Prolactin-inducible Proteins in Human Breast Cancer Cells*

The mechanism of action of prolactin in target cells and the role of prolactin in human breast cancer are poorly understood phenomena. The present study ex- amines the effect of human prolactin (hPRL) on the synthesis of unique proteins by a human breast cancer cell line, T-47D, in serum-free medium containing bovine serum albumin. [35S]Methionine-labeled proteins were analysed by sodium dodecyl sulfate-polyacryl-amide slab gel electrophoresis and fluorography. Treatment of cells with hPRL (1-1000 ng/ml) and hydrocortisone (1 pg/ml) for 36 h or longer resulted in the synthesis and secretion of three proteins having molecular weights of 11,000,14,000, and 16,000. Nei- ther hPRL nor hydrocortisone alone induced these proteins. Of several other peptide hormones tested, only human growth hormone, a hormone structurally and functionally similar to hPRL, could replace hPRL in causing protein induction. These three proteins were, therefore, referred to as prolactin-inducible proteins (PIP). Each of the three PIPs was purified to homogeneity by preparative sodium dodecyl sulfate-polyac- rylamide gel electrophoresis, and specific antibodies were generated to them in rabbits. By immunoprecipitation and immunoblotting (Western blot) of proteins secreted by T-47D cells, it was demonstrated that the three PIPs were immunologically identical to one an- other. In addition, the 16-kDa and 14-kDa proteins (PIP- 16 and PIP- 14), and were against distilled water at 4 “C, and lyophilized. The lyophilized pro- teins were redissolved in a solution of phosphate-buffered saline (PBS) containing 8 M urea. The dissolved proteins were fractionated at room temperature on a Sephadex G-100 column equilibrated with phosphate-buffered saline, 8 M urea, which was also used for elution. This gel filtration step was essential to separate the bulk of the proteins (e.g. BSA which was introduced with the hormones) from the prolactin-induced proteins. The prolactin-induced proteins in each fraction were monitored by SDS-PAGE and fluorography, as described earlier. The prolactin-induced proteins eluted after the main BSA peak. The fractions that contained the prolactin-induced proteins were pooled, dialyzed against water, lyophilized, redissolved in SDS mixture (1-2 ml), and, finally, heated. The entire sample was run on preparative slab SDS-PAGE (3 mm thick). The gel was stained with 0.1% Coomassie Blue in 7% acetic acid and, finally destained with 10% methanol in 7% acetic acid. The prolactin-inducedproteins, 16-, 14-, and ll-kDa bands, were prominently stained. Each band was excised with a sharp blade, after which each polyacrylamide gel segment was cut into small fragments (1-2 mm). The fragments from each band were soaked with agitation at room temperature for 10 min with 2 ml of SDS mixture (minus bromphenol blue), and this solution discarded. A second 2-ml buffer was added, and the contents were heated in a boiling water for 10 min. After boiling, the entire content was packed into a cylindrical glass tube with the bottom end plugged with glass wool and wrapped with a closed dialysis tubing (M, cut-off 3,500). The protein was electrophoresed (5 h at 10 mA/ tube) out of the gel into the reservoir made up of the dialysis tubing.

The mechanism of action of prolactin in target cells and the role of prolactin in human breast cancer are poorly understood phenomena. The present study examines the effect of human prolactin (hPRL) on the synthesis of unique proteins by a human breast cancer cell line, T-47D, in serum-free medium containing bovine serum albumin.
