BRCA1 accumulates in the nucleus in response to hypoxia and TRAIL and enhances TRAIL-induced apoptosis in breast cancer cells Latricia D. Fitzgerald 1, *, Charvann K. Bailey 1 , Stephen J. Brandt 2,3 and Marilyn E. Thompson 1,3 1 Division of Cancer Biology, Meharry Medical College, Nashville, TN, USA 2 Departments of Medicine, Cell and Developmental Biology, and Cancer Biology, Vanderbilt University Medical Center and VA Tennessee Valley Healthcare System, Nashville, TN, USA 3 Vanderbilt-Ingram Cancer Center, Nashville, TN, USA The breast cancer susceptibility gene BRCA1 encodes a tumor suppressor protein that is located predominantly within the nucleus. However, BRCA1 also contains two nuclear export sequences [1,2] and shuttles between the nucleus and cytoplasm. BRCA1 has been implicated in several cellular processes, including cell cycle regulation [3], transcription [4], DNA repair [5] and apoptosis [6]. Full understanding of BRCA1’s role in these processes is lacking; however, it is likely that some functions overlap. For example, the ability of BRCA1 to regulate the cell cycle and enhance apopto- sis could be due to its role as a transcriptional coacti- vator. BRCA1 induces the transcription of GADD45a [7], which is involved in the cell cycle G 2 –M check- point and in the control of apoptosis. BRCA1 can also enhance apoptosis via a pathway involving H-Ras, mitogen-activated protein kinase kinase kinase 4, Jun N-terminal kinase, Fas ligand ⁄ Fas and caspase-9 [8]. The ability of BRCA1 to enhance apoptosis mediated through Jun N-terminal kinase Keywords apoptosis; BRCA1; hypoxia; localization; TRAIL Correspondence M. E. Thompson, Division of Cancer Biology, WBSC 2137, Meharry Medical College, 1005 D.B. Todd Blvd, Nashville, TN 37208-3599, USA Fax: +1 615 327 6442 Tel: +1 615 327 6787 E-mail: methompson@mmc.edu *Present address University of California Davis Medical Center, Sacramento, CA, USA (Received 25 April 2007, revised 1 August 2007, accepted 7 August 2007) doi:10.1111/j.1742-4658.2007.06033.x A major contributing factor to the development of breast cancer is decreased functional expression of breast cancer susceptibility gene 1, BRCA1. Another key contributor to tumorigenesis is hypoxia. Here we show that hypoxia increased the nuclear localization of BRCA1 in MCF-7 and MDA-MB-468 human breast cancer cell lines without changing its steady-state expression level. Nuclear accumulation of BRCA1 was not evident in MCF-12A or HMEC (human mammary epithelial cell) nonma- lignant mammary epithelial cells under the same conditions. Hypoxia also increased the cell surface expression of TRAIL on MDA-MB-468 cells. Neutralization of TRAIL precluded the hypoxia-induced accumulation of BRCA1 in the nucleus, whereas exogenously administered TRAIL mim- icked the effect. Treatment of MDA-MB-468 cells with TRAIL resulted in a dose- and time-dependent increase in apoptosis. Furthermore, TRAIL- induced apoptosis in HCC1937 cells, which harbor a BRCA1 mutation, increased synergistically when wild-type BRCA1 was reconstituted in the cells, and downregulation of BRCA1 expression in MDA-MB-468 cells reduced the apoptotic response to TRAIL. These data provide a novel link between hypoxia, TRAIL and BRCA1, and suggest that this relationship may be especially relevant to the potential use of TRAIL as a chemothera- peutic agent. Abbreviations BRCA1, breast cancer susceptibility gene 1; FACS, fluorescence-activated cell sorting; GADD45, growth arrest and DNA damage-inducible gene; HIF, hypoxia-inducible factor; HMEC, human mammary epithelial cell; shRNA, short hairpin RNA; TRAIL, tumor necrosis factor-related apoptosis-inducing ligand. FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS 5137 [8,9], a stress-activated kinase, as well as data from studies with genotoxic agents [10], suggest a role for BRCA1 in the cellular response to stress. Stress induced by genotoxic drugs or UV radiation decreases the steady-state mRNA levels and enhances the phos- phorylation of BRCA1 [11]. Thus, these data provide a precedent for the involvement of BRCA1 in the cel- lular response to stress. One stress common to multiple solid tumors is hypoxia. During tumor growth, increased oxygen con- sumption in the presence of inadequate oxygen deliv- ery results in decreased intratumoral pO 2 [12]. In addition, malignant cells express elevated levels of hypoxia-inducible factor (HIF)-1a [13]. Under norm- oxic conditions, the expression of HIF-1a is low, due to its rapid degradation by the ubiquitin–proteasome pathway. Upon exposure to hypoxia, HIF-1a is stabi- lized, translocates to the nucleus and heterodimerizes with HIF-1b. The heterodimer binds to the hypoxia- responsive element