|Home | About | Journals | Submit | Contact Us | Français|
Resistance to chemotherapy is a major obstacle for successful treatment of breast cancer patients. Given that prolactin (PRL) acts as an anti-apoptotic/survival factor in the breast, we postulated that it antagonizes cytotoxicity by chemotherapeutic drugs. Treatment of breast cancer cells with PRL caused variable resistance to taxol, vinblastine, doxorubicin and cisplatin. PRL prevented cisplatin-induced G2/M cell cycle arrest and apoptosis. In the presence of PRL, significantly less cisplatin was bound to DNA, as determined by mass spectroscopy, and little DNA damage was seen by γ-H2AX staining. PRL dramatically increased the activity of glutathione-S-transferase (GST), which sequesters cisplatin in the cytoplasm; this increase was abrogated by Jak and mitogen-activated protein kinase inhibitors. PRL upregulated the expression of the GSTμ, but not the π, isozyme. A GST inhibitor abrogated antagonism of cisplatin cytotoxicity by PRL. In conclusion, PRL confers resistance against cisplatin by activating a detoxification enzyme, thereby reducing drug entry into the nucleus. These data provide a rational explanation for the ineffectiveness of cisplatin in breast cancer, which is characterized by high expression of both PRL and its receptor. Suppression of PRL production or blockade of its actions should benefit patients undergoing chemotherapy by allowing for lower drug doses and expanded drug options.
Worldwide, 1.3 million women are diagnosed annually with breast cancer, accounting for 10% of all new cancers and >400000 will die from the disease (1). Chemotherapy is the mainstay treatment for advanced and metastatic disease. DNA-damaging agents have a long and proven record as anticancer drugs. Cisplatin, a platinum-based drug, is one of the most potent antitumor agents that is highly effective against lung, ovarian and prostate cancer but not breast cancer (2). Cisplatin interacts with DNA, forms adducts and intrastrand cross-links and induces cell cycle arrest. Damaged DNA can either be repaired via several repair mechanisms or the cell is destined to die (3).
While the choice of chemotherapeutic agents has increased in recent years, tumor resistance remains a major obstacle. Some tumors are intrinsically resistant to drugs, whereas others acquire resistance following treatment. Resistance results from many causes, including drug efflux by transporters, inactivation by detoxifying enzymes, altered expression of pro-/anti-apoptotic proteins or tumor suppressors and increased DNA repair mechanisms (4,5). Among detoxification enzymes, glutathione-S-transferase (GST) conjugates electrophilic drugs to glutathione, rendering them targets for extrusion by an adenosine triphosphate-dependent pump (6). GST inactivates platinum drugs, thereby increasing resistance to cisplatin in breast cancer (7) and also reduces the efficacy of doxorubicin, cyclophosphamide and etoposide but not anti-microtubule drugs, by inactivating c-jun N-terminal kinase 1 (6).
Although breast tumors are among the few cancers that are hormone sensitive, the role of protein hormones in chemoresistance has received little attention. Prolactin (PRL) is a 23 kDa protein hormone that binds to a single-span membrane receptor, a member of the cytokine receptor superfamily, and exerts its action via several interacting pathways, including Jak2-Stat5a/b, mitogen-activated protein kinase (MAPK) and phosphoinositide-3 kinase (PI3K)/Akt (8). In humans, PRL is produced by the pituitary as well as by many non-pituitary sites, where it acts as a paracrine/autocrine factor (9). Both PRL and the prolactin receptor (PRL-R) are expressed in breast tissue and in many breast cancer cell lines (8,10). Autocrine PRL is mitogenic, as shown by the suppression of T47D cell proliferation following inhibition of PRL (11,12). In nude mice, growth of tumors derived from T47D cells is suppressed by a PRL-R antagonist (13), whereas PRL-overexpressing MDA-MB-435 cells generate faster growing tumors (14).