[35S]Methionine-labeled proteins were analysed by sodium dodecyl sulfate-polyacrylamide slab gel electrophoresis and fluorography. Treatment of cells with hPRL (1-1000 ng/ml) and hydrocortisone (1 pg/ml) for 36 h or longer resulted in the synthesis and secretion of three proteins having molecular weights of 11,000,14,000, and 16,000. Neither hPRL nor hydrocortisone alone induced these proteins. Of several other peptide hormones tested, only human growth hormone, a hormone structurally and functionally similar to hPRL, could replace hPRL in causing protein induction. These three proteins were, therefore, referred to as prolactin-inducible proteins (PIP). Each of the three PIPs was purified to homogeneity by preparative sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and specific antibodies were generated to them in rabbits. By immunoprecipitation and immunoblotting (Western blot) of proteins secreted by T-47D cells, it was demonstrated that the three PIPs were immunologically identical to one another. In addition, the 16-kDa and 14-kDa proteins (PIP-16 and PIP-14), and not the 11-kDa protein (PIP-1 l), incorporated [3H]glycosamine. Furthermore, 2deoxyglucose (2 mM) and tunicamycin (0.5 pg/ml), two compounds known to inhibit glycosylation, blocked the production of PIP-16 and PIP-14, with a concomitant increase in the accumulation of PIP-11. These results indicate PIP-16 and PIP-14 are glycosylated variants of PIP-1 1. Finally, in vitro translation of poIy(A)+ messenger RNA followed by immunoprecipitation revealed a 12.5-kDa protein, possibly the precursor form of PIPs. In addition, T-47D cells treated with hPRL plus hydrocortisone contained 10-fold more mRNA for PIPs than control cells, suggesting that the hormones' action is at the level of gene expression. Our finding represents a first demonstration of prolactin regulation of gene expression in human target cells. The human breast cancer cells, T-47D, appear to be an excellent model to afford future studies on the molecular action of prolactin and on the possible role of prolactin in human breast cancer. The role that the pituitary hormone, prolactin, plays in promoting the growth of mammary tumors in experimental animals is well known (1). However, its significance in human mammary tumorigenesis is unclear (2). In addition, large gaps still exist in our knowledge concerning the molecular mechanism of action of prolactin in target cells (3). In order to facilitate these studies, a prolactin-responsive human cell model is needed. Our past efforts have established that many human breast cancer cell lines maintained in tissue culture possess specific cell membrane receptors for prolactin (4). Furthermore, prolactin synergizes with glucocorticoid in causing cell rounding, loss of adhesion, and increased lipid synthesis in one human breast cancer cell line, T-47D (5). In the present study, we show for the first time that prolactin, in the presence of glucocorticoid, increased the accumulation of mRNA for, and the synthesis of, unique secretory proteins by the T-47D cells. This finding represents a first demonstration of an induction of specific proteins via regulation of gene expression by prolactin in human target cells.

MATERIALS AND METHODS
Cell Line-The human breast cancer cell line, T47D, was derived from the pleural effusion of a patient with disseminated carcinoma of the breast (6). This cell line contains the highest concentration of prolactin receptors of 11 breast cancer cell lines tested (4, 7). The nontumor human breast cell line, HBL-100, was generated from human breast milk (8). Cell lines were routinely maintained in Dulbecco's modified Eagle's medium supplemented with insulin (10 pg/ml), glutamine (4 mM), glucose (4.5 g/liter), streptomycin (50 pg/ ml), penicillin (50 units/ml), and fetal bovine serum (lo%, v/v). This medium is referred to as complete medium (CM). All the above reagents were purchased from Gibco. Cells were kept in a humidified atmosphere of 95% air, 5% CO, a t 37 "C.
Hormones and Antisera-Purified human prolactin (hPRL') was kindly provided by Drs. Henry G. Friesen and Ian G. Worsley, University of Manitoba. Purified human growth hormone (hGH), ovine growth hormone, ovine prolactin, and human luteinizing hormone were gifts of the National Pituitary Agency. National Institutes of Health. Hydrocortisone was purchased from Sigma. Antiserum to human milk was obtained from Cappel Laboratories. Antisera to human a-lactalbumin and casein were gifts of Dr. C. Kleinberg, Columbia University, and Dr. J. Kulski, University of Western Australia. All the antisera were raised in rabbits. In addition, antisera were generated to three prolactin-inducible proteins which will be described in a later section.