of target genes that encode proteins involved in various cellular processes, including inhibi- tion of apoptosis. Hypoxia has been reported to inhibit apoptosis induced by tumor necrosis factor-related apoptosis- inducing ligand (TRAIL) [14,15], which selectively induces apoptosis in malignant, but not nonmalignant, cells [16]. TRAIL binds to death receptors, DR4 and DR5, and recruits Fas-associated death domain pro- tein to the receptors. Caspase-8 is activated and subse- quently activates downstream caspases. Hypoxia inhibits TRAIL-mediated apoptosis in HCT116 human colon carcinoma cells by inhibiting the translocation of Bax from the cytosol to the mitochondria, thereby pre- venting the release of cytochrome c and progression of apoptosis [14]. Studies have also demonstrated that hypoxia decreases the BRCA1 mRNA level in prostate cells [17]. Hypoxia-induced downregulation of BRCA1 expression is due to changes in the relative distribution of activating and repressive E2Fs on the BRCA1 pro- moter [18]. Furthermore, in response to hypoxia, BRCA1 is phosphorylated in a checkpoint kinase 1-dependent manner [19] and DNA repair is impaired [17]. Here, we report that under conditions that do not decrease BRCA1 expression, hypoxia increases the nuclear content of BRCA1 in malignant mammary epi- thelial cell lines while having no effect on nonmalig- nant cell lines. This localization change was dependent on TRAIL, and exogenous TRAIL mimicked the effect. Furthermore, TRAIL, but not hypoxia, induced apoptosis, and TRAIL-mediated apoptosis was enhanced by BRCA1. Results Hypoxia induces nuclear accumulation of BRCA1 in breast cancer cells To investigate the effects of hypoxia on the localiza- tion of BRCA1 in mammary epithelial cells, we sub- jected malignant (MDA-MB-468 and MCF-7) and nonmalignant (MCF-12A and HMEC) cells to hypoxia and examined the protein and relative nuclear and non-nuclear levels of BRCA1. As it had been previ- ously reported that a 48 h exposure to 0.01% oxygen resulted in decreased BRCA1 expression [18], we used a milieu of 1% oxygen for 24 h to assess changes in BRCA1 localization in response to hypoxia without concurrent changes in its expression. In MDA-MB-468 and MCF-7 cells exposed to normoxia, BRCA1 local- ized within the nucleus, cytoplasm and in a concen- trated area around the nucleus (Fig. 1A). Under low-oxygen conditions, the nuclear content of BRCA1 increased, whereas the non-nuclear (cytoplasmic and perinuclear) level decreased. In MCF-12A and HMEC cells, there was no difference in BRCA1 localization in cells exposed to hypoxia as compared to cells exposed to normoxia. We next performed subcellular fractionation of cells, followed by western blot analysis, to substantiate and quantify the changes between the nuclear and non- nuclear compartments in malignant cells. Exposure of MDA-MB-468 cells to hypoxia for 24 h resulted in an approximately 70% increase in nuclear levels of BRCA1 relative to those in cells exposed to normoxia (Fig. 1B). Non-nuclear levels decreased in hypoxic cells to approximately 65% of the level in normoxic cells. There was no change in the steady-state level of Fig. 1. Hypoxia induces nuclear accumulation of BRCA1 in breast cancer cells but not in nonmalignant cells. MDA-MB-468 cells, MCF-7 cells, MCF-12A cells or HMECs were exposed to normoxia or hypoxia (1% O 2 ) for 24 h prior to (A) immunostaining for BRCA1 (magnifica- tion, 400·) or (B) nuclear ⁄ non-nuclear fractionation. One hundred micrograms of whole cell lysate (wcl) or nuclear protein or 200 lg of non- nuclear protein was immunoblotted for BRCA1. Densitometric values were normalized to actin. *P < 0.05; mean ± SE; n ¼ 3. N, normoxia; H, hypoxia. The value for each normoxic sample was arbitrarily set at 1. (C) One hundred micrograms of protein from each cell line was im- munoblotted for HIF-1a. Membranes were stripped and probed for MSH-6, a DNA mismatch repair protein, as a loading control as previously described [24]. Novel association between hypoxia, TRAIL and BRCA1 L. D. Fitzgerald et al. 5138 FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS A B C L. D. Fitzgerald et al. Novel association between hypoxia, TRAIL and BRCA1 FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS 5139 BRCA1. Similar results were obtained with MCF-7 cells. Hypoxia enhanced the nuclear levels of BRCA1 by 40% and decreased the level of non-nuclear BRCA1 by 20% when compared to levels in cells exposed to normoxia. There was no change in the steady-state protein level in these cells as well. Interest- ingly, neither MCF-12A nor human mammary epithe- lial cell (HMEC) cell lines exhibited a change in the expression level or nuclear ⁄ non-nuclear