The proto-oncogene Bcl3 induces immune checkpoint PD-L1 expression, mediating proliferation of ovarian cancer cells

The proto-oncogene Bcl3 induces survival and proliferation in cancer cells; however, its function and regulation in ovarian cancer (OC) remain unknown. Here, we show that Bcl3 expression is increased in human OC tissues. Surprisingly, however, we found that in addition to promoting survival, proliferation, and migration of OC cells, Bcl3 promotes both constitutive and interferon-γ (IFN)-induced expression of the immune checkpoint molecule PD-L1. The Bcl3 expression in OC cells is further increased by IFN, resulting in increased PD-L1 transcription. The mechanism consists of an IFN-induced, Bcl3- and p300-dependent PD-L1 promoter occupancy by Lys-314/315 acetylated p65 NF-κB. Blocking PD-L1 by neutralizing antibody reduces proliferation of OC cells overexpressing Bcl3, suggesting that the pro-proliferative effect of Bcl3 in OC cells is partly mediated by PD-L1. Together, this work identifies PD-L1 as a novel target of Bcl3, and links Bcl3 to IFNγ signaling and PD-L1-mediated immune escape.

The proto-oncogene Bcl3 induces survival and proliferation in cancer cells; however, its function and regulation in ovarian cancer (OC) remain unknown. Here, we show that Bcl3 expression is increased in human OC tissues. Surprisingly, however, we found that in addition to promoting survival, proliferation, and

