Human Prostate Cell Lines: Derivation, Molecular Characteristics and Appropriate Use

Research on prostate cancer and benign prostatic hyperplasia depends on a small number of immortalized human cell lines established between 1977 and 2001, most of them from advanced metastatic disease. The cell lines differ from one another in androgen receptor status, prostate-specific antigen secretion, p53 status, karyotype and growth requirements, and those differences determine which questions each cell line can answer.

Derivation reports, repository entries and subsequent characterization studies were retrieved for each human prostate cell line in current use. Sources were identified by database searching and from the reference lists of earlier reviews, and the primary derivation report was used in preference to any secondary description. Where two sources report a characteristic differently, both are given together with the assay used in each.

The lines are described grouped by tissue compartment and androgen receptor status rather than by date of derivation, and for each line the derivation, molecular and cytogenetic characteristics, culture requirements, behavior in vivo and documented limitations are set out. Reports that conflict with one another, including the androgen receptor status of PNT1A, WPMY-1, BPH-1 and DU145 and the Bax status of DU145, are largely explained by differences in assay sensitivity. Developments since the last comprehensive reviews are covered, including patient-derived organoid culture, the recombinant origin of the retrovirus reported in 22Rv1, the reassignment of NCI-H660 to a prostatic origin, and the adoption of short tandem repeat profiling as a condition of publication. Lines derived from men of African ancestry, the xenograft-maintained lines that are not adapted to plastic, and the in vivo systems used where an intact host is required are included.

Several descriptions repeated in the secondary literature are not supported by the primary characterization of the lines concerned. Selection of a line from the research question rather than from availability, verification of the relevant target in the working stock, authentication by short tandem repeat profiling, and full reporting of passage number and culture conditions would remove a substantial part of the variability presently seen in this literature.