Evaluation of angiogenesis in cervical cancer using CD34 as a biomarker and its correlation with pathoanatomical features

Aim: CD34 is an antigen found in hematopoietic progenitor cells and endothelial cells. Anti-CD34 antibody is a highly sensitive marker for neovascularisation and has also been studied as a marker for vascular tumours. However, there are few studies relating it to cervical carcinoma. The aim of our study is to evaluate the CD34 expression as a marker of angiogenesis in cervical cancer and to correlate the microvessel count and Microvessel Density (MVD) with histopathological type and grade of tumour and lymph node metastasis. Method: Total 75 malignant cases (65 cases of squamous cell carcinoma and 10 cases of adenocarcinoma) were stained with anti CD34 antibody to highlight the blood vessels. In each case, areas of highest neo-vascularization were identi ﬁ ed using scanner lens. Blood vessels were counted in 10 microscopic ﬁ elds at 400X magni ﬁ cation. Average number of non-canalized micro-vessels count per ﬁ eld was calculated (MVD). Result: The total number of microvessels stained per case ranged from 54 to 216. No statistically signi ﬁ cant differences were observed for MVD with respect to the age of patient and histological type of carcinoma (p > 0.05). The mean vascular density in differentiated carcinoma and undifferentiated carcinoma, was 10.56 and 14.14 respectively. This clearly shows that the distribution of microvessels was higher among undifferentiated carcinomas, as compared to differentiated carcinoma (p < 0.05). Furthermore, the MVD in lymph node positive cases was 12.84, slightly higher as compared to lymph node negative cases where it was 11.34. However, the relationship was statistically nonsigni ﬁ cant (p > 0.05). Conclusion: CD 34 can be used as a novel marker for estimation of MVD in cervical cancer, however further studies are required on large group of patients to establish its role as a therapeutic target against angiogenesis.


Introduction
The cervical carcinoma represents one of the most frequent malignant tumours in female. Great advancements were made in the understanding of the mechanisms of the appearance and progression of the tumours, and also in the early diagnosis of them. In the present time, the discussions are focused on the angiogenesis and the neovascularisation. Angiogenesis is the growth of new blood vessels toward and within a tumour, and it has been shown that tumours do not proliferate beyond a size of 2 to 3 mm 2 unless they are able to develop new capillaries from the existing vascular network [1]. Previous researches have demonstrated that microvessel density in carcinomas cervix is a factor associated with poor prognosis. It also tends to be related with pathological features indicative of a greater risk of recurrence of tumour [2]. The purpose of this study was to determine in patients with cervical cancer whether the intensity of tumour angiogenesis, expressed as microvessel Citation: Aijaz  Various researchers used different markers to highlight blood vessels in cervical carcinoma in order to measure angiogenesis. The different antibodies that were used are factor VIII or Von Willebrand factor, European I lecithin Ulex, and anti-CD31. Fewer studies were found using anti-CD34 in cancer of the uterine cervix, thus we used anti-CD34 as a marker of angiogenesis in our study, owing to the ability of anti-CD34 antibody to highlight blood vessels more strongly as compared to normal vessels.
CD34 is a transmembrane, highly glycosylated protein expressed by hematopoeitic stem/progenitor cells [3] and endothelial cells [4]. On endothelial cells, CD34 acts as a ligand for L-selectin [5], although this is not the case in HPSCs, where its function remains unclear.
Besides these, CD34+ fi brocytes also known to be present in the connective tissue compartment of various tissue such as the skin [6], gastrointestinal [7] and urogenital tracts [8], pancreas [9], breast [10], thyroid [11] and peripheral nervous system [12]. In addition, CD34 is known to be present in stromal fi broblasts of the uterine cervix [8]. Thus CD34 reactivity of normal endocervical and other uterine stromal tissues should be kept in mind while evaluating angiogenesis, as it may contribute to falsely elevated MVD. It is therefore necessary for the pathologist to try to identify the tissue morphology carefully on H & E stained slides, so that an adequate description of microvessels is possible.

