Prognostic significance of stem cell marker CD133 determined by promoter methylation but not by immunohistochemical expression in malignant gliomas

Increasing evidence suggests that there is a small subset of cells, called cancer stem cells (CSC), that are responsible for cancer initiation and development [1]. The CSC model of tumor development suggests that the clinical behavior of a tumor will be largely determined by a subpopulation of cells that are characterized by their ability to initiate new tumors [2]. Recently, CSC has been described in several solid tumors,including brain tumors [3, 4]. These studies used putative stem cell markers or side populations to isolate unique subsets of cancer cells from different types of tumors. These markers included CD24, CD44, CD133, and CD166 that are also expressed in normal cells. Among these markers, CD133 were widely used for isolating CSC from solid tumors [5‐8].

Surface membrane protein CD133, which is normally expressed in a subset of putative neural stem/precursor cells in the normal brain, have been identified in brain tumors [4, 5, 9, 10]. CD133-positive tumor cells can initiate neurospheres, which exhibit self-renewal, differentiation, and proliferation resembling that of normal NSCs. The transplantation of CD133-positive tumor cells into nonobese diabetic/severe combined immunodeficient mice is sufficient to produce tumors phenotypically identical to the patient’s original tumors [4, 11]. CD133-positive glioma cells were also reported to possess enhanced chemoresistance and radioresistance, resulting in tumor progression and recurrence [12‐15]. Therefore, the regulatory mechanisms of CD133 expression could help to elucidate the development of gliomas.

Changes in DNA methylation patterns are an important hallmark of tumor development and progression [16]. Although the role of hypermethylation in the silencing of tumor suppressor genes is now well-documented, abnormal methylation contributes to neoplastic progression in numerous types of human cancer, including glioblastomas [17‐19]. CD133 expression in glioblastomas is determinated by methylation status of the promoter region 1–3 in vitro [20]. Strong methylation of CD133 promoter displayed in CD133-negative tumor cells, but less in the CD133-positive fraction [21]. Surprisingly, however, apart from the observation of aberrant methylation of CD133 in glioblastoma [21], a direct link between CD133 methylation and patient outcome thus far has not been established. To address this question, we studied CD133 expression and promoter methylation in gliomas, and investigated the correlation of CD133 expression and promoter methylation with patient outcome.

CD133 has played a pivotal role in the identification and isolation of brain tumor stem cells. However, the correlation between CD133 expression in tumor tissues with patients survivals is still controversial [23, 24]. This study represents, to our knowledge, the first investigation aimed at evaluating the prognostic significance of CD133 in a series of patients with gliomas in which both tumor CD133 expression and promoter methylation status were simultaneously evaluated. Our results show that CD133 promoter methylation is independently associated with a longer overall survival in patients with anaplastic gliomas, whereas CD133 expression in IHC has no prognostic implications. The strength of these results relies mainly on the fact that CD133 predictive value has been evaluated after adjusting this variable for well-recognized clinicopathologic prognostic factors.

In our study, CD133 methylation pattern was heterogeneous, ranging from unmethylation to methylation. Differences in the methylation status of the CD133 promoter in tumor cell subpopulations may explain this heterogeneity. Other potential explanations for this variability include monoallelic promoter methylation, or loss of heterozygosity in 4p15. Whereas the presence of contaminating normal cells may not be ruled out. The presence of CD133 CpG island methylation in glioma patients was associated with grades of glioma. In low-grade tumors (WHO grade 2), the percentage of methylated and unmethylated CD133 were 74.2 % (46 of 62) and 25.8 % (16 of 62), respectively. With the progression to high grade glioma (WHO grade 3 and 4), the percentage of methylation decreased to 44.4 % (62 of 108), and the percentage of unmethylation increased to 55.6 % (60 of 108). This findings was consistent with the others study, which reported that promoter methylation of CD133 was lower in advanced colorectal carcinomas and aggressive breast cancer [25, 26]. This is concordant with the biologically aggressive nature of high grade glioma, poor overall prognosis, and limited therapeutic targets. More important, our study first reported that the methylation status of the CD133 promoter may have prognostic value. We found CD133 hypermethylation to be significantly (almost two-fold) more common in long-term survivors (74 %) as compared to in short-term survivors (43 %) not only in high grade glioma, but in low grade glioma. Multivariate analysis indicated CD133 methylation was a significant prognostic factor in gliomas, independent of tumor grade, extent of resection, and patient age. This findings may suggest that methylation of CD133 promoter may directly relevant to the stem cell state of CSC. In recent study, Gopisetty et al. demonstrate a role for the promoter CpG island in critically regulating CD133 expression in glioma stem cell (GSC) [27]. When the GSC cells were sorted into CD133+ve and −ve fractions, CD133+ve cells are significantly resistant to chemotherapeutic agents compared with autologous CD133−ve cells and to radiation therapy. CD133 promoter was hypermethylated in CD133−ve GSC and glioma cells, but unmethylated in CD133+ve ones. Promoter methylation of CD133−ve fraction can be reversible by demethylation agents and reproduced CD133 strongly supporting a functional role for methylation in the repression of CD133.

