Introduction. Gastric cancer is considered to be the fourth most common malignancy worldwide and the second cause of cancer deaths.
Regarding the cancer stem cells (CSCs) theory, they are a small group of tumor cells with unrestricted self-renewal and differentiation abilities
that help tumor formation. There is an interest in the utility of CD133 as a promising marker to detect the tumor stem cell population
for a variety of solid malignancies including gastric cancer. Tumors that express stem cell markers such as CD133 are found to be more
aggressive tumors with poor prognosis and high liability for recurrence. This study aimed to evaluate the immunohistochemical expression of
CD133 in invasive gastric carcinoma and study the relation between CD133 immunohistochemical expression and different clinicopathological
Material and methods. 77 cases of gastric carcinoma were collected from the surgical pathology unit at the Gastroenterology Center, Mansoura
University, Egypt. CD133 expression in tumor tissue was evaluated by immunohistochemistry.
Results. CD133 expression positively correlated with tumor metastasis and recurrence. Multivariate analysis revealed CD133 positivity to be
an independent prognostic factor for tumor recurrence (P = 0.03).
Conclusion. CD133 is a good marker that can predict tumor recurrence and metastasis in gastric carcinoma. Even though, studies regarding
CSCs are still in their initial stages especially those related to CD133 in gastric cancer.


Gastric cancer is one of the most aggressive human malignancies. It is considered by the WHO as the fourth most common cancer worldwide and the second leading cause of cancer deaths 1. According to The Egyptian National Cancer Institute, it is the fourteenth most common cancer representing 1.8% of cases in both sexes 2. Tumor recurrence and metastasis are the main causes of death in gastric cancer patients, regardless of surgical intervention and post-operative adjuvant therapy 3. The overall 5-year survival rate is less than 20% as most cases are usually diagnosed late and are incompatible for curative surgery. Therefore, there is a critical need to find more effective methods to eliminate cancer cells 4.

CSCs correspond to a subpopulation of cells within the tumor defined by self-renewal, asymmetrical division, and differentiation properties, giving rise to the more or less differentiated cells composing the tumor mass 5. CSCs were first known in the early half of the 2000s in solid cancer. A treatment targeting CSCs has become a challenge for radical cure of cancer. Therefore, it is important to identify and isolate CSCs using specific markers known as CSCs markers 6.

Different markers have been found to be expressed on the surface of CSCs. CD133 is one of these markers that has retained much attention and importance 7. CD133 receptors on gastrointestinal cancer cells was firstly recognized by Smith et al. 8. It was found that an antibody-drug conjugate against this receptor cause delayed growth response in gastric cancer cells 9.

Materials and methods


This retrospective study was performed using gastrectomy specimens from 77 cases of gastric carcinomas that were collected during the period from January 2012 until September 2015 from the archive of the pathology department. None of the patients had received preoperative adjuvant therapy, chemotherapy or radiotherapy. The protocol was approved by the Ethical Committee of Mansoura University, and performed in accordance with the principles of the Declaration of Helsinki. Written informed consents were obtained.


The paraffin blocks were re-cut in 5-μm-thick sections and stained with hematoxylin and eosin and were independently reviewed by two pathologists for confirmation of diagnosis.


For immunohistochemical staining for CD133, 5-μm-thick tissue sections were cut and mounted on coated slides. The sections were deparaffinized in xylene then rehydrated in a series of decreasing concentration of ethanol. After that, heat-induced antigen retrieval was done using a pressure cooker and EDTA buffer (PH 9) and immersed in peroxidase-blocking solution (3% hydrogen peroxide) for 15-20 minutes to inhibit endogenous peroxidase activity, the slides were washed in phosphate buffer saline (PBS). The slides were – then – incubated with the Anti-CD133, mouse monoclonal antibody (1:40, clone AC133, Miltenyi Biotech, Bergisch Gladbach, Germany) that was used as primary antibody for CD133 detection for 1 hour at room temperature, followed by incubation with the secondary antibody (poly horseradish peroxidase conjugate for mouse/rabbit) for 15 minutes at room temperature. Diaminobenzedene (DAB) was the chromogen used to visualize the stain. The slides were then counterstained with hematoxylin, dehydrated, coversliped, and finally mounted with DPX 10. Both normal gastric mucosa and normal kidney tissue were used as replace internal bypositive 11 12. Negative control was done by substituting the primary antibody with phosphate buffer saline 13.


CD133 show membranous and cytoplasmic pattern of expression 14. We consider the pattern of expression as the more dominant type was seen in the specimen 6. CD133 results were scored to as the sum of CD133 intensity of staining reaction in tumor cells plus CD 133 percent of expression in tumor cells. CD133 intensity was graded from zero to three at which (0 for negative expression, 1 for mild intensity, 2 for moderate intensity and 3 for strong intensity) and CD133 percent of expression also was graded from zero to three at which (0 for negative percent, 1 for > 10% of tumor cells, 2 for 10% to < 50% of tumor cells and 3 for ≥ 50% of tumor cells) 15. The receiver operating characteristic (ROC) curve was used to set the cut-off point (Fig. 1) 16 17. ROC curves facilitate the detection of the threshold value that corresponds to maximum sensitivity with minimal loss of specificity and above which a test result should be considered positive for some outcome 16.


