Expression of steroidogenic enzymes and metabolism of steroids in COS-7 cells known as non-steroidogenic cells

The COS-7 (CV-1 in Origin with SV40 genes) cells are known as non-steroidogenic cells because they are derived from kidney cells and the kidney is defined as a non-steroidogenic organ. Therefore, COS-7 cells are used for transfection experiments to analyze the actions of functional molecules including steroids. However, a preliminary study suggested that COS-7 cells metabolize [3H]testosterone to [3H]androstenedione. These results suggest that COS-7 cells are able to metabolize steroids. Therefore, the present study investigated the expression of steroidogenic enzymes and the metabolism of steroids in COS-7 cells. RT-PCR analyses demonstrated the expressions of several kinds of steroidogenic enzymes, such as cytochrome P450 side-chain cleavage enzyme, 3β-hydroxysteroid dehydrogenase/Δ5-Δ4 isomerase, cytochrome P450 7α-hydroxylase, cytochrome P450 17α-hydroxylase/17,20-lyase, 17β-hydroxysteroid dehydrogenase, 5α-reductase, cytochrome P450 21-hydroxylase, cytochrome P450 11β-hydroxylase, and cytochrome P450 aromatase in COS-7 cells. In addition, steroidogenic enzymes 3β-HSD, P4507α, 5α-reductase, P450c17, P450c21, P450c11β, and 17β-HSD actively metabolized various steroids in cultured COS-7 cells. Finally, we demonstrated that 17β-HSD activity toward androstenedione formation was greater than other steroidogenic enzyme activities. Our results provide new evidence that COS-7 cells express a series of steroidogenic enzyme mRNAs and actively metabolize a variety of steroids.

Statistical analysis. Data were statistically analyzed with a one-way ANOVA (when a normal distribution was found) and a post hoc Tukey-Kramer test. A significant difference was set at P < 0.05. All results were expressed as the mean ± SEM.

Cholesterol does not convert to pregnenolone in COS-7 cells.
To investigate the metabolism of cholesterol in COS-7 cells, RT-PCR analyses were used to detect the expression of steroidogenic and related enzymes, such as P450scc and StAR. RT-PCR analyses demonstrated the expression of P450scc mRNA, but not StAR mRNA in COS-7 cells (passages 3 to 15; Fig. 1A and Supplementary Fig. S1). Sequencing the amplified cDNA band verified that it was an authentic fragment of P450scc (GenBank accession no. XM_008015897).
To investigate cholesterol metabolism, COS-7 cells (passages 3 to 15) were incubated with [ 3 H]cholesterol and the radioactive metabolites were analyzed by reversed-phase HPLC. As shown in Fig. 1B and C, no radioactive metabolites were detected.
To investigate steroid formation from pregnenolone, COS-7 cells (passages 3 to 15) were incubated with [ 3 H] pregnenolone and the radioactive metabolites were analyzed by reversed-phase HPLC. Radioactive metabolites corresponding to 7α-hydroxypregnenolone and progesterone were detected (Fig. 2B,C). The concentration of these metabolites was reduced by treatment with ketoconazole, an inhibitor of P450s (Fig. 2B), and by trilostane, an inhibitor of 3β-HSDs (Fig. 2B).
To investigate steroid formation from progesterone, COS-7 cells (passages 3 to 15) were incubated with [ 3 H] progesterone and the radioactive metabolites were analyzed by reversed-phase HPLC. Radioactive metabolites corresponding to 5α-dihydroprogesterone, cortisol, and androstenedione were detected (Fig. 3C,D). The concentration of these metabolites was reduced by treatment with ketoconazole, an inhibitor of P450s (Fig. 3C), and by dutasteride, an inhibitor of 5α-reductases (Fig. 3C).
Androstenedione is metabolized to 5α-dihydrotestosterone in COS-7 cells. To investigate the metabolism of androstenedione in COS-7 cells (passages 3 to 15), RT-PCR analyses were performed to detect the expression of 17β-HSD type I and 17β-HSD type III. RT-PCR analyses demonstrated the expression of 17β-HSD type I, but not 17β-HSD type III (Fig. 4A and Supplementary Fig. S1). Sequencing the amplified cDNA band verified that it was an authentic fragment of 17β-HSD type I.
To investigate steroid formation from androstenedione, COS-7 cells (passages 3 to 15) were incubated with [ 3 H]androstenedione and the radioactive metabolites were analyzed by reversed-phase HPLC. A radioactive metabolite corresponding to 5α-dihydrotestosterone was detected (Fig. 4B,C), but testosterone, a precursor of 5α-dihydrotestosterone, was not detected (Fig. 4B,C).

