Nuclear expression of β-catenin and stem cell markers as potential prognostic indicators in medulloblastoma
Correspondence Address: Source of Support: None, Conflict of Interest: None DOI: 10.4103/0028-3886.103192
Source of Support: None, Conflict of Interest: None
Aims: To study the prognostic role of β-catenin and stem cell markers in medulloblastoma (MB). Materials and Methods: Sixty cases of MB were retrospectively analyzed to study the expression of β-catenin, CD15, and CD133 by immunohistochemistry. Their expression was correlated with histological subtypes and event-free survival (EFS). Patients were divided into Group 1 and 2 based on non-occurrence and occurrence of events during the follow-up period. Results: Fifty of the 60 cases were of classic type of MB while nine were of desmoplastic subtype and one case showed chondroid and rhabdomyoblastic differentiation. Immunoreactivity for β-catenin was observed as nuclear and/or cytoplasmic positivity within the tumor cells. Forty-one (68.3%) cases showed cytoplasmic positivity, while nuclear positivity was seen in 21 (35%) cases. There was a significant correlation between nuclear expression of β-catenin and different histological subtypes by Chi-square test (P value<0.05). A statistically significant positive correlation of β-catenin nuclear positivity with EFS was observed. Among 60 cases, 37 cases (67.3%) showed presence of CD15+ tumor cells with percentage of positivity varying between 0.1 to 17.1%. Overall, 42 of 60 (70%) cases showed presence of CD133+ cells. The percentage of positivity varied between 0.1 to 16.5%. A statistically significant negative correlation of CD15 and CD133 positivity with EFS was observed. Conclusions: Nucleopositive β-catenin cases were associated with a favorable outcome on univariate analysis. Both CD15 and CD133 positivity were associated with a worse outcome on univariate analysis.
Keywords: β-catenin, CD133, CD15, event-free survival, medulloblastoma
Medulloblastoma (MB) is the most common malignant central nervous system tumor in childhood. Despite a marked increase in the 5-year survival rates of 85% in average-risk group and 70% in high-risk group,  the long-term sequelae remain a major handicap. Current challenges lie in disease risk stratification and finding novel targeted therapies. Identification of prognostic markers is a crucial step toward achieving this.
Recent studies have documented that nuclear accumulation of β-catenin protein is associated with activation of Wnt signaling pathway and its nuclear expression as an independent marker of favorable outcome. The Wnt/β-catenin pathway is aberrantly activated in approximately 10 to 15% of MB. 
Additionally, it is now widely believed that tumor growth depends on a subset of tumor cells with self renewal capacity and multipotency, termed cancer stem cells (CSCs).  Two neural stem cell (NSC) markers, CD133 and CD15, have been used in recent past for identifying CSCs in MB. , Most studies identifying NSC have been carried on models of MB. Studies validating the results of these experiments on human MB are still limited. We conducted this study to analyze the expression of β-catenin and CSC subpopulation in MB and correlate their expression with the survival outcome. Identification of these chemoresistant cells would provide insights into the signals required for their survival, maintenance, and growth, thereby permitting development of more effective targeted therapies.
Sixty MB cases diagnosed during 2004-2010 with follow up data and sufficient material for immunohistochemistry were retrieved. Recurrent tumors already treated by surgery and/or radiotherapy were excluded. Clinical details and survival outcome were retrieved from the radiotherapy records. Indirect immunoperoxidase staining was performed using 4 μm thick sections. Antigen retrieval was done by heat-induced epitope retrieval method followed by application of primary antibodies (monoclonal)--anti-β-catenin (Clone-12F7, isotype-IgG1k, Santa Cruz Biotechnology), anti-CD133 (Clone- mAbcam27699, isotype-IgG2bk, Abcam), and anti-CD15 (Clone- C3D-1, isotype- IgMk, Dako) at 1:50, 1:200, and 1:20 dilutions, respectively. Goat anti-rabbit/mouse antibody was used as secondary antibody. Positive control was included with each batch [Figure 1]a and b.
Immunoexpression of β-catenin was analyzed based on presence or absence of nuclear positivity in tumor cells. This was further assigned a semi-quantitative score based on percentage of positive cells [Table 1]. 