Statistical analysis was done by using IBM SPSS, version 20.0. The relation between CD133 expression and the clinicopathological parameters were assed as a univariate analysis. Multivariate analysis was done using binary logistic regression test to detect the predictors of CD133 expression.



Seventy-seven cases of gastric carcinomas were involved in this study. The mean age of the studied cases was 54.44 years, and the youngest patient at time of diagnosis was 24 year old, while the oldest was 81 years. The cases showed male predominance as it included 45 males (58.4%) and 32 females (41.6%). The tumor size of the excised biopsies ranged from 1 cm to 18 cm in its maximum diameter with mean tumor size was (6.05 ± 3.3 cm). Grade II adenocarcinoma was the most common histology. The intestinal type was also the predominant according to Laurén classification. Fungating tumor growth pattern was the commonest. The tumor depth of invasion in most of the cases was T3 (75.3%), lymph node metastasis were positive in 57 cases (74.0%), while 10 cases only (13.0%) out of total number of the cases showed distant metastasis. Regarding TNM staging most of the cases were present in stage III (39%). Nineteen cases were positive for lymphovascular invasion (24.7%) and 17 cases were positive for perineural invasion (22.1%), 11 cases (14.3%) showed tumor recurrence (Tab. I).


ROC curve analysis was used to determine the clinically relevant threshold for CD133 positivity, the significant cut off was < 3.5, 32 cases showed CD133 score (0-3) that consider negative and about 45 cases showed CD133 score (4-6) that consider positive. Regarding CD133 pattern of expression ten cases were negative. Thirty five cases showed membranous pattern and thirty two cases showed cytoplasmic pattern (Tab. II). Examples of the staining patterns are shown in Figure 1.

Table III describes the relation between CD133 score and some of the clinico-pathological features. There was a significant relationship between CD133 score and distant metastasis as nine out of ten cases with distant metastasis showed high CD133 score (4-6) with p value = 0.03. Also there was a significance between CD133 score and tumor recurrence as ten out of elven cases with tumor recurrence showed high CD133 score with P value (0.02). Multivariate analysis using binary logistic regression test demonstrated that cases with tumor metastasis have 7.75 time incidence of expression of CD133 (P = 0.058), while cases with tumor recurrence have 8.86 time incidence of CD133 expression (p = 0.04), CD133 is an independent prognostic factor for tumor recurrence (Tab. IV).


CD133 was considered as one of the most prominent CSC markers that was found to be over expressed in hepatocellular carcinoma, gastric cancers, colorectal cancer, pancreatic cancer and other types of cancer. In the current study, CD133 positivity was found in 58.4% of cases. It was found that presence of CD133 positive staining in minor populations of cells may be indicative of cancer stem-like cells which lead to tumor recurrence and metastasis. Our study aimed to study the relation between CD133 immunohistochemical expression and different clinicopathological parameters.

This study showed a significant relation between CD133 score and tumor distant metastasis and recurrence and this agreed with the study by Zhao et al. that reviewed 336 gastric carcinoma cases, 46 cases showing distant metastasis, 38 cases of which showed CD133 overexpression 10. It also agreed with the results of a meta-analysis based study including 26 studies with a total of 4729 cases, a relation between CD133 overexpression and tumor metastasis was found 18. Also, significant relation was found between CD133 score and tumor recurrence as 10 of 11 cases with tumor recurrence had score (4-6) with P value = 0.02, it was also found that CD133 is an independent prognostic factor for tumor recurrence.

This positive relationship between high CD133 expression and tumor aggressiveness suggests that CD133-positive cells have more cancer stem-like cells. These cancer stem-like cells with their self-renewal and differentiation abilities could be responsible for tumor recurrence and metastasis even after traditional treatment 19. It indicates that CD133 may be a promising prognostic markers in gastric adenocarcinoma 10.

This study found no significant relation between CD133 score and the sex and age. This finding agreed with that of Ishigami et al.; Zhao et al.; Poddar et al. 20 10 21 and in contrary to that of Nosrati et al. that reported positive correlation between CD133 and age over 65 years with P < 0.001 that is statistically highly significant 11. Moreover, there was insignificant relation between CD133 expression and tumor size, histological type and grade and TNM staging. This finding disagreed with some studies including Ishigami et al., Zhao et al., Nosrati et al. and Poddar et al. that found significant relation between CD133 expression and tumor size, histological type, grade and TNM staging 20 10 11 21. This discrepancy in significance may be related to different patient characteristics, genetics, race, environmental factors, and the application of different scoring methods for CD133 expression. A non-significant relation between lymphovascular invasion and CD133 expression was found in the current study that agreed with that of Nosrati et al. 11, but in contrast to that of Ishigami et al.; Yu et al. and Poddar et al. that reported significant relation between CD133 expression and lymphovascular invasion with p value = (0.027), (0.000) and (0.01) respectively that was considered significant 4 20 21.

