar age range as CLL patients, data not shown) undergoing hip replacement. Primary bone biopsy stromal cell cultures were established using a previously published technique (Kay, et al 2007). With each passage, the BMSCs were given a passage number and the stromal cells used in the studies described here were from Passage 1 to Passage 6.
Approximately 5 × 104 stromal cells were then stained for surface expression of CD105, CD73, CD90, CD146, CD3, CD19, CD14, CD123, CD34, CD45 and HLA-DR by using directly conjugated phycoerythrin (PE), fluorescein isothiocyanate (FITC) or allophycocyanin (APC) antibodies (Becton-Dickinson BioSciences, San Jose, CA) and analyzed for reactivity using CELL-QUEST software.
To determine the ability of MSC to differentiate into adipocytes, chondrocytes or osteocytes, MSC differentiation kit (R&D system, Minneapolis, MN) were used according to the manufacturer’s instructions and previously published methods (Colter, et al 2001).
Freshly-isolated CLL peripheral blood mononuclear cells (PBMCs) were co-cultured with MSCs at a ratio of approximately 25–50:1. The purity of CLL PBMCs used in our in vitro experiments was typically greater than 90% (90.4 ± 3.7) CLL B-cells, as determined by staining for coexpression of CD19 and CD5. MSCs were cultured in 10% α minimal essential medium until confluence and then placed in AIM-V at 24 h prior to the addition of CLL B-cells. For spontaneous apoptosis assays, CLL B-cells were co-cultured with MSCs for 5 days. For analysis of drug-induced cell death, fludarabine (2-Fluoroadenine-9-β-D-arabinofuranoside, F-ara-A, Sigma) was added at various doses (10 μM, 25 μM and 75 μM) for 48 h as described below. Leukemic cells from the suspension (designated as the light-adherent fraction, [LA fraction]) were then collected and examined for viability/apoptosis status by staining with Annexin V- FITC and propidium iodide (PI) as previously described (Lee, et al 2004, Shanafelt, et al 2005). Lymphocytes adhering to the stromal layer (designated as the tight adherent fraction, [TA fraction]) were obtained by trypsinizing the co-culture; the cells were then collected analysed for viability/apoptosis by staining with Annexin V-FITC and PI.
Lymphocytes (5–10, 000 events) were distinguished by analysis of forward and side scatter and CD19 staining as collected by using CELL-Quest software. In some experiments, MSCs cultured alone were detached from the plate by trypsinization, and their survival was also analyzed by Annexin V-FITC and PI staining.
The surface expression of CD38, CD25, CD69, CD71, and CD70 on fresh CLL B-cells was tested with flow cytometry using APC-CD38, PE-CD25, FITC-CD69, FITC-CD71, PE-CD70, FITC-CD19 and APC-CD19 (BD Biosciences, San Jose, CA) before co-culture with MSCs. After co-culture in AIM-V for 2 weeks, at ratios from 25–50:1 between CLL B-cells and MSC, where it was either through direct contact or through a 0.4 μ-pore size transwell, the surface expression of CD38, CD25, CD69, CD71 and CD70 on CLL B-cells (CD19+) was then examined with flow cytometry and the percentage of positive cells were determined. These values were compared to baseline percentage values of the CD antigens on CLL B-cells determined prior to culture. The total mean fluorescence intensity (MFI) of each surface marker or isotype control was determined by analyzing viable CD19+ cells.
Fresh isolated CLL B-cells were labelled with carboxyfluorescein diacetate succinimidyl ester (CFSE) (Molecular Probes, Invitrogen, Carlsbad, CA) before the co-culture. Subsequently, CLL B-cells were cultured in the presence or absence of MSCs for 5 days. At the end of the culture period, CLL B-cells that might be in proliferation were analyzed with flow cytometry by assessing for CD19+ cells, which have decreased CFSE labeling. The percentage of proliferating CD19 negative (−) lymphocytes was also determined in a similar fashion.
The following parameters were analyzed for each patient: age, gender, disease stage at diagnosis according to modified Rai criteria, CD38 expression, ZAP-70 expression (Rassenti, et al 2004), IGHV gene mutational status (Jelinek, et al 2001), CLL FISH analysis (Dewald, et al 2003), and history of therapy.
