MCO Molecular and Clinical Oncology 2049-9450 2049-9469 D.A. Spandidos 10.3892/mco.2013.195 mco-02-01-0111 Articles Clinicopathological significance of CD133 in lung cancer: A meta-analysis TANYAOXI CHENBO XUWEI ZHAOWEIHONG WUJIANQING Department of Geriatrics, The First Affiliated Hospital of Nanjing Medical University, Nanjing, Jiangsu 210029, P.R. China Correspondence to: Professor Jianqing Wu, Department of Geriatrics, The First Affiliated Hospital of Nanjing Medical University, 300 Guangzhou Road, Nanjing, Jiangsu 210029, P.R. China, E-mail: jwuny@njmu.edu.cn 01 2014 30 09 2013 2 1 111 115 08 01 2013 09 07 2013 Copyright © 2013, Spandidos Publications 2014

CD133 is one of the most commonly used markers of lung cancer stem cells (CSCs), which are characterized by their ability for self-renewal and tumorigenicity. However, the clinical value and significance of CD133 in lung cancer remains controversial. Due to the limited size of the individual studies, the association between CD133 and the clinicopathological characteristics of lung cancer had not been fully elucidated. A meta-analysis based on published studies was performed with the aim of evaluating the effect of CD133 on the clinicopathological characteristics of lung cancer and to investigate the role of CSCs in the prognosis of lung cancer. A total of 15 eligible studies were included in this meta-analysis and our results indicated that a positive CD133 expression was significantly associated with poor differentiation and lymph node metastasis, although it was not associated with tumor stage or histological type. Therefore, CD133 may be considered as a prognostic maker of lung cancer. Further clinical studies, with larger patient samples, unified methods and cut-off levels to detect CD133 expression, classified by tumor stage, therapeutic schedule, follow-up time and survival events, are required to determine the role of CD133 in clinical application and the association between CD133 and the prognosis of lung cancer.

 lung cancer cancer stem cells CD133 clinicopathological characteristics prognosis meta-analysis Introduction

Lung cancer remains the most common cause of cancer-related mortality worldwide (1). Patients affected by lung cancer exhibit a poor prognosis. Despite the significant advances in surgical techniques, chemotherapy and radiotherapy, the relapse rate is high and only a few patients achieve long-term survival, with an overall 5-year survival rate of ∼15% (2). Evidence supports the cancer stem cell (CSC) hypothesis, according to which CSCs may be responsible for tumor initiation, metastasis, recurrence and resistance to treatment (3,4). Due to the characteristics of CSCs, the conventional therapies are unable to effectively eliminate these cells. The residual CSCs continue to proliferate, leading to the relapse of cancer. A variety of molecules have been investigated as putative markers of CSCs in malignancies, including lung cancer (5). Among the various markers, CD133 is one of the most commonly used. It is widely expressed in a number of malignancies, such as glioblastomas (6), hepatocellular (7), ovarian (8), colon (9) and lung (10) carcinomas.

A growing number of CD133-positive cancer cells have been identified in lung cancer (11,12). Eramo et al (10) observed that a rare population of CD133-positive cancer stem-like cells were able to self-renew and generate an unlimited progeny of non-tumorigenic cells, whereas the CD133-negative cancer cells lacked this potential. However, the association between CD133 expression and the clinicopathological characteristics of lung cancer remains unknown. Attempts to elucidate the association between CD133-positive cancer cells and clinicopathological characteristics in previous studies yielded controversial results (13,14). The limited sample availability resulted in discrepancies regarding the clinical significance determined by different CSCs studies. Therefore, we performed a systematic review of the literature with a meta-analysis to determine the association between CSCs marker CD133 and the clinicopathological characteristics of lung cancer and to investigate the role of CSCs in the prognosis of lung cancer.

 Materials and methods <sec> <title>Literature search

Studies were identified via an electronic search through Medline, EMBASE and the China National Knowledge Infrastructure (CNKI) databases, using the key words ‘lung cancer’ and ‘CD133’, completed by the personal bibliography of two of the authors (Yaoxi Tan and Bo Chen). The bibliographies reported in all the identified studies were used to complete the study search. Review articles were scanned to identify additional eligible studies. The search was completed on November 15, 2012. To be eligible for inclusion in this systematic review, a study was required to meet the following criteria: i) it only included patients with primary lung cancer; ii) it investigated the association between CD133 and clinicopathological characteristics; iii) it was published as a full-text article in English or Chinese and; iv) it reported the number of CD133-positive and -negative patients. When duplicate studies were published, only the most recent or most informative was included in the analysis, to avoid overlap between cohorts.

