CREB1 promotes expression of immune checkpoint HLA-E leading to immune escape in multiple myeloma

Abstract

Multiple myeloma (MM) cells effectively escape anti-tumoral immunity to survive in the tumor microenvironment (TME). Herein, we identify non-classical major histocompatibility complex (MHC) class I molecule HLA-E as a major contributing factor in immune escape. Clinically, HLA-E expression correlates with aggressive disease features such as t(4;14) and CD56 expression and is induced by IFN-gamma (IFN-γ) in the TME. We discovered that HLA-E is regulated by cAMP responsive element binding protein 1 (CREB1) transcription factor by direct promoter binding; genomic and pharmacological inhibition of CREB1 reduced HLA-E levels even in the presence of IFN-γ or IFN-γ activating agents, such as immunomodulatory drugs and panobinostat. HLA-E binds to natural killer group 2A (NKG2A), delivering an inhibitor signal to natural killer (NK) cells. Treatment with a CREB1 inhibitor was able to restore NK cell-mediated cytotoxicity against MM cell lines and patient samples. In conclusion, our results strongly demonstrate that CREB1 inhibition promotes anti-tumoral immunity in MM by limiting HLA-E expression and enhancing the activity of NK cells.

Fig. 1. HLA-E is expressed in MM cells and induced by IFN-γ.

A HLA-E log₂ mRNA expression values in bone marrow plasma cells from healthy individuals (NPCs, n = 22), individuals with monoclonal gammopathy of undetermined significance (MGUS, n = 44), or individuals with smoldering multiple myeloma (SMM, n= 12). 2000904_at probe: ANOVA p value = 0.0004; NPCs vs MGUS p = 0.0036, **; NPCs vs SMM p = 0.0005, ***. 200905_x_at probe: ANOVA p value = 0.0005; NPCs vs MGUS p = 0.0003, ***; NPCs vs SMM p = 0.0179, *. Solid blue lines indicate the median values; black dotted lines represent the 25^th and 75^th percentile. Data are derived from GSE5900 dataset. B HLA-E log₂ mRNA expression values in bone marrow plasma cells from healthy individuals (NPCs, n = 12) or patients with MM (MM, n = 65); 2000904_a probe: p < 0.0001, ****; 200905_x_at probe: p < 0.0001, ****. Solid blue lines indicate the median values; black dotted lines represent the 25^th and 75^th percentile. Data are derived from GSE4452 dataset. C Representative immunohistochemistry staining for HLA-E (20×) in one newly diagnosed patient with MM. Red arrows show MM cells positive for HLA-E. D HLA-E mRNA fold change in a panel of MM cell lines (U266, RPMI-8226, OPM-2, and H929) treated with solvent or IFN-γ 1 ng/mL for 24 h. n = 3, t test, two-tailed; U266 p = 0.0096, **; RPMI-8226 p = 0.0016, **; OPM-2 p = 0.0141, *; H929 p = 0.0031, **. Treated cells are normalized to each control. E HLA-E mean fluorescence intensity (MFI) fold change in a panel of MM cell lines (U266, RPMI-8226, OPM-2, and H929) treated with solvent or IFN-γ 1 ng/mL for 24 h. n = 2, t test, two-tailed; U266 p = 0.0014, **; RPMI-8226 p = 0.0455, *; OPM-2 p = 0.0311, *; H929 p < 0.0001, ****. Treated cells are normalized to each control. F Western blot analysis for HLA-E and GAPDH in a panel of MM cell lines (U266, RPMI-8226, OPM-2, and H929) treated with solvent, IFN-γ 1 ng/mL, or IFN-γ 50 ng/mL for 24 h.

Fig. 2. CD56 and CREB1 modulate HLA-E expression in MM.