migration of OC cells, Bcl3 promotes both constitutive and interferon-␥ (IFN)-induced expression of the immune checkpoint molecule PD-L1. The Bcl3 expression in OC cells is further increased by IFN, resulting in increased PD-L1 transcription. The mechanism consists of an IFN-induced, Bcl3-and p300-dependent PD-L1 promoter occupancy by Lys-314/315 acetylated p65 NF-B. Blocking PD-L1 by neutralizing antibody reduces proliferation of OC cells overexpressing Bcl3, suggesting that the pro-proliferative effect of Bcl3 in OC cells is partly mediated by PD-L1. Together, this work identifies PD-L1 as a novel target of Bcl3, and links Bcl3 to IFN␥ signaling and PD-L1mediated immune escape.
The proto-oncogene Bcl3 is a member of IB family that was first identified in patients with chronic lymphocytic leukemia (1,2). However, unlike other IB proteins in cancer cells, Bcl3 is a predominantly nuclear protein, which contains a transactivation domain, and can be recruited to NF-B-responsive promoters, resulting in transcriptional activation or repression, depending on the subunit composition of NF-B complexes, and other transcription factors and regulators present in the transcription complexes (3)(4)(5)(6)(7)(8)(9). High expression of Bcl3 has been reported in a number of hematological malignancies (10 -16), as well as in several solid tumors, including breast cancer, nasopharyngeal carcinoma, and colorectal and cervical cancer (17)(18)(19)(20)(21)(22)(23). Consistent with its oncogenic function, Bcl3 can transform cells and induce their proliferation and tumor growth (24). Recent studies have shown that miR-125b, which targets Bcl3, is down-regulated in ovarian cancer (OC) 2 tissues (25,26), suggesting an increased Bcl3 expres-sion in ovarian cancer. However, the Bcl3 expression in OC has not been investigated, and its function in OC cells remains unknown.
Epithelial ovarian cancer (EOC) is the most common gynecological cancer in women, with poor survival and high mortality rates. As many other types of cancer, EOC is characterized by an increased activity of the transcription factor NF-B (27)(28)(29), which promotes expression of anti-apoptotic and pro-angiogenic genes. However, recent studies have shown that in addition to inducing expression of anti-apoptotic and pro-inflammatory genes, NF-B induces transcription of the immune checkpoint molecule, programmed death ligand 1 (PD-L1; B7-H1, CD274) (30 -35). PD-L1 expression on tumor cells is induced by interferon-␥ (IFN). By binding to programmed cell death-1 (PD-1) expressed on cytotoxic T cells, PD-L1 then induces T cell apoptosis and tolerance, thus inhibiting the antitumor immunity. However, tumor PD-L1 has also tumor-intrinsic effects that include increased cancer cell survival and proliferation, regulation of tumor glucose utilization, and inhibition of autophagy (36 -38). PD-L1 is expressed on the surface of OC cells, and its increased expression correlates with poor prognosis in OC patients (38 -41); however, the mechanisms that regulate the PD-L1 expression in OC cells are not known.
Here, we show that Bcl3 expression is increased in human EOC tissues, and Bcl3 overexpression promotes survival, proliferation, and migration of OC cells. Remarkably, however, our results show that in addition to promoting survival and proliferation, Bcl3 induces both constitutive and IFN-induced PD-L1 expression in OC cells. The mechanism consists of Bcl3-and p300-mediated recruitment of Lys-314/315 acetylated p65 NF-B to the PD-L1 promoter in IFN-treated cells. In OC cells overexpressing Bcl3, neutralization of the induced PD-L1 decreases cell proliferation, indicating that the pro-proliferative effect of Bcl3 is partly mediated by PD-L1. These data identify PD-L1 as a novel target of Bcl3, and suggest that in addition to promoting cell proliferation, Bcl3 regulates immune escape in cancer cells.
To explore the functional significance of Bcl3 in OC cells, we first examined the effect of Bcl3 suppression on OC cell apoptosis, proliferation, and migration. Suppression of Bcl3 by siRNA decreased Bcl3 mRNA ( Fig. 2A) and protein (Fig. 2, B and C) levels in SKOV3 and OVCAR3 cells by about 50% compared with control scramble siRNA. Of note, in whole cell extracts (WCE) of OC cells, Bcl3 runs as a doublet of an approximate 50 kDa on SDS gels (Fig. 2B), consistent with previous reports demonstrating Bcl3 phosphorylation (24,(45)(46)(47). Importantly, Bcl3 suppression significantly increased apoptosis, evaluated by nucleosome release into the cytoplasm (Fig.  2D) (48) and by caspase-3 assay (Fig. 2E), and decreased proliferation of SKOV3 (Fig. 2F) and OVCAR3 (Fig. 2G) cells. Furthermore, Bcl3 suppression by siRNA significantly reduced migration of OC cells (Fig. 3).
To validate the above data, we suppressed and overexpressed Bcl3 in SKOV3 cells using CRISPR knockout and activation plasmids. Suppression of Bcl3 by CRISPR/Cas9 reduced both Bcl3 mRNA (Fig. 4A) and protein levels (Fig. 4, B and C). Importantly, Bcl3 suppression significantly increased apoptosis (Fig.  4D) and decreased proliferation (Fig. 4E) in SKOV3 cells. Conversely, Bcl3 overexpression decreased apoptosis (Fig. 4D) and increased cell proliferation (Fig. 4F). To confirm these data, we generated SKOV3 cells stably transfected with Bcl3 shRNA. Compared with the control SKOV3 cell line transfected with 3 Please note that the JBC is not responsible for the long-term archiving and maintenance of this site or any other third party hosted site.

Bcl3 induces PD-L1 in ovarian cancer cells
empty expression vector, SKOV3 cells stably transfected with Bcl3 shRNA express significantly decreased Bcl3 mRNA (Fig. 4A) and protein (Fig. 4, B and C) levels. Importantly, these cells exhibit increased apoptosis (Fig. 4D) and reduced proliferation ( Fig. 4G) compared with control SKOV3 cells. These data demonstrate that Bcl3 promotes OC cell survival, migration, and proliferation.

Bcl3 mediates constitutive PD-L1 expression in OC cells
Because Bcl3 regulates NF-B-dependent transcription, we analyzed expression of NF-B-dependent genes cIAP1, BclxL, TGF␤1, and IB␣ in SKOV3 and OVCAR3 cells transfected with Bcl3 siRNA. In addition, because PD-L1 promotes OC growth and cell proliferation (38,49), and is regulated by NF-B (30 -35), we wondered whether Bcl3 might regulate PD-L1 expression in OC cells. Remarkably, whereas Bcl3 suppression by siRNA did not have a substantial effect on cIAP1, BclxL, TGF␤1, and IB␣ mRNA levels, it significantly reduced PD-L1 expression in SKOV3 and OVCAR3 cells (Fig. 5A). In addition, Bcl3 suppression by siRNA signifi-cantly decreased the intracellular PD-L1 protein levels in both cell types (Fig. 5, B and C).