distribution of BRCA1 when exposed to hypoxia. Western blot analy- sis of HIF-1a (Fig. 1C) demonstrated that each cell line responded to hypoxia with increased expression of HIF-1a. As similar changes were observed in both malignant cell lines and no change occurred in either nonmalignant cell line, subsequent experiments were performed with MDA-MB-468 and MCF-12A cells as the malignant and nonmalignant models, respectively. Hypoxia-induced BRCA1 localization changes are TRAIL-dependent The selective cytotoxicity of TRAIL to tumor cells in the absence of an effect in many nontransformed cells has been widely reported. Therefore, as hypoxia induced changes in BRCA1 localization in malignant, but not in nonmalignant, mammary epithelial cells, we sought to determine whether TRAIL was involved in the localization change. First, we assessed whether there was a change in TRAIL expression in MDA- MB-468 cells in response to hypoxia by analyzing its cell surface expression. Immunofluorescent analysis of TRAIL in nonpermeabilized cells demonstrated an  40% increase in TRAIL fluorescence in cells exposed to hypoxia relative to cells exposed to normoxia (Fig. 2A). These data indicate that hypoxia upregulates cell surface expression of TRAIL. Second, we reasoned that if the hypoxia-induced change in BRCA1 localization was TRAIL-dependent, blocking the TRAIL signaling pathway should abro- gate this change. To test this, we preincubated MDA- MB-468 cells with a neutralizing antibody against human TRAIL, 2E5, before exposure to hypoxia, and evaluated the nuclear⁄ non-nuclear distribution of BRCA1. Neutralization of TRAIL blocked the nuclear accumulation of BRCA1 induced by hypoxia in MDA-MB-468 cells (Fig. 2B), resulting in a 35% decrease in nuclear BRCA1 relative to the level in normoxic cells in the presence of 2E5. The decrease in AB C Fig. 2. Hypoxia-induced changes in BRCA1 localization are mediated through TRAIL. (A) MDA-MB-468 cells were exposed to normoxia or hypoxia for 24 h prior to fixation and immunostaining for the extracellular C-terminus of TRAIL. Cells were not permeabilized. Fluorescence intensities of cells in captured images were quantified using ALPHAIMAGER 2000 software. The values graphed are the average integrated den- sity values of the cells. *P < 0.05; mean ± SE; n ¼ 90 cells. Magnification, 400·. (B) One microgram of neutralizing antibody against TRAIL was added to MDA-MB-468 cells for 1 h before initiation of hypoxia for 24 h. Cells were fractionated and analyzed by western blot. Densito- metric values were normalized to actin. *P < 0.05; mean ± SE; n ¼ 3. Values for each normoxic sample in the presence of normal IgG or anti-TRAIL IgG were arbitrarily set at 1. (C) 50 ngÆmL )1 KillerTRAIL was administered to MDA-MB-468 cells for 1.5 h before harvesting, frac- tionation, and western blot analysis. Densitometric values were normalized to actin. * P < 0.05; mean ± SE; n ¼ 3. wcl, whole cell lysate; V, vehicle; T, TRAIL. The value for the normoxic sample was arbitrarily set at 1. Novel association between hypoxia, TRAIL and BRCA1 L. D. Fitzgerald et al. 5140 FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS non-nuclear expression of BRCA1 in response to hypoxia was also attenuated by 2E5. Finally, we hypothesized that if hypoxia was activating a TRAIL-dependent mechanism leading to enhanced nuclear BRCA1 levels, exogenously adminis- tered TRAIL under normoxic conditions should mimic the effects of hypoxia. To determine whether human soluble recombinant TRAIL could regulate BRCA1 subcellular localization, we treated MDA-MB-468 or MCF-12A cells with KillerTRAIL for 1.5 or 4 h, respectively. TRAIL resulted in an  1.7-fold increase in nuclear BRCA1 (Fig. 2C) with a significant decrease in non-nuclear BRCA1 in MDA-MB-468 cells. Expo- sure to TRAIL did not alter the steady-state level of BRCA1 protein. Likewise, in MCF-12A cells, TRAIL had no effect on the steady-state levels of BRCA1; however, there also was no effect on the subcellular distribution of the protein (data not shown). Collec- tively, the data in Fig. 2 support the involvement of a TRAIL-dependent process in mediating hypoxia- induced changes in BRCA1 localization. TRAIL, but not hypoxia, induces apoptosis in breast cancer cells As our data suggested that hypoxia was stimulating a TRAIL signaling pathway and TRAIL activates apop- totic processes, we assessed the ability of hypoxia to induce apoptosis in our model systems. Approximately 4% of MDA-MB-468 cells were present within the subdiploid population of cells as determined by flow cytometry of normoxic cells. This did not change fol- lowing exposure to hypoxia for 8 or 24 h (Table 1). As hypoxia failed to induce apoptosis