The clinicopathological significance of neurogenesis in breast cancer

One hundred and ninety-six cases were retrieved from the files of the Departments of Pathology in Shanghai Third People's Hospital, Shanghai Jiaotong University School of Medicine and Shanghai First Maternity and Infant Healthy Hospital, Tongji University. The most histological type was invasive ductal carcinoma (144 cases), followed by ductal carcinoma in situ (DCIS, 18 cases), fibroadenoma (14 cases) and then normal breast tissue (20 cases). The pathological parameters, including tumor size, differentiation and the presence of nodal metastasis, were carefully reviewed. The histological grade and stage were evaluated by a modified Bloom-Richardson grading system and American Joint Committee on Cancer (AJCC), respectively. Out of 144 invasive ductal carcinoma (age range = 32-71 years; average age = 48.06 years), 62.5% (90/144) were equal to/more than 3 years of disease free survival. All the patients with IDC received adjuvant therapy. In addition, the patients with distant metastasis were not enrolled in this study. One representative paraffin block from each case was used for the study. The study was approved by the Ethical Review Boards of Shanghai Third People's Hospital, Shanghai Jiaotong University School of Medicine and Shanghai First Maternity and Infant Healthy Hospital, Tongji University. No consent from patients involved in this study was needed because the required consent was waived by Ethical Review Boards. The information about patients involved in this study is kept confidential at all times.

Immunohistochemical assays were performed on formalin-fixed paraffin-embedded tissues. Sections (5 μm thick) were cut, deparaffinized in xylene and rehydrated in graded alcohols. Slides were boiled in citrate buffer (pH 6.0) at 95 ~ 100°C for 5 minutes and were cooled for 20 minutes. Endogenous peroxide was blocked by 3% hydrogen peroxide in methanol for 10 minutes. Sections were incubated with rabbit anti-human PGP 9.5 (1:500, DAKO, Carpinteria, CA, USA), rabbit anti-human S100 (1:200, DAKO, Carpinteria, CA, USA) and mouse anti-human CD34 antibodies overnight at 4°C. Immunohistochemical staining was performed using EnVision + HRP DAB system (DAKOCytomation, Carpinteria, CA, USA). All sections were counterstained with Meyer’s Hematoxylin. The sections processed without the primary antibodies were used as negative control.

A pathologist read all immunostained slides. Each slide was marked at points in which positive PGP 9.5/S100 immunostaining was shown. A single digital image was created with the Olympus BLISS HD virtual microscopy system at × 400 magnification. The diameters of nerve fascicles were measured with the Optimas 6 Image Analysis Suite (Optimas Corp.) Most of the diameters of nerve fascicles were less than 100 μm (94.7%). Nerve density was evaluated by counting the number of nerve fascicles with diameters of < 100 μm in 20 continuous fields at × 200 magnification. Nerve density results were grouped into 3 categories: 1) negative, no nerve fascicles or nerve fibers, 2) weak expression, 1 to 10 nerve fascicles, and 3) moderate/strong expression, > 10 nerves fascicles. Intratumoral microvessel density (MVD) was recorded by counting CD34-positive vessels in the most vascularized area in four × 200 fields [10]. Blood vessels with a lumen diameter exceeding approximately eight red blood cells were excluded. For estrogen receptor (ER) and progesterone receptor (PR), we defined cases with more than 5% positive tumor cells of moderate intensity as positive. Immunohistochemistry results were analyzed by three independent pathologists under a multihead microscope in cases of disagreement.

All statistical analyses were carried out using SPSS software (SPSS Ver. 11.0, USA). Some data was presented as absolute numbers and percentages, other data was presented as mean ± SD. χ ²-test was used to examine the association between PGP 9.5 expression and the various clinicopathological characteristics. The relationship between PGP 9.5 expression and MVD was evaluated using the Mann-Whitney test. Reported P-values less than 0.05 were considered as significant.

PGP 9.5 or S100 expression was identified in normal breast tissue control cases and in high percentage of breast cancer cases. The pattern of neurogenesis in IDC was shown in Figure 1. In the most of cases the fragmented nerve fascicles were distributed in the tumor stroma. In some cases, the scattered fine nerve fibers were seen surrounding the blood vessels. Furthermore, some fine nerve fibers were sporadically located around the cancer cells. Nerve fibres were found in all the normal breast tissue control cases. However, there is no PGP 9.5 positive nerve fiber identified in the stroma of cases of fibroadenoma and DCIS. Overall, 61.8% (89/144) of invasive ductal carcinoma cases exhibited evidence of neurogenesis. PGP 9.5 positive nerve fibers were observed in all normal breast tissue controls (Table 1). There is no difference in diameters of the nerve fibers between normal breast tissue control (23.4 ± 8.2 μm) and in invasive ductal carcinomas (20.8 ± 10.4 μm).

While analyzing invasive ductal carcinomas of different MBR grade, we found that PGP 9.5 positive rate was 29.4% (10/34) in grade I cases, but PGP 9.5 positive rate was 71.8% (79/110) in grade II and III cases, the difference was statistically significant (p < 0.0001). The PGP 9.5 positive rate was significantly higher in cases of less than 3 years of disease-free survival (64.8%, 35/54) compared to cases of equal or more than 3 years of disease-free survival (46.7%, 42/90; p = 0.035). However, no significant difference in PGP 9.5 expression was found between tumor groups of different stage (stage I: 56.5%; stage II: 63.5%; stage III: 61.8%). Likewise, no significant difference of PGP 9.5 expression was found in cases of different lymph node, ER and PR status (Table 2).

As shown in Table 3, moderate/strong PGP 9.5 expression (more than 10 nerves) was found to be significantly related with tumor MVD (p = 0.014). Intriguingly, analysis of subgroups of ER and node status revealed that moderate/strong expression of PGP 9.5 was significantly associated with higher MVD in the ER-negative (p = 0.045) and node-negative (p = 0.039) subgroups. No significant association was found between PGP 9.5 expression and MVD in ER-positive and node-positive subgroups.