IFN induces Bcl3 expression in OC cells
PD-L1 expression on tumor cells, including OC cells, is induced by IFN produced by CD8 T cells (50 -52). Because our

Bcl3 induces PD-L1 in ovarian cancer cells
data showed that Bcl3 promotes PD-L1 expression in OC cells (Fig. 5), we asked whether IFN might regulate the Bcl3 expression. In this regard, Bcl3 expression was reported to be up-regulated by pro-inflammatory cytokines including TNF␣, IL-1␤, and IL-6 (20). However, to our knowledge, there is no available evidence showing that IFN induces Bcl3 expression. Indeed, human recombinant IFN significantly increased Bcl3 mRNA levels in SKOV3 and OVCAR3 cells (Fig. 6A). In addition, in agreement with a previous study demonstrating increased surface expression of PD-L1 in IFN-treated OC cells (52), IFN significantly increased PD-L1 mRNA levels in both OC cell lines (Fig. 6B). Importantly, IFN also increased intracellular Bcl3 and PD-L1 protein levels in SKOV3 and OVCAR3 cells (Fig. 6, C and D). Together, these data demonstrated that IFN induces Bcl3 expression in OC cells, and suggested a link between IFN, Bcl3, and PD-L1 signaling.

Bcl3 mediates IFN-induced PD-L1 expression in OC cells
Having shown that Bcl3 promotes the basal PD-L1 expression in OC cells (Fig. 5), and that IFN increases the intracellular levels of Bcl3 and PD-L1 (Fig. 6, A-D), we wanted to determine whether Bcl3 mediates also the IFN-induced PD-L1 expression in OC cells. To this end, we analyzed PD-L1 expression in IFNtreated SKOV3 and OVCAR3 cells transfected with control and Bcl3 siRNA. Bcl3 suppression (Fig. 6E) significantly attenuated the IFN-induced PD-L1 expression in SKOV3 and OVCAR3 cells (Fig. 6F), indicating that Bcl3 mediates the IFN-induced PD-L1 expression in OC cells.

IFN induces PD-L1 promoter occupancy by p65, Lys-314/315 acetylated p65, and p300
Because recent studies have shown that PD-L1 expression is regulated by p65 NF-B (30 -35), we wanted to determine whether the Bcl3-mediated PD-L1 expression in IFN-treated OC cells is associated with an increased p65 promoter occupancy. Furthermore, because Lys-314/315 acetylation of p65 regulates its transcriptional activity in OC cells (53), we analyzed the PD-L1 promoter occupancy by Lys-314/315 ac-p65. The human PD-L1 promoter contains several putative NF-Bbinding sites: B1 site (GGAAAGTCCA) (30) located at position Ϫ65 upstream from the transcription start site (TSS), B2 site (GGGGGACGCC) (34) located Ϫ358 from TSS, B3 site (GGGAAGTTCT) located Ϫ600 from TSS (30), and B4/B5 sites containing an identical putative NF-B binding sequence (GGGAAGTCAC) located Ϫ1256 and Ϫ1283 from TSS (Fig.  7A). So far, p65 recruitment to the B2 site has been demonstrated in non-small cell lung cancer and triple negative breast cancer cells (34,35), and p65 was also recruited to the B3 site in lipopolysaccharide-stimulated macrophages (30). However, it is not known whether NF-B binds to the B1 and/or B4/ B5 sites, or whether NF-B occupies the PD-L1 promoter in OC cells.
To determine whether the increased PD-L1 promoter occupancy by Lys-314/315 ac-p65 is associated with an increased occupancy of a histone acetyltransferase (HAT), we analyzed recruitment of the HATs cAMP-response element-binding protein (CBP) and p300, known to acetylate p65 (56). Although CBP was not significantly recruited, p300 was heavily recruited to all B sites in the PD-L1 promoter, and this recruitment was further enhanced by IFN treatment (Fig. 7, B-E). In addition, we tested whether the PD-L1 promoter is occupied by Bcl3;

Bcl3 induces PD-L1 in ovarian cancer cells
however, we did not observe any significant recruitment (Fig. 7, B-E).