in MDA-MB- 468 cells, we assessed the ability of TRAIL to do so. Exposure of MDA-MB-468 cells to increasing concen- trations of TRAIL from 5 to 100 ngÆmL )1 resulted in a dose-dependent increase in apoptosis, with the maxi- mally effective dose being 50–100 ngÆmL )1 (Fig. 3A). Treatment of cells with 50 ngÆmL )1 TRAIL resulted in a time-dependent increase in apoptosis that was detect- able within 2 h of treatment and stabilized at 6 h (Fig. 3B). To confirm that the subdiploid population of cells detected via flow cytometry was an accurate index of apoptosis, we measured caspase-3 activity (Fig. 3C). Cells treated with TRAIL exhibited a six- fold increase in caspase-3 activity relative to that in cells treated with vehicle. As it has been previously reported that MDA-MB-468 cells do not respond to TRAIL [20], we assessed the selectivity of Killer- TRAIL. MCF-12A cells were treated with 50 ngÆmL )1 KillerTRAIL and assessed for apoptosis by flow cytometry and caspase-3 activity. Neither assay Table 1. Effect of hypoxia on the induction of apoptosis in MDA- MB-468 cells. Percentage of apoptotic cells 8 h 24 h Normoxia 4.49 ± 1.41 3.48 ± 1.05 Hypoxia 4.52 ± 0.89 3.53 ± 0.36 Fig. 3. TRAIL induces apoptosis in breast cancer cells. MDA-MB-468 cells were treated with (A) increasing concentrations of KillerTRAIL for 4 h or (B) with 50 ngÆmL )1 KillerTRAIL for varying lengths of time. Cells were analyzed by FACS. Mean ± SE; n ¼ 3. (C) Caspase-3 activity confirmed apoptosis in MDA-MB-468 cells. **P < 0.01, mean ± SE; n ¼ 3. (D, E) MCF-12A cells were treated with 50 ngÆmL )1 KillerTRAIL for 4 h, harvested, and analyzed by FACS (D) or assayed for caspase-3 activity (E). Mean ± SE; n ¼ 3. L. D. Fitzgerald et al. Novel association between hypoxia, TRAIL and BRCA1 FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS 5141 revealed an induction of apoptosis in TRAIL-treated cells, indicating that the TRAIL effect on MDA-MB- 468 cells was selective for malignant cells. TRAIL does not alter the localization of truncated BRCA1 or induce apoptosis in a cell line expressing mutated BRCA1 The data in Figs 2 and 3 suggest that TRAIL can alter BRCA1 localization and induce apoptosis in breast cancer cells. Therefore, we assessed the requirement for functional BRCA1 in TRAIL-stimulated apoptosis. HCC1937 cells have a 5382insC mutation in BRCA1, resulting in the expression of a truncated, nonfunc- tional protein [21]. HCC1937 cells were exposed to hypoxia for 24 h (Fig. 4A) or treated with 50 ngÆmL )1 TRAIL for 4 h (Fig. 4B), and BRCA1 localization was assessed. There was no change in protein level or sub- cellular distribution of BRCA1 under these conditions. Cells were also exposed to TRAIL for up to 10 h, and apoptosis was not increased above that in cells exposed to vehicle during this time period (Fig. 4C). BRCA1 enhances TRAIL-mediated apoptosis Having demonstrated that HCC1937 cells were resis- tant to TRAIL-mediated apoptosis, we investigated whether introduction of wild-type BRCA1 could sensi- tize the cells to TRAIL. Routinely, we were able to obtain  20% transfection efficiency of BRCA1 into HCC1937 cells. This resulted in an increase in BRCA1 expression as evidenced by a more intense band recog- nized by BRCA1 antibody Ab-1 in cells transfected with pEGFPC1BRCA1 when compared to the inten- sity in cells transfected with empty vector (Fig. 5A). The slightly slower migration of the predominant band in BRCA1-transfected cells was consistent with the expression of a tagged, full-length protein. Transfec- tion of BRCA1 into these cells followed by treatment with vehicle did not alter the percentage of apoptotic cells relative to that in cells transfected with vector alone and treated with vehicle (Fig. 5A). Interestingly, in contrast to untransfected cells, the administration of TRAIL to cells transfected with vector alone resulted in a consistent, although not significant, increase in apoptosis relative to cells treated with vehicle. This effect was routinely observed, and we attribute it to the transfection procedure. Nevertheless, the expres- sion of BRCA1 in HCC1937 cells resulted in a 22-fold enhancement in apoptosis relative to cells expressing BRCA1 in the absence of TRAIL and a three-fold enhancement of apoptosis over that induced by TRAIL in the presence of the empty vector. Thus, the data indicate that expression of full-length BRCA1 augments TRAIL-induced apoptosis. We also investigated the effects of reduced BRCA1 expression on TRAIL-mediated apoptosis in MDA- MB-468 cells. Figure 5B demonstrates the decrease in BRCA1 in cells transfected with BRCA1-targeted short hairpin RNA (shRNA) relative to the level in cells transfected with control shRNA. The