In conclusion, CD133 is a good molecular biomarker that can predict tumor recurrence and metastasis in gastric carcinoma. Regardless, studies regarding CSCs are still in their initial stages especially those related to CD133 in gastric cancer. Further studies are required to confirm the role of CD133 in tumor recurrence and metastasis and consequently promote finding more appropriate treatment modalities involving targeted gene therapy.

Figures and tables

Fig. 1..

Tab. I..

Age N %
≤ 55 37 48.1
< 55 40 51.9
Male 45 58.4
Female 32 41.6
Tumor size
Median ± SD 6.05 ± 3.3
Mean (min-max) 5.0 (1.0-18.0)
Tumor depth of invasion
T1 3 3.9
T2 14 18.2
T3 58 75.3
T4 2 2.6
Regional LN metastasis
Negative (N0) 20 26.0
Positive (N1, N2, N3) 57 74.0
Distant metastasis
M0 67 87.0
M1 10 13.0
P TNM stage
I 10 13
II 27 35.1
III 30 39
IV 10 13.0
Lymphovascular invasion
Positive 19 24.7
Negative 58 75.3
Perineural invasion
Positive 17 22.1
Negative 60 77.9
Tumor recurrence
Positive 11 22.2
Negative 66 77.8
Histological type and grade
Grade 1 adenocarcinoma 5 11.1
Grade 2 adenocarcinoma 34 44.1
Grade 3 adenocarcinoma 15 19.4
Mucoid adenocarcinoma 6 7.79
Signet ring carcinoma 11 14.28
Undifferentiated carcinoma 6 7.79
Lauren classification
Intestinal 60 77.9
Diffuse 17 22.1
Ming classification
Fungating 35 45.5
Infiltrating 14 18.2
Malignant ulcer 28 36.4
Clinicopathological features of the studied cases.

Tab. II..

CD 133 score N %
0-3 32 41.6
4-6 45 58.4
CD133 pattern
Negative 10 13
Membranous 35 45.5
cytoplasmic 32 41.6
Immunohistochemical characteristics.

Tab. III..

Clinicopathological parameters CD 133 0-3 (n = 32) CD133 4-6 (n = 45) test of significance
Age N % n %
≤ 55 14 43.8 23 51.1 χ 2 = 0.41 p = 0.52
< 55 18 56.2 22 48.9
Male 22 68.8 23 51.1 χ 2 = 2.39 p = 0.12
Female 10 31.2 22 48.9
Tumor size
Median (Min-Max) 5.0(1.0-18.0) 6.0(1.0-18.0) Z = 0.037 P = 0.97
Tumor depth of invasion
T1 2 6.2 1 2.2 MC P = 0.39
T2 4 12.5 10 22.2
T3 26 81.2 32 17.1
T4 0 0.0 2 4.4
Regional LN metastasis
Positive NO 8 25.0 12 26.7 χ 2 = 0.03 P = 0.87
Negative N1,N2,N3 24 75.0 33 73.3
Histological type and grade
Grade 1 adenocarcinoma 0 0.0 5 11.1 MC P = 0.15
Grade 2 adenocarcinoma 14 43.8 20 44.4
Grade 3 adenocarcinoma 6 18.8 9 20.0
Mucoid adenocarcinoma 3 9.4 3 6.7
Signet ring carcinoma 4 12.5 7 15.6
Undifferentiated carcinoma 5 15.6 1 2.2
Lauren classification
Intestinal 23 71.9 37 82.2 χ 2 = 1.16 p = 0.28
Diffuse 9 28.1 8 17.8
Ming classification
Fungating 15 46.9 20 44.4 χ 2 = 0.09 P = 0.95
Infiltrating 6 18.8 8 17.8
Ulcerative 11 34.4 17 37.8
Distant metastasis
M0 31 96.9 36 80.0 χ 2 = 4.7 P = 0.03 *
M1 1 3.1 9 20.0
I-II 15 46.9 22 48.9 χ 2 P = 0.86
III-IV 17 53.1 23 51.1
Lymphovascular invasion 8 25.0 11 24.4 χ 2 = 0.003 p = 0.96
Perineural invasion 9 28.1 8 17.8 χ 2 = 1.16 p = 0.28
Tumor recurrence
Negative 31 96.9 35 77.8 χ 2 = 5.57 p = 0.02 *
Positive 1 3.1 10 22.2
The relation between CD133 score and the clinicopathological parameters.

Tab. IV..

Predictors Odds ratio 95% CI P
Tumor recurrence (positive) 8.86 (1.07-73.18) 0.04 *
Tumor metastasis (positive) 7.75 (0.93-64.6) 0.058
Binary logistic regression to detect predictors of CD133 positive cases.