|Home | About | Journals | Submit | Contact Us | Français|
The membrane glycoprotein MRC OX-2 (CD200) is expressed in several lymphoid malignancies. However, the diagnostic utility and potential prognostic importance of CD200 expression have not been rigorously examined. We show that CD200 is uniformly expressed in chronic lymphocytic leukemia (CLL) and absent in mantle cell lymphoma (MCL). Importantly, expression of CD200 is retained even in those CLLs with immunophenotypic aberrancies, making CD200 a particularly useful marker for discrimination between these cases and MCL. CD200 is expressed in nearly all precursor B lymphoblastic leukemias, with aberrant over- or underexpression when compared to normal B-cell progenitors in 55% of cases. Over 70% of plasma cell myelomas (PCMs) expressed CD200, and loss of CD200 expression in PCM may be associated with more clinically aggressive disease. In summary, CD200 is expressed in several hematolymphoid neoplasms. Analysis of its expression has several diagnostic, and potentially prognostic, applications in the flow cytometric evaluation of lymphoid malignancies.
Membrane glycoprotein MRC OX-2 (also known as MOX2 or CD200; hereafter referred to as CD200) is a member of the immunoglobulin (Ig) superfamily and is encoded by a gene residing at chromosome 3q12.1 CD200 is normally expressed by endothelial cells and neurons and by B-cells and a subset of T-cells within the hematolymphoid system.2 CD200 interacts with the CD200 receptor (CD200R), which has limited expression on granulocytes (neutrophils and basophils), monocytes and cells of the reticuloendothelial system, including dendritic cells and macrophages.3,4 Under most circumstances, CD200-mediated ligation of CD200R delivers an inhibitory signal to the target cell.4-6 Consistent with this, macrophages from CD200-deficient mice display an activated phenotype, and these mice manifest enhanced susceptibility to experimentally induced autoimmune disease.7
Expression of CD200 in hematologic malignancies was first reported for chronic lymphocytic leukemia (CLL).8 CD200 subsequently has been shown to be differentially expressed in CLL and mantle cell lymphoma (MCL); thus, evaluation of CD200 expression may have diagnostic utility for these CD5+ B-cell lymphomas.9 In addition, CD200 expression, as assessed primarily by gene expression profiling, has been associated with a superior clinical outcome in acute myeloid leukemia and plasma cell myeloma (PCM).10,11
In this study, we report the analysis of CD200 expression by flow cytometry in several types of lymphoid malignancy. We confirm the differential expression of CD200 in CLL and MCL and importantly show that such differential expression is maintained in those lymphomas with an atypical or indeterminate immunophenotype. We also show that while its expression is nearly uniform in precursor B-lymphoblastic leukemia (B-ALL), there is aberrant under- or overexpression of CD200 in most cases of B-ALL when compared to normal bone marrow B-cell progenitors. Finally, we demonstrate that although absent in normal bone marrow PCs, CD200 is expressed in 70% of PCMs, and negativity for this marker may be associated clinically with progression to more aggressive disease.
Routine clinical bone marrow, peripheral blood, body fluid and lymph nodes samples arriving in the clinical flow cytometry laboratory at the University of Arkansas for Medical Sciences, Little Rock were included in this study. Samples were screened morphologically and immunophenotypically to confirm the presence of neoplastic cells. Samples were then further analyzed for CD200 expression by flow cytometry as described below. Lymphoid malignancies were diagnosed as per the criteria delineated in the 2008 World Health classification.12 This study was approved by the institutional review board at the University of Arkansas for Medical Sciences.
EDTA- or heparin anticoagulated peripheral blood or BM aspirate specimens were washed with PBS and resuspended in PBS containing 2% fetal calf serum. Cell suspensions were then incubated for 15 minutes with various cocktails of fluorochrome-conjugated antibodies. These included CD200 PE, CD19, CD20, CD10, CD5, CD34, CD138 and CD38. All antibodies were obtained from BD Biosciences (San Jose, CA). Isotype controls were included in all analyses. Red blood cells were then lysed with either FACS Lyse (BD Biosciences) or ammonium chloride solution, rinsed with PBS, and resuspended in PBS containing 1% formaldehyde. Analysis was performed on a FACSCanto II flow cytometer using FACSDiva software (BD Biosciences). Gating on lymphoid cells and lymphoblasts was based on CD45 versus side scatter analysis. PCs were identified based on CD138 versus side scatter analysis. CD200 expression was evaluated semiquantitatively by comparison with the isotype PE control antibody and designated as either negative or 1+ (< 1 log shift in mean fluorescence intensity [MFI] compared to isotype control), 2+ (1-2 log shift in MFI) or 3+ (more than 2 log shift in MFI).