MSCs were cultured until reaching 80% confluence and the initial medium was replaced with AIM-V for 24 h. CLL B-cells were then added to the cultured MSCs at a ratio from 20–50:1 for direct co-culture or separated by a transwell (0.4 μ-pore size) for different time periods (typically 30 min to 24 h). CLL B-cells were then collected by vigorous pipetting from these co-cultures to minimize residual CLL B-cells, washed in phosphate-buffered saline (PBS) and lysed in 1% Nonidet P-40 (NP-40) lysis buffer. MSCs were also washed once with cold PBS and then lysed in 1% NP-40 buffer. Approximately 5 μg of protein from MSC cell lysate or 20–25 μg protein from CLL lysates was used for Western blot analysis as described before(Kay, et al 2007). The membranes were probed with antibodies specific for various signaling proteins (i.e., anti-pErk, anti-pAkt, anti-Erk, anti-Akt, Cell Signaling Technology, Danvers, MA) and anti-apoptotic proteins (i.e., anti-Mcl-1 from Chemicon, anti-XIAP and anti-Bcl-2 from BD Pharmingen/Transduction Laboratories). To confirm equal loading of the blots, the membrane was re-probed with monoclonal β-actin antibody (Novus, Littleton, CO). In control experiments, CLL B-cells were harvested from transwell co-culture experiments with MSC in order to immunoblot for Akt and Erk activation in CLL B-cells. The condition medium (CM) of CLL PBMCs were generated by collecting the supernatant of CLL PBMC (purity: higher than 95% CLL B-cells by CD19/CD5 staining) cultured for approximately 4 days in AIM-V at a concentration of roughly 5 × 106/ml.
Statistical significance differences were analyzed by paired Student’s t test or non-parametric Wilcoxon matched pairs test using two tails design for comparing the means of the cell survival, surface activation marker expression and cell proliferation where appropriate in experiments. The statistical significance of the observed differences in the upregulation of cell surface antigens in different prognostic groups was deduced by means of the Mann-Whitney test. P values ≤ 0.05 were considered significant.
Bone marrow stromal cells (BMSC) were generated from the core biopsies of CLL patients as previously described (Kay, et al 2007). Before the first passage, the stromal cells were labeled as Passage zero (P0). Along with each passage, the stromal cells were labeled accordingly based on their passage number. Fig. 1A illustrates the stromal layer morphology at the P0 and P1 passage, respectively. To further determine if these were functional MSC, BMSCs were subsequently tested for their differentiation capacity by culturing the BMSCs in the differentiation medium for adipocytes, osteocytes and chondrocytes. At the end of differentiation culture, cells were stained with Oil red, Alizarin red or Alcian blue for determination of differentiation into adipocytes, osteocytes or chondrocytes lineage (Fig. 1B) (Colter, et al 2001). BMSCs displayed the features of differentiated adipocytes, osteocytes and chondrocytes when placed in appropriate differentiation medium (Fig 1). BMSCs were positive for CD90, CD73, CD105 and CD146, but negative for CD45, CD34, CD14, CD123, CD19, and CD3 (Fig. 1C and Table 2). This immunophenotype is entirely consistent with a mesenchymal stem cell (MSC) phenotype (Phinney and Prockop 2007).
The stromal cells generated in our culture model thus met the criteria of MSC (Dominici, et al 2006) in that they were adherent in nature, expressed a typical immunophenotype and were capable of differentiation into several lineages. MSCs from healthy controls were found exhibit the same immunophenotype as CLL MSCs (Table 2) and had the same culture passage capacity (data not shown).
