Haploinsufficiency for NR3C1, the gene encoding the glucocorticoid receptor, in blastic plasmacytoid dendritic cell neoplasms

Abstract

Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare and highly aggressive leukemia for which knowledge on disease mechanisms and effective therapies are currently lacking. Only a handful of recurring genetic mutations have been identified and none is specific to BPDCN. In this study, through molecular cloning in an index case that presented a balanced t(3;5)(q21;q31) and molecular cytogenetic analyses in a further 46 cases, we identify monoallelic deletion of NR3C1 (5q31), encoding the glucocorticoid receptor (GCR), in 13 of 47 (28%) BPDCN patients. Targeted deep sequencing in 36 BPDCN cases, including 10 with NR3C1 deletion, did not reveal NR3C1 point mutations or indels. Haploinsufficiency for NR3C1 defined a subset of BPDCN with lowered GCR expression and extremely poor overall survival (P = .0006). Consistent with a role for GCR in tumor suppression, functional analyses coupled with gene expression profiling identified corticoresistance and loss-of-EZH2 function as major downstream consequences of NR3C1 deletion in BPDCN. Subsequently, more detailed analyses of the t(3;5)(q21;q31) revealed fusion of NR3C1 to a long noncoding RNA (lncRNA) gene (lincRNA-3q) that encodes a novel, nuclear, noncoding RNA involved in the regulation of leukemia stem cell programs and G1/S transition, via E2F. Overexpression of lincRNA-3q was a consistent feature of malignant cells and could be abrogated by bromodomain and extraterminal domain (BET) protein inhibition. Taken together, this work points to NR3C1 as a haploinsufficient tumor suppressor in a subset of BPDCN and identifies BET inhibition, acting at least partially via lncRNA blockade, as a novel treatment option in BPDCN.

Figure 1

Targeting of the NR3C1 gene by 5q deletion/translocation in BPDCN. (A) Schematic representation of 5q alterations in BPDCN patients (n = 17/47), as indicated (left). aCGH image of case UPN 33 showing focal deletion of the NR3C1 gene (right). (B) Kaplan-Meier cumulative survival curves in BPDCN according to 5q alteration status (for simplicity, this is labeled del(5q) [UPNs 1, 2, 8, 12, 14, 21, 22, 24, 26, 27, 29, 31, 35, and 45] and non-del(5q) [UPNs 3, 9, 10, 16-19, 25, 27, 36, 37, and 40-42]; upper) and to NR3C1 deletion status (UPNs 1, 8, 14, 22, 24, 26, 27, 31, 35, and 45) (lower), determined by FISH or aCGH; ΔNR3C1). P values were determined by log-rank test. (C) FISH analysis of the t(3;5)(q21;q31) in the BPDCN cell line GEN2.2 (UPN 1) with chromosome 5q probes CTB-88J14 (green) and RP11-278J6 (red) (supplemental Table 2). Note a split signal for RP11-278J6 (NR3C1) on the der(3) and der(5) chromosomes. (D) Affymetrix-derived NR3C1 expression in BPDCN patients presenting del(5q) targeting NR3C1 or not. P value was determined by Wilcoxon test. (E) aCGH image showing 5q/NR3C1 deletion in DNA from a skin biopsy sample with leukemic infiltrate (UPN 22) compared with control DNA (Ctrl), and analysis in 2 BPDCN cases without 5q/NR3C1 deletion (UPNs 19 and 21). (F) Overview of immunostaining (diaminobenzidine and hematoxylin) of GCR on BPDCN dermal localization in a del(5q) case: NR3C1(+) (left) and a del(5q)/NR3C1 case: NR3C1(−) (upper right). Epidermal squamous cells showing positive nuclear staining are marked with thin gray arrows; leukemia infiltrate cells are marked with thick black arrows. A higher magnification of the clonal infiltrate in dermis of the same case as the upper right image shows mainly negative tumor cells (red arrow) and some positive cells (thick black arrow) (lower right). BM, bone marrow.