Bcl3 and p300 mediate IFN-induced Lys-314/315 ac-p65 recruitment to PD-L1 promoter
Even though Bcl3 was not directly recruited to PD-L1 promoter, we wanted to test whether it might mediate the IFNinduced Lys-314/315 ac-p65, p65, and p300 occupancy. In addition, because p300 was recruited to the PD-L1 promoter, we analyzed whether it might facilitate the Lys-314/315 ac-p65 promoter occupancy. To this end, we measured Lys-314/315 ac-p65, p65, and p300 recruitment to PD-L1 promoter in SKOV3 cells transfected with control, Bcl3, and p300 siRNA and treated with IFN (0 and 50 ng/ml) for 6 h. Interestingly, both Bcl3 and p300 silencing significantly suppressed the IFNinduced PD-L1 promoter occupancy by Lys-314/315 ac-p65 (Fig. 8A). In contrast, p65 recruitment to PD-L1 promoter was not Bcl3-or p300-dependent (Fig. 8B); however, given the relatively low occupancy of p65 at the PD-L1 promoter, it was difficult to accurately assess the role of Bcl3 and p300 in p65 recruitment. The occupancy of p300 at the PD-L1 promoter in both untreated and IFN-treated cells was also not suppressed by Bcl3 silencing (Fig. 8C), indicating that p300 resides on the PD-L1 promoter even in the absence of Bcl3. Together, these data suggest that the PD-L1 promoter in OC cells is permanently occupied by p300, and upon IFN stimulation, Bcl3 facilitates Lys-314/315 p65 acetylation and promoter occupancy, resulting in increased PD-L1 transcription.

PD-L1 mediates Bcl3 pro-proliferative effect in OC cells
Because in addition to suppressing the anti-tumor activity of cytotoxic T cells, tumor PD-L1 has tumor-intrinsic effects (36 -38), we asked whether the Bcl3 pro-proliferative effect in OC cells might be mediated by PD-L1. To address this question, we analyzed proliferation of SKOV3 and OVCAR3 cells transfected with Bcl3 overexpression or control plasmids, in the presence of PD-L1 neutralizing antibody, or isotype-matched control IgG. The results demonstrated that compared with

Bcl3 induces PD-L1 in ovarian cancer cells
control IgG, PD-L1 neutralizing antibody significantly reduced proliferation of SKOV3 cells, both in cells transfected with control plasmid, and in cells transfected with Bcl3 overexpression plasmid (Fig. 9A). Similar results were observed in OVCAR3 cells (Fig. 9B), indicating that the pro-proliferative effect of Bcl3 in OC cells is partly mediated by PD-L1. Because IFN increases the Bcl3 expression in OC cells (Fig.  6), and promotes OC tumor growth in mice (52), we tested whether OC cell proliferation in IFN-treated cells also depends on PD-L1. Incubation of SKOV3 (Fig. 9C) and OVCAR3 (Fig.  9D) cells with IFN in the presence of control IgG increased cell proliferation, but this effect was observed only during later incubation times. Importantly, the OC cell proliferation in IFNtreated cells was significantly reduced in the presence of PD-L1 neutralizing antibody. These data are consistent with the recent in vivo study by Abiko et al. (52) demonstrating that the IFN-induced OC tumor growth is PD-L1 dependent. Together, these results indicate that IFN induces Bcl3 expression, resulting in the increased PD-L1 transcription and OC cell proliferation (Fig. 10).