Having established that the marrow stromal cells are MSC in terms of phenotype and function as generated from the core bone biopsies of CLL patients, we next tested if MSCs could protect CLL B-cells from undergoing spontaneous apoptosis. CLL PBMCs were used in co-culture with either CLL-MSC (cMSC) or normal MSC (nMSC) for 5 days. Interestingly, two cohorts of CLL B-cells were found among the co-culture; a light adherent (LA) fraction, which was easily collected from the supernatant and a tight adherent (TA) fraction, which could be collected only after detaching the adherent stroma-CLL layer with trypsinization. Spontaneous apoptosis was significantly reduced for the MSC-exposed CLL B-cells from both the LA and TA fractions compared to those cultured in medium alone (Fig. 2A). Importantly, normal MSC and CLL MSC were able to rescue CLL B-cells based on the leukemic cell’s apoptotic levels to a similar degree.
We subsequently evaluated the survival of CLL B-cells in the presence or absence of MSCs where fludarabine (F-ara-A) was present for 48 h. The CLL B-cells in the TA fraction following co-culture with both normal and CLL MSC had significantly more resistance to fludarabine compared to CLL B-cells cultured in medium alone (p<0.001, Fig. 2B). There was also increased resistance to fludarabine for CLL B-cells in LA fractions compared to CLL B-cells cultured alone (p=0.011 for LA-nMSC, p=0.001 for LA-cMSC, Fig. 2B). Moreover, TA CLL B-cells had significantly increased resistance to fludarabine compared to LA CLL B-cells (p<0.001, Fig. 2B). Of note, fludarabine did not induce significant apoptosis in MSC when added alone to the MSC, suggesting that the ability of MSC to function is probably not compromised even in the presence of this drug. Thus, in the presence of fludarabine, MSC remained alive at levels comparable to MSC without exposure to the drug as determined by Annexin-PI analysis throughout the co-culture periods (data not shown).
In order to explore the potential mechanisms for the apoptosis protection of CLL cells mediated by MSC, we tested the expression of Bcl-2 family proteins including Mcl-1, XIAP and Bcl-2 in the CLL cells harvested after 24 h of co-culture with MSC. As expected, the expression of Mcl-1 and XIAP were elevated after co-culture compared to the basal level of expression (Fig. 2C). Bcl-2 expression in CLL cells after co-culture appeared to be equivalent to the basal level. The expression of phosphorylated Akt in CLL cells after co-culture with MSC for 30 min, 1 h, 16 h and 24 h were tested. Akt activation did not appear to be different from basal level of expression (Fig. 2C).
To evaluate the role of MSCs on CLL B-cell activation, initially CD38 expression was investigated on CLL B-cells from 23 patients before and after co-culture with CLL-MSCs. Among the 23 CLL patients, 8 patients had a baseline CD38 expression of less than 1%. Of these 8 CLL samples, all but 1 patient had no significant increase of CD38 expression after co-culture with CLL-MSC (p= 0.312, data not shown). However, 15 CLL patients with initial CD38 expression of more than 1% of CLL B-cells did have a significant increase in the percentage of the CD38+ CLL B-cells after 2 wk of co-culture (p=0.006, Fig. 3A). Among the 15 CLL samples, 5 patients had more than a 3-fold increase of CD38 percent expression. The remaining 9 CLL patients had modest (< 2-fold) increases of CD38 percentage. The average increase of CD38 percentage for all 15 CD38+ (i.e., greater than 1% CD38 + at baseline) CLL B-cells was found to be 1.7-fold. The MFI of CD38 expression in CLL B-cells with an initial CD38 expression of more than 1% increased significantly when directly co-cultured with MSCs (p=0.008, Fig. 3B). The increased CD38 activation needed direct interaction between MSC and CLL B-cells. Separation of these 2 cell types by transwell resulted in non-significant changes in either CD38 percentage or MFI expression (p=0.141 for CD38 percent change, p=0.126 for CD38 MFI change, data not shown). We believed that the less than 1% positive CD38+ cells were truly negative and not able to respond to MSC activation with increases in CD38. That any CD38+ CLL B-cells are a truly positive cell is supported by our earlier finding that only CD38+ cells respond to interefron γ with increases in CD38 levels (Pittner, et al 2005) and the finding that the presence of a distinct CD38+ population, irrespective of its size within the leukemic clone, correlates with IGHV mutational status and identifies CLL patients with more progressive disease (Ghia, et al 2003).
