Introduction
Diffuse large B-cell lymphoma (DLBCL) accounts for approximately 30% of B-cell lymphoma cases [
1]. Molecular profiling of DLBCL cell lines and patient tumors has led to the identification of distinct subtypes, which has been a useful tool in predicting patient survival and therapeutic response [
2]. Genome-wide studies have shown that approximately 30% of DLBCL tumors harbor mutations in two highly related histone acetyltransferase (HAT) genes,
EP300 and
CREBBP[
3‐
7].
EP300 and
CREBBP encode related HATs, p300 and CBP, respectively, that have widespread genomic effects on chromatin structure and gene expression as well as non-genomic effects on protein function [
8]. These HATs serve as coactivators for many transcription factors, either through acetylation of lysine residues on histones to modify DNA structure at sites of active transcription or through acetylation of transcription factors to modify their activity. In both cases, the centrally-located, catalytic HAT domain is required for these effects on transcription. Consistent with its broad role in transcriptional control, p300 can directly interact with a wide variety of transcription factors, including NF-κB [
9,
10], p53 [
11,
12], MyoD [
13], HIF-1α [
14], BRCA1 [
15], and Ets-1 [
16]. In addition, p300 and CBP contain several protein-protein interaction domains and can exhibit HAT-independent functions; for example, p300 can enhance transcription simply by recruiting proteins to transcriptional start sites, including members of the transcription pre-initiation complex and the RNA polymerase holo-enzyme [
8,
17].
Most p300/CBP mutations identified in DLBCL are point mutations, nonsense mutations, or deletions that disable HAT activity [
3,
5,
10,
18]. In some epithelial cancers where a truncated p300/CBP protein is expressed, the wild-type allele is silenced or otherwise inactivated [
19], and ectopic expression of wild-type p300 in some HAT-deficient p300 cancer cell lines slows cell growth [
7,
20]. Such results have led p300 to be classified as a tumor suppressor, arising from the hypothesis that it is the loss of wild-type p300 activity which contributes to oncogenesis.
We have previously shown that, due to a 3’ alteration in one copy of the
EP300 gene, the DLBCL cell line RC-K8 expresses a C-terminally truncated HAT-deficient p300 protein (herein called p300ΔC-1087). Even though the other copy of the
EP300 locus appears intact, RC-K8 cells express low to undetectable levels of wild-type p300 mRNA and protein [
10,
18]. We previously reported that the RC-K8 p300ΔC-1087 could not act as a coactivator for the REL transcription factor [
18]. Of note, knockdown of p300ΔC-1087 expression reduces the proliferation and soft agar colony-forming ability of RC-K8 cells [
18], and re-expression of wild-type p300 is tolerated in RC-K8 cells, but sensitizes them to the cell killing effects of small-molecule BCL6 inhibitors [
7]. Other studies have demonstrated that expression of a HAT domain mutant of p300 results in increased proliferation of hematopoietic stem and progenitor cells, whereas complete loss of p300 does not [
21]. Such findings suggest that p300 HAT activity normally limits B-cell proliferation, and that expression of p300 proteins with an inactive catalytic domain contributes to B-cell growth, survival, and tumorigenesis.
In this report, we have characterized a truncated p300 protein expressed in the DLBCL cell line SUDHL2. We show that this C-terminally truncated and HAT-deficient p300 mutant is a weak transcriptional coactivator, and that its expression is required for the optimal growth of SUDHL2 cells. These results and others suggest that expression of C-terminally truncated p300 coactivators defines a subset of DLBCL that utilize distinct oncogenic pathways.
