Medullary carcinoma of the breast: a tumour lacking keratin 19

Diagnostic and prognostic impact of cytokeratin 19 expression analysis in human tumors: a tissue microarray study of 13,172 tumors

To evaluate cytokeratin 19 (CK19) expression in normal and cancerous tissues, 15,977 samples from 122 tumor types and 608 samples of 76 normal tissue types were analyzed by immunohistochemistry (IHC). In normal tissues, CK19 expression occurred in epithelial cells of most glandular organs but was strictly limited to the basal cell layer of nonkeratinizing squamous epithelium and absent in the skin. CK19 expression in ≥90% of cases was seen in 34% of the tumor entities including the adenocarcinomas of the pancreas (99.4%), colorectum (99.8%), esophagus (98.7%), and stomach (97.7%), as well as breast cancer (90.0%-100%), high-grade serous (99.1%) or endometrioid (97.8%) ovarian cancer, and urothelial carcinoma (92.6%-100%). A low CK19 positivity rate (0.1-10%) was seen in 5 of 122 tumor entities including hepatocellular carcinoma and seminoma. A comparison of tumor versus normal tissue findings demonstrated that upregulation and downregulation of CK19 can occur in cancer and that both alterations can be linked to unfavorable phenotypes. CK19 downregulation was linked to high grade (p = 0.0017) and loss of estrogen receptor- and progesterone receptor-expression (p < 0.0001 each) in invasive breast carcinoma of no special type. CK19 upregulation was linked to nodal metastases in neuroendocrine tumors and papillary thyroid carcinomas (p < 0.05 each) and to poor grade in clear cell renal cell carcinoma (p < 0.05). CK19 upregulation was particularly common in squamous cell carcinomas. We concluded that CK19 IHC might separate primary liver cell carcinoma from liver metastases, seminoma from other testicular tumors, and helps in the detection of early neoplastic transformation in squamous epithelium.

Medullary carcinoma, provocative now as then

The recent observation that studies of BRCA1-associated tumors contain a high proportion of medullary carcinomas and ductal carcinomas with medullary features has re-introduced pathologists to an old diagnostic problem. The term "medullary carcinoma" dates to the 19th century, but the modern entity was introduced in 1949 by Moore and Foote, who described a carcinoma with a lymphoid infiltrate, a favorable prognosis, and low frequency of metastasis. Almost three decades later, Ridolfi et al proposed specific criteria for diagnosis, resulting in an entity with an even more favorable prognosis and a lower incidence. The reproducibility and clinical relevance of the diagnosis have been questioned recently, and new criteria have been proposed and compared. The tumors typically express cytokeratin 7, often vimentin and S100-protein, but not cytokeratin 20. The usual ones are positive for p53 and negative for estrogen receptor, Her2/neu, and bcl-2. Medullary carcinomas express e-cadherin and beta-catenin more often than ordinary high-grade ductal carcinomas, and the former have genetic differences from the latter. The lymphoid infiltrate of medullary carcinomas is related to beta-actin fragments exposed by apoptotic cells. The present review discusses historical and recent developments and emphasizes diagnostic criteria.

The cytokeratin profile of medullary carcinoma of the breast

AIMS

The cytokeratin (CK) phenotype and vimentin expression of 31 medullary carcinomas was studied using commercially available antibodies on archived material. Comparing the phenotype of typical and atypical tumours and the phenotype of metastases, the biological significance of cytokeratin and vimentin expression in medullary carcinomas of the breast was determined.

METHODS AND RESULTS

Antibodies to CK4, CK5 and 6, CK7, CK14, CK8 and 18, CK19, CK20 and to vimentin were used. All the typical and atypical medullary carcinomas and the metastases (10 cases) stained negatively for CK4 and positively for CK8-18 (CAM5.2). Almost all the tumours were CK7 and CK19 positive and CK20 negative. Twelve per cent of the tumours contained CK14. Twenty-five per cent of the typical, 43% of the atypical and 20% of the metastatic medullary carcinomas showed CK5-6 positivity. No association between the cytokeratin-vimentin profile of the tumours and axillary node metastases, tumour size or oestrogen receptor status was found but instability of CK expression was demonstrated by comparing the primary tumours with their metastases.

CONCLUSIONS

: Medullary carcinomas of the breast express all the glandular type CKs including CK19 and additionally a proportion of the tumours expresses some of the CKs typical for myoepithelial cells. There was no correlation with prognostic factors.

Effects of epidermal growth factor on the invasion activity of the bladder cancer cell line

PURPOSE

Epidermal growth factor (EGF) is excreted in high concentrations in the urine and stimulates urothelial cell growth. The cultured bladder cancer cell line KU-1 was used to study the molecular mechanisms by which EGF affects urothelial tumor growth and invasion activity.

MATERIALS AND METHODS

KU-1 cells were grown in cell culture in the presence or absence of EGF. Anchorage-independent cell growth assays and Matrigel invasion assays were performed. Expression of cytokeratins was examined by Northern and Western blot analyses. Chloramphenicol acetyltransferase assays were used to determine whether EGF stimulated matrix metalloproteinase expression.

RESULTS

EGF enhanced anchorage-independent growth in soft agar and increased the number of cells penetrating into a Matrigel membrane. A transient transfection assay revealed that EGF increased the promoter activities of the matrix metalloproteinase 1 and 9 genes in KU-1 cells. Moreover, the morphology of KU-1 cells changed after the addition of EGF to the culture medium. Western and Northern blot analyses demonstrated that EGF decreased cytokeratin 19 expression, but did not affect expression of cytokeratin 8 or 18.

CONCLUSION

EGF increased the invasive activity of KU-1 bladder cancer cells in part by increasing the secretion of matrix metalloproteinases. Morphologic changes may result from altered composition of cytoskeletal proteins.

Medullary breast carcinoma vs. poorly differentiated ductal carcinoma: an immunohistochemical study with keratin 19 and oestrogen receptor staining

Sixty breast carcinomas previously indexed as medullary carcinomas over a 24-year-period were reviewed and reclassified according to definitions suggested by Ridolfi et al. as typical medullary carcinoma, atypical medullary carcinoma, and non-medullary carcinoma. Paraffin sections of tumour tissue were examined by an avidin-biotin complex method using two keratin 19-specific monoclonal antibodies (BA17, DAKO and clone 170-2-14, Boehringer) and a monoclonal oestrogen receptor antibody (DAKO). For comparison 52 ductal carcinomas of grade II and grade III were immunostained as well. The results showed that all 60 tumours with medullary features and all 52 ductal carcinomas reacted moderately to strongly positive with anti-keratin 19 (Boehringer). The staining was diffuse in all cases, except one case of ductal carcinoma (grade III), which stained focally. Immunostaining with the second keratin 19 antibody (BA17) revealed similar results with positive staining in 59 (95%) cases of carcinomas with medullary features and 51 (98%) cases of ductal carcinomas. Only one case in each group did not express keratin 19 (BA17), one re-classified case of non-medullary carcinoma with neuroendocrine features and one case of ductal carcinoma of grade III. None of the 13 cases of typical medullary carcinoma were oestrogen receptor positive and only seven (12%) of the carcinomas with medullary features (2 atypical, 5 non-medullary) were oestrogen receptor positive with quantitative values from 20 to 100%. The 52 ductal carcinomas of grade II and III were oestrogen receptor positive in 56% and 47% of cases. It is concluded that keratin 19 staining is of no particular value in differentiating medullary from poorly differentiated ductal carcinoma. A carcinoma with positive oestrogen receptor staining is not likely to be a typical medullary carcinoma.