NR4A3 Suppresses Lymphomagenesis through Induction of Proapoptotic Genes

NR4A1 (Nur77), NR4A2 (Nurr1), and NR4A3 (NOR-1) belong to the nuclear orphan receptors of the Nur77 family. NR4A1, NR4A2, and NR4A3 are widely expressed in different types of tissues, such as skeletal muscle, adipose tissue, heart, kidney, T cells, liver, and brain. They are immediate early- or stress-response genes and can be induced by a wide range of physiological signals, such as fatty acids, stress, prostaglandins, growth factors, calcium, inflammatory cytokines, peptide hormones, and neurotransmitters. Further, NR4A1 and NR4A3 play a central role in negative selection of T lymphocytes, as well as IgM-mediated and virally induced B-cell apoptosis (1–4). NR4A1 and NR4A3 were identified to function as tumor suppressors in acute myeloid leukemia (AML). Deletion of both nuclear receptors led to rapid development of AML in mice. Loss of NR4A1 and NR4A3 was also found in leukemic blasts from human AML patients, irrespective of karyotype (5). Additionally, NR4A1 and NR4A3 hypoallelic mice (NR4A1^+/− and NR4A3^−/−; NR4A1^−/− and NR4A3^+/−)—mice with a reduced NR4A1 and NR4A3 expression—develop a chronic myeloid malignancy that recapitulates the pathologic features of myelodysplastic/myeloproliferative neoplasm with progression to AML in rare cases (6).

In our previously published study (7), we found a significant reduction of both, NR4A1 and NR4A3, in B-chronic lymphatic leukemia (CLL; 71%), follicular lymphoma (FLIII; 70%), and in diffuse large B-cell lymphoma (DLBCL; 74%) compared with normal controls. Survival analysis revealed that low NR4A1 expression is associated with poor cancer-specific survival. Further, overexpression of NR4A1 caused apoptosis in several lymphoma cell lines. SuDHL4 lymphoma cells stably transduced with an inducible (tet-off) NR4A1 expression construct and subcutaneously injected in NOD/SCID/IL-2rγnull (NSG) mice led to suppressed lymphoma outgrowth after NR4A1 induction by withdrawal of doxycycline (7).

Because it has been reported that the NR4A3 expression levels correlate positively with therapy success in whole-genome expression analysis of 58 DLBCL patients (8), that NR4A3 and NR4A1 are functionally redundant in T-cell apoptosis (9), and that NR4A3 correlates with NR4A1 expression in our cohort of aggressive lymphoma patients (7), we aimed to functionally characterize NR4A3 in aggressive lymphoma cells. Here, we show for the first time that overexpression of NR4A3 in aggressive lymphoma cells causes apoptosis in vitro and leads to reduced tumor growth in a xenograft model. Additionally, the induction of NR4A3 in aggressive lymphoma cells by thapsigargin (TG), a NR4A3 inducing agent (10), and BF175—a pinacolyl boronate-substituted stilbene derivate (11, 12)—leads to immediate lymphoma cell apoptosis. Collectively, these results define NR4A3 as novel gene with tumor-suppressive properties in lymphoid malignancies.

To test the implication of NR4A3 expression levels in overall survival of human lymphoma samples, we measured the mRNA expression levels of NR4A3 in a cohort of 92 histologically confirmed aggressive lymphoma patients. The patients received a rituximab-containing standard treatment regimen at the Division of Haematology, Medical University of Graz between 2000 and 2010 (with last follow-up until May 2016). Patients were categorized to a low- and high-expression group according to a cutoff value generated by receiver operating curve (ROC) analysis. The low- and high-expression groups were related to the endpoint overall survival.