Discussion
Our study shows, rather surprisingly, that in addition to promoting cell survival and proliferation, the proto-oncogene Bcl3 induces expression of PD-L1 in ovarian cancer cells. In addition, our findings demonstrate that Bcl3 expression is increased in OC tissues, and is induced by IFN in OC cells. The mechanism of how Bcl3 induces PD-L1 transcription in IFN-stimulated cells involves an increased, Bcl3-and p300-dependent recruitment of Lys-314/315 ac-p65 to PD-L1 promoter. Because blocking PD-L1 with neutralizing antibody reduces proliferation of OC cells overexpressing Bcl3 or treated with

Bcl3 induces PD-L1 in ovarian cancer cells
IFN, these results suggest that the pro-proliferative effect of Bcl3 in OC cells is partly mediated by PD-L1. Together, these data link Bcl3 to IFN␥ and PD-L1 signaling, and suggest that in addition to mediating cell survival and proliferation, Bcl3 promotes immune escape in cancer cells (Fig. 10).
Bcl-3 was originally identified as a candidate proto-oncogene up-regulated in B-cell chronic lymphocytic leukemia (1,2); later studies demonstrated its increased expression also in other hematological malignancies, as well as in several types of solid cancer (10 -23). The link between Bcl3 overexpression and malignant transformation was suggested to stem from its transcriptional up-regulation of cyclin D1 (57), increased expression of HDM2, the main negative regulator of p53 (58), and regulation of DNA damage response (59). In addition, recent studies have shown that Bcl3 induces expression of proinflammatory cytokines IL-8 and IL-17 in cutaneous T cell lymphoma cells (16), and TGF␤ signaling in breast cancer (60). Our present findings demonstrate that Bcl3 promotes expression of PD-L1, indicating that in addition to regulating NF-Bdependent genes involved in cell survival and proliferation, Bcl3 controls genes involved in immune escape. However, the regulation of NF-B-dependent transcription by Bcl3 is gene specific; whereas Bcl3 induces transcription of PD-L1, it does not have a significant effect on the expression of NF-B-regulated genes cIAP1, BclxL, TGF␤1, or IB␣, in OC cells (Fig. 5).

Bcl3 induces PD-L1 in ovarian cancer cells
What determines the specificity of the transcriptional regulation by Bcl3? Because Bcl3 contains a transactivation domain, it can modulate transcription depending on the transcription factors and co-regulators present in the transcription complexes (6 -9, 16, 61, 62). In this context, Bcl3 was shown to interact with the NF-B subunits p50 and p52, the AP-1 transcription factors c-Jun and c-Fos, STAT1, STAT3, PPAR␥, class I histone deacetylases, and the HATs CBP and p300 (6 -9, 16, 22, 24, 57, 62-66). Our results demonstrate that even though Bcl3 is not directly recruited to the PD-L1 promoter, it mediates, together with the HAT p300 present at the PD-L1 promoter (Fig. 7), the promoter occupancy by Lys-314/315 acetylated p65 NF-B in IFN-treated OC cells (Fig. 8A). Interestingly, previous studies have reported that acetylation of p65 at Lys-314/315 is mediated by p300, and results in a genespecific regulation of NF-B-dependent genes in TNF-stimulated cells (54,55). Together, our data support a model in which the PD-L1 promoter in OC cells is occupied by p300, and upon IFN stimulation, Bcl3 promotes PD-L1 transcription by facilitating the promoter-specific occupancy by Lys-314/315 ac-p65. Future studies should determine whether IFN induces p65 acetylation on Lys-314/315 by p300, and/or

Bcl3 induces PD-L1 in ovarian cancer cells
whether it induces histone deacetylases's removal. In addition, it will be interesting to determine whether the high promoter occupancy by Lys-314/315 ac-p65 in IFN-treated OC cells is unique for PD-L1, or whether IFN induces Lys-314/315 ac-p65 recruitment to other NF-B-dependent promoters as well.
Little is known about the signaling pathways inducing Bcl3 expression in cancer cells. In line with the reported induction of Bcl3 by p65 NF-B (68), Bcl3 expression was shown to be upregulated by pro-inflammatory cytokines including TNF␣, IL-1, and IL-6 (20). Our study is the first to demonstrate that the Bcl3 expression is induced also by IFN␥ (Fig. 6). The induction of Bcl3 by IFN is intriguing, especially because our data also show that the IFN-induced Bcl3 expression promotes expres-sion of PD-L1 in IFN-stimulated cells (Fig. 6), thus linking Bcl3 to IFN and PD-L1 signaling.
Increased PD-L1 expression in OC tissues promotes tumor growth (39 -41), but the regulation of PD-L1 expression in OC cells is little understood. Our study demonstrates that the PD-L1 expression in OC cells is regulated by the proto-oncogene Bcl3. Analysis of four different public datasets, together containing 26 control ovarian tissues and 898 OC samples, has revealed that the Bcl3 gene expression is statistically increased in OC tissues (p Յ 0.016; Fig. 1). Interestingly, the Bcl3 expression was most significantly increased in ovarian serous surface papillary carcinoma (Fig. 1C) (43). Even though the sample size was relatively small, these data suggest that this type of OC has a significantly higher Bcl3 expression compared with other

