The protocol for study of human blood was approved by the Institutional Review Board (approval number: 17-0075). PBMCs were purified from leukopack (NY Blood center) as described previously (14). To prepare MO-DCs, CD14+ monocytes were isolated from MO-DCs by EasySep Human CD14 positive selection kit II (StemCell Technologies) according to the manufacturer's protocol. CD14+ monocytes were cultured with RPMI1640 supplemented with 10% heat-inactivated fetal bovine serum (FBS), 1% penicillin-streptomycin (P/S), 1% L-glutamine, 100 ng/ml of recombinant human granulocyte-macrophage colony-stimulating factor (GM-CSF) (PeproTech), and 50 ng/ml of recombinant human IL-4 (PeproTech) for 7 days. Cultures were kept at 37°C in a humidified atmosphere and 5% CO₂. On day 7, MO-DCs were collected only from the non-adherent cells and the purity of MO-DCs was confirmed by flow cytometry with antibodies which were purchased from eBioscience (anti-HLA-DR-FITC: LN3 and anti-CD209-PE/Cy7: eB-h209) (23). Over 85% of HLA-DR+CD209+ MO-DCs were obtained consistently. We excluded adherent cells since cells shows mixed population with CD209+ and CD209- with various degrees (Figure S1A).

The HEK-293 cell line was purchased from ATCC (ATCC CRL-1573) and maintained in DMEM with 10% FBS, 1% P/S, and 1% L-glutamine. The human myeloma cell lines U-266, RPMI-8266 and Daudi (a gift from Dr. Chiorazzi, FIMR, NY) were maintained in a 5% CO₂ atmosphere in RPMI 1640 supplemented with 15% FBS, 1% P/S, and 1% L-glutamine.

Co-IP was performed as described previously (24). Briefly, 2–5 μg of PRDM1 rabbit mAb (Cat# OAR03181, Aviva Systems Biology or cat# 9115s, Cell Signaling Technology) or normal rabbit immunoglobulin G (IgG) (Cat# 2729, Cell Signaling Technologies) were coupled to protein G or A-magnetic beads (DynaBead, Thermo Scientific). Anti-flag M2 magnetic beads (Milipore) were used to pull-down flag tagged RPDM1 in some experiments. Nuclear protein was extracted from PRDM1 transfected MO-DCs with a NE-PER nuclear and cytoplasmic extraction reagents kit (Thermo Scientific) and incubated with a bead-conjugated PRDM1 antibody or control IgG overnight at 4°C. The beads were washed and proteins bound by antibody were eluted by elution buffer and stored at −80°C until used for either immunoblotting or mass spectrometry. Mass spectrometry was performed, and analyses were done at cold spring harbor laboratory shared resources as described previously (25).

Western blot was performed as described (24). Cell extracts or eluted proteins were separated by 4–12% Bis-Tris polyacrylamide gel electrophoresis (PAGE) (Invitrogen). Proteins were transferred to polyvinyliden difluoride (PVDF) membrane (GE Amersham, Hybond-C or Millipore, Immobilon-FL) and blocked for 1 h at room temperature with 5% non-fat dry milk in TBS-T buffer (20 mM Tris, 150 mM NaCl, 0.1% Tween20, pH 7.4). The membranes were then incubated with primary antibodies to HDAC1 (Cat# 5356s, Cell Signaling Technology), HDAC2 (Cat# 5113s, Cell signaling Technology), PRDM1 (Cat# 9115s), hnRNPM (Cat# SAB1404107, Sigma Aldrich), TP53BP1 (Cat# 4937s, Cell Signaling Technology), β-Actin (Cat# ab8226, Abcam), V5 (Cat# MA5-15253, Sigma Aldrich), and Flag (Cat# F1804, Sigma Aldrich) overnight at 4°C. Proteins bound by antibody were visualized by ECL (Thermo Scientific, # 34580 or Advansta, K-12045) and sapphire biomolecular imager (Azure Biosystems).