Discussion
In this report, we have characterized molecular properties of the HAT-deficient p300ΔC-820 protein from the human DLBCL cell line SUDHL2. This is only the second truncated p300 mutant that has been functionally characterized in a human DLBCL cell line [
10,
18]. We show that p300ΔC-820 is the only form of p300 protein expressed in SUDHL2 cells and that p300ΔC-820 contributes to SUDHL2 cell growth, as knockdown of p300ΔC-820 expression compromised the liquid media and soft agar growth of SUDHL2 cells. Like wild-type p300, p300ΔC-820 localizes to the nucleus and can interact with NF-κB family member REL, but p300ΔC-820 has a reduced ability to enhance REL-dependent transactivation in reporter assays. As such, p300ΔC mutants have the potential to attenuate expression of transcription factor-specific target genes by preventing the interaction of transcription factors with other functionally intact coactivators. Indeed, knockdown of p300ΔC-1087 in RC-K8 cells resulted in increased expression of NF-κB target genes
A20,
CCR7,
NFKBIA,
TRAF1 and
TNFα, as well as an increase in A20 and IκBα protein expression. Finally, the RC-K8 and SUDHL2 cell lines, which have reduced expression of wild-type p300, had generally reduced levels of acetylation of histone H3 K14 and K18 among a panel of B-lymphoma cell lines.
Like wild-type p300 and the p300ΔC-1087 protein from RC-K8 cells, p300ΔC-820 showed a punctate pattern of nuclear staining by immunofluorescence, which has been associated with sites of active transcription for wild-type p300 [
29]. Using reporter assays, p300ΔC-820 and p300ΔC-1087 are both weak transcriptional coactivators for REL in A293 cells (Figures
2c,d). Because p300 acts as a transcriptional coactivator through both HAT-dependent and HAT-independent mechanisms, the limited coactivator activity retained by these two C-terminally truncated p300 mutants is likely a function of protein-protein interactions that result in recruitment of transcriptional machinery to the transcription start site (i.e., HAT-independent activities). In some promoter contexts, such HAT-independent activities may suffice to maintain normal p300 function. For example, it has been shown that HAT deletion mutants of p300 can still enhance MyoD-dependent transcription, possibly by stabilizing a ternary complex between MyoD and other coactivators [
30]. Thus, it is likely that HAT-deficient p300 proteins have altered p300 activity in some settings (e.g., REL-dependent transactivation), but not in others (e.g., MyoD-dependent transactivation). Therefore, we suggest that cells expressing mutant p300 proteins are distinct from p300-null cells.
A20 is a tumor suppressor and a target gene of NF-κB, is biallelically inactivated in approximately 30% of DLBCL, and is mutated in the SUDHL2 and RC-K8 cell lines [
28,
31,
32]. Knockdown of p300ΔC-1087 resulted in increased expression of A20 in RC-K8 cells (Figures
4d,e). That observation and the presence of p300ΔC-1087 at the
A20 promoter (Figure
4f) suggest that p300ΔC-1087 directly reduces
A20 gene expression in RC-K8 cells, leading to reduced A20 protein. Reduced A20 protein activity appears to be essential for RC-K8 and SUDHL2 survival, as re-expression of wild-type A20 induces apoptosis in both cell types [
28]. Therefore, it appears that A20 activity is reduced in SUDHL2 and RC-K8 cells by both mutation [
28] and transcriptional repression mediated by mutant p300.
Knockdown of p300ΔC-1087 in RC-K8 cells also resulted in increased IκBα expression (Figures
4e,f). We have previously shown that RC-K8 cells have inactivating mutations in three of four copies of the
NFKBIA gene, express little wild-type IκBα protein, and consequently show high levels of both nuclear REL DNA-binding activity and REL target gene expression [
27]. Forced expression of wild-type IκBα protein slows the growth of RC-K8 cells, presumably due to inhibition of REL [
27].
