In preclinical models, established molecular determinants of cellular sensitivity to cyclophosphamide, long a mainstay of chemotherapeutic regimens used to treat breast cancers, include the aldehyde dehydrogenases that catalyze the detoxification of this agent, namely, ALDH1A1 and ALDH3A1. As judged by bulk quantification of relevant catalytic activities, as well as of relevant proteins (ELISAs), tissue levels of these enzymes vary widely in primary and metastatic breast malignancies. Thus, interindividual variation in the activity of either of these enzymes in breast cancers could contribute to the wide variation in clinical responses obtained when such regimens are used to treat these malignancies. Direct evidence for this notion was sought in the present investigation. Cellular levels of ALDH1A1 and ALDH3A1 in 171 repository human breast tumor (122 primary and 49 metastatic) samples were semiquantified using immunocytochemical staining. Clinical responses were retrieved from the archived medical records of each of 48 metastatic breast cancer sample donors, 26 of whom had been treated with a cyclophosphamide-based chemotherapeutic regimen subsequent to tumor sampling and 22 of whom had not. The premise that cellular levels of ALDH1A1 and/or ALDH3A1 predict clinical responses to cyclophosphamide-based chemotherapeutic regimens was submitted to statistical analysis. Confirming an earlier report, ALDH1A1 and ALDH3A1 levels varied widely in both primary and metastatic breast tumor cells. When measurably present, each of the enzymes appeared to be evenly distributed throughout a given tumor cell population. Retrospective analysis indicated that cellular levels of ALDH1A1, but not those of ALDH3A1, were (1) significantly higher in metastatic tumor cells that had survived exposure to cyclophosphamide than in those that had not been exposed to this drug, and (2) significantly higher in metastatic tumors that did not respond (tumor size did not decrease or even increased) to subsequent treatment with cyclophosphamide-based chemotherapeutic regimens than in those that did respond (tumor size decreased) to such regimens. The therapeutic outcome of cyclophosphamide-based chemotherapy corresponded to cellular ALDH1A1 levels in 77% of cases. The frequencies of false-positives (cyclophosphamide-based chemotherapy not effective when a low level of ALDH1A1 predicted it would be) and false-negatives (cyclophosphamide-based chemotherapy effective when a high level of ALDH1A1 predicted it would not be) were 0.00 and 0.43, respectively. Thus, partial or complete responses to cyclophosphamide-based chemotherapy occurred 2.3 times more often when the ALDH1A1 level was low than when it was high. Given (1) the wide range of ALDH1A1 levels observed in malignant breast tissues, (2) that ALDH1A1 levels in primary breast tumor tissue, as well as those in normal breast tissue, directly reflect ALDH1A1 levels in metastatic breast tumor cells derived therefrom, and (3) the findings reported here, measurement of ALDH1A1 levels in primary breast malignancies and/or normal breast tissue prior to the initiation of chemotherapy is likely to be of value in predicting the therapeutic potential, or lack of potential, of cyclophosphamide and other oxazaphosphorines, e.g. ifosfamide, in the treatment of primary, as well as metastatic, breast cancer, thus providing a rational basis for the design of individualized therapeutic regimens for this disease. Failure to observe the expected inverse relationship between clinical responses to cyclophosphamide-based chemotherapeutic regimens and ALDH3A1 levels was probably because even the highest breast tumor tissue ALDH3A1 level thus far reported appears to be below the threshold level at which ALDH3A1-catalyzed detoxification of oxazaphosphorines becomes pharmacologically meaningful. However, ALDH3A1 levels in certain other malignancies, e.g. those of the alimentary tract and lung, may be of a sufficient magnitude in that regard.