Immunoreactivity of CD15 and CD133 was evaluated taking into account the percentage positivity of tumor cells with each case being scored semi-quantitatively [Table 2]. The percentage was calculated after examining 10 representative hpf (40x) and counting the number of cells showing membranous positivity.
Data were analyzed statistically using Chi-square test, Kaplan-Meier survival analysis, Log rank test, and Pearson's coefficient of correlation. Level of significance was fixed at 0.05 level.
The age ranged from 1 to 47 years with mean of 12.4 years and M:F ratio of 4:1. Cerebellum was the primary site of tumor in 40 cases, roof of fourth ventricle in six cases, and in remaining 14 cases, it grossly involved the posterior fossa structures and the exact site could not be determined. Histological spectrum included 50 cases of classic subtype, nine desmoplastic MB and one case of MB with chondroid and rhabdomyoblastic differentiation [Figure 2]a-d.
Twenty-six patients had undergone total excision of tumor while the rest had near total excision [Table 3]. Of 60 cases, 48 cases completed the course of craniospinal irradiation (CSI), nine patients received adjuvant chemotherapy (CR) following CSI, two patients received chemotherapy (CT) alone, while one patient did not receive any form of postoperative therapy. The CSI comprised of delivering megavoltage linear accelerator irradiation to the entire neuraxis, with posterior fossa boost. The dose delivered was in the range between 30 000 to 40 000 rads to the craniospinal axis, followed by a 14 000 - 18 000 rads local tumor boost. CR consisted of etoposide and cisplatin with or without vincristine after the completion of CSI, whereas CT included administration of drugs alone without radiotherapy.
Based upon the non-occurrence and occurrence of events (R-Recurrence/Metastasis, D-Death), patients were divided into two prognostic groups. In Group 1, event-free survival (EFS) ranged from 4 to 63 months (median, 24.6 months). EFS in Group 2 (with occurrence of above events) ranged from 2 to 17 months (median, 7.1 months). EFS is defined as the duration between tumor excision and occurrence of any of the above events.
Immunoreactivity for β-catenin was observed as nuclear and/or cytoplasmic positivity in tumor cells. Forty-one (68.3%) cases showed cytoplasmic positivity, while 19 (31.7%) cases were negative. Nuclear positivity was seen in 21 (35%) cases [Figure 3] a-d, while it was absent in 39 (65%) cases. The scores of nuclear expression of β-catenin were compared in different histological subtypes. Twenty-one of 50 classic MB showed nuclear positivity. Neither the desmoplastic subtype nor MB with chondroid and rhabdomyoblastic differentiation showed nuclear positivity. Association between β-catenin nuclear expression and histological subtypes by Chi-square test revealed a statistically significant difference in nuclear expression of β-catenin in the classic versus desmoplastic subtype ('P' value = 0.035).
Overall, 19 of 40 (47.5%) cases of Group 1 showed β-catenin nucleopositivity. Nine (22.5%), six (15%), four (10%) cases among Group 1 showed focal, intermediate, and extensive nuclear expression of β-catenin, respectively. Among Group 2 patients, 2 of 20 (10%) cases showed focal nuclear expression. None of these cases showed intermediate or extensive nuclear positivity. β-catenin nucleopositivity was associated with better EFS which was statistically significant ('P' value Chi-square test = 0.029, Log rank test = 0.008).
On evaluation of stem cell markers, 37 of 60 cases (67.3%) showed presence of CD15+ tumor cells [Figure 4]a and b. The percentage of positivity varied between 0.1 to 17.1%. There was no statistically significant difference in the expression of CD15 in the classic vs desmoplastic MB on Chi-square analysis ('P' value>0.05). However, a statistically significant correlation of CD15 positivity with EFS was observed with CD15 positivity being associated with a worse outcome on univariate analysis (Chi-square and log rank test, P value<0.05). The median EFS in CD15+ cases was 13.2 months, while it was 27.7 months in negative cases.
Overall, 42 of 60 (70%) cases showed presence of CD133+ cells [Figure 4]c and d with percentage of positivity varying between 0.1 to 16.5%. No significant statistical difference in expression of CD133 in classic vs desmoplastic MB was observed (Chi-square test, 'P' value>0.05). Similar to CD15, a statistically significant correlation of CD133 positivity with EFS was observed. CD133 positivity was associated with a worse outcome on univariate analysis (Chi-square and log rank test, P value < 0.05). The median EFS in CD133+ cases was 14.3 months while it was 29.2 months in negative cases.