Some CLL B-cells have been demonstrated to express a phenotype of activated B-cells (Damle, et al 2002). To investigate if MSCs generated from CLL patients were capable of modulating other activation markers, CD71, CD69, CD25 and CD70 expression levels were evaluated on CLL B-cells pre and post co-culture with MSCs. Indeed, CD71, CD69, CD25 and CD70 expression increased significantly after 2 weeks of co-culture of CLL B-cells with CLL-MSCs (Fig. 4A–D). Of interest, co-culture of MSC with CLL B-cells though transwell mediated the upregulation of CD71 and CD69 to a similar degree as direct interaction, suggesting these activation antigens were mediated by soluble factors. However, CD25 and CD70 activation, when assessed after transwell-separation of the CLL B-cells and MSCs, demonstrated non-significant increases of both CD25 (p= 0.077, Fig. 4C) and CD70 (p= 0.105, Fig. 4D) on the CLL B-cells. This latter activation was seen in both CD38+ and CD38- CLL B-cells implying that these above activation antigens are modulated in CLL B-cells by different mechanisms from CD38.
Having demonstrated that MSCs are capable of generating a more generalized activation phenotype of CLL B-cells, we next evaluated the effect of MSC on the in vitro proliferation of CLL B-cells. Freshly isolated PBMCs from CLL patients were cultured for 5 days in the presence or absence of CLL-MSC. PBMC were labeled with CFSE immediately prior to co-culture with MSC. After CLL cells were cultured with or without MSC for 5 days in the absence of stimulation, cell proliferation was assessed by analyzing the decreased CFSE staining. There were variable levels of proliferation detected in CLL cells with a general trend towards increasing proliferation in CD19+ cells in the mixed population (p=0.01, n=17, Fig. S1A). Of interest, among the 17 CLL patients tested, 3 samples had approximately a 2-fold increase of CLL B-cell proliferation. Another 2 samples had more than 1.5-fold increase of CLL B-cell proliferation. Finally, there were 2 CLL patients who appeared to have a decreased level of proliferation. However, we found that the average increase of CLL proliferation after co-culture with MSC was subtle (1.3-fold). In contrast, the proliferation level of the CD19− lymphocytes appeared to be suppressed by MSC interaction (p=0.03, n=17, Fig. S1B). Thus, seven out of 17 samples had approximately 50% reduction in the proliferation of CD19- lymphocytes with an average fold change for the CD19− lymphocytes of 0.6. These findings are consistent with prior work on MSC showing suppression of T cell function (Bacigalupo, et al 2005, Meisel, et al 2004). These data do suggest that the observed increase of CD38 is a result of the combination of increased surface expression of CD38, as demonstrated by the increase of CD38 MFI on CLL B-cells, and is probably associated with the increased proliferative capacity of the co-cultured CLL cells.
In order to determine if the bone marrow microenvironment modulates the cell activation differently in different CLL B-cell subsets, we tested if the upregulation of CD38, CD71, CD69, CD25 and CD70 were different when the tested CLL patients were defined by ZAP-70 expression, CD38 expression or IGHV mutation status. We used the commonly accepted 20% expression in CLL B-cells as the cut-off for ZAP-70 positive expression and 30% expression in CLL B-cells as the cut-off for CD38 positive expression. The CLL cases were classified as unmutated if their IGHV differed by less than 2% from the germ line gene sequence. Indeed, we found that fold increase of CD71 percent expression after co-culture with MSC demonstrated significant increases in the CLL patients where high-risk parameters were present (ZAP-70 positive, p = 0.02, CD38 positive, p = 0.05 and IGHV unmutated, p = 0.05, Fig. 5A) compared to the CLL patients where these parameters were absent. The fold change of other activation markers including CD38 after co-culture did not appear to be different when compared to the prognostic parameter status of the leukemic clone (Fig. 5B–D).