In the present study, a heterogeneous staining pattern of CD133 was observed within the individual tumor and between different tumors of the same grade. This is in agreement with other studies which show expression of CD133 was considerable various in gliomas on frozen section [23, 28] and by flow cytometry [4]. We found that 85.0 % of the anaplastic astrocytomas and 75.4 % of the glioblastomas were negative for CD133, which was much lower than that of other studies who found that less than 40 % of anaplastic gliomas and less than 2 % of glioblastomas were negative for CD133 [23, 24]. Use of different CD133 antibody clones might explain the difference between our findings and the results obtained in other studies. In other studies, the monoclonal antibody clone AC133 and W6B3C1 were from Miltenyi Biotec [3, 29], and goat polyclonal antibody was from Santa Cruz were used [28], whereas we used the monoclonal antibody clone CD133-1 from Miltenyi Biotec. Another explanation was use of different tissue fixation methods. In other studies, cryosections were used to stain CD133, whereas we used formalin-fixed paraffin sections because the morphology is better preserved in paraffin sections making it easier to evaluate a membrane staining like CD133 [29].

Comparing the CD133 expression assessed by ICH with overall patient, we found CD133 protein expression in glioma was insufficient correlated with patient survival. Therefore, in our study, CD133 IHC does not prove to be a clinically usable tool in the prognostic assessment of glioma. Previously published studies on diffuse glioma [23, 30, 31] reported a significant association of immunohistochemically assessed CD133 expression and patient survival. However, this association was not confirmed our present study in a larger patient cohort. This result may be explained by the bona fide cancer stem cell being a subpopulation of the CD133+ tumor cells, but it is also likely that other non tumor stem cells apart from endothelial cells express CD133. Other explanations may be methodological differences (e.g., different pretreatment of sections prior to immunostaining and different type of sections (paraffin sections vs. frozen sections) [23].

Very few studies have investigated the relationship between CD133 promoter hypermethylation and protein expression in human glioma samples. In the present study, there was an inconsistent correlation between aberrant promoter methylation and loss of protein expression. To date, this inconsistency was not limited to the CD133 gene [32‐35]. Promoter methylation is clearly involved in the inactivation of CD133 gene in numerous tumors and cancer cell lines [20, 21], but regulation of CD133 expression is a more complex phenomenon in which abnormal methylation of the promoter is not the only determining factor [21, 36, 37]. Indeed, several studies indicate that grade of methylation both in the promoter region and in neighboring sequences may regulate gene expression [20, 37, 38]. As discussed previously, another possible explanation for this finding is CD133 protein expression in entrapped pre-existing endothelial cells, which is confirmed by flow cytometry showing that CD133+ cells could have either blood vessel or glioma origin [39, 40]. Therefore, there is increasing evidence that immunohistochemically assessed CD133 expression is a poor indicator of CD133 promoter methylation status in glioma.

However, the importance of this study goes beyond showing the putative clinical benefit of CD133 methylation status as a prognositic marker. Currently, little data exist on the clinical relevance of CSCs. Properties of CSCs could well explain many clinical features of cancer, such as recurrence, and therapy resistance [1, 41‐49]. Some studies also tested whether the prevalence of putative CSCs in the tumor was relevant to patient outcome, but the results were controversial [23, 24, 30, 31, 50‐54]. In light of the results reported here, the discrepancy may seem not surprising because different CSC markers, heterogenous expression, and observation variability all effect on the results. Here, for the first time, we presented a direct link between the methylation of a CSC marker and patients’ outcome. These data provide strong supportive evidence for the CSC model and the clinical relevance of the CD133 methylation status in gliomas.

To conclude, in our study, CD133 promoter methylation status in glioma is closely correlated with patient survival, which suggest CD133 promoter methylation pattern is a promising tool for prognostic purposes. While, lack of association with CD133 IHC and with patient survival impede the use of CD133 IHC as a clinically useful biomarker for routine purposes and clinical decision making.