CD15 and CD133 co-expression was observed in 35 cases (59.3%). CD15 alone was expressed in two (3.3%) cases and CD133 alone was expressed in seven (11.7%) cases. Statistical analysis using Pearson's
co-efficient of correlation revealed a positive correlation with significant co-expression of CD15 and CD133 ('P' value <0.01).
Eight cases (13.3%) showed co-expression of CD15 and β-catenin, while 10 (16.7%) cases were negative for both. CD15 alone was expressed in 29 (48.3%) cases and β-catenin alone was expressed in 13 (21.7%) cases. It was noted that there was a negative correlation with no significant co-expression of CD15 and β-catenin by Pearson's co-efficient of correlation analysis ('P' value <0.01).
Ten cases (16.7%) showed co-expression of CD133 and β-catenin, while seven (11.7%) cases were negative for both. CD133 alone was expressed in 32 (53.3%) cases and β-catenin alone was expressed in 11 (18.3%) cases. It was seen that there was a negative correlation with no significant co-expression of CD133 and β-catenin by Pearson's co-efficient of correlation analysis ('P' value <0.01).
Overall, 12 of 60 (20%) cases showed co-expression of both SC markers and β-catenin, seven (11.7%) were negative for both, nine (15%) showed only nuclear expression of β-catenin, and 32 (53.3%) cases showed expression of only SC markers. Kaplan-Meier survival curves revealed a significant difference in EFS, with cases showing only nuclear expression of β-catenin having the best outcome and case expressing only SC markers having the worst prognosis ('P' value= 0.001) [Figure 5].
Currently, we are faced with two major challenges in our research on MB. First, circumventing the cognitive side effects of adjuvant therapies and second is to improve the EFS rates. Both these can be achieved by de-escalating radiotherapy, besides simultaneously using targeted therapies. Here lies the importance of risk stratification of MB. The current clinical risk stratification of MB is primarily based on clinical parameters like age, extent of surgical resection, and metastatic status.  Recently, a lot of emphasis has been laid on identification of molecular markers and their incorporation into risk stratification based upon their correlation with prognosis and/or response to therapy. This is believed by many to be a step toward the refinement of treatment strategies.
Molecular markers proven to be associated with poor prognosis include loss of TP53, amplification of MYCC, MYCN,  and low TRKCmRNA expression levels.  However, till now none of these markers have been successfully used in disease stratification.
Numerous studies ,, have reported activation of Wnt/β-catenin signaling pathway, evidenced by nuclear expression of β-catenin, as a factor for favorable outcome in MB patients. Gene expression profiles in two independent series of 62 and 46 MB has revealed five distinct molecular subgroups including one with over-representation of Wnt/β-catenin pathway genes (9/62 and 6/46 cases, respectively). ,
We sought to analyze the immunohistochemical profiles of 60 cases of MB for their expression of β-catenin, CD15, and CD133 and correlated this to its histological subtypes and EFS.
β-catenin nucleopositivity was identified in 21 of 60 cases of MB. This is comparable to initial observations of nine of 72 (12.5%),  33 of 206 (16%),  27 of 109 (25%),  and 13 of 32 (40%)  in four independent studies. Use of different methodologies might have contributed to difference in reported percentage of nucleopositivity. In this study, 6.7% cases showed extensive nuclear positivity while 18.3% showed focal nucleopositivity. Unlike previous studies, we included an additional intermediate grade of nucleopositivity (10-49.9%) with 10% cases falling in this category.
Various independent studies have consistently shown that nuclear expression of β-catenin is associated with a favorable outcome in MB. ,, Nineteen of 40 Group 1 patients showed β-catenin nucleopositivity, and only 2 of 20 Group 2 cases showed nuclear staining of β-catenin. Group 1 patients had EFS of 24.6 months, while Group 2 cases had EFS of 7.1 months. β-catenin nucleopositivity was associated with better EFS which was statistically significant. Based on the above evidences, MB with β-catenin nucleopositivity and without clinical, pathological, or other molecular adverse indicators could reasonably be stratified to a low-risk group. De-escalation of CSI in such individuals has the potential to ameliorate long-term adverse effects while maintaining cure rates.