We next evaluated whether CLL B-cells were able to regulate the activation status of MSC. In order to test this, we assessed the expression of Erk and Akt in the MSC after their exposure to CLL B-cells through either direct co-culture or separated from the CLL B-cells by transwell. Indeed both Erk and Akt were found to be phosphorylated in MSC within 30 min of exposure to CLL B-cells (Fig. 6A – B). There was also robust activation of Erk and Akt in MSC in the transwell cultures, suggesting that this activation could be mediated effectively by soluble factor(s) released from CLL B-cells (Fig. 6B). In addition, when MSCs were exposed to the condition medium of cultured CLL cells, potent activation of both Erk and Akt were detected within 30 min (data not shown). These results confirmed that soluble factors released from CLL B-cells were capable of inducing activation of the signal molecules Akt and Erk in MSC.
In order to exclude the possibility that CLL cell lysates were affecting the Erk and Akt levels seen in MSC when both cell types were directly co-cultured, the CLL cells were collected as stringently as possible from both the supernatant and the MSC monolayers of the co-cultures. After vigorous washing of MSC to collect more tightly bound CLL B-cells, CLL lysates were made from the harvested cells obtained from the LA and TA fractions. We then tested these lysates in different protein concentrations (25 μg from CLL lysates vs. 5 μg from MSC lysates, respectively) for a possible activation status of Erk and Akt protein. Erk (data not shown) was not phosphorylated in the CLL lysates, making it highly unlikely that the Erk phosphorylation seen in MSC was coming from CLL B-cells. Basal Akt phosphorylation was detected from CLL lysates. However, the level of phosphorylated Akt in CLL lysates at both baseline and after coculture with MSC appeared much weaker compared to Akt activation detected in activated MSC, making it unlikely that Akt activation in MSC lysates is a pure contaminant of CLL lysates.
This report further extends our analysis of bone biopsy cultured stromal cells from CLL patients (Kay, et al 2007), their nature and the extent of interaction and cross-talk between the bone marrow stromal cells and the leukemic CLL B-cells. We have uncovered novel features of this interaction. First, we find that bone biopsy cultured stromal cells are MSC, and these marrow-derived MSC can enhance the survival of two cohorts of CLL B-cells: the LA and TA components. Additionally, it was found that the MSCs can activate resting CLL B-cells to increase their expression of CD38, as well as promote activation of CD71, CD69, CD25 and CD70. This latter phenomenon can occur by potentially different mechanisms. Moreover, the CD71 upregulation after co-culture was significantly increased in ZAP-70, CD38 positive and IGHV unmutated CLL B-cells. An additional unique finding here is that the interaction between MSC and CLL B-cells is not just one-way but that robust levels of activation of both Erk and Akt can occur in MSC by direct contact with CLL B-cells and their soluble factors. We believe that these findings have implications for further understanding of the CLL B-cell disease process.
Although CLL is primarily a disease of accumulating clonal B-cells defective in apoptosis, CLL cells also rely on the microenvironment to maintain survival. In this study, we characterized the bone marrow stromal cells generated from bone biopsies in our system. These cells exhibited the typical features of mesenchymal stem cells: adherent in nature, typical immunophenotype (CD90+, CD73+, CD105+, CD146+ and negative for other hematopoietic and immunological markers) and with the capacity to differentiate into three lineages - adipocytes, osteocytes and chondrocytes. To our knowledge, this is the first study to define the MSC features of cultured CLL bone marrow stromal cells and MSC represents a unique stromal cell system to study. As an adult stem cell, MSC is known to be present in multiple tissues including bone marrow, tumor stroma, adipose tissue, amniotic fluid, fetal liver and lung, etc.(Campagnoli, et al 2001, De Ugarte, et al 2003, Erices, et al 2003, in ‘t Anker, et al 2003, Tsai, et al 2004). Importantly, we have found similar immunophenotype and functions of MSC isolated from lymph node and spleen of CLL patients (data not shown) and speculate that MSC interaction with CLL B-cells is widespread in secondary lymphoid tissue. Although this study has not found significant differences between normal MSC and CLL MSC in terms of their interaction with clonal B-cells (data not shown), we continue to assess both sources of MSC to more completely investigate this aspect.