Figure 2

The t(3;5)-encoded GCR-lincRNA-3q fusion protein is associated with glucocorticoid resistance. (A) Schematic representation of the genomic organization surrounding the normal 3q and der(3) breakpoint regions, indicating the resulting fusion transcript NR3C1-lincRNA-3q. The blue filled circle denotes the GATA2 superenhancer. In black, 3q sequences. In gray, 5q sequences. The NR3C1-lincRNA-3q fusion transcript likely derives from a splicing event between NR3C1 exon 2 and a cryptic splice acceptor site located in the antisense orientation in lincRNA-3q exon 2 (chromosome 3:129,832,478). (B) Schematic representation of the structure of wild-type GCR (top left) and predicted GCR-FP (bottom left). GCR and ACTIN western blot analyses showing expression of an abnormal GCR isoform (GCR-FP) exclusively in GEN2.2 cells compared with t(3;5)-negative cell lines, as indicated (right); * denotes nonspecific band. (C) Western blot using anti-GCR and anti-ACTIN antibodies in the parent CAL-1 cell line (nontransduced; NT) compared with CAL-1 cells transduced with empty GFP vector (−) or with the GFP-tagged GCR-FP expression vector (+), as indicated (n = 2). (D) Western blot using anti-GCR and anti-ACTIN antibodies in the parent CAL-1 cell line (NT) compared with CAL-1 cells transduced with shCtrl or shGCR (n = 2 independent hairpins, shGCR1 and 2), as indicated (right). Affymetrix NR3C1 messenger RNA expression in shGCR1 compared with shCtrl-transduced cells (experimental haploinsufficiency, n = 3) (left). (E) Evaluation of drug sensitivity of CAL-1 cells overexpressing GCR-FP compared with control CAL-1 cells after 16 hours of treatment with 10 μM etoposide (Eto), 72 hours of treatment with 10 μM dexamethasone (Dex), or a combination of both treatments. Specific apoptosis was measured by AnnexinV/propidium iodide staining and flow cytometry, and calculated as follows: (% drug-induced cell death − % control cell death) ÷ (100 − % control cell death) × 100. P value determined by Wilcoxon test; n = 3. (F) Evaluation of drug sensitivity of CAL-1 cells shGFP compared with shCtrl as described in panel E. aa, amino acid; AF, activation function domain; cen, centromere; DBD, DNA binding domain; LBD, ligand binding domain; ex, exon; NTD, N-terminal domain; tel, telomere; TSS, transcription start site; ZnF, zinc finger domain.

Figure 3

Misregulated GCR and EZH2 signaling as a consequence of NR3C1 alterations in BPDCN. (A) Luciferase reporter assay for glucocorticoid transcriptional transactivation activity in COS7 cells transduced with empty pcDNA3.1 vector, pcDNA3.1-GCR, and pcDNA3.1-GCR-FP, as indicated, after a 6-hour treatment with either dimethyl sulfoxide (DMSO) or 100 nM dexamethasone (n = 3). (B) Heatmap representation of the common differentially expressed genes among CAL-1-GCR-FP (GCR-FP[+]), CAL-1 control cells (GCR-FP[−]), and shGCR-transduced cells (shGCR[+]) compared with controls (shGCR[−]) after treatment with 100 nM dexamethasone for 6 hours (n = 3 for each group). Black arrows indicate genes cited in the text. (C) GSEA plots showing gene regulatory circuits that are differentially expressed between dexamethasone-treated CAL-1-GCR-FP or shGCR compared with respective control CAL-1 cells (CAL-1-GFP or CAL-1 shCtrl). (D) Western blot analysis and quantification of global H3K27me3 levels in CAL-1-GCR-FP (GCR-FP[+]) and CAL-1 shGCR (upper) and respective CAL-1 control cells (CAL-1 GFP [GCR-FP(−)] and CAL-1 shCtrl) (lower) upon 6-hour or 24-hour treatment with 100 nM dexamethasone (n = 3). (E) ChIP for H3K27me3 followed by qPCR analysis for enrichment on HOXA gene promoters, as indicated, in CAL-1-GCR-FP (GCR-FP[+]), CAL-1 shGCR, and respective CAL-1 control cells (CAL-1 GFP [GCR-FP(−)] and CAL-1 shCtrl) treated for 24 hours with 100 nM dexamethasone (n = 3). (F) GSEA plots showing gene regulatory circuits that are differentially expressed between BPDCN patients presenting, or not, del(5q) abnormalities that target NR3C1 (bone marrow or skin, as indicated; unmutated EZH2 and ASXL1 cases only). (G) Tumor bioluminescence (7 and 14 days postinjection) for xenotransplanted nude mice bearing tumors derived from CAL-1 shCtrl and shGCR cell lines (left). P value determined by Wilcoxon test; n = 6 for each group. Bioluminescence imaging for representative mouse (right). D, day; GSEA, gene set enrichment analysis.