Karpas-422 and SuDHL4 as model for germinal center-(GCB-) DLBCL and RI-1 and U2932 as model for activated B cell-(ABC-) DLBCL were used for functional characterization of NR4A3. SuDHL4, RI-1 and U2932 were maintained in Iscove's Modified Dulbecco's Medium (IMDM, Thermo Fisher Scientific) with 10% fetal bovine serum (FBS, Thermo Fisher Scientific), Karpas-422 was cultured in IMDM with 20% FBS. To the culture media, penicillin (50 U/mL) and streptomycin (50 μg/mL) were added. Cells were periodically checked for mycoplasma by PCR and were found to be negative. All cell lines were treated in a range from 1 × 10⁻¹ to 1 × 10⁻⁷ mol/L with TG (Sigma-Aldrich), a NR4A3-inducing agent (10) and BF175. NR4A1- and NR4A3-overexpressing cell lines were additionally treated with 20 nmol/L leptomycin B to inhibit shuttling of the nuclear proteins into cytoplasm. The identity of the DLBCL cell lines was confirmed by STR analysis using Power Plex 16 System (Promega) and verified at the online service of the DSMZ cell bank (http://www.dsmz.de; Supplementary Table S1).

Sequence verified full-length PCR product—generated from peripheral blood B-cell cDNA using specific PCR primers (Supplementary Table S2)—was subcloned into the pTight Tet-Off viral expression system (Clontech) for lentiviral transduction and into pEZ-M61 (Genecopoeia) for transient transfection.

Details on the generation of lentiviral vectors, transduction, and functional assays of NR4A3 induced lymphoma cells are provided in the Supplementary Methods section.

Thapsigargin- and BF175-treated lymphoma cells and their controls were plated at a density of 10,000/mL and cultured for 72 hours. Seven replicates of the CellTiter 96 AQueous One Solution Cell Proliferation Assay (Promega) were done using 3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium, inner salt (MTS). The absorbance was recorded by a BioRad spectrophotometer at 490 nm.

Total RNA was extracted using TRIzol (Invitrogen) according to the manufacturer's protocol. cDNA was synthesized using the RevertAid H Minus First-Strand cDNA Synthesis Kit (Fermentas).

Real-time semiquantitative PCR (RQ-PCR) for NR4A1, NR4A3, Bim 1,6, Bim 9, Puma, BCL2, BCLX, MCL1, Trail, FasL, DR4, DR5, and Fas (primers are listed in Supplementary Table S2) was performed using an ABI Prism 7000 Detection system (Applied Biosystems). PCR reaction and data analysis were performed as previously described by our group (7, 13, 14). GAPDH, PPIA, and HPRT1—known to exhibit the lowest variability among lymphoid malignancies (15)—served as housekeeping genes.

Cells were transfected by electroporation using AMAXA Kit V (Lonza). The programs used were O-017 for SuDHL4 and X-001 for Karpas-422, RI-1 and U2932, respectively. Briefly, 5 × 10⁶ cells of SuDHL4, Karpas-422, RI-1, and U2932 were transfected with 2.5 μg pEZ-M61 carrying NR4A1 or NR4A3 or empty pEZ-M61 vector in two replicates for each cell line. Cells were seeded at 5 × 10⁵ cells per mL and functional assays were performed after 24, 48, and 72 hours.

Cells (4 × 10⁶) of SuDHL4 and U2932 were transfected with 175 ng siRNA targeting NR4A3 (Qiagen) or scramble controls. Cells were seeded 5 × 10⁵ per mL, after 12 hours treated with 10 μmol/L TG (Sigma-Aldrich) or BF175 and functionally tested.

NSG-mice were purchased from The Jackson Laboratory. Ten male NSG mice received subcutaneous flank injections of 1 × 10⁷ transduced SuDHL4 cells containing the empty pLVX plasmid as vector control on the right flank and the inducible NR4A3 pLVX construct on the left flank resuspended in 200 μL Matrigel (BD). Doxycycline (Clontech) was administered to five animals through their drinking water in a concentration of 200 μg/mL to suppress induction of NR4A3 as controls. Tumor burden was assessed weekly by tentative inspection. At day 20, tumor volume was estimated by ultrasonic testing, and all tumors were harvested for histologic analysis. Tumor volume and histologic stains were compared between the doxycycline-administered mice and mice without doxycycline administration. All animal work was done in accordance with a protocol approved by the Institutional Animal Care and Use Committee at the Medical University of Graz (Graz, Austria).