In PCMs where gene expression profiling (GEP) was performed, RNA was extracted from CD138 purified PCs and analyzed using an Affymetrix-based platform with hybridization to U133 Plus2.0 microarrays. Molecular classification and derivation of a 70 gene risk score was performed as previously described.13,14 Molecular classification of a given case was achieved through the use of a class predictor. This predictor is comprised of significantly overexpressed and underexpressed genes for each molecular subgroup and was generated using prediction analysis for microarrays (PAM) software (PAM, Stanford, CA), as described.14 GEP-based assessment of MOX2 expression, the gene encoding CD200, was achieved by analysis of the two probe sets for MOX2 (209582_s_at; 209583_s_at) which are present on the U133 Plus2.0 microarray.
For the analysis of plasma cell myeloma cases, a linear model was fit to assess differences in the mean GEP 70-gene score among the three CD200 groups (CD200 negative, CD200 1+, and CD200 2+). To compare the distribution of each of the myeloma molecular subtypes (CD1, CD2, HY, LB, MF, MS, PR), a chi-square test was performed for each molecular subtype to determine whether there was a significant difference in the distribution of that particular molecular subtype within each of these CD200 expression groups.
Analysis of normal peripheral blood and lymph node demonstrated somewhat variable CD200 expression in B- and T-cells. In blood, CD3+ T-cells were generally CD200−, whereas CD19+ B-cells were usually CD200+, with uniform positivity in some cases and more variable expression in others. In benign or reactive lymph node, similar patterns of CD200 expression were observed with relatively uniform expression in B-cells and more heterogeneous expression in T-cells (Image 1).
Expression of CD200 was evaluated in several hematolymphoid malignancies by flow cytometry, the results of which are summarized in Table 1. While T-ALLs were uniformly negative for CD200, 95% of B-ALLs analyzed were variably positive for CD200. The relative level of CD200 expression was not tightly associated with any particular cytogenetic abnormality (data not shown). CD34+ B-ALLs expressed a significantly higher level of CD200 than did CD34− cases (average MFI of 1950 versus 686). Although brightly expressed in many B-ALLs, the level of expression of CD200 was variable with some being only dimly CD200+ (Image 2). These observations led us to assess the expression of CD200 in normal bone marrow B-cell progenitors or hematogones.
Stages of hematogone development are defined by the differential expression of CD10 and CD20: stage 1, CD10bright/CD20−; stage 2, CD10+/CD20variable; stage 3, CD10+/CD20bright.15,16 As shown in Table 2, CD200 was expressed at a relatively constant level throughout all stages of hematogone development. Interestingly, the highest level of CD200 is expressed in CD34+/CD19+ B-cell progenitors, which normally comprise a minor subset of CD10bright/CD20− cells and represent the most immature B-cell progenitors. Thus, when compared to normal CD34+/CD19+ B-cell progenitors, CD200 is significantly over- or underexpressed (MFI at least 2-fold greater or less than that of B-cell progenitors) in 55% of CD34+ B-ALLs.
Confirming previous reports,9 CD200 expression was observed in 100% of CLLs and was uniformly absent in MCL (Table 1). CD200 expression was bright in all CLL cases analyzed, with at least a one log shift in the MFI compared to that of the isotype control antibody (Image 3a). Although most CLLs and MCLs manifest immunophenotypes that permit their reliable distinction by flow cytometry, an aberrant immunophenotype is observed in a minority of cases, precluding a definitive diagnosis on the basis of flow cytometry alone.17-20 Therefore, CD200 expression in such cases was evaluated to determine whether it might enhance diagnostic accuracy by flow cytometry. Two cases of CLL with either aberrantly bright CD20 or surface Ig light chain expression were both strongly CD200+ (Image 3B). In addition, one case of MCL which aberrantly coexpressed CD23 was CD200−. Thus, the differential CD200 expression in CLL and MCL appears to be retained even in those cases with an otherwise indeterminate immunophenotype.
To evaluate CD200 expression in normal PCs, bone marrow samples from 4 patients without a history of a PC dyscrasia were analyzed. Normal PCs were uniformly negative for CD200 (Image 4). In contrast, CD200 was expressed by 71% of analyzed cases of PCMs (Table 1) similar to previous reports.10 CD200 positive cases manifested variable levels of expression with approximately half showing weak expression and the remaining cases being moderately-strongly positive (Image 3).
At our institution, gene expression profiling (GEP) is routinely performed in new patients with PCM, from which a molecular subtype is assigned and a 70-gene risk score is derived, the latter being highly predictive of clinical behavior.13,14 GEP-defined molecular subtypes include: CD-1 and CD-2, in which there is dysregulated CCND1 expression due to the t(11;14); hyperdiploid (HY), characterized by gains primarily of odd-numbered chromosomes; MF in which there is dysregulated expression of either MAF or MAFB; low bone (LB), associated with low expression of DKK1 and characterized clinically by fewer MRI-defined focal lesions; MS, characterized by dysregulated FGFR3 and MMSET expression due to the t(4;14); and proliferative (PR), characterized by overexpression of several genes associated with cell cycle regulation and proliferation.