PRL functions as a survival factor in both malignant and non-malignant cells (15–17). This raises the prospect that PRL antagonizes apoptosis by anticancer drugs, thereby contributing to chemotherapeutic resistance. Indeed, a PRL-R antagonist enhanced cisplatin-induced apoptosis in T47D cells (18) and increased the suppressive effects of doxorubicin and taxol on colony formation in MCF-7 cells (19). Furthermore, breast cancer cells that produce PRL were more resistant to ceramide-induced apoptosis than those with low or no PRL (20). However, previous studies have not addressed the mechanism by which PRL confers chemoresistance, which was the major objective of the present studies. To this end, we used MDA-MB-468 (468) cells, which have very low endogenous PRL and found that exogenous PRL confers resistance against several classes of anticancer drugs. Upon focusing on cisplatin, we discovered that PRL antagonizes drug cytotoxicity by increasing GST activity resulting in reduced entry of cisplatin into the nucleus. These data provide a rational explanation for the ineffectiveness of cisplatin in breast cancer, which is characterized by high expression of both PRL and its receptor. These observations also suggest that suppression of PRL production or blockade of its action should benefit breast cancer patients undergoing chemotherapy.
The 468 and T47D cells were obtained from the American Type Culture Collection (Manassas, VA). The 468 cells were maintained in Dulbecco's modified Eagle's medium (Hyclone, Logan, UT) supplemented with 10% fetal bovine serum (Hyclone). Cells were plated in Dulbecco's modified Eagle's medium with 3% charcoal-stripped serum (CSS; Hyclone) and treated with Dulbecco's modified Eagle's medium with 1% CSS. T47D cells were maintained in RPMI (Hyclone) supplemented with 10% fetal bovine serum, 5 μg/ml insulin and 10 mM N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid. Cells were plated in RPMI with 5% CSS and insulin transferrin selenium+ supplement (BD biosciences, Bedford, MA) and treated in RPMI with 1% CSS and insulin transferrin selenium+.
Cells were plated at 6000 or 8000 cells per well in 96-well plates in plating medium. The following day, cells were incubated with PRL for 24 h in treatment medium. When using probenecid (Sigma–Aldrich, St Louis, MO), an inhibitor of multidrug resistance-associated proteins (MRPs), and ethacrynic acid (Sigma–Aldrich), an inhibitor of GST, they were added 1 h before PRL. Following PRL pretreatment, cisplatin was added for an additional 1–4 days, as specified in the various experiments. PRL remained on the cells during drug treatment. Cytotoxicity was determined by the 3-(4,5-dimethyl-thiazol-2-yl)2,5-diphenyl tetrazolium bromide method (21).
Total cell proteins were electrophoresed onto 12% sodium dodecyl sulfate–polyacrylamide gel electrophoresis gels. For phospho-Stat5, lysates were immunoprecipitated with Stat5 antibody and captured with protein A agarose beads. After transfer to polyvinylidene difluoride membranes, samples were incubated with the following antibodies: phospho-extracellular signal-regulated kinase (ERK)1/2, ERK1/2, phospho-Akt, Akt (Cell Signaling, Danvers, MA; 1:1000), phosphotyrosine (Santa Cruz Biotechnology, Santa Cruz, CA; 0.75 μg/ml) or Stat5 (Cell Signaling; 1:1000). After incubation with horseradish peroxidase-conjugated secondary antibody (Amersham, Piscataway, NJ), products were developed using SuperSignal chemiluminescence reagents (Pierce, Rockford, IL).
Cells were centrifuged and the pellets were resuspended in phosphate-buffered saline containing 0.1% Triton X-100, 0.02 mg/ml propidium iodide (PI) (Sigma–Aldrich) and 0.2 mg/ml RNase A and incubated for 15 min at 37°C. Flow cytometry data were acquired on a Coulter Epics XL (Beckman Coulter, Miami, FL) and 10000 gated events were collected per experiment. Cell cycle analysis was performed using ModFit LT (Verity Software House, Topsham, ME).
Cells were resuspended in buffer containing 5 μl allophycocyanin Annexin V (BD Pharmingen, San Jose, CA) and 1 μg/ml PI. Samples were analyzed by flow cytometry at 488 and 633 nm using a BD LSRII instrument (Becton Dickinson, San Jose, CA). Log fluorescence was collected for PI and allophycocyanin using a 610/20 or 660/20 band pass filter, respectively; 10000 gated events were collected per experiment.
Cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100. Coverslips were incubated for 1 h with anti-phospho-histone H2AX (Ser139) antibody or anti-phospho-histone H3 (Ser10) antibody (Cell Signaling). After mounting in Gelmount with 4′,6-diamidino-2-phenylindole (Fisher Scientific, Pittsburgh, PA), images were obtained with Zeiss Axioplan Imaging 2 microscope (Zeiss, Oberkochen, Germany). Quantitation of H3-stained cells was done using several fields under the microscope.
DNA was isolated using the DNeasy kit (Qiagen, Valencia, CA). Platinum content was determined with an Aligent 7500ce mass spectrometer (Aligent Technologies, Santa Clara, CA). Samples or platinum standards (SpexCertiPrep, Metuchen, NJ) were introduced at a flow rate of 1 ml/min using a 100 μl injection volume. Instrument parameters were forward power of 1450 W plasma gas flow rate of 15.0 l/min, auxiliary gas flow rate of 1.0 l/min, carrier gas flow rate of 0.97 l/min, makeup gas flow rate of 0.14 l/min, octopole bias of −18 V, quadrupole bias of −16 V and 194Pt and 195Pt-monitored isotopes.
Cells were lysed in 100 mM potassium phosphate and 2 mM ethylenediaminetetraacetic acid and protein concentrations were determined by a Pierce protein assay. Lysates (50 μg) were incubated with 1 mM of 1-chloro-2,4,dinitrobenzene (Sigma–Aldrich) substrate and 2.5 mM of L-glutathione (Sigma–Aldrich). The rate of conjugation of 1-chloro-2,4,dinitrobenzene to glutathione by GST was measured spectrophotometrically at 340 nm every 30 s for 5 min. Total GST activity was calculated from a standard curve.
Total RNA was isolated using Tri-Reagent (Molecular Research Center, Cincinnati, OH) and complementary DNA was synthesized as described (22). Polymerase chain reaction was performed on 200 ng complementary DNA using intron-spanning primers for GSTμ (forward 5′-TCGTGTGGACATTTTGGAGA-3′ and reverse 5′-GGGCTCAAATATACGGTGGA-3′) and GSTπ (forward 5′-GGAGACCTCACCCTGTACCA-3′ and reverse 5′-GGGCAGTGCCTTCACATAGT-3′); β2-microglobulin was used as a reference gene (forward 5′-GGCATTCCTGAAGCTGAC-3′ and reverse 5′-GAATCTTTGGAGTACGCTGG-3′).
Statistical differences were determined by one-way analysis of variance followed by Newman–Keuls post hoc analysis. P values <0.05 were considered significant. All experiments were performed at least three times.
We first examined whether PRL antagonizes cytotoxicity by the microtubule-altering drugs taxol and vinblastine and the DNA-damaging drugs cisplatin and doxorubicin. As shown in Figure 1A, pretreatment of 468 cells with a low PRL dose (25 ng/ml) completely protected the cells from 5 ng/ml taxol, a dose that reduced cell viability by 75%, but was only partially effective against vinblastine. PRL antagonized all tested doses of doxorubicin. Cisplatin-induced cytotoxicity was then compared in 468 and T47D cells. Unlike the strong suppression effect of cisplatin on 468 cells (Figure 1B), it was only moderately effective in T47D cells (Figure 1C). Notably, PRL completely antagonized all doses of cisplatin in T47D cells, as opposed to its partial efficacy against higher cisplatin doses in 468 cells. We next found that all tested concentrations of PRL, well within its physiological range, antagonized cisplatin in 468 cells (Figure 1D). To explore the mechanism by which PRL confers chemoresistance, we selected the prototypical DNA-damaging agent cisplatin and took advantage of the low endogenous PRL in 468 cells. In experiments using a shorter time course than 4 days, usually 200 or 800 ng/ml cisplatin was used so as to obtain meaningful cytotoxic effects.