Analysis of Proteins Synthesized by Cells-Human breast cancer cells, T-47D (1 X 105/35-mm culture dish), were seeded in CM. Two days later, the medium was aspirated, the cells were washed with insulin-free and fetal bovine serum-free medium (referred to as DM), and each dish finally received 2 ml of DM. Hormones were added as 100 X concentrates (20 p1/2 ml) in DM that contained 1% (w/v) bovine serum albumin (BSA). For those dishes that received only one hormone or no hormone, 20 or 40 pl of DM/BSA was added. All dishes contained the same amount were incubated for another 6 days at which time 100 pCi of L-[%] methionine (>1200 Ci/mmol, Amersham Corp.), or 15 pCi of D-[1,6-3H]glucosamine hydrochloride (30-60 Ci/mmol, New England Nuclear) was added to each dish. Eighteen hours later, the medium from each dish was collected and centrifuged at 500 X g for 5 min to remove any cells present. The medium was then extensively dialyzed at 4 "C against glass-distilled water and lyophilized. The dry proteins were dissolved in an SDS mixture (10 mM sodium phosphate buffer, pH 7.2, containing 2% sodium dodecyl sulfate, 5% 0-mercaptoethanol, 10% glycerol, and trace amount of bromophenol blue). Each sample was heated in a boiling water bath for 5-10 min. Proteins in each sample were separated by SDS-polyacrylamide slab gel electrophoresis (9). All reagents for electrophoresis were obtained from Bio-Rad. The radiolabeled proteins on the gel were detected by fluorography (10). The molecular mass marker 14C-proteins were: myosin (200,000), phosphorylase (94 kDa), bovine serum albumin (68 kDa), ovalbumin (43 kDa), carbonic anhydrase (30 kDa), and cytochrome c (12 kDa).
Purification of Prolactin-inducible Proteins and Production of Specific Antisera-Twenty T-150 culture flasks were each seeded with lo6 T-47D cells in CM. Two days later, the cells were washed twice with DM, and each flask was replaced with DM. Human prolactin (or human growth hormone) and hydrocortisone were added, each to a final concentration of 1 pg/ml. After 5 days of hormone treatment, three of the flasks received ['%]Met (50 pCi/ml). The next day, the media from all the flasks (600 ml) were collected, pooled, dialyzed against distilled water at 4 "C, and lyophilized. The lyophilized proteins were redissolved in a solution of phosphate-buffered saline (PBS) containing 8 M urea. The dissolved proteins were fractionated at room temperature on a Sephadex G-100 column equilibrated with phosphate-buffered saline, 8 M urea, which was also used for elution. This gel filtration step was essential to separate the bulk of the proteins (e.g. BSA which was introduced with the hormones) from the prolactin-induced proteins. The prolactin-induced proteins in each fraction were monitored by SDS-PAGE and fluorography, as described earlier. The prolactin-induced proteins eluted after the main BSA peak. The fractions that contained the prolactin-induced proteins were pooled, dialyzed against water, lyophilized, redissolved in SDS mixture (1-2 ml), and, finally, heated. The entire sample was run on preparative slab SDS-PAGE (3 mm thick). The gel was stained with 0.1% Coomassie Blue in 7% acetic acid and, finally destained with 10% methanol in 7% acetic acid. The prolactin-inducedproteins, 16-, 14-, and ll-kDa bands, were prominently stained. Each band was excised with a sharp blade, after which each polyacrylamide gel segment was cut into small fragments (1-2 mm). The fragments from each band were soaked with agitation at room temperature for 10 min with 2 ml of SDS mixture (minus bromphenol blue), and this solution discarded. A second 2-ml buffer was added, and the contents were heated in a boiling water for 10 min. After boiling, the entire content was packed into a cylindrical glass tube with the bottom end plugged with glass wool and wrapped with a closed dialysis tubing (M, cut-off 3,500). The protein was electrophoresed (5 h at 10 mA/ tube) out of the gel into the reservoir made up of the dialysis tubing. The content of each dialysis bag was dialyzed against water at room temperature overnight. This procedure resulted in the complete purification of all the three prolactin-induced proteins. The homogeneity of each protein was established by analytical SDS-PAGE (see "Results"). Using the above procedure, 40-60 pg of each prolactininduced protein could be obtained from 1 liter of conditioned medium.