Statistical analysis was performed by using IBM SPSS Statistics 21.0 (SPSS Inc). The nonparametric Mann–Whitney U test was used to analyze differences in the functional assay and expression analysis. The Spearmen correlation test was performed to examine any correlation of the NR4A1 and NR4A3 mRNA levels to apoptotic and antiapoptotic genes. When the P value was lower than 0.05, a significant value was reached. Overall survival was defined as the time in months from the date of diagnosis to death by any cause. Patients' overall survival was calculated with the Kaplan–Meier method, and differences were tested by the log-rank test.

Because aggressive B-cell lymphomas exhibited the lowest NR4A3 expression levels in our previous study (7), we decided to explore the influence of mRNA expression levels on survival in a cohort of 92 patients diagnosed with DLBCL and FLIII at our institution. A marked (more than 2-fold) downregulation of NR4A3 was detected in the vast majority of DLBCL-GCB, DLBCL-NGCB, and FLIII patients in comparison the germinal center B cells (CD19⁺, CD77⁺ cells) on mRNA levels, respectively (Fig. 1A, P < 0.05 for DLBCL-GCB and DLBCL-NGCB, P < 0.1 for FLIII). To determine whether reduced NR4A3 mRNA expression translates to reduced protein levels, Western blot analyses for NR4A3 was performed in selected lymphoma—and control samples as previously described and a significant positive correlation was observed by comparing densitometrically quantified protein levels to mRNA levels (Spearman ρ = 0.565 for NR4A3, P < 0.01; ref. 7). In addition, a significant association between low NR4A3 expression and poor survival was observed (P = 0.043, log-rank test, Fig. 1B). Patients with high NR4A3 expression reached a plateau in their survival approximately after 48 months.

To investigate whether promoter hypermethylation, deletions, or mutations in the promoter (1 kb upstream and downstream of the transcriptional initiation site) or coding sequence (CDS) occur, direct sequencing, gene copy number and methylation-specific PCR analyses of NR4A3 in aggressive lymphomas were performed on selected cases. However, besides some already described single-nucleotide polymorphisms derived from publicly available databases (http://www.ncbi.nlm.nih.gov/snp), no alterations were detected (Supplementary Table S3).

To functionally characterize NR4A3, the SuDHL4 lymphoma cell line was stably transduced with an inducible lentiviral construct coding for NR4A3. The transduction was performed in duplicates (SuDHL4 pLVX NR4A3-1 and -2). Transduction with the viral vector without insert (SuDHL4 pLVX empty) served as vector control. The removal of doxycycline from the culture medium led to a ∼28-fold NR4A3 induction (P < 0.003), whereas NR4A3 levels remained unchanged in the vector control (Fig. 2A and B). After 48 hours of doxycycline removal, a higher percentage of NR4A3-induced cells stained positive for Annexin V compared with vector control (48.3% vs. 5.3%, Fig. 2C, P = 0.035). Additionally, after 48 and 72 hours of doxycycline removal, a higher proportion of SuDHL4 pLVX NR4A3 1-2 cells exhibited an increased cleaved caspase-3/7 activity (relative activity: 3.1 vs. 0.95 after 48 hours and 3.05 vs. 0.98 after 72 hours, Fig. 2D, P < 0.05), and after 72 hours, a significantly higher SubG1 peak compared with their vector controls (18.5% vs. 5.59%, Fig. 2D; P < 0.01) was detectable.

Additionally, NR4A3 induction resulted in significantly reduced cell proliferation after 48 hours as estimated by BrdU incorporation (13.9% vs. 55.1%; Fig. 2E, P < 0.001) and cell growth after 72 hours as determined by the MTS assay (Fig. 2E, P = 0.0008).