GEP analysis was also performed in 49 of the cases which had been evaluated by flow cytometry. As shown in Table 3, there is good correlation between the expression of MOX2 mRNA, as assessed by GEP analysis, and that of CD200 protein detected by flow cytometry. Furthermore, the relative level of CD200 expression is associated with molecular subtype. Over 90% of LB and HY PCMs were CD200+, with uniformly high CD200 expression among LB PCMs. Somewhat surprisingly given previously published data10, 7/9 (78%) cases of PR PCM lacked CD200 expression, and as such, comprised 50% of CD200− cases. The difference in the frequency of LB PCM and PR PCM within the CD2002+ and CD200− groups, respectively, was highly statistically significant (p < 0.009 and p < 0.001, respectively). By contrast, the frequency of other myeloma subtypes (e.g., CD-2 and MS) was not statistically different among the three CD200 subgroups, indicative of more variable CD200 expression.
Most PR PCMs harbor genetic lesions conferring aggressive clinical behavior and are associated overall with a significantly higher 70-gene risk score than other molecular subtypes.13,21 Not surprisingly then, CD200− PCMs had a significantly higher gene risk score than those with 2+ CD200 positivity (p <0.0006; Image 5). There is no reported data on the stability of CD200 expression in PCM, and we were not able to serially evaluate CD200 expression by flow cytometry. However, 27 of the patients in our series did undergo two or more GEP analyses. Using a two-fold change in the expression for both MOX2 probe sets as a threshold, 81% of patients showed no significant variation in MOX2 expression (data not shown). Of the 5 cases in which there was significant modulation of MOX2 expression, one showed up-regulated and 4 showed down-regulated expression of MOX2.
CD200 is a cell surface glycoprotein expressed in normal B-cells and some T-cell subsets. The expression of CD200 has been analyzed in various lymphoid malignancies. In this report, we confirm several of these previously reported findings. Importantly, through the use of a flow cytometry-based approach, we have made several novel observations, some of which may have diagnostic and prognostic importance.
We have confirmed previous reports that CD200 is uniformly expressed in CLL, whereas its expression is not detected in MCL. While both are defined by CD5 expression, several immunophenotypic differences exist between CLL and MCL which readily allow for their distinction by flow cytometry in typical cases.22 These include characteristic CD23 coexpression and the dim expression of CD20 and surface Ig in CLL, in contrast to the bright expression of CD20 and surface Ig and CD23 negativity typically present in MCL. However, the immunophenotype in CLL not infrequently deviates from this “classic” immunophenotype, making its distinction from MCL in such cases difficult by flow cytometry alone.17,18,20 Common immunophenotypic variations include unusually bright expression of CD20 or surface light chain in CLL and aberrant CD23 expression in MCL. Although we were able to analyze only a limited number of cases, our findings indicate that the differential expression of CD200 is retained in immunophenotypically atypical CLL and MCL. Thus, CD200 analysis appears to have particular diagnostic utility when confronted with such immunophenotypically indeterminate cases. It is important to note that CD200 expression alone does not equate to malignancy, since this antigen is expressed in most normal peripheral blood lymphocytes and in a subset of lymph node B-cells. Therefore, the evaluation of CD200 expression must be interpreted in conjunction with that of other cell markers, including those for surface immunoglobulin light chains.
PCM is a genetically heterogeneous disease. While the investigation of underlying cytogenetic abnormalities has provided significant insight into its pathogenesis, more recent GEP-based analyses have significantly advanced our understanding of the molecular pathogenesis of PCM. Thus, seven major molecular subtypes of PCM have been identified through GEP analysis, including CD-1, CD-2, HY, MF, LB, MS, and PR. Some of these PCM subtypes have been well characterized due to their underlying cytogenetic abnormalities, e.g., the CD-1 and CD-2 subtypes, both of which are associated in most instances with the t(11;14) resulting in aberrant cyclin D1 overexpression. Importantly, GEP analysis has facilitated the identification of novel PCM subtypes which lack defining recurrent cytogenetic abnormalities, including the LB and PR subtypes. At our institution, GEP analysis is routinely performed on all myeloma patients, from which molecular subtyping and a 70-gene risk score are derived.14,21 The latter has been confirmed to be a strong predictor of poor clinical outcome.