levels of active caspase-3 between cells treated with vehicle and those treated with TRAIL were assessed via immuno- fluorescence of cells transfected with either negative control shRNA or BRCA1-targeted shRNA. Transfec- tion efficiency was  10%. Integrated density values of ABC Fig. 4. Neither hypoxia nor TRAIL alters BRCA1 localization or induces apoptosis in HCC1937 cells. Cells were (A) exposed to normoxia or hypoxia for 24 h or (B) treated with 50 ngÆmL )1 KillerTRAIL for 4 h prior to harvesting, fractionation, and western blot analysis. Densitometric values were normalized to actin. Mean ± SE; n ¼ 3. wcl, whole cell lysate; N, normoxia; H, hypoxia; V, vehicle; T, TRAIL. Values for normox- ic and vehicle samples were arbitrarily set at 1. (C) Cells were treated with vehicle or 50 ngÆmL )1 KillerTRAIL for various lengths of time prior to FACS analysis. Mean ± SE; n ¼ 2. Novel association between hypoxia, TRAIL and BRCA1 L. D. Fitzgerald et al. 5142 FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS active caspase-3 fluorescence in cells treated with vehi- cle or TRAIL are shown in Fig. 5B. In cells transfect- ed with control shRNA, administration of TRAIL resulted in a three-fold increase in active caspase-3 flu- orescence relative to fluorescence in vehicle-treated cells. This increase was not evident in cells transfected with BRCA1 shRNA. Thus, these data support the hypothesis that BRCA1 enhances TRAIL’s ability to mediate apoptosis in breast cancer cell lines. Discussion There is increasing interest in the association between hypoxia and BRCA1. Recent studies showed that pro- longed exposure (48 h) of cells to severe hypoxia (0.01% oxygen) resulted in decreased levels of BRCA1 [18] and impaired ability of cells to repair damaged DNA [17]. However, no data are available on the effects of short-term exposure (£ 24 h) to hypoxia, which does not alter protein levels, on the localization or function of BRCA1 in mammary cells. As solid tumors grow, their oxygen consumption can exceed oxygen delivery, resulting in decreased intratumoral oxygen. The reported pO 2 values within breast tumors range from 0 to 100 mmHg, with the highest frequency occurring between 0 and 16 mmHg. In noncancerous tissue, the highest frequency of values occurred between 60 and 80 mmHg [12]. We exposed cells to 1% oxygen, which would exert a pressure of 7.6 mmHg at sea level. Thus, our studies are consistent with oxygen levels in the range of values occurring with high frequency in breast tumors. We monitored the sustained oxygen concentration in all experiments using an oxygen sensor within the Billups–Rothenburg hypoxia chamber. Furthermore, if cells were responsive to a low-oxygen environment, oxygen-sensitive mark- ers would be expected to increase. We observed an increase in HIF-1a protein, and thus were confident that the cells were being maintained in a low-oxygen environment and were responsive to low oxygen. There is also increasing interest in the effects of hypoxia on TRAIL-mediated apoptosis. Available data suggest that hypoxia blocks TRAIL-mediated apopto- sis by blocking the translocation of Bax to the mitochondria [14] and by activation of lysosomal cath- epsins [20]. In contrast, Lee et al. [22] reported that hypoxia synergistically enhanced TRAIL-mediated apoptosis in the DU-145 prostate cancer cell line. Furthermore, Kwon & Choi [23] reported that hydrogen peroxide increases the expression of TRAIL AB Fig. 5. BRCA1 enhances TRAIL-mediated apoptosis. (A) Approximately equimolar amounts of pEGFPC1 or pEGFPC1BRCA1 were transfected into HCC1937 cells. Expression levels were assessed by immu- noblotting lysates with BRCA1 antibody Ab-1. Membrane was stripped and reprobed for MSH-6 to ensure equal loading. Forty- eight hours post-transfection, cells were treated with KillerTRAIL for 4 h before being harvested and analyzed by FACS. *P < 0.05 compared to pEGFPC1 vehicle, pEGFPC1 TRAIL or pEGFPC1BRCA1 vehicle; mean ± SE; n ¼ 3. (B) Negative control or BRCA1-targeted shRNA was transfected into MDA-MB-468 cells. Twenty-four hours later, cells were treated with vehicle or 50 ngÆmL )1 TRAIL before fixation and immu- nofluorescence of active caspase-3. The val- ues graphed are densitometric values for fluorescence intensity of active caspase-3 in cells from at least 12 fields. BRCA1 immuno- blot shows BRCA1 levels in cells transfected with control or BRCA1-targeted shRNA. Loading variability was assessed by probing for MSH-6 as previously reported [24]. L. D. Fitzgerald et al. Novel association between hypoxia, TRAIL and BRCA1 FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS 5143 in astroglial cells, providing a precedent for oxidative stress-mediated