MSC from both normal healthy donors and CLL patients were able to protect CLL B-cells from undergoing spontaneous and drug-induced apoptosis. In our study, CLL B-cells in suspension (designated as LA) or in tight contact with MSC (TA) were protected from spontaneous apoptosis. However, under fludarabine exposure, the TA fraction of CLL B-cells appeared to have more protection from apoptosis compared to LA cells. These results suggested to us that there are important differences in how the CLL B-cell can interact with a stromal cell. Thus, close contact between CLL B-cells and MSC is capable of mediating the most effective drug-resistance, and it is this latter interaction that could be the most important in providing a niche for residual CLL B-cells post treatment. Recent studies demonstrated the significance of CD49d (α4 integrin) in the prognosis of CLL disease (Gattei, et al 2008, Shanafelt, et al 2008), as well as its biological role of regulating matrix metalloproteinase-9(Redondo-Munoz, et al 2006, Redondo-Munoz, et al 2008). These observations imply that α4β1 integrin could be a critical mediator of tight interactions between MSC and CLL B-cells and MSC-mediated CLL protection ( Nowakowski et al 2005). We also believe there may be a dynamic cycling between LA and TA cells when interacting with MSCs. We speculate that, TA cells, after tight contact with MSC, may become LA cells and this scenario could mimic the proposed in vivo cycling of CLL cells between bone marrow and peripheral blood (Chiorazzi 2007, Chiorazzi and Ferrarini 2006).
We next explored the potential mechanisms of the apoptosis protection of CLL cells mediated by MSC. As expected, the Bcl-2 family proteins including Mcl-1 and XIAP in CLL B-cells were found to be up-regulated by MSC. However, AKT activation in CLL cells after co-culture did not change compared to baseline AKT phosphorylation in leukemic cells. These results suggest that the apoptosis protection mediated by MSC to CLL cells is probably through up-regulation of these two rapidly modified anti-apoptotic protein Bcl-2 family members. We do not think that CLL B-cell Akt activation is involved in the apoptosis protection found for CLL B-cells when cultured over 24 h r with MSC.
Survival was not the only biological feature to be altered on CLL B-cell interaction with MSC, as increases were also observed in certain membrane receptors including CD38. CD38 expression on CLL B-cells can be regulated by activated T cells (Patten, et al 2008) and cytokines, e.g., interferon γ (Pittner, et al 2005) and interleukin-2 (Deaglio, et al 2003). CD38 expression is higher in secondary lymphoid tissues containing pseudofollicles (Patten, et al 2008), and its expression is higher in lymph node than in peripheral blood in CLL patients (Jaksic, et al 2004). We found that co-culture of marrow MSC with CLL cells could induce CD38 activation but only in CLL B-cells that were clearly CD38 positive. This is consistent with our prior findings (Pittner, et al 2005) and with the data reported by Ghia et al (2003) that any CD38 positivity may confer high risk on CLL patients in terms of disease progression. In this current study, we detected both CD38 percentage and MFI increase in CLL B-cells after their co-culture with MSC. CLL B-cells were also found to demonstrate very low level of CFSE staining without co-culture of MSC, consistent with the published low rate of proliferation of CLL cells in vivo (Messmer, et al 2005). After CLL cells were co-cultured with MSC for 5 days, there were variable levels of proliferation detected in CLL cells with a general trend towards increasing proliferation in CD19+ cells in the mixed population. In contrast we found that the proliferation for the CD19− cells appeared to decrease. These data support that the observed increase of CD38 is a result of the combination of increased surface expression of CD38 and is probably associated with the increased proliferative capacity of the co-cultured CLL cells. In support of our observations, Patten et al (2008) have recently demonstrated that CD38 in CLL cells increased after co-culture with activated T cells. Our finding supports the notion that CD38 on CLL B-cells can be regulated in the tissue microenvironment. However, our system is different from the co-culture of CLL cells with activated T cells because we did not use any stimulus in the co-culture system of MSC with CLL cells. This is probably more robust proliferation of CLL cells was not detected, and the level of proliferation varied among individual leukemic samples. We understand these finding are preliminary and probably will need more testing of this aspect in a larger group of CLL patients. Our finding is in support of the notion that some CLL B-cells are capable of proliferating and being activated in the bone marrow microenvironment after appropriate interaction with different cellular components.