Taken together, these results suggest that C-terminally truncated p300 proteins contribute to the oncogenic state in SUDHL2 and RC-K8 cell lines, at least in part, by reducing expression of both A20 and IκBα, which allows for tolerable and optimal levels of nuclear NF-κB activity that promote cell growth. Indeed, both cell lines belong to the ABC-subtype of DLBCL, which is characterized by constitutive NF-κB activity and sensitivity to NF-κB inhibitors [
33]. Overall, we propose that the high levels of nuclear REL-driven transactivation of target genes that is unleashed by mutations in the REL/NF-κB inhibitors A20 and IκBα in RC-K8 and SUDHL2 cells is tempered by expression of p300ΔC proteins, which act as muted REL coactivators. The model that moderate, chronic increases in REL-driven target gene expression are optimal for B-lymphoid cell transformation is reminiscent of the mutation-driven activation of the lymphoid cell-specific oncoprotein v-Rel, which is a chronic low level activator of target gene expression as compared to c-Rel [
34].
The CH1 domain of p300 is retained in both p300ΔC-1087 and p300ΔC-820, and is required for the interaction of p300 with REL (Figure
3b). Thus, the CH1 domain and interaction with REL may be important for the growth-promoting activity of truncated p300 proteins in DLBCL. In support of this hypothesis, Kimbrel
et al. [
21] used a mouse
in vivo reconstitution system to show that expression of a HAT domain mutant of p300
increased the proliferative potential of hematopoietic stem and progenitors cells, whereas expression of a CH1 domain mutant resulted in severe defects in hematopoiesis.
We have found that DLBCL cell lines with reduced expression of wild-type p300 generally have low levels of H3K14 and H3K18 acetylation (Figure
5, Additional files
2 and
3). It has been shown that p300 and CBP are able to acetylate H3K14 and H3K18
in vitro and that p300 and CBP are required for H3K18 acetylation
in vivo[
35,
36]. Additionally, hypoacetylation of H3K18 by inhibition of p300 and CBP stimulates cell cycling in quiescent human cells and has been associated with recurrence of low-grade prostate cancer in patient studies [
36‐
38]. Developmental studies in mice have shown that acetylation of H3K14 is associated with gene activation [
39], suggesting that its reduction in RC-K8 and SUDHL2 cells prevents expression of target genes specifically related to growth inhibition and/or apoptosis. Consistent with this hypothesis, H3K14 acetylation at the promoter of the cell cycle inhibitor p21 is upregulated 10-fold in response to treatment with the topoisomerase II inhibitor doxorubicin, and is required for stress-induced cell-cycle arrest in human cancer cell lines [
40]. We suggest that expression of truncated p300 and the associated reduction of wild-type p300 is one mechanism that can lead to reduced acetylation of H3K14 and H3K18, which contributes to DLBCL cell growth. Of note, SUDHL2 and RC-K8 cells are sensitive to apoptosis induced by treatment with two HDAC inhibitors [
41].
Our findings contradict a previously published report on the lack of full-length p300 protein in the BJAB, SUDHL8, and Farage DLBCL cell lines [
3]. We were able to show that multiple cell line stocks, including ones used in the conflicting report, express easily detectable levels of full-length p300 protein. Thus, we believe that the lack of full-length p300 protein in these three cell lines reported in Pasqualucci
et al. [
3] was due to a technical error. Northern blotting data reported by Pasqualucci
et al. [
3] further support our results, in that the BJAB, SUDHL8, and Farage cell lines express detectable levels of full-length p300 mRNA, whereas SUDHL2 cell do not. Moreover, full-length p300 protein expression in BJAB and Farage cells has been reported by several others [
7,
22‐
25].
Materials and methods
Plasmids
DNA manipulations were carried out by standard methods [
42]. Complete details of subclones and primers used in this study are described in supplementary information and at
http://www.nf-kb.org. GST-p300-CH1 and GST-p300-N340 expression plasmids were kindly provided by Andrew Kung and have been described previously [
14]. All recombinant DNA and human cell line work was conducted with BSL-2 level approval of the Boston University Institutional Biosafety Committee (approval number 11-072).
