In a study published in The Journal of Pathology, when researchers activated the PI3K pathway and removed the tumor-suppressor gene p53 in a specific group of mouse mammary cells (those that express keratin 15), the mice developed breast tumors that looked similar to those seen in metaplastic breast cancer. To do this, they used a genetic model (Krt15-CrePR1) that switched on a cancer-causing version of the Pik3ca gene and deleted p53 in Krt15-positive cells. The mice developed breast lesions even without chemical induction. Importantly, cancer only formed if both copies of p53 were lost. In female mice, losing both copies of p53 caused tumors to become detectable about a month earlier than if only one copy was lost. In male mice, tumors developed only when both copies of p53 were missing
The tumors that developed included several types: squamous cell carcinoma, carcinosarcoma, sarcoma, and carcinoma. When stained for the progesterone receptor, only the epithelial parts of the tumors were positive; the sarcomas and most squamous tumors did not show progesterone receptor staining. Keratin staining revealed distinct patterns: keratin 5 was found at the edges where tumor cells met supporting tissue, keratin 8 was present in luminal (inner) areas, and keratin 15 was seen in carcinoma regions but absent in sarcomatous areas.
In human studies, researchers reviewed 73 cases of metaplastic breast cancer and found that mutations in both PIK3CA and TP53 often occurred together, especially in squamous metaplasia. They also created cell lines from the mouse tumors: one from squamous cell carcinoma (G1330R), two from carcinosarcoma (G1319, G1330L), and two from sarcoma (M935C, C1362). These cell lines kept the same appearance and marker expression as the original tumors. The squamous cell carcinoma and carcinosarcoma lines expressed E-cadherin, keratin 5, keratin 8, and keratin 14, while the sarcoma lines did not express E-cadherin. All of the lines expressed vimentin. Neither estrogen receptor nor progesterone receptor was detected, but some lines expressed the androgen receptor (AR). Treatment with the AR blocker enzalutamide had only limited effect.
The Pik3ca-mutant lines were more sensitive to alpelisib, a p110α-selective PI3K inhibitor already approved by the FDA for PIK3CA-mutant cancers. This suggests that targeting the PI3K pathway could be a potential therapeutic option in metaplastic breast cancers with PIK3CA mutations, even though hormone-based therapies (estrogen, progesterone, or androgen receptor–directed) may not be effective.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