To clarify whether the proapoptotic property of NR4A3 is mediated via transactivation of its target genes (nuclear localization) or its mitochondrial targeting (cytoplasmic localization), inducing mitochondrial cell death (16), we treated under doxycycline withdrawal SuDHL4 pLVX NR4A3-1 and SuDHL4 pLVX empty cells with leptomycin B (LMB)—an inhibitor of the nuclear export system (17) causing nuclear retention of NR4As proteins (18). Apoptosis rates estimated by Annexin V staining of SuDHL4 pLVX NR4A3-1 cells were detected to a similar extent (62.3% vs. 60.7%; Fig. 2F, P = 0.83) and NR4A3 expression was uninfluenced by LMB treatment (Fig. 2G), suggesting that the proapoptotic property of NR4A3 is mediated by its nuclear function.

To confirm the specificity of the NR4A3-mediated apoptosis, we transfected SuDHL4 pLVX NR4A3-1 cells with siRNA targeting NR4A3 under doxycycline withdrawal (Fig. 2h). Silencing of NR4A3 under concomitant induction abrogated the apoptosis rate of NR4A3 silenced lymphoma cells in comparison with scramble controls as estimated by Annexin V staining (11.3% vs. 68.7%; Fig. 2I, P < 0.001).

In order to support the tumor suppressive function of NR4A3 in vivo, stably transduced SuDHL4 lymphoma cell lines were further investigated in the NSG mouse model. Control cells (SuDHL4 pLVX—containing no insert) were subcutaneously injected into the right flank and the same number of SuDHL4 pLVX NR4A3-1 lymphoma cells into the left flank of male NSG mice (n = 10). Five of 10 mice did not receive any doxycycline to induce NR4A3 expression, whereas doxycycline was administered to the residual five mice to suppress NR4A3 expression. Within 20 days, all mice developed visible tumors in their right flanks, whereas tumors in the left flanks were detected only in the doxycycline-administered mice. Macroscopic inspection (Fig. 3A) showed a clear size difference: Mice inoculated with SuDHL4 NR4A3-1 cells without doxycycline developed no lymphomas compared with doxycycline-administered mice or mice inoculated with isogenic empty vector control (Fig. 3B, P < 0.01). Histologic analysis also demonstrated that only mice without doxycycline from SuDHL4 NR4A3-1 lymphoma cells remained tumor free (Fig. 3C).

Taken together, these data imply that NR4A3 possesses tumor suppressor functions in aggressive lymphoma cells.

To investigate the effects of pharmacologic activation of NR4A3 with the NR4A3-inducing agent TG in lymphoma cells, we used Karpas-422 and SuDHL4 (as GCB-DLBCL model) and RI-1 and U2932 (as ABC-DLBCL model) cell lines. After 72 hours of TG treatment, a concentration-dependent growth inhibition in all investigated cell lines was detected by the MTS assay (Fig. 4A). NR4A3 expression was induced by treatment with 1.0 × 10⁻⁵ mol/L TG in all lymphoma cell lines (Fig. 4B, P < 0.01) after 4, 12, 24, 48, and 72 hours of TG treatment on mRNA levels. Furthermore, Western blot analysis confirmed TG-mediated NR4A3 induction on the protein level (Fig. 4C). In order to investigate whether TG-induced growth inhibition is mediated by apoptosis, we determined the cleavage of caspase-3, the sub-G₁ peak and the positivity for Annexin V. After 24 hours of treatment, TG led to a significantly higher number of Annexin V–positive cells compared with their untreated controls (Fig. 4D; P < 0.005). Additionally, the percentage of cells exhibiting cleaved caspase-3 after 24 hours and a sub-G₁ peak was also higher in TG-treated lymphoma cells compared with their untreated controls (Fig. 4E, P < 0.008). Finally, to investigate whether the observed apoptotic effects are specific to TG action via induction of NR4A3, siRNA-mediated silencing was performed in SuDHL4 and in U2932 followed by TG treatment. Silencing of NR4A3 entirely abrogated the apoptotic effects of TG (Fig. 4F–H, P < 0.01). Reduction of apoptosis was also observed after 24, 48, and 72 hours TG treatment in NR4A3-silenced lymphoma cells (Fig. 4I, P < 0.01). Additionally, the number of viable cells significantly increased in NR4A3-silenced lymphoma cells in comparison with the respective scramble controls, in which the number of viable cells decreased under TG treatment (Fig. 4J, P < 0.001). After 72-hour treatment, the number of viable cells of scramble controls were just 4.2% for U2932 and 1.8% for SuDHL4 of the NR4A3-silenced lymphoma cells compared with scramble controls.

