Characteristics of separated epithelial and stromal subfractions of prostate: I. Rat ventral prostate

Abstract

These studies were initiated with the objective of isolating epithelial and stromal cells of human prostatic tissue in undamaged state, in order to study the cellular distribution of steroid receptors in benign prostatic hyperplasia (BPH) relative to normal prostate. Initial experiments showed that when BPH tissue immersed in tissue culture media was progressively fragmented by various cutting procedures, epithelial elements were selectively released as clumps of variable size and individual cells, but that a large percentage of these cells were damaged, as evidenced by their failure to exclude trypan blue (TB). These observations suggested that if tissue framentation were carried out under defined conditions that minimize cell damage, BPH subfractions might be obtained containing a large percentage of undamaged cells. To determine conditions of tissue fragmentation which result in maximal recovery of epithelial cells which exclude TB, rat ventral prostate (RVP) was chosen as a model system. Experiments with RVP revealed that maximal yields of such cells were obtained in “large” epithelial clumps (>30 cells per clump) released under the following conditions: (1) chopping the tissue with razor blades in a large volume (2 ml/100 mg RVP) of a Ca²⁺‐free tissue culture medium (Joklik's‐MEM) containing 1% casein, (2) carrying out the entire fractionation procedure in the cold, and (3) maintaining a 1% casein concentration in the medium during chopping, as well as in subsequent washing procedures, to protect cells from proteolytic activity. In large epithelial clumps, cells in the interior of the clump were not stained by TB but the cells at the periphery of the clump were freely permeable to TB. Single epithelial cells and small epithelial clumps (3–10 cells) released by razor blade fragmentation were also permeable to TB. When large epithelial clumps were incubated at 20°C for 90 min, the clumps disaggregated into smaller clumps and morphologically intact single cells, which did not exclude TB. The residual tissue fragments remaining after chopping contained the bulk of stromal cells plus some epithelial elements. The latter could be removed by gentle rubbing of the fragments on a sieve in the presence of medium. The stromal fraction thus obtained consisted of stromal cells, embedded in mesenchymal matrix, which were not stained by TB and appeared normal when examined histologically by light microscopy. The androgen receptor content per unit DNA (AR/DNA) of epithelial clumps and stromal fractions from castrate (1 day) and normal intact rats were compared to control RVP (coarsely minced and washed, but not further dissociated) using [³H]‐R‐1881 as ligand. In RVP from castrate animals (where AR is unoccupied and cytosolic), the mean AR/DNA ratio in large epithelial clumps was 17% of the control value. However, on a protein basis, AR per milligram cytosolic protein in epithelial clumps was 80% of the unfractioned control. After disaggregation of large epithelial clumps at 20°C, the recovered cells no longer contained detectable levels of AR. The mean AR/DNA ratio in stromal cell fractions from castrate RVP was about half that found in epithelial clump fractions. In RVP from normal intact animals (where AR is occupied with endogenous DHT and predominantly localized in nuclei), AR was measured by a [³H]‐R‐1881 exchange assay. The mean AR/DNA ratio in epithelial clumps from normal RVP was about 45% of the value found in control tissue; in stromal fractions the mean AR/DNA ratio was about 60% of the found in epithelial clumps. Estradiol receptor (ER) in normal intact rats, measured using [³H]‐R‐2858 as ligand was found to be unoccupied, cytosolic, and present in low concentration relative to unoccupied, cytosolic AR in castrate RVP. Highest ER/DNA ratios were found in the stromal cellular fraction with successively lower values in unfractionated RVP and epithelial clumps. Arginase activity per unit DNA in epithelial clumps was about 28% of that found in control tissue. The mean arginase/DNA ratio in stromal fractions was about 80% of the value found in the control and consistently higher than in epithelial clumps. Thus in RVP, arginase appears to be a marker for “viability” of stromal as well as epithelial cells, rather than a specific marker for epithelial cells. The present results demonstrate that AR and ER are both present in epithelial and stromal subfractions of RVP. Since AR/DNA ratios were higher in epighelial clumps than in stromal fractions in both normal and castrate RVP, (despite the greater degree of cell damage in epithelial elements relative to stroma, as suggested by arginase content) this finding indicates that AR concentration is much higher in epithelial than in stromal cells. IN contrast, Er appears to be predominantly localized in stromal cells.