The PR subtype comprises 18% of all newly diagnosed PCM and is characterized by overexpression of several proliferation and cell cycle-related genes.14 PR PCMs manifest metaphase cytogenetic abnormalities in 70-80% of cases, whereas such abnormalities are detected in only about 20% of other molecular subtypes.14 Finally, several genes mapping to chromosome 1q, a region commonly amplified during disease progression and associated with poor prognosis, are frequently overexpressed in PR PCM.23-25 Thus, PR PCM manifests several genetic features associated with poor clinical outcome and is associated with a significantly increased 70-gene risk score in most cases.
While not detectable on normal PCs, we found that CD200 is expressed in 71% of PCMs, a frequency similar to that reported by others.10,26,27 Interestingly, the PR type is significantly overrepresented among CD200− PCM, comprising 50% of such cases, whereas PR myelomas comprise only 6% of CD200+ cases. Not surprisingly, the CD200− group had a 70-gene risk score that was significantly higher than that of CD200+ PCMs.
These findings appear to contradict a previous report in which patients with CD200+ PCM, who comprised 78% of patients, were shown to have inferior survival compared with those with CD200− disease.10 However, several differences exist between our analysis and that of Moreaux et al. In the latter study, CD200 expression analysis was performed only at initial diagnosis, whereas the majority of our analyses were performed on patients who had received previous anti-myeloma therapy. Furthermore, in our study, flow cytometry was used to evaluate the expression level of CD200 on the cell surface of neoplastic plasma cells. By contrast, CD200 expression in the study by Moreaux et al was achieved primarily through GEP-based analysis of MOX2 expression. Although well-correlated overall, several cases in our series expressed significant levels of MOX2 mRNA by GEP yet lacked detectable protein expression as assessed by flow cytometry (data not shown). This observation may account for the somewhat lower incidence of CD200+ myelomas in the current study as compared to that of Moreaux et al. Finally, although we were not able to perform serial flow cytometric analysis of CD200 expression, repeat GEP analyses suggest that MOX2 expression is dynamically modulated during the course of disease development in a subset of patients. Furthermore, serial GEP analysis of several patients showed that the down-regulation of MOX2 expression occurred concurrent with disease evolution to the PR subtype. The basis for this down-regulation is presently unknown; however, it is unlikely due to gene deletion as the gene encoding CD200, MOX2, is located on located at chromosome 3q12, a region infrequently deleted in PCM.28 Nevertheless, these preliminary findings suggest that loss of CD200 expression may serve as a marker for disease progression in PCM.
We found differential expression of CD200 in precursor lymphoid malignancies, with T-lineage ALL being uniformly CD200−, whereas B-lineage ALLs were variably positive for this marker in 95% of cases. Precursor B-lymphoblastic leukemia/lymphoma is readily distinguished from normal B-cell progenitors in most instances. However, in the setting of low-level bone marrow involvement, such distinction may be problematic. Immunophenotypic aberrancies have been detected in the majority of B-ALLs.29 While nearly all B-ALLs were CD200+ in our analysis, over half aberrantly over- or underexpressed CD200 in comparison to normal CD34+/CD19+ B-cell progenitors. Thus, evaluation of CD200 may be of utility in distinguishing between leukemic lymphoblasts and normal hematogones.
Interestingly, the four B-ALLs with the strongest CD200 overexpression were either hyperdiploid or TEL-AML1+, two ALL subtypes which comprise approximately one-half of all pediatric ALL and which are typically associated with an excellent clinical outcome.30-32 Although the analysis of additional cases and further confirmatory studies are needed, one implication of our findings is that CD200 may represent a potential therapeutic target in B-ALL, particularly in those subtypes manifesting high levels of CD200 expression. A humanized anti-human CD200 antibody, ALXN6000, has recently been developed and is currently in a phase I/II clinical trial in patients with CLL and PCM (NCT00648739; ClinicalTrials.gov). An immunotherapeutic approach might obviate the significant toxicity associated with the multiagent chemotherapy presently used for ALL. Such an approach would be particularly appealing for patients with hyperdiploid or TEL-AML1+ B-ALL, both of which are associated with an excellent prognosis with current therapy and thus represent candidates for reduced intensity chemotherapeutic approaches.
We thank the members of the UAMS flow cytometry lab for their invaluable assistance with this project and Dr. Neslihan Cetin for her help in obtaining cytogenetic data.
JDS is co-founder of Myeloma Health LLC and owns stock in the company; he receives royalties from Novartis, Genzyme, and Myeloma Health, and he is a paid consultant to Novartis, Myeloma Health, Genzyme, Array BioPharma, Onyx, Millennium and Celgene.
BB has received research funding from Celgene and Novartis. He is a consultant to Celgene and Genzyme. He has received speaking honoraria from Celgene and Millennium. Dr. Barlogie is a co-inventor on patents and patent applications related to use of GEP in cancer medicine.