Cell culture
A293, A293T, BOSC23 human embryonic kidney cells, DF-1 chicken fibroblasts, and RC-K8 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS) (Biologos, Montgomery, IL, USA) as described previously [
43]. BJAB cells were cultured in DMEM supplemented with 20% FBS. SUDHL2, Ramos, and Farage cells were cultured in RPMI supplemented with 10% FBS. SUDHL8 cells were cultured in RPMI supplemented with 20% FBS. The human B-lymphoma cell lines are classified as follows: DLBCL (BJAB [
3,
16,
44], Farage [
2,
45], RC-K8 [
8], SUDHL2 [
3], SUDHL8 [
3,
19]), and Burkitt’s lymphoma (Ramos).
Transfection of A293, A293T, BOSC23, and DF-1 cells was performed as described previously [
18]. For Western blotting and indirect immunofluorescence, cells were processed 48 h after addition of the transfection mix.
Design of control and
EP300 short hairpin RNAs (shRNA), generation of virus stocks, and infections were performed as described previously [
18]. Two days after infection, SUDHL2 and RC-K8 cells were selected with 1 μg/ml puromycin (Sigma, St. Louis, MO, USA), respectively, for 2-4 weeks and maintained in puromycin throughout all experiments.
Western blotting and indirect immunofluorescence
Whole-cell lysates were prepared in AT buffer containing protease inhibitors as described previously [
46] and were analyzed by Western blotting according to standard methods [
10]. High molecular weight proteins (full-length p300 and CBP) were transferred at 260 mA for 2.5 h at 4°C using a modified large-protein transfer buffer (20 mM Tris, 150 mM glycine, 0.05% SDS, 10% methanol) as described previously [
10]. Western blots were quantified using ImageJ software [
47].
The following antisera were used: rabbit anti-p300 (1:200; anti-N-terminal, sc-584, Santa Cruz Biotechnology, Santa Cruz, CA, USA), rabbit anti-REL (1:200; obtained from Nancy Rice [
43]), mouse anti-CBP (1:200; sc-7300, Santa Cruz Biotechnology), mouse anti-A20 (1:200, 550859, BD Pharmingen, Franklin Lakes, NJ, USA), mouse anti-IκBα (1:1000, 4814, Cell Signaling Technology), and rabbit anti-β-tubulin (1:500; sc-9104, Santa Cruz Biotechnology).
Indirect immunofluorescence was performed as described previously [
18,
43] using anti-p300 (1:50; sc-584, Santa Cruz Biotechnology) primary antibody and fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG (1:80; Sigma) secondary antibody. Nuclei were also stained with 4’,6-diamidino-2-phenylindole (DAPI). Cells were visualized using a fluorescent microscope (Olympus FLUOVIEW Laser Scanner Microscope BX 50, Center Valley, PA, USA).
Co-immunoprecipitation
For co-immunoprecipitation of overexpressed proteins, A293 cells in 100-mm dishes were co-transfected with 10 μg pcDNA-Flag-REL and 10 μg pCMVβ-p300ΔC-820. Two days later, nuclear extracts were prepared using a Nuclear Complex Co-IP Kit (cat no. 54001, Active Motif, Carlsbad, CA, USA) according to the manufacturer’s instructions. Nuclear extracts containing 300 μg of protein were incubated with anti-p300 antiserum or rabbit pre-immune serum for 3 h at 4°C. 100 μl of a 50% slurry of Protein A Sepharose CL-4B (GE Healthcare Life Sciences, Pittsburgh, PA, USA) was added and the sample was incubated for an additional 1 h at 4°C. The beads were then washed in PBS and proteins were eluted by heating at 95°C in SDS sample buffer. Proteins were electrophoresed on a 7.5% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane as described above. One percent of the amount of nuclear extract used for one immunoprecipitation (3 μg) was included on the gel as an input lane. The membranes were then subjected to anti-REL Western blotting.