 Data extraction

The following information was extracted from each study: i) year of publication and first author’s name; ii) sample size, test method and cut-off level; iii) tumor data including stage, grade, histological type and lymph node metastasis. Information was carefully extracted from all the eligible studies independently by two of the authors of the present study (Yaoxi Tan and Bo Chen). Differences in the extraction of data were assessed by a third investigator (Jianqing Wu).

 Statistical analysis

To assess the association between CD133 and the clinicopathological characteristics of lung cancer including stage, grade, histological type and lymph node metastasis, odds ratios (ORs) with 95% confidence intervals (CIs) were calculated. The heterogeneity of combined ORs was initially evaluated by graphical examination of the forest plots. Statistical assessment was then performed using a χ2-based test of homogeneity and evaluation of the inconsistency index (I2) statistic. The I2 statistic was defined as the percentage of variability due to heterogeneity rather than chance, with values >50% representing the possibility of substantial heterogeneity (15). P<0.05 was considered to indicate a statistically significant difference. If no obvious heterogeneity existed, the OR was calculated by the fixed-effects model (the Mantel-Haenszel method) and χ2 tests. Otherwise, the random-effects model (the DerSimonian-Laird method) was used. In addition, evidence of publication bias was determined using the Egger’s (16) and Begg’s methods (17). All the calculations were performed using the Stata statistical software version 12.0 (StataCorp, College Station, TX, USA).

 Results <sec> <title>Study characteristics

In total, 15 studies published between 2008 and 2012 were selected for this systematic review. The study sample size ranged from 30 to 145 subjects. All studies investigated the association between CD133 and the clinicopathological characteristics of lung cancer. Cortes-Dericks et al (18) used reverse transcriptase-polymerase chain reaction (RT-PCR) to detect the expression of CD133, whereas Herpel et al (19) used tissue microarray, Tirino et al (20) used flow cytometry and the remaining studies used immunohistochemistry (IHC) with different cut-off levels. A total of 11 studies investigated non-small-cell lung cancer (NSCLC) alone, one study investigated adenocarcinoma and one squamous cell carcinoma. One study included adenocarcinoma, squamous cell carcinoma and small-cell lung cancer (SCLC). Li et al (21) conducted the study on patients with neuroendocrine lung cancer. The main characteristics of the 15 eligible publications are presented in Table I.

 Main results of the meta-analysis Correlation of CD133 with tumor stage

Among the selected studies, 8 analyzed the association between CD133 and tumor stage. Cortes-Dericks et al (18) and Li et al (21) observed that positive CD133 expression exhibited a significant association with tumor stage. Of the 8 studies, one study (22) was limited to stage I and one study (19) assessed patients with stage I–II disease. Two studies (13,21) were performed on stage I–IV lung cancer of different histological types. Excluding the four studies mentioned above, we performed a meta-analysis of the studies which investigated stage I–III cancer. A comparison of stage I–II to stage III revealed that there was no association between a positive CD133 expression and tumor stage (pooled OR=0.78, 95% CI: 0.44–1.41 and P=0.411) (Fig. 1). There was no evident publication bias (Egger’s test, P=0.089), a finding also supported by the Begg’s funnel plots (figure not shown).

 Correlation of CD133 with tumor differentiation

Eleven studies investigated the association between CD133 and tumor differentiation. Three of those studies (13,14,18) concluded that positive CD133 expression was associated with poorly differentiated tumors. Significant heterogeneity (pooled OR=1.17, 95% CI: 0.68–2.00 and P=0.567; random effects, I2=55.3% and P=0.013) existed when we analyzed all 11 studies. Excluding the studies using RT-PCR (18) and tissue microarray (19), we performed a subgroup analysis among the studies using IHC. However, significant heterogeneity (pooled OR=1.39, 95% CI: 0.81–2.37 and P=0.229; random effects, I2=52.2% and P=0.033) was still evident. We excluded one study (13) which was conducted on European patients and analyzed the studies that investigated Asian patients using IHC. The comparison of poor to high tumor differentiation revealed that a positive CD133 expression was significantly correlated with poor differentiation (pooled OR=1.66, 95% CI: 1.15–2.40 and P=0.006) without significant heterogeneity (I2=23.5% and P=0.242) (Fig. 2). No evident publication bias existed (Egger’s test, P=0.684), a finding also supported by the Begg’s funnel plots (figure not shown).