Furthermore, we identified an agent that induced NR4A3 in aggressive lymphoma cell lines (Fig. 5A and 5B). Similar to TG treatment, BF175—a pinacolyl boronate-substituted stilbene derivate (11, 12)—treatment resulted in concentration-dependent growth inhibition of Karpas-422, RI-1, and U2932 cells (Fig. 5C). Additionally, treatment with 1.0 × 10⁻⁵ mol/L of BF175 induced NR4A3 expression (Fig. 5A and 5B, P < 0.01) accompanied by higher Annexin V positivity (Fig. 5D, P < 0.008) and by an increased cleavage of caspase-3 and significantly higher sub-G₁ peak (Fig. 5E, P < 0.01) in BF175-treated cells compared with untreated controls. Furthermore, silencing of NR4A3 in U2932 inhibited the apoptotic effects of BF175 (Fig. 5F–5H). Diminished apoptosis was also observed after 24, 48, and 72 hours of BF175 treatment in NR4A3-silenced lymphoma cells (Fig. 5I, P < 0.009). Likewise, the number of viable cells significantly increased in NR4A3-silenced lymphoma cells compared with controls, which showed a decreased number of viable cells under BF175 treatment (Fig. 5LJ, P < 0.001). After 72-hour treatment, the number of viable cells of scramble controls was just 1.8% of the NR4A3-silenced lymphoma cells.

Importantly, NR4A1 mRNA expression levels were also induced by TG (Supplementary Fig. S1) and BF175 (Supplementary Fig. S2) treatment. However, the extent of NR4A1 induction was remarkably lower compared with NR4A3.

Together, these data indicate that the apoptotic effects of TG and BF175 in aggressive lymphoma cells are NR4A3 mediated.

NR4A1 and NR4A3 were separately overexpressed in aggressive lymphoma cell lines (Karpas-422, SuDHL4, RI-1, and U2932) followed by expression analysis and various apoptotic assays. As shown in Fig. 6A and B, NR4A1 and NR4A3 were significantly overexpressed at similar levels in the respective cell lines (P < 0.003). A marked increased Annexin V positivity was detected after 48 hours in all cell lines caused by NR4A1 overexpression (Fig. 6C; P < 0.01) and by NR4A3 overexpression (Fig. 6C, P < 0.01). A similar increase in caspase-3/7 activity (Fig. 6D, P < 0.01) and sub-G₁ peak positive cells (Fig. 6E, P < 0.01) was detected for NR4A1 and NR4A3 overexpression in all tested cell lines.

LMB treatment of NR4A1 overexpressing SuDHL4 (pEZ-M46 NR4A1) and its controls (pEZ-M46 empty) did not influence apoptosis estimated by Annexin V staining (59% vs. 60.3% for pEZ-M46 NR4A1 with and without LMB, Fig. 6F and G, P = 0.73), suggesting that the proapoptotic property of NR4A1 is mediated by its nuclear function.

To confirm the specificity of NR4A1-mediated apoptosis, we transfected SN1 III cells, which carry an inducible NR4A1 construct (Tet off -pLVX) and were previously described by our group (7), with siRNA targeting NR4A1. The apoptosis rate estimated by Annexin V staining of NR4A1 silenced lymphoma cells was significantly diminished compared with scramble controls (6.6% vs. 52.8%; Fig. 6H and I, P < 0.001).

Expression analysis of potential NR4A apoptotic and antiapoptotic target genes demonstrated that NR4A1 and NR4A3 overexpression caused a similar strong induction of proapoptotic Puma, TRAIL, BID, BIK, isoform 1 and 6 of Bim and BAK (Fig. 7, P < 0.001 for all six genes), whereas expression levels of their inhibitors (BCL2, BCLX, and MCL1), and their receptors (Fas, DR4, and DR5) remained unchanged after 48 hours (data not shown).