Bcl3 induces PD-L1 in ovarian cancer cells
types of OC. Alternatively, as the tumor samples are often heterogeneous, a subset of cancer cells may express higher levels of Bcl3. It will be important to correlate these data in future with Bcl3 protein levels in individual cells. In addition, because the Bcl3 transcriptional activity is regulated by phosphorylation (24,(45)(46)(47), future studies should analyze the phosphorylation status of nuclear Bcl3 in OC tissues. In addition to inhibiting anti-tumor cytotoxic T cells, the tumor-expressed PD-L1 has tumor-intrinsic effects that include the regulation of cancer cell survival and proliferation, autophagy, and regulation of glucose metabolism and mTOR signaling (36 -38). Our results show that Bcl3 has prosurvival and pro-proliferative effects in OC cells (Figs. 2-4). However, because blocking the Bcl3-induced PD-L1 by neutralizing antibody decreases proliferation in Bcl3-overexpressing cells (Fig.  9), these data indicate that the Bcl3 prosurvival effect in OC cells is, at least partly, mediated by the Bcl3-up-regulated PD-L1. In summary, our study identifies PD-L1 as a novel target of Bcl3, indicating that Bcl3 regulates not only cancer cell proliferation and survival, but also immune escape.

Transfection with siRNA and CRISPR knockout and overexpression plasmids
Human Bcl3 (sc-29789) and nonsilencing (sc-37007) siRNAs were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Prior to transfection, 2 ϫ 10 5 cells were seeded into a 6-well plate and incubated in a humidified 5% CO 2 atmosphere at 37°C in antibiotic-free RPMI medium supplement with 10%

Bcl3 induces PD-L1 in ovarian cancer cells
FBS for 24 h to about 80% confluence. For each transfection, 80 pmol of either nonsilencing siRNA control or Bcl3 siRNA were used. Cells were transfected 7 h in transfection medium with siRNA transfection reagent according to the manufacturer's instructions (Santa Cruz Biotechnology). After transfection, fresh medium with antibiotics was added, and cells were grown for 24 h before treatment. Bcl3 CRISPR/Cas9 knockout (KO) plasmid (sc-400740), control CRISPR/Cas9 plasmid (sc-418922), Bcl-3 CRISPR activation plasmid (sc-400740-ACT), and control CRISPR activation plasmid (sc-437275) were obtained from Santa Cruz Biotechnology. Prior to transfections, 2 ϫ 10 5 cells were seeded into a 6-well plate and incubated in antibiotic-free RPMI medium supplement with 10% FBS for 24 h to 80% confluence. For each transfection, 3 g of Bcl3 CRISPR/Cas9 KO or activation plasmids, or the corresponding control plasmids were used. Cells were transfected 24 h in plasmid transfection medium according to the manufacturer's instructions (Santa Cruz Biotechnology). After transfection, fresh medium with antibiotics was added, and cells were grown for 24 h before treatment.
For stable transfection, Bcl3 shRNA (sc-29789-SH) and control shRNA (sc-108060) plasmids were obtained from Santa Cruz Biotechnology. For each transfection, 2 g of Bcl3 shRNA or control shRNA plasmid were used, and cells were transfected using shRNA plasmid transfection medium (sc-108062) and transfection reagent (sc-108061) according to the manufacturer's instructions (Santa Cruz Biotechnology). Transfected colonies were selected using 3 g/ml of puromycin.