Ten micrograms of each protein was emulsified with complete Freund's adjuvant, and injected intradermally at multiple sites in a rabbit. Two and four weeks later, the same amount of protein was given via the same route in incomplete Freund's adjuvant. The rabbits were bled 2 weeks after the final injection. Antisera were obtained after removal of the clots and cells. Pre-immune serum was also obtained by bleeding each of the rabbits before immunization with the protein.
Immunoprecipitation-Each sample containing [35S]Met-labeled proteins was allowed to incubate at 4 "C overnight with 5 pl of antiserum or preimmune serum in 0.05 M Tris-HC1, pH 7.8, containing 0.1 M NaCI, 0.1% SDS, 0.5% sodium deoxycholate, 100 units/ml trasylol, and 0.5% Nonidet P-40 (referred to as immunoprecipitation buffer) in a final volume of 100 pl. The next day, 50 pl of PansorbinTM (Staphylococcus aureus suspension, Calbiochem-Behring), washed previously with 4 M urea in immunoprecipitation buffer and resuspended in immunoprecipitation buffer, was added. One hour later, the bacteria-antibody-antigen complexes were washed once with 4 M urea followed by two washes with immunoprecipitation buffer. SDS mixture (20-50 pl) was added to each of the pellets, heated in boiling water for 10 min, and centrifuged. The protein(s) recovered in the supernatant was analyzed by SDS-PAGE and visualized by fluorography. For each sample, a preimmune serum was always used to account for proteins that were nonspecifically precipitated. However, such results were not always shown for clarity of figures.
Electrophoretic Transfer of Proteins from Polyacrylamide Gels to Nitrocellulose Papers (Western Blot) and Immunolocalization of Proteins-Protein transfer was carried out using the method of Towbin et al. (11). A Trans-Blot cell and an Immune-Blot (GAR-HRP) assay system (Bio-Rad) were used according to instructions supplied by the manufacturer. Briefly, after SDS-PAGE, protein bands were transferred to nitrocellulose paper electrophoretically. The nitrocellulose paper was sequentially treated with 3% BSA (to reduce nonspecific adsorption of immunoglobulins), antiserum (or preimmune serum) used at dilution of 1:100, peroxidase-conjugated goat anti-rabbit IgG (1:3000), and finally a solution of 4-chloro-l-naphthol/hydrogen peroxide.
In Vitro Translation of Poly(A)+ mRNAs Isolated from T-47D Cells-Three T-150 flasks of T-47D cells treated with hGH plus hydrocortisone and another three flasks not treated with hormones (control) were grown for 5 days as described in a previous section. Cells were lysed with Nonidet P-40 and cytoplasmic RNA was extracted with phenol by using a protocol similar to that described by Maniatis et al. (12). Poly(A)+ messenger RNA was isolated from the total RNA preparations by oligo(dT) (Collaborative Research) affinity chromatography as described (12). Poly(A)+ mRNA was translated with a [35S]methionine in vitro reticulocyte translation kit (New England Nuclear). Equal aliquots from each translation reaction were used to react either with antiserum or with preimmune serum, as described in a previous section. The immunoprecipitated product(s) was analyzed by SDS-PAGE and fluorography, and quantitated by densitometric tracing of the fluorograph. Fig. 1 illustrates the effects of various hormone combinations on [35S]methionine-labeled proteins secreted by T-47D cells.