 Correlation of CD133 with histological type

A total of 12 studies were found to be eligible for the analysis of the association between a positive CD133 expression and histological type. Of these, Bertolini et al (13) observed a correlation between a positive CD133 expression and adenocarcinoma. We excluded two more studies, one of which (18) investigated adenocarcinoma alone, while the other (21) investigated neuroendocrine lung cancer. We analyzed the remaining 9 studies which were conducted on patients with NSCLC. There was no significant association between positive CD133 expression and histological type (adenocarcinoma vs. non-adenocarcinoma), with pooled OR=0.97, 95% CI: 0.71–1.33 and P=0.86 (Fig. 3). No evident publication bias existed (Egger’s test, P=0.143), a finding also supported by the Begg’s funnel plots (figure not shown).

 Correlation of CD133 with lymph node metastasis

Eleven publications investigated the association between CD133 expression and lymph node metastasis. One study (23), which was limited to patients with N2 or N3 NSCLC undergoing induction chemoradiotherapy (CRT), was excluded. We observed that CD133 was associated with nodal status (pooled OR=2.30, 95% CI: 1.36–3.89 and P=0.002), with a positive CD133 expression in tumors with lymph node metastasis (N+ compared to N0) (Fig. 4). No evident publication bias existed (Egger’s test, P=0.639), a finding also supported by the Begg’s funnel plots (figure not shown). However, there was significant heterogeneity. We performed a subgroup analysis among the studies that used IHC and observed that positive CD133 expression was associated with nodal status (pooled OR=2.98, 95% CI, 2.04–4.35 and P<0.001), without significant heterogeneity (Fig. 5).

 Discussion

Accumulating evidence indicates that specific subpopulations of cancer cells with stem cell characteristics within the majority of tumors may be crucial in the pathogenesis of malignant tumors, including lung cancer. Several methods were used to identify and enrich CSCs from lung cancer, e.g., side populations and stem cell markers of normal tissue (24). CD133 is one of the most extensively used markers in lung cancer, although its function in lung cancer has not been fully elucidated. It remains controversial whether CD133 is associated with clinicopathological characteristics and prognosis of lung cancer. The aim of this current meta-analysis was to investigate the correlation between CD133 expression and the clinicopathological characteristics in patients with lung cancer, since therapeutic decisions are directly associated with clinicopathological characteristics, such as tumor type and differentiation. Our results indicated that a positive CD133 expression was significantly associated with differentiation and lymph node metastasis, although it was not associated with tumor stage or histological type.

Several studies (14,18,21,23,25–27) reported that CD133 expression was correlated with poor prognosis in lung cancer, whereas other authors (13,19,22,28) observed no such association. The patient populations varied among these studies. For example, Herpel et al (19) conducted the study on previously untreated stage I–II NSCLC patients, while Shien et al (23) investigated patients with locally advanced N2 or N3 NSCLC, who underwent induction CRT followed by surgery. The methods used to detect CD133 expression were also different, including IHC, RT-PCR and tissue microarray. Even among the studies using IHC, the cut-off level and the antibodies used varied widely. In addition, certain studies reported disease-free survival, whereas others reported overall survival with different follow-up times. Therefore, due to these considerable differences, we were not able to directly perform a systematic review to assess the correlation between CD133 expression and the prognosis of lung cancer.

The studies mentioned previously provided important information regarding the indirect prognostic value of CD133 and the therapy targeting CSCs in lung cancer. Eramo et al (10) observed that CD133-positive stem-like cancer cells were capable of self-renewal. The injection of immunocompromised mice with CD133-positive lung cancer cells readily generated tumor xenografts phenotypically identical to the original tumor, leading to the conclusion that CD133 was a reliable marker of lung cancer. We inferred that there was a larger percentage of CSCs in the majority of lung cancer tissue, with a higher number of CD133-positive lung cancer cells. One of the properties of CSCs was the ability to undergo asymmetrical division, leading to pluripotential differentiation and metastasis (28). This finding was consistent with the results of our meta-analysis, suggesting that a positive CD133 expression was clearly associated with poor tumor differentiation and lymph node metastasis. Poor differentiation and metastasis were significantly associated with poor survival of cancer. In brief, positive CD133 expression was most likely correlated with poor prognosis of lung cancer. It should also be noted that, since CD133 positivity was correlated with the expression of resistance-related proteins (28), the proportion of CSCs in patients who had received radiotherapy and/or chemotherapy was higher compared to that in patients who received surgery alone. As a result, it was not clear whether CD133 expression was of higher prognostic value in patients who received radio-therapy and/or chemotherapy. Further studies are required to elucidate the direct association between CD133 expression and the prognosis of lung cancer.