Human PRDM1 Tagged ORF Clone PRDM1 (RC217363L1V) and human small interfering Ribonuclic acid (siRNA) oligo duplex exogenous (SR300437; PRDM1 and SR321120; NonO) were purchased from Origene. FLAG-NonO (pCMV-myc-Flag-p54), FLAG-TP53BP1 (pcDNA5-FRT/T0-Flag-53BP1), and V5-hnRNPM (pT7-V5-SBP-C1-HshnRNPM) expressing plasmids were purchased from Addgene. Transfections were prepared as described in previous study (24). For transient transfection to HEK-293 cells, 1–2 μg of plasmid was transfected to 70% confluent monolayered HEK-293 cells by Lipofectamin (Invitrogen). After 24 h, medium was replaced with complete medium and cells were further cultured for 2 days. 200 nM siRNA or 1–2 μg of plasmid was transfected to 10⁶ MO-DCs at day 5 during differentiation by Human Dendritic Cell Nucleofector™ Kits (Lonza). After transfection, MO-DCs were further differentiated for 2 days and cells were harvested for experiments. 10⁶ myeloma cells were transfected by using Nucleofector™ Kits (R kit for U-266, T kit for RPMI-8226, and V kit for Daudi) with 400 nM siRNA (Origene), according to manufacturer's instructions.

The in situ PLA was performed on fixed MO-DCs with Duolink in situ Detection Reagents Red (Sigma Aldrich) according to the manufacturer's instructions. Briefly, cells were fixed with 4% paraformaldehyde (PFA) at room temperature and washed with PBS. Samples were permeabilized with 0.5% Triton-X-100 in PBS and blocked by blocking solution (provided by the kit) for 1 h at 37°C. Primary antibodies against NonO (Cat# sc-376865, Santa Cruz), hnRNPM, TP53BP1, PRDM1, V5, Flag or normal rabbit IgG (Cat# 2729, Cell Signaling Technology) were incubated overnight at 4°C. The samples were washed twice for 5 min with buffer A (provided by the kit), followed by incubation with the PLA probes (Sigma Aldrich) for 60 min at 37°C. Subsequent ligation for 30 min at 37°C and amplification for 100 min at 37°C were performed. Finally, the samples were mounted using Duolink in situ Mounting Medium with DAPI (Sigma Aldrich). Z-Stacks Images were captured using a 60X oil objective on Zeiss Apotome 2 microscope and LSM 880 confocal microscopy (Carl Zeiss Microscopy). Three-dimensional foci counting analysis was performed with Imaris software (Imaris v8.0.2).

ChIP assays were performed as previously described (14). 5 × 10⁶ MO-DCs were cross-linked with 1% formaldehyde for 10 min at room temperature and quenched with 125 mM glycine. Cells were washed with ice-cold 1X DPBS twice. Cell pellets were lysed in 300 μl ChIP Lysis Buffer I (50 mM HEPES.KOH, pH 7.5, 140 mM NaCl, 1 mM EDTA, pH 8.0, 10% Glycerol, 0.5% NP-40, 0.25% Triton X-100), ChIP Lysis Buffer II (10 mM Tris-HCl, pH 8.0, 200 mM NaCl, 1 mM EDTA, pH 8.0, 0.5 mM EGTA, pH 8.0), then ChIP Lysis Buffer III (10 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 0.1% sodium deoxycholate, 0.5% N-lauroylsarcosine). All three lysis buffers were supplemented with complete proteinase inhibitor (Roche), and each lysis was performed for 10 min at 4°C with rotation. After lysis, chromatin was sheared by sonication (7 cycles of 30 s ON and 60 s OFF by Q500 sonicator) (Fisher), which generated fragments ranging from 200 to 800 bp. Ten percent Triton X-100 was added to sonicated chromatin (nuclear membrane and lipids were removed by centrifuge). Ten percent of sonicated chromatin supernatant was saved as input control. Sonicated chromatin was incubated with 2 μg of antibody-Protein G and A Dynabeads (Invitrogen) complex overnight at 4°C. Unbound chromatin was removed with RIPA Buffer (50 mM HEPES.KOH, pH 7.5, 100 mM LiCl, 1 mM EDTA, 1% NP-40, 0.7% Sodium Deoxycholate), followed by one time washing with 10 mM pH 8.0 Tris elution buffer. Chromatin elution was done by incubation with elution buffer (50 mM Tris-HCl, pH 8.0, 10 mM EDTA 1% SDS) at 70°C for 10 min. DNA and chromatin de-crosslinking was done by incubation at 65°C for overnight in elution buffer. DNA elute was cleaned by PCR purification kit (Qiagen) and kept at −20°C until PCR or library prep for sequencing. To detect binding to IL6 promoter regions, primer sets that detect each PRDM1 consensus sequence was used for PCR. The PCR condition was as followed: 94°C for 5 min; 94°C for 30 s, 60°C for 30 s, and 72°C for 1 min for 40 cycles.