Identified steroidogenic pathways and comparison of steroidogenic enzyme activities in COS-7 cells.
Thus, the active steroidogenic enzymes in cultured COS-7 cells were identified as 3β-HSD, P4507α, 5α-reductase, P450c17, P450c21, P450c11β, and 17β-HSD (Fig. 6). To compare the activities of steroidogenic enzymes in COS-7 cells (passages 3 to 15), we analyzed their metabolites by reversed-phase HPLC. COS-7 cells were cultured in serum-free DMEM containing 70 nmol [ 3 H]pregnenolone, 70 nmol [ 3 H]17α-hydroxyprogesterone, or 70 nmol [ 3 H]testosterone for 6 h. After incubation, the extracted steroids were subjected to HPLC analyses to measure metabolites. The activity of 17β-HSD that metabolites testosterone to androstenedione was significantly higher than those of other steroidogenic enzymes in COS-7 cells (Fig. 7). To investigate the effect of molecules in FBS on steroid formation, COS-7 cells (passages 3 to 15) were cultured in 10%FBS supplemented DMEM or FBS-free DMEM to 90% confluence. At 90% confluence, COS-7 cells were cultured in serum-free DMEM containing 70 nmol [ 3 H]pregnenolone, 70 nmol [ 3 H]17α-hydroxyprogesterone, or 70 nmol [ 3 H]testosterone for 6 h. After incubation, the extracted steroids were subjected to HPLC analyses to measure metabolites. The absence of FBS did not alter the activity of steroidogenic enzymes significantly in COS-7 cells (Supplementary Fig. S3).
In addition, the number of passages of the cells is an important factor for their function. Therefore, the activities of steroidogenic enzymes in cultured COS-7 cells were measured. In COS-7 cells between passages 30 and 40, the enzymatic activities of 3β-HSD and 5α-reductase were increased compared to those of COS-7 cells between passages 3 and 15 ( Supplementary Fig. S4).
The kidney is generally considered to be a non-steroidogenic organ 9,10 ; however, steroidogenesis in kidney tissue has been reported by some groups. Expression of P450c11β protein has been shown by western blotting and immunohistochemistry in normal human kidney 19 . In addition, northern blot analysis has detected 17β-HSD type XI mRNA expression in the human kidney 21 . In male and female rat kidneys, P450scc 24 and 3β-HSDs 22 mRNAs and 3β-HSD protein 22   metabolized to 5α-dihydroprogesterone, 11-deoxycorticosterone, 17α-hydroxyprogesterone, androstenedione, and testosterone in the kidney tissue, from rats of both sexes 23,24 . These previous studies suggest that the kidney has the ability for local steroid production. These reports also support our findings of steroid metabolism in COS-7 cells.
In this study, we found that 5α-dihydrotestosterone, a metabolite of testosterone, was produced from androstenedione, a precursor of testosterone, in COS-7 cells. However, we could not detect a testosterone peak by reversed-phase HPLC in this system (Fig. 4B). Furthermore, testosterone was depleted from the medium after 6 h of incubation (Fig. 5B). Our results suggest that the activity of 17β-HSD type II and IV is higher than that of 17β-HSD type I in COS-7 cells (Fig. 6). In addition, a previous study has shown that the activity of 5α-reductase, an enzyme that converts testosterone into dihydrotestosterone, is high in the kidney 23,24 . Furthermore, testosterone is produced in the rat kidney 23,24 . These results suggest that testosterone is produced from androstenedione in COS-7 cells, but the activities of 17β-HSD type II and IV and 5α-reductase were high enough to deplete testosterone from COS-7 cell culture media. However, we could not detect estradiol-17β in cultured COS-7 cells after the incubation with testosterone although P450arom are expressed in COS-7 cells (Fig. 5). It is considered that COS-7 cells may not convert androgen to estrogen. Further studies are needed to confirm this conclusion. COS-7 cells are used for transfection experiments to analyze the function of steroidogenic genes [11][12][13] , the function of steroid receptors [14][15][16] , and the effects of steroids on functional molecules 17,18 . However, our findings  here suggest that testosterone is actively converted to androstenedione or 5α-dihydrotestosterone in COS-7 cells. Thus, it should be noted that COS-7 cells are unfit to use for analyzing the effects of testosterone by testosterone addition in vitro. Past researchers probably inferred from their experience that testosterone is inactivated in COS-7 cells. In fact, they used synthetic, non-metabolizable androgen instead of testosterone in COS-7 cells 32,33 .
In the present study, steroidogenic enzymes 3β-HSD, P4507α, 5α-reductase, P450c17, P450c21, P450c11β, and 17β-HSD were active in cultured COS-7 cells. However, as shown in Fig. 7, the steroidogenic enzyme activities of 3β-HSD, P4507α, 5α-reductase, P450c21, and P450c11β were low compared to the activity of 17β-HSD in cultured COS-7 cells. Thus, the activities of these enzymes might not influence the analysis of the function of steroidogenic genes. In fact, in a previous study 30 , we failed to detect enzymatic activity of P4507α in COS-7 cells. This discrepancy may be due to differences in the cell cycle stage of COS-7 cells. A future study is needed to confirm this hypothesis.