Effects of Hormones on Protein Synthesis in T-47D Ce1l.s-
In the absence of hormones (Fig. 1, lune A ) , T-47D cells synthesized and secreted at least 18 proteins with molecular weights ranging from approximately 400,000 to 10,000. The cancer cells that were exposed to hydrocortisone alone (Fig.  1, lane B ) were essentially similar to those produced by untreated cells, with the exception that three high molecular mass proteins (300-440 kDa) were reduced. Upon the treatment of cells with hPRL plus hydrocortisone, (Fig. 1, lane C), three proteins, 160, 75, and 36 kDa, disappeared (or were substantially reduced). In addition, three new proteins were induced: 16,14, and llkDa with the 14-kDa protein being the most prominent. hGH plus hydrocortisone (Fig. 1, lane D ) produced an identical protein pattern as hPRL plus hydrocortisone. In contrast, in the absence of hydrocortisone, treatment with hPRL or hGH alone (Fig. 1, lanes E and F ) resulted in the disappearance of only the 160-kDa protein, but new protein induction did not occur. This result indicates that the induction of the 16-, 14-, and ll-kDa proteins required the synergistic actions of hPRL (or hGH) and hydrocortisone. Furthermore, the observation that hGH was equipotent with hPRL is consistent with our previous observations that hGH and hPRL bind to the same receptors, and that these two "lactogenic" hormones are equipotent in their ability to induce other biological responses in the T-47D cells (4, 5).
The hormone specificity and the dose-response of induction of 16-, 14-, and ll-kDa proteins were also examined. Fig. 2 shows that, in the presence of hydrocortisone, 1 ng/ml hGH (lane D ) caused a significant induction of the proteins; maximal induction was achieved with 0.1-1 pg/ml hGH (lanes A and B ) . When the radioactivity in the protein bands was quantitatively determined (see legend to Fig. 2), it was found that 1 ng/ml and lpg/ml hGH caused a 2-and 10-fold induction, respectively, of each of the three proteins. In the presence of hydrocortisone, hPRL produced a similar dose response (not shown). Thus, the induction of proteins in T-47D cells occurs at physiological concentrations of the hPRL and hGH, suggesting that this phenomenon is a physiologically relevant kDa A B C D E F G

11-
event. In contrast, in the presence of hydrocortisone, human luteinizing hormone (lane F), ovine growth hormone (lane E), and ovine prolactin (not shown), at 1 pg/ml, failed to induce the three proteins. Since the three proteins were induced specifically by hPRL and by hGH acting through the prolactin receptors (4, 5), we propose to call them "prolactin-inducible proteins" or "PIPS"-coded as PIP-11 (11 kDa), PIP-14 (14 kDa), and PIP-16 (16 kDa). Hormone treatment of 36 h was found to be required to observe a significant increase in protein induction. The maximum rate of synthesis of PIPs occurred, however, 72 h after the addition of hormones, and this rate could be maintained for 1 week, the longest time interval tested (data not shown).
Purification of PIPs and the Generation of Antibodies-In an effort to establish the relationship of the three PIPs with each other and possibly with other proteins produced by T-47D cells (see Fig. l), as well as to understand the mode of induction of the PIPs by prolactin, we decided to generate specific antibodies to each of the three PIPs. This was accomplished first by purifying each of them to homogeneity in two steps (gel filtration chromatography and preparative SDS-PAGE; see "Materials and Methods"). Each of the three purified PIPs was shown to be homogeneous (Fig. 3). Each of the PIPs was used to immunize a rabbit, and the resulting antisera were used for the following studies.