We also found that several other activation markers including CD71, CD69, CD25 and CD70 can be increased with MSC co-culture. It is unclear if the mechanisms for increases in these membrane antigens are linked to a common signaling pathway but it is of interest that CD71 was the only one that was found more often to be significantly increased in high-risk CLL as defined by novel prognostic factors. The observed increases in CD71, CD69, CD25 and CD70 expression irrespective of the CD38 status of CLL B-cells does suggest that the mechanism of action for those surface antigen inductions are separate from that of CD38. CD71 is transferrin receptor 1 and its expression is up-regulated in a relatively late fashion compared to CD69 in normal B-cells upon stimulation (Biselli, et al 1992). CD71 was demonstrated to be elevated in CLL patients compared to normal B-cells (Damle, et al 2002, Smilevska, et al 2006) and the percentage of CD71 expression is higher in mutated CLL cases (Damle, et al 2002). Interestingly, we found that the CD71 fold increases in CLL B-cells after MSC co-culture was significantly increased in CLL patients with high-risk features. More specifically, it was seen on CLL B-cells post co-culture with MSC in ZAP-70 positive, CD38 positive and IGHV unmutated patients vs. patients with negative prognostic features. These results imply that CD71 upregulation by bone marrow stroma may reflect an activated phenotype for CLL B-cells that could be related to disease progression. Indeed recent studies have identified that CD71 is a target gene of c-Myc and is critical for enhancing cellular proliferation and tumorigenesis (O’Donnell, et al 2006). In addition, we found approximately 6 CLL cases with high-risk features who had only modest CD71 upregulation after MSC co-culture. Interestingly, FISH analysis of these 6 cases revealed mostly favorable cytogenetic characteristics based on the published criteria (Dewald, et al 2003) (data not shown). However, further study is needed to investigate the role of CD71 in CLL activation and its relationship with other prognostic factors that are associated with CLL disease progression.
Another unique feature of this work is our finding that the activation of CLL B-cells was accompanied by activation of MSC. Thus, within 30 min of exposure to CLL cells, both Erk and Akt were phosphorylated in MSC. Given that the conditioned medium from CLL B-cells can induce very robust Erk and Akt in MSC, this implies that soluble factors released from CLL cells are likely to be largely responsible for this signal activation. Could this indicate that once CLL B-cells are activated their soluble factors not only mediate local but distant tissue site activation of stromal tissue? This could facilitate more favorable tissue site interactions for the leukemic B-cell, and this finding is consistent with our earlier report that MSC can undergo an angiogenic switch known to be favorable for CLL disease progression (Kay, et al 2007). In addition, the activation of MSC is possibly a critical step for further proliferation, migration and differentiation of MSC. We are also now in the process of determining if the induction of signaling pathway activation in MSC could modulate MSC function to subsequently provide an even more nurturing microenvironment for the CLL B-cells.
In summary, we believe, these results add to the evidence that in human tumors, including hematological malignancies, tissue site interaction with the tumor cell significantly impacts the critical features of both the stromal cell and the malignant cell. It is hoped that further interrogation of the microenvironment, using relevant models for this in CLL patients, will generate important insights both to understand disease progression and provide new clues for treatment.
This work was supported by a grant from the CLL Global Research Foundation, NIH grant, CA-1116237. We are grateful for the philanthropic support from Mr. Edson Spencer and the Donner Family Foundation. We acknowledge Dr. Prockop and his team’s support for assistance in the characterization of MSCs.
Conflict-of-interest disclosure: The authors declare no competing financial interests.
W.D. designed and performed the experiments, analyzed and interpreted the results, and wrote the paper. G.S.N designed the research, interpreted the results and provided statistical help. M.L.M, T.R.K, L.E.W performed the research. S.M.S and W.W provided the clinical and statistical support. A.B.D designed the research and interpreted the results. A.K.G, C.R.S, K.L.M, T.D.S, C.S.Z, T.G.C provided discussion and technical or clinical support. N.E.K designed research, interpreted the results and wrote the paper.