For 82 patients diagnosed with aggressive lymphomas (31 non–GCB-DLBCL, 18 GCB-DLBCL and 33 follicular lymphoma grade 3)—exhibiting reduced NR4A1 and NR4A3 expression levels as previously described by us (7)—expression levels of pro- and antiapoptotic genes were determined. A significant positive correlation of Puma (Spearman ρ = 0.460 for NR4A1 and Pearson ρ = 0.485 for NR4A3, P < 0.001; Supplementary Figs. S3A and S3B), isoform 1 and 6 of Bim (Spearman ρ = 0.415 for NR4A1 and Spearman ρ = 0.636 for NR4A3, P < 0.001, Supplementary Figs. S3C and S3D) and Trail (Spearman ρ = 0.492 for NR4A1 and Spearman ρ = 0.502 for NR4A3, P < 0.001, Supplementary Figs. S3E and S3F) were detected. These data suggest that NR4A3 and NR4A1 possess a redundant function by regulating proapoptotic genes to a similar extent.

To investigate whether NR4A1 and NR4A3 correlate with BAK, Puma, BIM, BID, and Trail, we performed database retrieval via The Cancer Genome Atlas (TCGA) and analyzed 12 different tumor entities, which contained more than 90 patients per group. All together, we were able to analyze 4,788 patient specimens. A statistical significant correlation between NR4A1 and NR4A3 expression could be shown in 10 of 12 tumor entities (Supplementary Table S4). Almost all analyzed proapoptotic genes significantly correlated with NR4A1 and NR4A3 expression in AML and colorectal adenocarcinoma (Supplementary Tables S4 and S5). Additionally, in low-grade glioma, all apoptotic genes correlated with NR4A3 expression (Supplementary Table S5). In the other cancer entities, a marked correlation of single proapoptotic genes to NR4A1 and/or NR4A3 was observed (Supplementary Tables S4 and S5). However, the correlation coefficients were by far lower. These data show that NR4A3 is functionally redundant to NR4A1 in many different cancer entities and that both NR4As possess a proapoptotic function at least in AML, colorectal adenocarcinoma, and low-grade glioma.

This study was designed to investigate the function of NR4A3 in aggressive lymphomas and to study whether NR4A3 has tumor suppressive properties. It was reported that NR4A3 expression levels correlated positively with therapeutic success in whole-genome expression analysis of 58 DLBCL patients (8) and that NR4A3 plays an important role in tumorigenesis (1, 5, 6, 19–21). Although we described a downregulation of NR4A3 together with NR4A1 in aggressive lymphomas (7), the function of NR4A3 in aggressive lymphomas is unknown. In our study, a significant association of low NR4A3 expression and poor survival indicated a tumor suppressive role of this protein in human lymphoma samples. To test this hypothesis, our in vitro experiments showed that overexpression of NR4A3 led to an induction of apoptosis in several aggressive lymphoma cell models. Likewise, xenograft experiments demonstrated that NR4A3 induction completely suppressed lymphoma formation. Until now, proapoptotic effects of NR4A3 have been exclusively described in T cells (9, 16), AML cells (22, 23), and breast cancer cell lines (24); however, data on the underlying mechanisms are lacking. Based on the functional redundancy of NR4A3 to NR4A1 (9) shown in other cell types, it is possible that either NR4A3 transactivates proapoptotic genes similar to NR4A1 (16) or it translocates to the mitochondria, leading to BCL2 rearrangement and ultimately to induction of apoptosis (25–27). As Puma, Trail, Bid, Bik, Bim, and Bak1 were upregulated by NR4A3 overexpression in aggressive lymphoma cell lines and correlated to NR4A3 expression in primary lymphomas, and LMB treatment did not influence the apoptosis rate in NR4A3-induced SuDHL4 cells, we speculate that the proapoptotic effects of NR4A3 are achieved by transactivation of proapoptotic genes.