in Breast Cancer
Characterization of the Three anti-PIP Antisera-Identical aliquots of [35S]methionine-labeled proteins secreted by T-47D cells treated with hPRL (or hGH) plus hydrocortisone were reacted with each of the antisera, and the immunoprecipitated products were analyzed by SDS-PAGE and fluorography (see "Materials and Methods"). Fig. 4A shows that all three PIPs were precipitated by each of the antisera. Anti-PIP-14 antiserum appeared to have the highest titer of antibodies. To exclude the possibility that the PIPs were tightly associated with one another, resulting in their coprecipitation by any one antiserum, the proteins secreted by T-47D were first separated by SDS-PAGE and then electrophoretically transferred to nitrocellulose paper (Western blot). The transferred proteins were reacted with anti-PIP-14 antiserum and visualized using an Immune-Blot assay system (Bio-Rad). The result of the Western blot was shown in Fig. 4B: all three PIPs were detected by the anti-PIP-14 antiserum. Thus, the data of Fig. 4, A and B, indicate that all three PIPs were immunologically related, possibly variant species of the same protein.
Based on two-dimensional gel electrophoretic data (not shown), charge heterogeneity of PIP-14 and PIP-16, but not PIP-11, was evident. Taken together with the immunological similarity of the three PIPs, it was suspected that PIP-14 and PIP-16 were glycosylated forms of PIP-11. To test this possibility, T-47D cells were labeled with [3H]glucosamine, and the proteins in the media were analyzed. Fig. 5 shows that PIP-14 and PIP-16, but not PIP-11, were labeled. In addition, [35S]methionine-labeled proteins made by hormone-treated T-47D cells in the presence of 2-deoxyglucose and tunicamycin, two compounds known to effectively inhibit glycosylation (13, 14), were analyzed after immunoprecipitation with anti-PIP-14 antiserum. Fig. 6 shows that both 2-deoxyglucose and tunicamycin were able to block the synthesis of PIP-14 and PIP-16. Concomitantly, there was an increased accumulation of PIP-11. It is concluded that PIP-14 and PIP-16 are glycosylated forms of PIP-11.
The relationship, if any, of PIPs to other proteins made by T-47D cells in the absence and presence of various hormone combinations was also studied. Fig. 7 (lanes 1-4) shows that anti-PIP-14 antiserum was only able to precipitate the three PIPs in the culture medium of T-47D cells treated with hGH plus hydrocortisone (lane 4 ) ; very little of them was secreted by cells not treated with hormone (lane I ) , or by cells exposed to only hydrocortisone (lane 2) or to hGH alone (lane 3). No other high molecular weight proteins were recognized by the antiserum. Similar results were obtained when intracellular proteins were analyzed (lanes [5][6][7][8]. 1) that the PIPs were not related to, and, therefore, not processed products of, some other large proteins, and (2) that hGH plus hydrocortisone stimulated the synthesis, and not secretion, of PIPs because PIPs were not found in the intracellular compartment of control cells or cells treated with hydrocortisone or hGH alone. Quantitation of Messenger RNA for PIPS-In order to gain insights into the mechanism of induction of PIPs in T-47D cells by hormones, we decided to quantitate the amount of translatable messenger RNA in control and hormone (hPRL plus hydrocortisone) -treated cells. Oligo(dT) -selected poly(A)+ RNAs were translated in vitro using a rabbit reticulocyte lysate system, and the products were immunoprecipitated, analyzed, and quantitated. Fig. 8 (top panel) shows that the only in vitro translated product that reacted with anti-PIP-14 antibodies was a peptide of 12.5 kDa. This is presumably the precursor form of PIP. Substantial amounts of translatable PIP mRNA were present in hormone-treated cells (lanes D-F), and very little was found in control cells (lane A-C). The intensity of the 12.5 kDa bands was quantitaed by densitometric tracing as shown in Fig. 8, lower panel: a 10-fold increase of PIP-mRNA was present in hormonetreated cells as compared to control cells. It is concluded from this result that hormone treatment induced the accumulation of PIP-mRNA in T-47D human breast cancer cells. Other Observations-We have also examined whether or not PIPs are milk components. Incubation of [35S]methionine-labeled proteins secreted by T-47D cells with antiserum to human whole milk, a-lactalbumin, and casein failed to immunoprecipitate any one of the PIPs (not shown). In addition, Western blot analysis of human whole milk using anti-PIP-14 antiserum failed to reveal the presence of PIPs in the milk sample (not shown). Finally, immunoprecipitation and Western blot analysis of proteins synthesized and secreted by a nontumor human breast epithelial cell line HBL-100 failed to detect the presence of PIPs (data not shown).