A meta-analysis is a quantitative approach in which individual study findings on the same topic are statistically integrated and analyzed. With more samples, the results of a meta-analysis are more reliable compared to those of a single study. However, the present meta-analysis had certain limitations. When analyzing whether CD133 expression was associated with tumor differentiation or lymph node metastasis, there was significant heterogeneity. The methods used to detect CD133 expression varied widely among the studies and heterogeneity was eliminated following exclusion of the studies that did not use IHC. Although IHC was the most commonly applied method, the cut-off level was defined differently among the studies. In addition, the IHC results were based on the primary antibody used and different antibodies were used by the eligible studies. The dilution of the antibody also varied, leading to differences in the sensitivity of the method, depending on the antibody concentration. Another factor was the effect of ethnicity. The majority of the studies included were conducted on Asian patients. There is the possibility that different ethnic groups exhibit differences in CD133 expression, leading to heterogeneity and bias.

Although there was no evident publication bias in this meta-analysis, we were not able to completely exclude biases. For example, the study was restricted to studies published in English and Chinese, which may lead to bias.

In conclusion, our meta-analysis suggests that CD133 expression is significantly associated with poor differentiation and lymph node metastasis in lung cancer. CD133 expression is most likely correlated with poor prognosis of lung cancer. However, further studies are required, with larger patient samples, unified methods and cut-off levels to detect CD133 expression, classified by tumor stage, therapeutic schedule, follow-up time and survival events, to confirm the findings of the present meta-analysis.

This study was a project funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions and also supported by the National Natural Science Foundation of China (no. 81272602).

 References JemalASiegelRXuJWardECancer statisticsCA Cancer J Clin602773002010 AlbergAJFordJGSametJMAmerican College of Chest PhysiciansEpidemiology of lung cancer: ACCP evidence-based clinical practice guidelines (2nd edition)Chest132Suppl 3S29S552007 DontuGLiuSWichaMSStem cells in mammary development and carcinogenesis: implications for prevention and treatmentStem Cell Rev12072132005 DonnenbergVSDonnenbergADMultiple drug resistance in cancer revisited: the cancer stem cell hypothesisJ Clin Pharmacol458728772005 ToyookaSMitsudomiTSohJMolecular oncology of lung cancerGen Thorac Cardiovasc Surg595275372011 SinghSKClarkeIDTerasakiMBonnVEHawkinsCSquireJDirksPBIdentification of a cancer stem cell in human brain tumorsCancer Res63582158282003 SuetsuguANagakiMAokiHMotohashiTKunisadaTMoriwakiHCharacterization of CD133+ hepatocellular carcinoma cells as cancer stem/progenitor cellsBiochem Biophys Res Commun3518208242006 FerrandinaGBonannoGPierelliLExpression of CD133-1 and CD133-2 in ovarian cancerInt J Gynecol Cancer185065142008 KojimaMIshiiGAtsumiNFujiiSSaitoNOchiaiAImmunohistochemical detection of CD133 expression in colorectal cancer: a clinicopathological studyCancer Sci99157815832008 EramoALottiFSetteGIdentification and expansion of the tumorigenic lung cancer stem cell populationCell Death Differ155045142008 JanikovaMSkardaJDziechciarkovaMIdentification of CD133+/nestin+ putative cancer stem cells in non-small cell lung cancerBiomed Pap Med Fac Univ Palacky Olomouc Czech Repub1543213262010 CuiFWangJChenDChenYJCD133 is a temporary marker of cancer stem cells in small cell lung cancer, but not in non-small cell lung cancerOncol Rep257017082011 BertoliniGRozLPeregoPHighly tumorigenic lung cancer CD133+ cells display stem-like features and are spared by cisplatin treatmentProc Natl Acad Sci USA10616281162862009 XuYHWangJMZhangGBHuHCExpression and clinical significance of CD133 and B7-H4 in non-small cell lung cancerJiangsu Med J44124152011 HigginsJPThompsonSGDeeksJJAltmanDGMeasuring inconsistency in meta-analysesBMJ3275575602003 EggerMDavey SmithGSchneiderMMinderCBias in meta-analysis detected by a simple, graphical testBMJ3156296341997 BeggCBMazumdarMOperating characteristics of a rank correlation test for publication biasBiometrics50108811011994 Cortes-DericksLGalettaDSpaggiariLSchmidRAKaroubiGHigh expression of octamer-binding transcription factor 4A, prominin-1 and aldehyde dehydrogenase strongly indicates involvement in the initiation of lung adenocarcinoma resulting in shorter disease-free intervalsEur J Cardiothorac Surg41e173e1812012 HerpelEJensenKMuleyTThe cancer stem cell antigens CD133, BCRP1/ABCG2 and CD117/c-KIT are not associated with prognosis in resected early-stage non-small cell lung cancerAnticancer Res31449145002011 TirinoVCamerlingoRFrancoRThe role of CD133 in the identification and characterisation of tumour-initiating cells in non-small-cell lung cancerEur J Cardiothorac Surg364464532009 LiHWangYYuLZhaoPClinicopathological significance of expression of CD133 protein in neuroendocrine lung carcinoma tissuesChin J Cancer Prev Treat1829312011 LiFZengHYingKThe combination of stem cell markers CD133 and ABCG2 predicts relapse in stage I non-small cell lung carcinomasMed Oncol28145814622011 ShienKToyookaSIchimuraKPrognostic impact of cancer stem cell-related markers in non-small cell lung cancer patients treated with induction chemoradiotherapyLung Cancer771621672012 RiveraCRiveraSLoriotYVozeninMCDeutschELung cancer stem cell: new insights on experimental models and preclinical dataJ Oncol20115491812011 WooTOkudelaKMitsuiHPrognostic value of CD133 expression in stage I lung adenocarcinomasInt J Clin Exp Pathol432422010 WeiYPWangMHuaPExpression of tumor stem cell marker CD133 in non-small cell lung carcinoma and its clinical significanceJ Sun Yat-Sen Univ (Med Sci)293123162008 YaoJWangZGTongXWLiZJYuZGExpression of tumor stem cell marker CD133 and CD44 in original tumor tissues and metastatic lymph nodes of lung cancerMed J Natl Defend Force Southwest China20130013032010 SalnikovAVGladkichJMoldenhauerGVolmMMatternJHerrICD133 is indicative for a resistance phenotype but does not represent a prognostic marker for survival of non-small cell lung cancer patientsInt J Cancer1269509582010 LinXLiuSLiuNYangXXuHWangEExpression and significance of stem cell markers CK19, Notch3, CD133, P75NTR, STRO-1 and ABCG2 in pulmonary squamous carcinomasZhongguo Fe123163212009(In Chinese). GuYPSunMMGuLQZhangHXieFExpression and significance of cancer stem cell marker CD133, ABCG2 and p75NTR in non-small cell lung carcinomaSuzhou Univ J Med Sci305135162010 ChengJRWangSQZuMRTZouJExpressions of CD133 and CD105 in lung cancer tissue and their clinical significanceTumor303343372010 SunHYYangMZhengMJRenZJLiuHExpression of CD133 and ALDH1 in non-small cell lung cancer and their clinical significanceJ Clin Exp Pathol288138152012 XuLZYangWTThe criteria of immunohistochemical reaction resultsChina Oncol62292311996 Figures and Table