Primers to amplify the IL-6 area (forward, 5'-CGATATAGCCGAGCTGGAAG-3'; reverse, 5'- AAACCAGACCCTTGCACAAC-3') yield 932-bp amplicon. The PCR condition was as followed: 94°C for 5 min; 94°C for 30 s, 60°C for 30 s, and 72°C for 1 min 15 s for 30 cycles. IL6 PCR product was cloned in pGL4.25 (Promega). 2 × 10⁴ HEK-293 cells were plated in 12-well culture plates in DMEM containing 10% FBS and 1% penicillin-streptomycin. IL6 promoter luciferase reporter construct and tk-Renilla luciferase construct was transfected by Lipofectamine 2000 (Invitrogen). At 48 h post-transfection, transfected cells were lysed and assayed for both firefly and Renilla luciferase activity using the Dual-GLO Luciferase Assay System (Promega). Luciferase activity was measured using a luminometer (Perkin Elmer Victor3). The relative luciferase activity was calculated by normalization to the level of Renilla luciferase.

MO-DCs were differentiated and indicated siRNAs were transfected at day 5 during differentiation. Cells were further cultured for 2 days. MO-DCs were stimulated with lipopolysaccharides (LPS) (1.0 μg/ml) for 6 h and washed. Naïve CD4+ T cells were isolated from PBMCs by using an Easysep human naïve CD4+ T cell isolation kit (Stemcell technologies). 1.3 × 10³ of LPS pre-stimulated MO-DCs or unstimulated MO-DCs were co-cultured with 4 × 10⁴ naïve CD4+ T cells for 6 days. NC (negative control) was naïve CD4+ T cell alone and PC (positive control) was naïve CD4+ T cells with T_FH differentiation cocktails [CD2/3/28 activation beads (Miltenyl Biotec), IL-6 (50 ng/ml, R&D systems) and IL-12 (20 ng/ml, R&D systems)]. T_FH cells were analyzed by flow cytometry with a Fortessa (BD Biosciences). Fixable Viability Dye eFluor 506 (FVD, eBioscience) was used to exclude dead cells. For flow cytometry, antibodies were purchased from BioLegend (anti-CXCR5-APC: J252D4), eBioscience (anti-PD1-pacific blue: EH12.2H7 and anti-CD11c-Amcyan: B-ly6), and BD Bioscience (anti-IL-21-PE: 3A3-N2 and IFN-γ-APC/Cy7: B27).

Total RNA was extracted with Direct-zol RNA Micro Prep (Zymo Research,CA) and RNA samples were treated with DNase I to remove genomic Deoxyribonuclic acid (gDNA) contamination. cDNA was prepared with the iScript cDNA synthesis kit (Bio-Rad). Gene-specific primers were purchased from Taqman (Life Technologies) and qRT-PCR was performed with a Light cycler 480 II (Roche). Taqman primers: Hs00153368_m1 (BCL6), Hs00172187_m1 (POLR2A), Hs99999902_m1 (RPLP0), Hs00174131_m1 (IL6), Hs00153357_m1 (PRDM1), Hs00939763_g1 (NonO), Hs00175407_m1 (CTSS). Relative expression of a gene of interest to housekeeping gene was calculated by ΔCt or ΔΔCt.

To measure the cytokine secretion, supernatants from the MO-DCs were collected and the level of IL-6 was measured by human IL-6 ELISA kit (DuoSet ELISA kit, R&D Systems, Minnesota, USA). The lower level of detection for the assay was 4.68 pg/ml.

Statistical analysis was calculated and determined by a non-parametric Man-Whitney test in the Prism 6 (Graphpad software). P < 0.05 were considered significant.