DISCUSSION
In the present study, we were able to demonstrate for the first time that hPRL, in the presence of glucocorticoid, was able to induce the synthesis of three unique proteins in the prolactin receptor-positive human breast cancer cell line, T-47D. These PIPs have molecular mass of 16,14, and 11 kDa. Only hGH, a hormone evolutionarily, structurally, and functionally related to hPRL, was able to cause the induction of PIPs also. This observation is consistent with our previous findings that 1) hGH binds to the prolactin receptors in the in Breast Cancer

T-47D cells, and 2) hGH is equipotent with hPRL in a number of measurable biological responses in T-47D cells (4,5).
To facilitate characterization of the PIPs and to gain better insight into the molecular mechanism of induction of PIPs by prolactin, we decided to purify each of the three PIPs and to generate polyclonal antibodies to each of them. Using these specific antibodies, we were able to establish that the three PIPs were immunologically related. We also demonstrated that only the two larger PIPs, 16 and 14 kDa, incorporated [3H]glucosamine, and that 2-deoxyglucose and tunicamycin, two inhibitors of protein glycosylation (13, 14), inhibited the appearance of PIP-14 and PIP-16 with concomitant accumulation of PIP-11. These results indicate that PIP-16 and PIP-14 are glycosylated forms of the PIP-11. Under the influence of hPRL (or hGH) and cortisol, the major species produced by T-47D cells was PIP-14.
In addition, we established that 1) the amount of PIPs secreted by the cells was a reflection of PIP content within the cells, and 2) PIPs were unrelated to any other larger protein species produced by T-47D cells whether treated or not treated with hormones. These observations indicate that neither alteration in protein processing nor alteration in the rate of protein secretion can account for the action of these hormones. Our observation that the combination of hGH and cortisol resulted in a 10-fold increase in mRNA for PIP suggests that the hormones' action(s) is at the levels of gene expression and/or RNA processing and stability.
The in vitro translation of poly(A)+ mRNA resulted in a single peptide having a molecular mass of 12.5 kDa. Since proteins translated in vitro retain their signal peptide and are not modified (e.g. glycosylation), it is reasonable to assume that the 12.5-kDa protein is the precursor form of the PIPs.
Our observations clearly showed that the induction of the final PIP products required the synergistic actions of prolactin and cortisol. This same hormone combination was previously shown to be required for the induction of casein mRNA in the rodent mammary gland (15-17). However, the exact role played by prolactin and glucocorticoid and their respective degree of contribution to the regulation of casein gene expression in the rodent mammary gland remain to be determined (15, 17). Nevertheless, it is probable that the two hormones regulate casein gene expression both at the transcriptional and post-transcriptional levels. The exact role of prolactin and glucocorticoid in the regulation of PIP gene expression in the human breast cancer cells, T-47D, awaits studies using a cloned PIP cDNA probe. The observation that prolactin and cortisol stimulate casein gene expression in the rodent mammary gland prompted us to examine if the PIPs made by T-47D human breast cancer cells are milk components and if hormone-treated nontumor human mammary epithelial cells (HBL-100) produced PIPs. Our data indicated that PIPs are not present in normal human milk and that PIPs are not made by nontumor breast cells. Thus, the identity and function(s) of PIPs remain to be elucidated. Nevertheless, our findings represent a first demonstration of prolactin regulation of specific gene product in human target cells. Furthermore, the human breast tumor cell line T-47D appears to be an excellent model to study the molecular mechanisms of action of prolactin. The gene coding for the prolactin-inducible proteins will provide a unique marker for these studies. The present study has also paved the path for the elucidation of the biological significance of the prolactin-inducible proteins, and the role of prolactin in human breast neoplasm.