Co-immunoprecipitation of endogenous REL and p300ΔC-820 in SUDHL2 cells was performed using the Nuclear Complex Co-IP Kit as described by the manufacturer (cat. no. 54001, Active Motif). Three micrograms of normal rabbit IgG (sc-2027, Santa Cruz Biotechnology) or anti-p300 antiserum (sc-584, Santa Cruz Biotechnology) was incubated with 250 μg of nuclear extract in IP Low Buffer (Active Motif) for 3 h at 4°C. 50 μl of a 50% slurry of Protein A Sepharose CL-4B (GE Healthcare Life Sciences) was added, and samples were incubated for an additional 3 h. Beads were then washed with IP Low Buffer and proteins were eluted by heating at 95°C in SDS sample buffer. Proteins were electrophoresed on a 6% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane as described above. Ten percent of the amount of nuclear extract used for one immunoprecipitation (25 μg) was included on the gel as an input lane. The membranes were then subjected to anti-REL or anti-p300 Western blotting.
GST pulldown assays
GST pulldown assays followed by Western blotting were performed as described previously [
10]. One percent of the amount of extract used for each pulldown (30 μg) was included on the gel as an input lane. The membrane was then subjected to anti-p300 or anti-REL Western blotting.
Luciferase reporter assays
Luciferase reporter assays were performed using the Luciferase Assay System (Promega, Madison, WI, USA) as described previously [
18]. A293 cells in 35-mm plates were transfected with 0.5 μg of reporter plasmid pGL2-3×-κB-luciferase and 0.5 μg of normalization plasmid pRSV-βgal. Cells were co-transfected with 0.5 μg of pcDNA-REL or pcDNA3.1 vector alone, along with 0.5 μg of pCMVβ-p300, pCMVβ-p300ΔC, or vector alone. In titration experiments (Figure
2d), cells were transfected with 0.5 μg of pcDNA-REL, and increasing amounts of pCMVβ-p300, pCMVβ-p300ΔC-1087, or pCMVβ-p300ΔC-820. Increasing amounts of each p300 plasmid were titrated in until luciferase activity reached a plateau. For all luciferase reporter assay experiments, total DNA per transfection was kept constant by including varying amounts of pcDNA3.1 vector. Luciferase and β-galactosidase activities were determined, and values were normalized to the relevant vector control (1.0).
Statistical analyses were performed using a paired one-tailed t-test and p < 0.05 was considered significant.
Cell proliferation and soft agar assays
Cell proliferation and soft agar colony assays were performed as described previously [
18,
48]. Equal numbers (2000 and 5000) of SUDHL2 cells expressing the indicated shRNA were placed in soft agar containing RPMI with 20% FBS and 0.3% Bacto Agar (Difco, Franklin Lakes, NJ, USA), and plates were incubated at 37°C in a humid incubator with 5% CO
2. Macroscopic colonies were counted 14 days after plating.
Chromatin immunoprecipitation assays and qPCR
For ChIP assays, approximately 10
8 RC-K8 cells were fixed with 3% formaldehyde for 20 min at room temperature. Cells were then rinsed three times with ice-cold PBS and nuclear lysate was prepared as described previously [
49]. Samples containing 350 μg of protein were then incubated at 4°C overnight with either rabbit anti-p300 antiserum or pre-immune serum. The next day, 50 μl of a 50% slurry of protein A beads was added and the reaction was incubated for 3 h at 4°C. Beads were washed with RIPA buffer (50 mM Tris-HCl pH 7.2, 150 mM NaCl, 1% deoxycholic acid, 1% Triton X-100, 0.1% SDS) and then TE supplemented with 50 mM NaCl. Beads were eluted in TE with 2% SDS for 15 min at 65°C. Crosslink reversal and DNA purification were performed as described previously [
49]. Purified DNA was then subjected to qPCR using primers to amplify the
TNFAIP3 (
A20) promoter. PCRs were performed in the ABI Prism 7900HT Sequence Detection System (Applied Biosystems, Foster City, CA, USA) using 40 cycles of 94°C for 15 s and 60°C for 1 min. The primers used were 5’-CAGCCCGACCCAGAGAGTCAC and 5’-TTCGTGGCGGGCCAAG [
50]. Cycle threshold (C
t) values for p300 were normalized to the pre-immune serum control values (1.0). Error bars represent standard error.
