Detection of human cytomegalovirus in normal and neoplastic breast epithelium

Abstract

Introduction: Human cytomegalovirus (HCMV) establishes a persistent life-long infection, and can cause severe pathology in the fetus and the immunocompromised host1. Breast milk is the primary route of transmission in humans worldwide, and breast epithelium is thus a likely site of persistent infection and/or reactivation, though this phenomenon has not previously been demonstrated. Increasing evidence indicates HCMV infection can modulate signaling pathways associated with oncogenesis. We hypothesized that persistent HCMV infection occurs in normal adult breast epithelium and that persistent viral expression might be associated with normal and neoplastic ductal epithelium.

Methods: Surgical biopsy specimens of normal breast (n = 38) breast carcinoma (n = 39) and paired normal breast from breast cancer patients (n = 21) were obtained. Specimens were evaluated by immunohistochemistry, in situ hybridization, PCR and DNA sequencing for evidence of HCMV antigens and nucleic acids.

Results: We detected HCMV expression specifically in glandular epithelium in 17/27 (63%) of normal adult breast cases evaluated. In contrast, HCMV expression was evident in the neoplastic epithelium of 31/32 (97%) patients with ductal carcinoma in situ (DCIS) and infiltrating ductal carcinoma (IDC) cases evaluated (p = 0.0009).

Conclusions: These findings are the first to demonstrate that persistent HCMV infection occurs in breast epithelium in a significant percentage of normal adult females. HCMV expression was also evident in neoplastic breast epithelium in a high percentage of normal and neoplastic breast tissues obtained from breast cancer patients, raising the possibility that viral infection may be involved in the neoplastic process.

Figure 1

Immunohistochemical detection of HCMV in normal breast and breast cancer. (a, b) Examples of HCMV IE immuonreactivity in isolated area of normal ductal epithelium in normal breast from a reduction mammoplasty patient are presented. Low (a, 40×) and higher (b, 100×) power views of the same area of immunoreactive epithelium demonstrate discrete perinuclear and cytoplasmic epithelial cell staining. No IE immunoreactivity is detected in two different specimens obtained from normal reduction mammoplasty (c and d, 40×; e is 100× magnification of d). IE immunoreactivity is demonstrated primarily in a nuclear distribution in matched non-neoplastic epithelium from a patient with infiltrative ductal carcinoma (f, 100×). Early and late (E/L) immunoreactivity is demonstrated in the tumor epithelial cells, but not the stroma, from an area of ductal carcinoma in situ (DCIS) in a patient with infiltrative ductal carcinoma (g, 40×, h, 100×). Positive control immunostaining for IE immunoreactivity is shown in HCMV infected pneumocytes from an AIDS patient with CMV pneumonia (i, 100×). Late antigen (L) immunoreactivity is shown in another patient with infiltrative ductal carcinoma (j, 40×; k, 100×). Negative control (IE isotype control antibody staining for CD34) immunoreactivity is seen only in vascular endothelial cells in an area of infiltrative ductal carcinoma (l, 40×).

Figure 2

Immunohistochemical detection of three different HCMV antigens in a serial sections from infiltrative ductal carcinoma. Immunoreactivity with monoclonal antibodies specific for HCMV immediate early (IE; a), early and late (E/L; b), and late (L; c) antigens is shown in serial sections from a single specimen of infiltrative ductal carcinoma (100×).

Figure 3

HCMV in situ hybridization in infiltrating ductal carcinoma. Two examples of HCMV ISH staining from two different patients with infiltrating ductal carcinoma are displayed. Low and high power images of the first patient (a, 40×; b, 100×) demonstrate heterogeneous pattern of detection of HCMV nucleic acids in the nuclei (dark blue) and cytoplasm (light blue) in infiltrating tumor cells but no significant nucleic acid detection in the intervening stroma. Two images from another example of infiltrating ductal carcinoma (c, d; 100×) also reveal areas of HCMV nucleic acid detection in nuclear (dark purple) and cytoplasmic (light purple) areas, without any significant nucleic acid detection in the intervening stroma. Positive control probe specific for DNA Alu repeats from the same case reveals intense (dark purple) nuclear probe hybridization with nucleic acids in infiltrating tumor nuclei and intervening stromal cells, without evidence of cytoplasmic staining (e, 100×). Hybridization signal of a negative control probe specific to insect nucleic acids is completely absent in tumor tissues (f, 100×). Positive control (HCMV infected lung) is positive for HCMV nucleic acid hybridization in scattered pneumocytes (blue cells in g, 100×), while negative control probe is not detected in the same specimens (h, 100×; light hematoxylin counterstain was used in g and h). Low power image (i, 40×) of normal breast epithelium (from reduction mammoplasty) that was negative for HCMV nucleic acid hybridization reveals no hybridization signal. High power (j, 100×) image of normal breast epithelium from reduction mammoplasty reveals faint purple hybridization signal in scattered normal ductal epithelial cells.