Treatment of aggressive lymphoma cell lines with TG—a NR4A3 inducing agent (10)—effectively induced apoptosis. TG is an endoplasmic reticulum stress inducer with apoptotic effects in various cells (28–30) and the molecular mechanisms of TG-mediated apoptotic effects are unknown. Importantly, our observation, that silencing of NR4A3 by siRNA reversed the induction of apoptosis, suggests that TG works, at least in part, through transcriptional induction of NR4A3.

Our in vitro experiments, in which aggressive lymphoma cell lines were treated with BF175, resulted in induction of NR4A3 accompanied by apoptosis. Silencing of NR4A3 abrogated the apoptotic effects of BF175. It was reported that BF175 affects the lipid metabolism, but it did not induce apoptosis in any previously investigated cell line—mainly liver cell lines—even at higher concentrations (11, 12). Thus, it seems that the apoptotic effect of BF175 is mediated by NR4A3 and is apparently exclusively found in aggressive lymphoma cell lines.

Overexpressing either NR4A3 or NR4A1 in aggressive lymphoma cell lines demonstrated that both NR4As possess proapoptotic effects in aggressive lymphomas. NR4A3 has been shown to be functionally redundant with NR4A1 in T-cell apoptosis (9) and overexpression of either NR4A1 or NR4A3 suppressed growth and survival of AML cells (22). Additionally, it was demonstrated that in AML cells NR4A1 and NR4A3 overexpression caused almost identical gene expression profiles (22). However, NR4A3 knockout mice are either embryonically lethal or show an inner ear defect with partial bidirectional circling behavior. In contrast, NR4A1 knockout mice show no obvious phenotype (31–33). Together with the fact that NR4A3 expression profiles significantly correlated with aggressiveness of lymphomas (7, 34), it might be hypothesized that NR4A1 and NR4A3 have a functional redundancy at least with regard to regulation of apoptosis in aggressive lymphoma cells. Additionally, based on our TCGA analysis, it might be hypothesized that NR4A3 also shows a tumor suppressive function redundant to NR4A1 in other tumor entities like AML, colorectal adenocarcinoma, and low-grade glioma.

In conclusion, we have demonstrated that NR4A3 has proapoptotic properties, which are functionally redundant to NR4A1 and define NR4A3 as novel tumor suppressor involved in aggressive lymphoma development. Hence, NR4A3 and NR4A3 inducing agents are novel promising future targets for drug development in lymphoma therapy.

No potential conflicts of interest were disclosed.

Conception and design: A.J.A. Deutsch, A.M. Krogsdam, P. Neumeister

Development of methodology: A.J.A. Deutsch, B. Rinner, M. Pichler, M.-T. Frisch, A. Prokesch, A.M. Krogsdam, C. Wang, C. Beham-Schmid, P. Neumeister

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): A.J.A. Deutsch, B. Rinner, M. Pichler, M. Bischof, K. Wenzl, C. Beham-Schmid, P. Neumeister

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): A.J.A. Deutsch, S. Hatzl, J. Feichtinger, K. Wenzl, H. Sill, G.G. Thallinger, H.T. Greinix, C. Beham-Schmid, P. Neumeister

Writing, review, and/or revision of the manuscript: A.J.A. Deutsch, B. Rinner, M. Pichler, K. Troppan, K. Fechter, J. Feichtinger, K. Wenzl, A. Prokesch, G.G. Thallinger, H.T. Greinix, C. Wang, P. Neumeister

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): A.J.A. Deutsch, B. Rinner, K. Troppan, K. Fechter, V. Stiegelbauer, A. Prokesch, A.M. Krogsdam, P. Neumeister

Study supervision: A.J.A. Deutsch, P. Neumeister

P. Neumeister was supported by grants of Fellinger Krebsforschung, Land Steiermark, Jubiläumsfond der ÖNB (N11181) and the OMICS Center Graz grant of the Austrian Ministry for Science, Research and Economy (G.G. Thallinger). A.J.A. Deutsch was supported by the START-Funding-Program of the Medical University of Graz and by a research grant of MEFOgraz.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.