PCR and real-time quantitative PCR
Two hundred ng of genomic DNA from SUDHL2 cells was subjected to PCR using forward and reverse primers specific for sequences surrounding exon 14 of EP300. The primers used were 5’- AGCATAGGCAGGCCCTAGA and 5’- TATGCTTGGGGGAGTATGGT. Sequencing of the amplified fragment was performed by Eurofins MWG Operon (Huntsville, AL, USA).
For qPCR of mRNA, total RNA was first isolated from RC-K8 cells using TRIzol Reagent (Invitrogen, Grand Island, NY, USA) according to the manufacturer’s protocol. The mRNA was reverse transcribed into cDNA using M-MLV reverse transcriptase (Promega) and random primers (Promega). One thirtieth of the synthesized cDNA was combined with gene-specific primers and Power SYBR Green PCR Master Mix (Applied Biosystems). PCRs were performed as described above. C
t values were obtained for each sample and normalized to C
t values for
GAPDH cDNA amplification (ΔC
t) and then to C
t values from control shRNA-expressing RC-K8 cells (ΔΔC
t) using methods described previously [
51]. The fold change in mRNA was normalized to the fold change in
GAPDH mRNA expression (1.0) between p300 and control knockdown RC-K8 cells. Primers used were
A20: 5’- CGCTCAAGGAAACAGACACA and 5’- CTTCAGGGTCACCAAGGGTA;
CCR7: 5’-TGAGGTCACGGACGATTACAT and 5’- GTAGGCCCACGAAACAAATGAT;
NFKBIA: 5’- CTCCGAGACTTTCGAGGAAATAC and 5’- GCCATTGTAGTTGGTAGCCTTCA;
TRAF1: 5’- TCCTGTGGAAGATCACCAATGT and 5’- GCAGGCACAACTTGTAGCC;
TNF: 5’- GAGGCCAAGCCCTGGTATG and 5’- CGGGCCGATTGATCTCAGC;
LTA: 5’- CATCTACTTCGTCTACTCCCAGG and 5’-CCCCGTGGTACATCGAGTG;
A1: 5’- TACAGGCTGGCTCAGGACTAT and 5’- CGCAACATTTTGTAGCACTCTG;
GAPDH: 5’- TGGTATCGTGGAAGGACTCATGAC and 5’- ATGCCAGTGAGCTTCCCGTTCAGC.
Statistical analyses were performed using a paired two-tailed t-test, and p < 0.05 was considered significant.
Quantification of histone acetylation via mass spectrometry
Cell lines were maintained in healthy conditions for several passages before histones were purified using the Active Motif Histone Purification Kit (cat no. 40025) according to the manufacturer’s instructions. Concentrations were determined using Nanodrop, 5 μg of each sample was chemically propionylated using 1.5 μl propionic anhydride, and ammonium hydroxide was used to quickly adjust the pH to approximately 8.0 [
52]. Samples were then incubated at 51°C for 1 h followed by trypsin digestion overnight at 37°C. The fraction of acetylated to unmodified at a given histone H3 site was performed as described previously [
53]. Means and 95% confidence intervals of acetylation values for different cell lines were calculated.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
LH performed all experiments except for full-length p300 detection in multiple DLBCL isolates (JL-P) and mass-spectrometry analysis of histone H3 acetylation (RAH, AJA). LH and TDG designed the studies and analyzed and interpreted the data. LH and TDG wrote the manuscript. All authors read and approved the final manuscript.