regulation of TRAIL expression. The results presented here provide the first evidence linking hypoxia, TRAIL signaling and BRCA1 in mammary epithelial cells. Our data demonstrate that hypoxia induces nuclear accumulation of BRCA1 with- out altering its protein level, and that this effect is mediated, at least in part, by TRAIL. Our results also indicate that hypoxia can upregulate TRAIL cell sur- face expression and that the apoptotic effects of TRAIL are enhanced by BRCA1. Thus, our studies are consistent with others demonstrating a role for BRCA1 in stress-induced apoptotic signaling. Our results further indicate that whereas both hypoxia and TRAIL effect a change in BRCA1 locali- zation, only TRAIL induces apoptosis. The inability of hypoxia to cause apoptosis was not surprising, consid- ering the number of reports demonstrating that hypoxia blocks TRAIL-induced apoptosis. There are two possible explanations for our results. First, hypoxia could stimulate TRAIL signaling, leading to the activation of two distinct pathways. One pathway might lead to increased nuclear localization of BRCA1 and the other to apoptosis. Hypoxia could block the apoptotic arm (Fig. 6A). Alternatively, hypoxia could stimulate TRAIL signaling, resulting in nuclear accu- mulation of BRCA1 and potentially leading to the induction of TRAIL-mediated apoptosis. However, hypoxia might also block this pathway downstream of the change in BRCA1 localization (Fig. 6B). Our data are also consistent with studies demon- strating TRAIL’s specificity for malignant cells [16]. However, it has been reported that MDA-MB-468 cells are resistant to TRAIL [20]. Our data contradict those findings, and there are several possible reasons for the discrepancy. Previous studies used 50 nm TRAIL, whereas our results were obtained with 50 ngÆmL )1 TRAIL, which is approximately five times higher. A dose–response curve demonstrated this concentration to be a near maximally effective dose. We also used commercially available KillerTRAIL, which does not require a crosslinker or enhancer for its biological activity. Finally, our experiments were performed in the presence of 1% serum, whereas other studies were performed in 10% serum. With the low level of in vivo toxicity reported for TRAIL and its selective cytotoxic- ity, it has been targeted as a desirable therapeutic tool, and its development as a therapeutic modality is underway. Our data suggest that BRCA1 expression may be critical to the success of TRAIL as a chemo- therapeutic agent. BRCA1 is mutated in approximately half of hereditary breast cancer cases, and its expres- sion is reduced in sporadic cancers. Therefore, the level of functional protein is decreased, and this may hinder the response of cells to TRAIL. In conclusion, this study provides novel evidence that BRCA1 is critical in the transduction of a TRAIL-mediated apoptotic signal. It also suggests that hypoxia may not only block apoptosis resulting from exogenously administered TRAIL, but also blocks apoptosis induced by endogenously expressed TRAIL. Investigation of the mechanism for this inhibition is ongoing. However, regardless of the mechanism, this article establishes a connection between hypoxia, TRAIL and BRCA1 in regulating apoptosis in breast cancer cells. Experimental procedures Cell culture MCF-7, MDA-MB-468, HCC1937 and MCF-12A cells were from the American Type Culture Collection (Manas- sas, VA, USA). Unless otherwise specified, cell culture reagents were from Invitrogen (Carlsbad, CA, USA) or Sigma (St Louis, MO, USA). MCF-7 cells were cultured as previously described [24]. MDA-MB-468 cells were cultured in DMEM and 10% Fetal Clone III (fetal bovine serum) (Hyclone, Logan, UT, USA). HCC1937 cells were cultured in RPMI 1640 (Hyclone), 1% insulin ⁄ transferrin ⁄ selenium A, and 10% fetal bovine serum. MCF-12A cells were cultured in DMEM ⁄ F-12, 20 ngÆmL )1 human epidermal growth factor, 100 ngÆmL )1 cholera toxin, 0.01 mgÆmL )1 bovine insulin, 500 ngÆmL )1 hydrocortisone, and 5% horse serum. HMECs (provided by S. Eltom, Meharry Medical College) were cultured in DMEM ⁄ F-12, 10 mm Hepes, 1 lgÆmL )1 bovine insulin, 1 lgÆmL )1 hydrocortisone, 10 lgÆmL )1 ascorbic acid, 12.5 ngÆmL )1 EGF, 10 lgÆmL )1 transferrin, 0.1 mm phosphoethanolamine, 2 nm b-estradiol, 10 nm triiodo-l-thyronine, 15 nm sodium selenite, 0.1 mm ethanolamine, 1 ngÆ mL )1 cholera toxin, 1% fetal bovine serum, and 35 lgÆmL )1 bovine pituitary extract. Fig. 6. Schematic representation of two possible mechanisms by which hypoxia could preclude TRAIL-mediated apoptosis. Novel association between hypoxia, TRAIL and BRCA1 L. D. Fitzgerald et al. 5144 FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS For hypoxia and TRAIL studies, 3 · 10 6 cells were pla- ted in 150 mm dishes for 24 h. Cells were incubated in media containing 1% serum for 1 h before continuation of normoxia or initiation of hypoxia (1% oxygen, 5% carbon dioxide, 94% nitrogen in a Billups–Rothenberg chamber; Billups Rothenberg, Inc., Del Mar, CA, USA) for 24 h or treatment with KillerTRAIL (Alexis Biochemicals, San Diego, CA, USA) or vehicle (20 mm Hepes, pH 7.4, 300 mm NaCl, 0.006% Tween-20, 1% sucrose, 0.05 mm dithiothreitol). For TRAIL neutralization experiments, monoclonal antibody to human TRAIL (2E5; Alexis Bio- chemicals) was added 1 h before hypoxia initiation. Hypoxia was initiated by flushing a Billups–Rothenburg chamber, which housed the cells, with a gas mixture of 1% oxygen, 5% carbon dioxide and 94% nitrogen until an oxy- gen sensor (Micropac O 2 ; Drager, Leubeck, Germany), which was also sealed within the chamber, stabilized at 1.0– 1.2% oxygen. The oxygen meter remained in the chamber throughout the experiment, and experiments were completed only if the oxygen level remained between 1.0% and 1.2%. Subcellular fractionation and western blotting Cell fractionation and BRCA1 immunoblotting were per- formed as previously described [24]. HIF-1a and MSH-6 antibodies (BD Biosciences, San Jose, CA, USA) were diluted 1 : 500. Caspase activity assays Cells were plated at 2–3 · 10 6 cells per 100 mm dish for 24 h before treatment. Two hundred micrograms of protein was assayed using a Caspase-3 Colorimetric Assay Kit (Sigma). Fluorescence-activated cell sorting (FACS) After treatment, cells and media were collected and centri- fuged at 500 g at 4 °C with an Eppendorf 5810 centrifuge and A-4-62 rotor. Cells were permeabilized in 0.1% sodium citrate, 0.1% Triton X-100, and 50 lgÆmL )1 propidium iodide, and analyzed by FACS using a Becton-Dickinson (Franklin Lakes, NJ, USA) FACScan Benchtop Analyzer and winlist software. Transfections Modified pEGFPC1 vector and pEGFPC1BRCA1 have been described previously [2]. HCC1937 cells were plated at 1 · 10 6 cells per 100 mm dish and transfected with approxi- mately equimolar amounts of pEGFPC1 (5 lg) or pEG- FPC1BRCA1 (10 lg) using GeneJammer (Stratagene, La Jolla, CA, USA). Forty-eight hours later, cells were treated and harvested for flow cytometry. Immunofluorescence Immunofluorescence and quantification were performed as previously described [24]. Immunofluorescence of cells in captured images was quantified using AlphaImager2000 (Alpha Innotech, San Leandro, CA, USA) or the image j processing and analysis program (NIH, Bethesda, MD, USA). For detection of cell surface TRAIL, Triton X-100 was omitted. TRAIL antibody (ProSci Incorporated, Poway, CA, USA) was diluted 1 : 100. shRNA studies Cells were plated at 120 000 cells per well in a six-well cul- ture dish and transfected with 1 lg of pEGFPC1 green fluorescent protein expression plasmid (Clontech, Mountain View, CA, USA) and 3 lg of negative control shRNA or BRCA1 shRNA (SuperArray Bioscience, Frederick, MD, USA) using GeneJammer. Twenty-four hours post-transfec- tion, cells were treated. Three hours after treatment, cells were fixed in 2% paraformaldehyde and immunostained for active caspase-3 (Promega, Madison, WI, USA; 1 : 500) using Cy-3-conjugated donkey anti-(rabbit IgG) (Jackson Immunoresearch, West Grove, PA, USA). Transfected cells were assessed by the presence of green fluorescence, and Cy-3 immunofluorescence was quantified by imagej software. Statistical analysis Student’s t-test and anova were performed using graphpad prism version 4 for Windows (GraphPad Software, San Diego, CA, USA). Acknowledgements We thank Drs Hal Moses, Lynn Matrisian and Lee Limbird for helpful comments. This work was sup- ported by Public Health Service Grants 2G12RR03032 from NCRR-supported Research Centers in Minority Institutions, KO1 CA89494 and U54 CA91408 from the National Cancer Institute, Research Initiative for Scientific Enhancement Grant 2R25 G59994, and T32 HL007735-11 from the National Heart, Lung and Blood Institute. References 1 Rodriguez JA & Henderson BR (2000) Identification of a functional nuclear export sequence in BRCA1. J Biol Chem 275, 38589–38596. 2 Thompson ME, Robinson-Benion CL & Holt JT (2005) An amino-terminal motif functions as a second nuclear L. D. Fitzgerald et al. Novel association between hypoxia, TRAIL and BRCA1 FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS 5145 export sequence in BRCA1. J Biol Chem 280, 21854– 21857. 3 Somasundaram K, Zhang H, Zeng Y-X, Houvras Y, Peng Y, Zhang H, Wu GS, Licht JD, Weber BL & El-Deiry WS (1997) Arrest of the cell cycle by the tumour-suppressor BRCA1 requires the CDK-inhibitor p21WAF1 ⁄ CiPl. Nature 389, 187–190. 