PRDM1 is known to regulate gene expression by recruitment of chromatin modifiers, including histone deacetylases (HDACs), lysine-specific demethylase1 (LSD1), protein arginine methyltransferase (PRMT5), and euchromatic histone-lysine N-methyltransferase 2 (EHMT2, also known as G9a) in PCs and primordial germ cells (16–19). PRDM1 also recruits polycomb repressive complex 2 (PRC2) by directly binding the enhancer of zeste homolog 2 (Ezh2) domain of PRC2 in murine plasmablasts (26). We investigated whether PRDM1 recruits the same chromatin modifiers in MO-DCs. Binding of HDAC1, HDAC 2, PRMT5 or G9A to PRDM1 was assessed by Co-IP; however, no significant binding of any of those molecules to PRDM1 in MO-DCs was found (Figure S1B and data not shown).

To identify binding proteins of PRDM1 in a non-biased way, relative and absolute quantitation (iTRAQ) MS was performed on the nuclear fraction of MO-DCs immunoprecipitated by PRDM1 antibody. Compared to the fraction immunoprecipitated by control IgG, 39 proteins were pulled-down specifically by the anti-PRDM1 antibody (cutoff >1.5-fold) (Table S1). Consistent with the Co-IP data, there were no HDACs or other known chromatin modifiers identified by mass spectrometry. Thus, PRDM1 does not recruit detectable chromatin modifiers for regulation of target gene expression in MO-DCs.

Among the PRDM1-associated proteins identified by MS, we chose three molecules, NonO, Tumor Protein P53 Binding Protein 1(TP53BP1), and Heterogeneous Nuclear Ribonucleoprotein M (hnRNPM), as candidate co-regulators of PRDM1 due to their known transcriptional regulatory functions. To verify the interaction between those three proteins and PRDM1, Co-IP was performed in HEK-293 cells. Since HEK-293 cells do not express PRDM1 endogenously, they were transiently transfected with vectors encoding PRDM1 and Flag-NonO, Flag-TP53BP1, or V5-hnRNPM. As expected, PRDM1 were not detected in input of non-transfected HEK-293 (Figure 1A, lane 1) and immunoprecipitated proteins from HEK-293 cell nuclear extract (PRDM1-negative) did not display any of the three proteins (NonO, TP53BP1, and hnRNPM) by Western blot (Figure 1A, lane 6). In contrast, anti-FLAG and anti-V5 immunoblotting, which detect Flag-NonO, Flag-TP53BP1, and V5-hnRNPM, showed an association between PRDM1 and NonO and hnRNPM in PRDM1-transfected cells (Figure 1A, lane 8 and 9) but no interaction between PRDM1 and TP53BP1 (Figure 1A, lane10).

PLA was used to verify these interactions in HEK-293 cells. PLA is an antibody-based detection technique that permits the assessment of colocalization between two proteins within < ~40 nm distance in a cell (27, 28). PLA complexes are depicted as red puncta and each punctum represents an interaction between PRDM1 and a candidate molecule. PLA-positive red clusters were not detected in the technical control, which included incubation with only anti-PRDM1 antibody (Figure 1B, top panel). All three candidates, NonO, TP53BP1, and hnRNPM led to PLA positive clusters with PRDM1 (Figure 1B). The signals from hnRNPM and NonO were predominantly nuclear while signal from TP53BP1 was detected in the cytoplasm, suggesting that the interaction of hnRNPM and NonO with PRDM1 may be involved in regulation of gene expression while an interaction of TP53BP1 and PRDM1 may regulate pathways in the cytoplasmic compartment. This observation explains the lack of association of TP53BP1 and PRDM1 in the Co-IP of nuclear extracts.

We further validated these results in MO-DCs, in which we did not need to overexpress PRDM1. No PLA signals were detected in MO-DCs with any single primary antibodies and normal IgG (Figure 2A, left panel). As expected, PLA signals were detected between PRDM1 and NonO in the nucleus of MO-DCs (Figure 2A, the bottom of right panel). There was no significant signal detected with hnRNPM and TP53BP1 in either the nucleus or the cytoplasm (Figure 2A, the top and middle of right panel). Quantitative analysis of the PLA signal between PRDM1-candidate proteins and negative control-IgG confirmed specific PLA signals between PRDM1 and NonO (Figure 2B, right graph).