Apoptosis, cell proliferation, and PD-L1 neutralization assays
Apoptosis was evaluated using a cell death detection ELISA kit that quantifies release of nucleosomes into the cytoplasm (Cell Death Detection ELISAPLUS, Roche Applied Science) (48), and by measuring caspase 3 activity using a human active caspase 3 ELISA kit (ab181418, Abcam, Cambridge, MA).
Cell proliferation was measured by CellTiter 96 One Solution Cell Proliferation Assay (Promega, Madison, WI). Transfected cells were seeded into 96-well plates at a density of 5000 cells/ 100 l of medium, and incubated at 37°C. At the indicated time points, 20 l of CellTiter 96 One Solution Reagent was added to each well, incubated for 4 h at 37°C, and absorbance at 490 nm was measured.

Wound healing assay
SKOV3 cells were seeded in 6-well plates (2 ϫ 10 5 cells/well) and transfected with control or Bcl3 siRNA as described above. Once the cells became confluent, a wound area was created by scraping the cell monolayer with a sterile 200-l pipette tip. After washing twice with PBS, RPMI medium without FBS was added to the wells. The scratch area was monitored under a phase-contrast microscope at 0, 24, and 48 h after transfection. The wound width was measured in five random fields using ImageJ software. All samples were tested in triplicates.

Real-time RT-PCR
Total RNA was isolated using RNeasy mini-kit (Qiagen, Valencia, CA). The iScript one-step RT-PCR kit with SYBR Green (Bio-Rad) was used as a Supermix and 20 ng/l of RNA was used as template on a Bio-Rad MyIQ Single Color Real-Time PCR Detection System (Bio-Rad). The primers used for quantification of human Bcl3, PD-L1, cIAP1, BclxL, TGF␤1, IB␣, p65, and actin mRNA were purchased from SA Biosciences (Frederick, MD). The mRNA values are expressed as a percentage of control or untreated samples, which were arbitrarily set as 100%.

Chromatin immunoprecipitation (ChIP)
ChIP analysis was performed as described (67). Briefly, proteins and DNA were cross-linked by formaldehyde, and cells

Bcl3 induces PD-L1 in ovarian cancer cells
were washed and sonicated. The lysates were centrifuged (15,000 ϫ g, 10 min, 4°C), and the supernatant extracts were diluted with ChIP dilution buffer and pre-cleared with Protein A/G-agarose (Santa Cruz Biotechnology) for 2 h at 4°C. Immunoprecipitations were performed overnight at 4°C, using p65 (MAB3026; Sigma), Lys-314/315 acetylated p65 (HW136; Signalway Antibody, College Park, MD), CBP (sc-7300; Santa Cruz Biotechnology), p300 (sc-585; Santa Cruz Biotechnology), Bcl3 (23959 -1-AP; Proteintech), and control IgG (sc-2025) antibodies that were pre-incubated (6 h, 4°C) with Protein A/G-agarose, and the immune complexes were collected by centrifugation (150 ϫ g, 5 min, 4°C), washed, and extracted with 1% SDS, 0.1 M NaHCO 3 . After reversing the cross-linking, proteins were digested with proteinase K, and the samples were extracted with phenol/chloroform, followed by precipitation with ethanol. Immunoprecipitated DNA was analyzed by real-time PCR (25 l reaction mixture) using the iQ SYBR Green Supermix and the Bio-Rad MyIQ Single Color Real-Time PCR Detection System (Bio-Rad). Each immunoprecipitation was performed at least three times using different chromatin samples, and the occupancy was calculated by using the human IGX1A negative control primers (SA Biosciences, Frederick, MD), which detect specific genomic ORF-free DNA sequence that does not contain a binding site for any known transcription factors. The results were calculated as fold-difference in occupancy of the particular protein at the particular locus compared with the IGX1A locus.

Statistical analysis
The results represent at least three independent experiments. Numerical results are presented as mean Ϯ S.E. Data were analyzed by using InStat software package (GraphPad, San Diego, CA). Statistical significance was evaluated by using Mann-Whitney U test, and p Ͻ 0.05 was considered significant. Levels of significance are indicated as *, p Ͻ 0.05; **, p Ͻ 0.01; and ***, p Ͻ 0.001.