PRL actions as a survival factor (19,20) can be mediated through several signaling pathways (8). Treatment of 468 cells with PRL caused activation of MAPK, Jak/Stat5 as well as PI3K pathways, albeit with different dynamics (Figure 2A). PRL induced a robust activation of ERK1/2 that began at 5 min and progressively increased thereafter. Stat5 was also immediately phosphorylated by PRL but was slightly decreased by 240 min. In contrast, Akt was only marginally activated by PRL. To determine which of these pathways mediate protection by PRL against cisplatin, we used selective inhibitors. Figure 2B shows that the cisplatin-induced decrease in cell viability was abrogated by pretreatment with PRL. However, PRL did not protect the cells when either the Jak or the MAPK pathways were blocked by AG490 and U0126, respectively. In contrast, inhibition of the PI3K pathway with wortmannin did not prevent PRL from antagonizing cisplatin. These results indicate that protection by PRL in 468 cells involves Jak-Stat and MAPK signaling rather than the PI3K–Akt survival pathway.
In response to DNA damage, cells can arrest at either G1, intra-S phase or G2/M cell cycle checkpoints to allow for repair or induce apoptosis. Therefore, we examined whether PRL overrides cisplatin-induced cell cycle arrest. Flow cytometry shows that 17–18% of control or PRL-treated cells were in the G2/M phase (Figure 3A). Treatment with cisplatin dramatically increased this number to 54%, whereas pretreatment by PRL decreased arrested cells to 29%, indicating a partial override of the cisplatin-induced G2/M cell cycle arrest by PRL. Next, we questioned whether the cells were arrested at the G2/M boundary or in mitosis. Figure 3B depicts cells stained with phosphorylated histone H3, which labels cells in mitosis. When normalized to the total number of cells (Figure 3C), cisplatin treatment reduces the number of cells in mitosis by 75%, suggesting an arrest at the G2 boundary. The percentage of mitotic cells returns to control levels following treatment with PRL.
Annexin V/PI staining in combination with flow cytometry was utilized to determine whether cisplatin-treated cells undergo apoptosis and whether PRL opposes cisplatin-induced cell death. This approach enables discrimination between live cells (no labeling), cells in early apoptosis (Annexin V positive), late apoptosis/necrosis (Annexin V + PI positive) and dead cells (PI positive). Figure 3D and E show that within 24 h, cisplatin decreased the percentage of live cells from 89.6% in controls to 69.2%; this was accompanied by a concomitant increase in cells in early and late apoptosis/necrosis. The number of live cells in response to treatment with both PRL and cisplatin was the same as in controls, confirming that PRL antagonizes cisplatin-induced apoptosis.
We next questioned whether PRL antagonizes cisplatin-induced DNA damage. To this end, cells were labeled with an antibody against phosphorylated histone H2AX (γ-H2AX), which detects double-strand breaks caused by drugs such as cisplatin (23). Figure 4A shows robust γ-H2AX staining following treatment with cisplatin. This staining was not observed when cells were pre-exposed to PRL, suggesting that PRL prevented cisplatin-induced DNA double-strand breaks. To determine whether PRL prevents cisplatin from binding to DNA, we used inductively coupled plasma mass spectroscopy to quantify cisplatin–DNA binding. Figure 4B shows that within 24 h, PRL reduced the amount of platinum bound to DNA by 50%, as compared with cisplatin alone.
To determine whether PRL actions were mediated by activating MRP transporters or GST, inhibitors of MRP (probenecid) or GST (ethacrynic acid) were utilized. Whereas PRL effectively antagonized cisplatin cytotoxicity in the presence of probenecid, it did not protect the cells when GST was blocked (Figure 5A), implicating GST, rather than MRP transporters, in its protective actions. Indeed, PRL treatment caused a 4-fold increase in GST activity, an effect that was abrogated by either Jak or MAPK inhibitors (Figure 5B). The involvement of GST as a mediator of PRL action was further supported by demonstrating that ethacrynic acid abolished the ability of PRL to decrease cisplatin binding to DNA (Figure 5C). Using real-time polymerase chain reaction, we found that PRL increased the expression of GSTμ by 3-fold without affecting GSTπ (Figure 5D).