Observations and results
The present study was conducted in the Department of Immunohistochemistry for CD34 was performed on paraffi n embedded sections using anti CD34 antibody (QBEnd/10) to highlight the blood vessels. All cases of cervical carcinoma were considered for this study (total 75 cases). Sections were examined and microvessel density and microvessel count was calculated for each case of cervical cancer. A vascular unit was identifi ed according to criteria established by Weidner, et al.
1991 [13]. In each case tumour sections were screened under scanner to fi nd out areas of highest neo-vascularization.
Using 400X magnifi cation blood vessels were counted in 10 microscopic fi elds ( Figure 1A & 1B). Average number of noncanalized micro-vessels count per fi eld was calculated. Any brown staining endothelial cell or endothelial cell cluster which was clearly separate from adjacent micro-vessels was considered a single, countable micro-vessel. Red blood cells in lumen were not used to defi ne a micro-vessel. Statistical analysis was carried out using the Mann-Whitney test to compare the distribution function of microvessel density according to pathoanatomic features.
The mean age of patients with cervical carcinoma was 51.14 years with a SD of 11.34 years. Seventy-One percent of the patients had clinical stage I cancers. Out of 75 malignant cases, 65 cases were of squamous cell carcinoma and 10 cases were of adenocarcinoma (Figures 2,3). The total number of microvessels (which is a sum of microvessels in 10 hpf) stained per case ranged from 54 to 216 (Table 1).
Microvessel density was compared with respect to the age of patient and histological type of carcinoma. No statistically signifi cant differences were observed for these categories (p > 0.05).
Mean vascular density was also compared with degree or grade of differentiation of tumour. For convenience, all cases of well differentiated and moderately differentiated squamous cell carcinoma were grouped as differentiated carcinoma (72.3% cases), whereas cases of poorly differentiated carcinoma were  considered as undifferentiated carcinoma. The mean vascular density in differentiated carcinoma was 10.56, whereas in undifferentiated carcinoma, it was found to be 14.14. Thus it is clearly evident that the distribution of microvessel density is higher among undifferentiated carcinomas, as compared to differentiated carcinoma. (p < 0.05) ( Table 1). All cases of adenocarcinoma showed glandular differentiation, so such comparison was not done in cases of adenocarcinoma.
In our study, thirty-six percent cases showed lymph node involvement. Furthermore, MVD was also compared between node negative and node positive cases in order to determine whether increased angiogenesis would increase the chances of metastasis to lymph nodes. The MVD in lymph node positive cases was found to be 12.84, whereas it was 11.34 in lymph node negative cases. It is slightly higher in node positive cases.
However, the relationship is statistically insignifi cant. (p > 0.05) (Table1).   [17]. However,   they counted the microvessels density in only three fi elds using 200X magnifi cation, as against our study where we counted 10 fi elds using 400X magnifi cation. They found the number of vessels to vary from 17 to 174.

Discussion
The MVD in undifferentiated carcinoma was 14.14 whereas in differentiated carcinoma MVD was 10.56. It was observed that the distribution of MVD was higher in undifferentiated carcinoma as compared to differentiated carcinoma and this relationship was found to be statistically signifi cant (p value <0.05). In the study by Vieira, et al. (2005) the median vascular density in undifferentiated carcinoma was 10.8 whereas in differentiated carcinoma was 9.3 and this was statistically signifi cant (p > 0.03) [18]. These results are in accordance with the fi ndings of our study. Wiggens et al. (1995) compared the microvessel density in benign cervical tissue with invasive cervical carcinoma and observed a higher density in the invasive group, which was statistically signifi cant (P = 0.013) [14].
Abulafi a et al. (1996) analysed the microvessel count between normal cervical epithelium, in situ carcinomas and invasive cervical carcinoma. He also found a signifi cantly lower microvessel density in the normal tissue (P < 0.005) [16]. Thus it can be inferred from these studies that with increase in invasive potential of tumour or decrease in grade of differentiation, the MVD is signifi cantly increases.
We also observed high microvessels count per high power Hence we also studied correlation of microvessel density with lymph node metastasis. We observed that MVD in lymph node positive case is higher as compared to lymph node negative cases. However; the p value was found to be insignifi cant (p< 0.19) an observation similar to Tjalma, et al. (1998) ;Hirakawa, et al. (19990;Vieira, et al. (2005). This is in contrast to study by Bremer, et al. (1996) where a statistically signifi cant relationship was found. (p < 0.002) [2,17,18,15].
When comparing the results of the different studies published about tumour angiogenesis and cervical carcinoma, a great variation was seen among them. We analysed that besides the choice of the endothelial cell antibody, the method of microvessel quantifi cation is also important for angiogenesis study. In almost all the studies the selection area for tumour quantifi cation was mainly based on the pioneering work of Weidner, et al. (1991) [13]. In most of the studies quantifi cation was done manually, however few authors used computer method for quantifi cation. In our study manual method was used. We also observed that in these studies the amount of fi elds counted varies from unknown to 15, the magnifi cation also varies from 100 to 400. The vessel marker, methodology and the technique used are probably the explanations for the noticeably different micro vessel densities in our study compared with those in the literature.
Regardless of the methodology, with the present study we found that in most studies the microvessel density is an independent prognostic marker. CD 34 can be used as a novel marker for quantifi cation of microvessel density.
Few studies showed high MVD is associated with invasive and undifferentiated carcinoma similar to our study. Others showed association of high microvessel count with positive lymph node, higher stage, lymphovascular invasion. Some of the authors like Hirakawa et al. (1999) failed to establish any relationship with these parameters [17].

Conclusion
Our study suggests that CD 34 can be used as a novel marker for estimation of MVD in cervical cancer. It was seen that anti-CD34 antibody immunoreactivity in cervical carcinoma is associated with pathoanatomical features which are indicative of a poorer prognosis. Further studies can also be undertaken on large group of patients to establish its role as a therapeutic target against angiogenesis in cases of cervical cancer.