Additionally, PRDM1 in nuclear extracts coprecipitated with NonO but not with hnRNPM and TP53BP1 (Figure S2). These inconsistent results obtained from primary MO-DCs and HEK-293 cells are likely due to the overexpression of PRDM1 in HEK293 cells. The data obtained from MS, Co-IP and PLA confirmed that NonO is a novel PRDM1 binding protein in the nucleus of MO-DCs.

Knowing that PRDM1 and NonO interact in the nucleus of MO-DCs, we further investigated whether NonO participates in the transcriptional function of PRDM1. Previous data showed that the level of the proinflammatory cytokine IL-6 was negatively regulated by PRDM1 in DCs in response to LPS stimulation (11, 14, 29). If NonO is required for PRDM1-mediated suppression of target gene expression, NonO-deficiency could lead to an increase in the level of IL-6 after LPS stimulation. We confirmed the binding of PRDM1 and NonO in MO-DCs after LPS stimulation (Figure 3A). Next, NonO or PRDM1 targeting siRNA or scrambled control siRNA was transfected into MO-DCs and the level of IL-6 was measured. Effective knockdown of NonO, PRDM1 or both was achieved; about 50% of either NonO or PRDM1 mRNA was present in NonO, PRDM1 or both NonO and PRDM1-siRNA compared to control siRNA transfected MO-DCs (Figure 3B). NonO expression was unchanged in PRDM1-deficient MO-DCs and PRDM1 expression was unchanged in NonO-deficient MO-DCs. To investigate whether IL-6 expression is regulated by the level of NonO, PRDM1 or both, the level of IL-6 was measured in the basal state and at 6 h after LPS stimulation. The basal level of IL-6 mRNA and IL-6 protein in the supernatant were minimal and no change was detected with knock down of NonO, PRDM1, or both (Figure S3). In contrast, following LPS stimulation, IL-6 induction (both transcripts and protein in the supernatant) was increased in siRNA-transfected MO-DCs with NonO, PRDM1, or both compared to control siRNA-transfected MO-DCs (Figure 3C). There was no synergistic effect observed in double-knock down MO-DCs, suggesting PRDM1 and NonO are in a same regulatory pathway. These data demonstrate that NonO-deficiency and PRDM1-deficiency lead to the up-regulation of IL-6 in LPS stimulated MO-DCs.

Previous reports demonstrated that NonO can regulate gene expression by binding to promoter regions (transcriptional regulation) or by binding to mRNA (post-transcriptional regulation) (30–32). Therefore, we investigated the mechanism of NonO-mediated IL-6 expression in MO-DCs. First, binding of NonO to PRDM1-binding sites in the IL6 promoter area was investigated. A search for the PRDM1 binding motif in the IL6 promoter area revealed eight potential PRDM1 binding sites which contain the consensus sequence (A/C)AG(T/C)GAAAG(T/C)(G/T) or (A/C)AG(T/C)GAAAT(T/C)(G/T) within 2,000 bp upstream from transcription start site (TSS) (33) (Figure 3D, #1–#8). We first performed ChIP-PCR using anti-NonO antibody or control antibody and the binding of NonO to PRDM1 binding sites in the IL6 promoter region was assessed. The ChIP efficiency was optimized by detection of P4Hα1, a known target gene of NonO in DNA precipitated with anti-NonO antibody compared to control IgG (data not shown) (32). We performed PCR analysis with primer sets at multiple sites throughout the IL-6 gene; region #5 (−1,247~–1,378 bp TSS) was significantly enriched in DNA precipitated with anti-NonO antibody (Figure 3E). Thus, the #5 region is recognized by NonO. The percent of input (%IP) was calculated from the quantitative PCR (qPCR) and 2–5 fold more enrichment was observed with anti-NonO antibody compared to control IgG (Figure 3F). To confirm whether PRDM1 can bind to the same recognition sequence as NONO, we performed ChIP with anti PRDM1 antibody. We could detect enrichment of PRDM1-binding to IL-6 promoter at region (#5), but the difference between control IgG and PRDM1 was not significant (Figure S4). Taken together, NonO-PRDM1 complexes are recruited to IL6 promoter region to suppress IL-6 expression in MO-DCs.