We are reporting that PRL antagonizes multiple anticancer drugs that act by different mechanisms. Whereas others reported that PRL-R antagonists enhanced cytotoxicity by taxol, doxorubicin and cisplatin (18,19), they have not offered a mechanism for PRL actions. We chose to focus on cisplatin because it is highly effective against many types of cancer but has shown little success in the treatment of breast cancer (2). Our data reveal that PRL decreases the amount of cisplatin entering the nucleus and binding to DNA by increasing the activity of GST, which inactivates cisplatin. Since doxorubicin is also a substrate for GST (24), it is probable that PRL antagonizes its cytotoxicity by activating GST. However, the manner by which PRL opposes the microtubule-altering drugs, which are not affected by the GST system, remains to be determined. We speculate that resistance against taxol and vinblastine may be determined by the ability of PRL to increase anti-apoptotic proteins. This is supported by the marked increase in expression of Bcl-2 in PRL-overexpressing breast cancer xenografts (14) and by the ability of PRL to reduce methotrexate toxicity in leukemia cells by upregulating Bcl-2 (25).
To establish chemoresistance by PRL, it was critical to ascertain that its effects are not restricted to a single cell line. Although we primarily used 468 cells, PRL also antagonized cisplatin in T47D cells (Figure 1), as well as in MDA-MB-231 cells (data not shown). Whereas cisplatin reduced the viability of 468 cells down to 17% of controls, T47D cells were more resistant to cisplatin, raising the possibility that this is due to their elevated level of endogenous PRL (12,26). Additionally, a moderate dose of PRL (25 ng/ml) completely antagonized all doses of cisplatin in T47D cells but PRL was only partially effective against the higher cisplatin doses in 468 cells. We postulate that this is due to the very high density of the PRL-R in T47D cells (27), whereas the less abundant PRL-R in 468 cells may become saturated more easily. An important issue is whether partial protection by PRL is clinically relevant. The use of very high drug doses in vitro is bound to kill most of the responsive cancer cells, an effect that may not be prevented by many chemoresistance agents. However, patients cannot be treated with overwhelming drug doses because a balance must be made between an effective eradication of the tumor cells while preserving the integrity of normal tissues and minimizing side effects. The intermediate doses of cisplatin, doxorubicin and taxol that reduced cell viability by 40–50% and were antagonized by PRL (Figure 1), are well within the range of blood levels of these drugs in treated patients (28–30).
Upon binding to its receptor, PRL can rapidly activate several pathways, including Jak-Stat, MAPK and PI3K (8). Our data show a more robust activation of Stat5 and ERK1/2 by PRL in 468 cells, as compared with Akt. The use of selective inhibitors indicated that signaling through the Jak and MAPK pathways, but not the PI3K survival pathway, are required for PRL to oppose cisplatin cytotoxicity in 468 cells. This is in contrast to protection by PRL against radiation-induced damage in T47D cells, which is mediated by the PI3K/Akt pathway (31). The roles of Jak-Stat5 and MAPK in the regulation of apoptosis have been amply demonstrated in several cell types (5,32,33).
Previous studies reported that PRL overcame gamma irradiation-induced cell cycle arrest in T47D cells (31). Since chemotherapeutic agents can suppress cell growth by a variety of mechanisms, including apoptosis, necrosis and mitotic catastrophe (32), it was important to verify that those cells arrested by cisplatin at the G2/M boundary eventually died. The use of flow cytometry reveals that cisplatin indeed causes cell death via apoptosis, an effect that was opposed by PRL. These data also verified that the cisplatin-induced decreases in cell viability, as determined by the 3-(4,5-dimethyl-thiazol-2-yl)2,5-diphenyl tetrazolium bromide method, correlate well with increased apoptotic cell death. Since DNA is the main target of cisplatin, we reasoned that PRL could inhibit cisplatin-induced apoptosis either by preventing DNA damage or by altering pro-/anti-apoptotic proteins following DNA damage. The first possibility was confirmed using γ-H2AX staining, whereas inductively coupled plasma mass spectroscopy revealed that reduced DNA damage in the presence of PRL was due to the lower amount of cisplatin bound to DNA. Inductively coupled plasma mass spectroscopy is commonly used to identify cellular platinum, and the quantities that we have detected are in good agreement with those reported by others (34,35). Collectively, these results demonstrate that PRL reduces cisplatin binding to DNA rather than antagonizing its actions through alterations in pro-/anti-apoptotic proteins that are downstream of DNA damage.