Figure 4

lincRNA-3q overexpression in BPDCN and AML drives G1/S and leukemia-driver gene expression programs. (A) RT-qPCR-derived lincRNA-3q expression profiles in normal bone marrow; normal pDC; BPDCN (CAL-1, GEN2.2), AML (U937, MUTZ-3), and chronic myeloid leukemia cell lines (K562); and in BPDCN and AML patient samples. AML patients are classified according to cytogenetic risk group. Cases presenting chromosome 3q abnormalities are identified in blue. (B) RT-qPCR and RT-PCR analysis of the subcellular localization of lincRNA-3q in CAL-1 BPDCN cells and U937 AML cells. RNA extracted from whole-cell and subcellular fractions (n = 2). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was used as a control. (C) RT-qPCR-derived lincRNA-3q expression in CAL-1 cells transduced with control (shCtrl) or lincRNA-3q-targeting short hairpin constructs (sh1 and 2) (left). Cell cycle analysis in CAL-1 cells transduced with control (shCtrl) or lincRNA-3q sh1 and 2 constructs (upper right). *P < .05 by Wilcoxon test; n = 6. Representative histogram representation of percentage of cells in cell cycle phases (n = 4) (lower right). (D) GSEA plots obtained by comparing gene expression profiles of CAL-1 cells transduced with control (shCtrl) or lincRNA-3q-targeting short hairpin (shlincRNA-3q) (right) and associated heatmap (n = 4 for each group) (left). Genes mentioned in the text are marked with an arrow. (E) E2F cyclin E–luciferase (luc) reporter assay in H1299 cells transiently transfected with control or siRNA targeting lincRNA-3q with or without addition of exogenous E2F1 (100 ng) (upper). *P < .05, **P < .01 by Wilcoxon test; n = 6. Western blot using anti-E2F1 antibody in H1299 cells transduced with siCtrl or silincRNA-3q, as indicated (n = 2) (lower). (F) Table presenting clone frequency and associated statistics derived from in vitro limiting dilution clonogenicity for CAL-1 either nontransduced (NT) or transduced with control (shCtrl) or lincRNA-3q-targeting (shlincRNA-3q) shRNA. Adv, adverse; cyt, cytosolic fraction; Fav, favorable; Interm, intermediate; nuc, nuclear fraction; tot, whole cell extract.

Figure 5

Abrogation of lincRNA-3q activity by BET inhibition in BPDCN and AML leukemia cells. (A) Tumor growth kinetics measured in nude mice engrafted with shCtrl or shlincRNA-3q-transduced CAL-1 cells (left); bioluminescence imaging for 2 representative mice (day 14) (middle); and quantification of tumor bioluminescence (7 and 14 days postinjection) for each group (right). *P < .05, **P < .01 by Student t test; n = 8 for each group. (B) Genomic organization of the lincRNA-3q locus and BRD4 ChIP-sequencing analysis (ArrayExpress accession number ERP004614) showing promoter occupancy by BRD4 in MUTZ-3 cell line (3q rearrranged AML) compared with K562 cell line (no 3q rearrangement). ENCODE data showing positions of active chromatin marks and phospho-RNA polymerase II (Pol II) binding, as indicated. (C) Anti-BRD4 and anti-AcH3 ChIP at the promoter region of lincRNA-3q in CAL-1 and U937 cells, as indicated. (D) RT-qPCR-derived lincRNA-3q expression in MUTZ3, K562, U937, and CAL-1 cells treated with 1 μM JQ1 at the time indicated. *P < .05, **P < .01 by Wilcoxon test; n = 3. (E) Venn diagrams (left) and heatmap (right) showing time course of differential gene expression of lincRNA-3q targets upon JQ1 treatment in CAL-1, as indicated. (F) Tumor growth kinetics measured in nude mice bearing tumors derived from CAL-1 cell lines after control or JQ1 treatment (left); bioluminescence imaging for 2 representative mice (day 14) (middle); and quantification of tumor bioluminescence (7 and 14 days posttreatment) for each group (right). *P < .05, **P < .01 by Student t test; n = 6 for each group. AcH3, acetylated histone H3; BET, bromodomain and extraterminal domain; Ig, immunoglobulin; LP, lincRNA ChIP Primer.

Figure 6

A model illustrating how NR3C1 haploinsufficiency and lincRNA-3q misregulation contribute to BPDCN pathogenesis. We postulate that attenuated GCR signaling and lincRNA-3q malfunction drive BPDCN disease pathogenesis through epigenetic reprogramming. This is proposed to favor emergence of clinically aggressive disease and predicted to occur progressively by 2 routes. In the first, altered GCR signaling drives a loss-of-EZH2 phenotype that rewires key downstream PRC2 targets (eg, the HOXA locus) and drives deregulation of pDC differentiation pathways and treatment resistance in BPDCN. In the second, BET-dependent desilencing of oncogenic lncRNA genes (lincRNA-3q) may occur as “collateral damage” downstream of altered GCR and EZH2 activity or through other mechanisms. Abnormal activity of the affected nuclear lncRNA (in this case, lincRNA-3q) would engage further rounds of epigenetic reprogramming, leading to misregulation of E2F activity and activation of leukemia stem cell programs.