To further investigate whether NonO can regulate the transcription of IL6, a luciferase reporter assay was performed. Since ChIP-PCR results showed that NonO/PRDM1 binding was enriched in the proximal region [−1.3~–2.2 kb] of the IL6 gene promoter, we engineered an IL6 promoter-Luc plasmid (pGL4.25) containing NonO/PRDM1-binding region of human genomic DNA. To modulate the level of PRDM1 and NonO in the HEK-293 cell line, HEK-293 cells were transfected with a PRDM1 expressing plasmid with control siRNA or with NonO siRNA. NonO siRNA led to a 30–60% decrease in NonO protein levels compared to levels in control siRNA transfected HEK-293 cells (Figure 4A). As expected, PRDM1 suppressed IL6 promoter activity; this suppressive effect was abrogated by a decrease in NonO (Figure 4B). These results demonstrate that NonO functions to enable transcriptional repressor of the IL6 gene by PRDM1. There is no significant change in IL6 promoter activity by NonO deficiency without PRDM1 expression; thus, the regulatory mechanism of NonO depends on the PRDM1 expression level.

IL-6 production is one of key factors for murine follicular helper T cell (T_FH) differentiation, and an increased production of IL-6 in DCs leads to an expansion of T_FH in Prdm1 CKO mice (11, 14, 29). To address the function of PRDM1-deficiency and NonO-deficiency in DCs on the differentiation of CD4⁺ T_FH cells, unstimulated or LPS pre-stimulated MO-DCs were co-cultured with naïve CD4+ T cells. After co-culture, surface phenotype and cytokine production by T cells were investigated by flow cytometry (Figure 5A). After co-culture, live CD4+ T cells were identified by exclusion of FVD-positive (dead cells) and CD11c-positive (MO-DCs). There was no difference in the expansion of T cells, and CXCR5-positive T cells were not strongly induced in any culture condition (data not shown). Interestingly, LPS-stimulated PRDM1-deficient or NonO-deficient MO-DCs induced a higher percent of IL-21+CXCR5-PD1+ cells compared to control siRNA-treated MO-DCs (Figure 5B). NonO-deficient MO-DCs also induced higher percent of IL-21+CXCR5-PD1+ T cells even in the absence of LPS stimulation, but this effect was not observed in PRDM1- or double deficient MO-DCs (Figure 5B). We do not know the mechanism, but NonO may regulate other regulatory molecules which positively regulate T cell differentiation or survival. We also compared IFN-γ production in T cells, but none of T cells were IFN-γ-positive (Figure 5B). To confirm the IL-21-producing T cells are T_FH cells, we measured the induction of B-Cell Lymphoma 6 (BCL6) which is a master transcription factor of Tfh cells (34). The BCL6 induction was small and not significantly higher than negative control group, and there was no difference of BCL6 levels among the groups (Figure 5C). Therefore, NonO, PRDM1, or PRDM1/ NonO deficient MO-DCs generate more IL-21 producing T_FH-like cells_. No synergistic effects of PRDM1 and NonO were observed in IL-21 producing T_FH-like cell differentiation. Increased production of IL-6 might contribute to this alteration.

Since plasma cells express a high level of PRDM1 and secrete IL-6, we wanted to know whether PRDM1 and NonO regulate IL-6 in the human myeloma B cell lines, U-266, and RPMI-8226 since both cell lines express a high level of PRDM1 and NonO (Figure S5A). Daudi, a non-myeloma B cell line with a high level of NonO but not PRDM1, was used as a negative control. Using the PLA assay, we found a NonO and PRDM1 interaction in U-266 and RPMI-8226 in nucleus but not in Daudi (Figures 6A,B). Next, we tested whether NonO participates in PRDM1-mediated IL-6 production in these cells. NonO knockdown mediated by siRNA was sufficient to decrease the NonO expression level in both U-266 and RPMI-8226 cells (Figure S5B). In contrast to MO-DCs, there is high level of endogenous IL-6 expression in myeloma cells which was not further increased by stimulation with LPS, and no significant induction of IL-6 when NonO levels were decreased (Figure 6C). Similarly, PRDM1-deficiency did not increase the expression level of basal or LPS stimulated IL-6 mRNAs in U-266 (Figure S5B). Hence, in myeloma cells, NonO could be recruited to a PRDM1 complex but no PRDM1-mediated regulatory effects on IL-6 expression by NonO-deficiency and PRDM1-deficiency were observed.