The next objective was to identify the mechanism by which PRL prevents cisplatin from binding to DNA. Potential candidates were (i) membrane transporters that extrude cisplatin from the cells and (ii) detoxification mechanisms that inactivate cisplatin. Both MRPs (36,37) and thiol-containing compounds, such as glutathione (7,24,38), play important roles in resistance to cisplatin. The use of ethacrynic acid and probenecid revealed that GST activity, but not transporter activity, is necessary for PRL to confer resistance. GSTs are best known as phase II detoxification enzymes that catalyze conjugation of glutathione to a wide variety of endogenous and exogenous compounds (6). In addition, certain GST isozymes serve as endogenous inhibitors of c-jun N-terminal kinase 1 via protein–protein interactions. This explains why high levels of GST can inactivate drugs that induce apoptosis via the MAPK pathway, even though the drugs themselves are neither subject to conjugation with glutathione nor substrates for GST.
Our data clearly showed that PRL increased GST activity in 468 cells and that ethacrynic acid blocked the ability of PRL to reduce cisplatin binding to DNA. Since ethacrynic acid alone did not increase nuclear cisplatin levels, the transport of cisplatin into the nucleus does not appear to be affected by unstimulated GST activity. Previously, Luquita et al. (39) reported that hepatic GST activity in rats was increased following PRL administration. GSTs are divided into several distinct classes, based on sequence similarities and substrate specificity (40). Much attention has been paid to the roles of GSTμ, GSTπ and GSTθ polymorphism in cancer incidence and response to therapy (40,41). We found that PRL increases expression of GSTμ without affecting GSTπ. Notably, GSTM1- and GSTT1-null genotypes are associated with increased survival in women with advanced breast cancer that were treated with several chemotherapeutic agents (42).
Figure 6 depicts a model that conceptualizes the mechanism by which PRL antagonizes cisplatin cytotoxicity. Cisplatin uptake by the cell occurs via passive diffusion and, in some cases, the drug can be immediately extruded by transporters. Cisplatin that remains in the cell enters the nucleus where it binds to DNA, forms adducts and intrastrand cross-links and induces apoptosis. Binding of PRL to its receptor results in receptor dimerization and activation of the Jak2-Stat5a/b and the MAPK-signaling pathways, which involve adaptor proteins such as Shc. Signaling through these pathways, either in parallel or via their conversion, leads to increased transcription of GST and higher enzyme activity. GST reduces intracellular cisplatin levels by promoting its conjugation to glutathione, culminating in the expulsion of the conjugate from the cells. Since less cisplatin is available for transport to the nucleus, the overall effect of PRL is suppression of cisplatin-induced apoptosis. Future studies should determine whether knockdown of specific GST isozymes abrogates the protective effects of PRL, examine the mechanisms by which PRL promotes resistance against drugs such as doxorubicin and taxol and also explore to what extent PRL affects radiation-induced tumor suppression.
Finally, drug resistance is a challenging issue for breast cancer patients, being the main reason for treatment failure in advanced and metastatic disease. Our data suggest that a reduction in circulating PRL levels or blockade of the PRL-R should improve the efficacy of chemotherapy in breast cancer patients. Indeed, in a small clinical study, Lissoni et al. (43) reported increased tumor responsiveness to taxotere in breast cancer patients that were also treated with bromocriptine, a dopamine agonist that suppresses PRL secretion. Long acting, orally administered dopamine agonists, such as bromocriptine and cabergoline, have an excellent record of safety and efficacy in the treatment of pituitary prolactinomas and in restoring fertility in PRL-dependent reproductive dysfunctions (44). Since these drugs are Food and Drug Administration approved, clinical trials for their effectiveness in breast cancer can commence without a delay. The projected benefits of lowering the PRL input to breast tumors include a wider choice of effective chemotherapeutic agents, increased effectiveness of commonly used anticancer drugs and decreased toxicity and side effects normally associated with high-dose chemotherapy.
National Institutes of Health (CA096613, ES016803); Department of Defense (BC05725); Susan G.Komen for the Cure (BCRT87406 to N.B.J.); National Institutes of Health training (5T32ES007250 to E.W.L.).
Conflict of Interest Statement: None declared.