Studies have observed that β-catenin nucleopositivity characterizes a molecular subgroup of MB associated with activation of the Wnt signaling pathway, CTNNB1 mutations, and monosomy of chromosome 6. The rates of detection of CTNNB1 mutation and monosomy of chromosome 6, which have been mentioned as components in the molecular subgrouping of MB,  in the β-catenin nucleopositivity cases vary between 60 to 100% in different studies. ,,, However, CTNNB1 mutation and monosomy of chromosome 6 were not detected in any of the β-catenin negative cases. Thus, immunohistochemistry for β-catenin captures all the cases of MB with CTNNB1 mutation and monosomy of chromosome 6 can be regarded as a standard test for identifying MB subgroup with activation of Wnt/β-?catenin pathway.
Identification of CSC marks a step toward finding new and effective ways to treat MB. The concept of CSC has constructive significance for clinical practices because it has been elucidated that CSC contribute to relapse and chemoresistance or radioresistance of brain tumors. catenin pathway.
Identification of CSC marks a step toward finding new and effective ways to treat MB. The concept of CSC has constructive significance for clinical practices because it has been elucidated that CSC contribute to relapse and chemoresistance or radioresistance of brain tumors. , Most studies exploring the property of CSC in brain tumors were built on the assumption that CSC express a cell surface marker, CD133.  However, many recent studies have shown the existence of both CD133-positive and -negative CSC in brain tumors. , Recent studies on mouse models of MB have used CD15 as a SC marker. , Therefore, it is more likely that there are at least two types of SC/CSC identified by CD133 and CD15. Although CD133 has been widely studied, the expression of CD15 has not been fully explored in CSC. Given the similarity between SC and CSC and recent disputes about CD133 as a CSC marker, we performed this study to further explore the expression of CD15 in CSC and to investigate if CD15 can be used to identify a group of CSC different from CD133+ cells.
The percentage positivity varied between 0.1 to 17.1% (mean, 1.59%) for CD15 and 0.1 to 16.5% (mean, 1.48%) for CD133, in comparison to values of 3 to 87%  and 3.5 to 46.3%,  respectively, in previous studies. Higher percentage of cells reported in previous studies is probably because of use of flow cytometry which is a more sensitive technique. The percentage of CD15 positivity is higher in the current as well as previous study, as it has been shown that CD15, in addition to detection of CSC, also detects progenitor cells. Additionally, CD15 and CD133 were staining different population of cells distributed throughout the tumor. Both CD15+ and CD133+ cells were predominantly noted in specific niche microenvironments around perivascular areas. Previous studies have shown that the brain tumor microvasculature forms a niche that is critical for the maintenance of CSC.  Evidence suggests that normal NSCs also exist in vascular niches, into which endothelial cells secrete factors that regulate NSC function.  However, it has been noted that vascular niches in brain tumors are abnormal and contribute directly to the generation of CSC and tumor growth. In addition to their role of regulating stem cell proliferation and cell-fate decisions, vascular niches also seem to play a protective role, shielding CSC from chemo- and radiotherapies, enabling these cells to reform a tumor mass following an initial clinical response.  Hence, targeting these microenvironments could prove to be more effective treatment.
We found no significant difference between classic and desmoplastic MB in their expression of CD15+ and CD133+ CSC. A significantly higher percentage of CD15+ and CD133+ CSC in Group 2 in comparison to Group 1 cases were noted, with a poor EFS in the former ('P' value<0.05). Thus, our study further validates the hypothesis that CSC play a major role in recurrence and metastasis by virtue of their chemo- and radio-resistance. However, different treatment modalities used in study groups could also influence the prognosis partly.
Targeted therapies against CSC could lead to de-escalation of CSI, resulting in amelioration of its toxic effects, thereby improving the survival rates. However, similarities observed between SC and CSC hints at a possibility of therapeutic toxicities. Overcoming this would be a major challenge in order to achieve the desired cure rates with minimal therapeutic side effects. Also, β-catenin is essential for maintenance and proliferation of neuronal progenitors, and in controlling the size of the cerebellar progenitor pool. Interestingly, our study revealed a negative correlation between nuclear expression of β-catenin and percentage of positivity of CD15 and CD133 ('P' value<0.05). A better understanding of the molecular link between cerebellar development and tumorogenesis will lead to therapies specifically targeting MB cells.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]
[Table 1], [Table 2], [Table 3]