4 Chapman MS & Verma IM (1996) Transcriptional acti- vation by BRCA1. Nature 382, 678–679. 5 Abbott DW, Thompson ME, Robinson-Benion C, Tomlinson G, Jensen RA & Holt JT (1999) BRCA1 expression restores radiation resistance in brca1-defec- tive cancer cells through enhancement of transcription- coupled DNA repair. J Biol Chem 274, 18808–18812. 6 Shao N, Chai Y, Shyam E, Reddy P & Rao V (1996) Induction of apoptosis by the tumor suppressor protein BRCA1. Oncogene 13, 1–7. 7 Jin S, Zhao H, Fan F, Blanck P, Fan W, Colchagie A, Fornace AJ Jr & Zhan Q (2000) BRCA1 activation of the GADD45 promoter. Oncogene 19, 4050–4057. 8 Thangaraju M, Kaufmann SH & Couch FJ (2000) BRCA1 facilitates stress-induced apoptosis in breast and ovarian cancer cell lines. J Biol Chem 275, 33487– 33496. 9 Harkin DP, Bean JM, Miklos D, Song Y-H, Truong VB, Englert C, Christians FC, Ellisen LW, Maheswaran S, Oliner JD et al. (1999) Induction of GADD45 and JNK ⁄ SAPK-dependent apoptosis following inducible expression of BRCA1. Cell 97, 575–586. 10 Yuan R, Fan S, Wang J, Meng Q, Ma Y, Schreiber D, Goldberg I & Rosen E (1999) Coordinate alterations in the expression of BRCA1, BRCA2, p300, and Rad51 in response to genotoxic and other stresses in human pros- tate cancer cells. Prostate 40, 37–49. 11 Scully R, Chen J, Ochs RL, Keegan K, Hoekstra M, Feunteun J & Livingston DM (1997) Dynamic changes of BRCA1 subnuclear location and phosphorylation state are initiated by DNA damage. Cell 90, 425–435. 12 Vaupel P, Schlenger K, Knoop C & Hockel M (1991) Oxygenation of human tumors: evaluation of tissue oxy- gen distribution in breast cancers by computerized O2 tension measurements. Cancer Res 51, 3316–3322. 13 Semenza GL (2003) Targeting Hif-1 for cancer therapy. Nat Rev Cancer 3, 721–732. 14 Kim M, Park S-Y, Pai H-S, Kim T-H, Billiar TR & Seol D-W (2004) Hypoxia inhibits tumor necrosis factor-related apoptosis-inducing ligand-induced apop- tosis by blocking Bax translocation. Cancer Res 64, 4078–4081. 15 Nagaraj N, Vigneswaran N & Zacharias W (2007) Hypoxia inhibits TRAIL-induced tumor cell apoptosis: involvement of lysosomal cathepsins. Apoptosis 12, 125– 139. 16 Wiley S, Schooley K, Smolak P, Din W, Huang C, Nicholl J, Sutherland G, Smith T, Rauch C, Smith C et al. (1995) Identification and characterization of a new member of the TNF family that induces apoptosis. Immunity 3, 673–682. 17 Meng AX, Jalali F, Cuddihy A, Chan N, Bindra RS, Glazer PM & Bristow RG (2005) Hypoxia down- regulates DNA double strand break repair gene expression in prostate cancer cells. Radiother Oncol 76, 168–176. 18 Bindra RS, Gibson SL, Meng A, Westermark U, Jasin M, Pierce AJ, Bristow RG, Classon MK & Glazer PM (2005) Hypoxia-induced down-regulation of BRCA1 expression by E2Fs. Cancer Res 65, 11597–11604. 19 Gibson S, Bindra R & Glazer P (2006) CHK2-depen- dent phosphorylation of BRCA1 in hypoxia. Radiation Res 166, 646–651. 20 Singh TR, Shankar S, Chen X, Asim M & Srivastava RK (2003) Synergistic interactions of chemotherapeutic drugs and tumor necrosis factor-related apoptosis- inducing ligand ⁄ Apo-2 ligand on apoptosis and on regression of breast carcinoma in vivo. Cancer Res 63, 5390–5400. 21 Tomlinson GE, Chen TT-L, Stastny VA, Virmani AK, Spillman MA, Tonk V, Blum JL, Schneider NR, Wi- stuba II, Shay JW et al. (1998) Characterization of a breast cancer cell line derived from a germ-line BRCA1 mutation carrier. Cancer Res 58, 3237–3242. 22 Lee Y, Moon M, Kwon S & Rhee J (2005) Hypoxia and low glucose differentially augment TRAIL-induced apoptotic death. Mol Cell Biochem 270, 89–97. 23 Kwon D & Choi I (2006) Hydrogen peroxide upregu- lates TNF-related apoptosis inducing ligand (TRAIL) expression in human astroglial cells, and augments apoptosis of T cell. Yonsei Med 47, 551–557. 24 Hinton CV, Fitzgerald LD & Thompson ME (2007) Phosphatidylinositol 3-kinase ⁄ Akt signaling enhances nuclear localization and transcriptional activity of BRCA1. Exp Cell Res 313, 1735–1744. Novel association between hypoxia, TRAIL and BRCA1 L. D. Fitzgerald et al. 5146 FEBS Journal 274 (2007) 5137–5146 ª 2007 The Authors Journal compilation ª 2007 FEBS . BRCA1 accumulates in the nucleus in response to hypoxia and TRAIL and enhances TRAIL- induced apoptosis in breast cancer cells Latricia D. Fitzgerald 1, *, Charvann. 22-fold enhancement in apoptosis relative to cells expressing BRCA1 in the absence of TRAIL and a three-fold enhancement of apoptosis over that induced by TRAIL in the presence of the empty vector. Thus, the data. and induce apoptosis in breast cancer cells. Therefore, we assessed the requirement for functional BRCA1 in TRAIL- stimulated apoptosis. HCC1937 cells have a 5382insC mutation in BRCA1, resulting