Correlation of CD133 with tumor stage.

Correlation of CD133 with tumor differentiation.

Correlation of CD133 with histological type.

Correlation of CD133 with lymph node metastasis.

Subgroup analysis of the correlation of CD133 with lymph node metastasis.

Main characteristics of the 15 eligible studies.

First author Year Histological type Method Cut-off level No. of patients No. of case groups Refs.
Shien 2012 NSCLC IHC >1% staining 30 9 (23)
Bertolini 2009 NSCLC IHC Strong staining 58 32 (13)
Salnikov 2010 NSCLC IHC Any staining 88 56 (28)
Cortes-Dericks 2012 AD RT-PCR Median 64 63 (18)
Li 2011 NSCLC IHC >1% staining 145 46 (22)
Herpel 2011 NSCLC Tissue microarray Diffuse expression or distinct staining in at least 1 out of 4 tissue cores per sample 86 13 (19)
Xu 2011 NSCLC IHC Any staining 103 51 (14)
Wei 2008 NSCLC IHC >10% staining 77 40 (26)
Li 2011 N/E LC IHC >10% staining 90 44 (21)
Yao 2010 LC IHC >10% staining 42 31 (27)
Lin 2009 SQ IHC Any staining 54 27 (29)
Gu 2010 NSCLC IHC >10% staining 44 30 (30)
Cheng 2010 NSCLC IHC >10% staining 65 45 (31)
Sun 2012 NSCLC IHC Method described by Xu and Yang (33) 67 42 (32)
Tirino 2009 NSCLC Flow cytometry NA 89 64 (20)

NSCLC, non-small-cell lung cancer; AD, adenocarcinoma; LC, lung cancer; SQ, squamous cell carcinoma; N/E LC, neuroendocrine lung cancer; IHC, immunohistochemistry; RT-PCR, reverse transcriptase-polymerase chain reaction; NA, not available.