A Schema of CD56-CREB1 signaling in MM. CD56 induces phosphorylation of RPS6KA3 (also known as RSK2), which in turn phosphorylates CREB1, leading to gene transcription. B Pathway analysis of the upregulated IFN_driven gene sets in the MMRF CoMMpass and GSE4452 datasets. Patients are divided based on median cutoff of CREB1 expression. NES normalized enrichment score; FDR false discovery rate; gene count includes the number of significant genes in the pathway. C HLA-E log₂ expression values in patients with low or high CREB1 expression based on median cutoff of CREB1 expression. CoMMpass MMRF dataset: n = 809, p < 0.0001, **** and GSE4452 dataset: n = 65, p = 0.0061, **. Dashed blue lines indicate the median values; black dotted lines represent the 25^th and 75^th percentile. D Regression studies to correlate HLA-E (probe 200904_at) as dependent variable to CREB1 (probe 204314_s_at). p = 0.0031; R = 0.13. E ChIP-sequencing tracks showing CREB1 signal on individual locus for HLA-E. The x-axis shows genomic coordinates. F Quantitative PCR of CREB1-ChIP enriched binding site to HLA-E promoter. n = 2, t test, two-tailed; p = 0.0317, *. G HLA-E MFI fold change in U266 cells overexpressing CREB1 compared with U266 control cells (CNT). n = 3, t test, two-tailed; p = 0.0001, ***. H HLA-E MFI fold change in U266 cells overexpressing CD56 compared with U266 control cells (CNT). n = 3, t test, two-tailed; p = 0.0010, **. I HLA-E MFI fold change in OPM-2 cells, H929 cells, and CD56⁺ CD138⁺ patient-derived MM cells treated with DMSO (D) or 666-15 (CRi) 0.3 μM for 48 h. n = 3, t test, two-tailed; OPM-2 p = 0.0153, *; H929 p = 0.04, *; MM patient samples (MM pts) p = 0.04, *. J HLA-E mRNA fold change in OPM-2 and H929 cells silenced for CREB1 (shCREB1) or with scrambled vectors, scr (n = 2, t test, two-tailed; p = 0.039, * and 0.0031, **) and in U266 cells overexpressing CREB1 or the control vector- CNT (n = 3, t test, two-tailed; p = 0.05, *). K HLA-E mRNA fold change in OPM-2 and H929 cells silenced for CD56 (shCD56) or with scrambled vectors, scr (n = 2, t test, two-tailed; p = 0.0008, *** and 0.0021, **) and in U266 cells overexpressing CD56 or the control vector-CNT (n = 3, t test, two-tailed; p = 0.04, *).

Fig. 3. Both STAT1 and CREB1 regulate HLA-E expression.

A Schema of IFN-γ signaling. IFN-γ signals via a receptor composed of two IFNGR chains (IFNGR1 and 2), which activate JAK1 and STAT1/STAT2. Phosphorylated STAT1/STAT2 complexes interact with IRF1 and IRF9, translocate to the nucleus, and bind to DNA on interferon sensitive response elements (ISRE). NLRC5 also binds to DNA to promote IFN-γ-related gene transcription. B STAT1 log₂ expression values in patients with low or high CREB1 expression based on median cutoff of CREB1 expression. CoMMpass MMRF database: n = 809, p = 0.0015, ** and GSE4452 database: n = 65, p = 0.0013, **. Dashed blue lines indicate the median values; black dotted lines represent the 25^th and 75^th percentile. C Regression studies to correlate STAT1 (probe 200887_s_at) as dependent variable to CREB1 (probe 204313_at). p < 0.0001; R = 0.4. D Western blot analysis for STAT1, IRF1, IRF9, and GAPDH in U266 control cells (CNT) or U266 cells overexpressing CREB1. E STAT1, IRF1, and IRF9 mRNA fold change in U266 control cells (CNT) or U266 cells overexpressing CREB1. n = 2, t test, two-tailed; STAT1 p = 0.0031, **; IRF1 p = ns; IRF9 p < 0.0001, ****. F Western blot analysis for STAT1, IRF1, IRF9, and GAPDH in OPM-2 and H929 cells treated with DMSO, 666-15 (CREBi) 0.3 μM, and 666-15 (CREBi) 1 μM for 48 h. G CREB1 mRNA fold change in U266 control cells (CNT) or U266 cells overexpressing STAT1. n = 3, t test, two-tailed; p = 0.05, *. H Western blot analysis for STAT1, phospho-CREB1, and GAPDH in U266 control cells (CNT) or U266 cells overexpressing STAT1. I Western blot analysis for phospho-STAT1, STAT1, HLA-E, CREB1, and GAPDH in U266 control cells (CNT) or U266 cells overexpressing CREB1 treated with solvent or IFN-γ 1 ng/mL for 24 h. J Western blot analysis for STAT1, CREB1, HLA-E, and GAPDH in U266 control cells (CNT), U266 cells overexpressing STAT1, CREB1, or CREB1 + STAT1. K HLA-E mRNA fold change in the same conditions reported in panel J. n = 2, t test, two-tailed; CREB1 versus CREB1 + STAT1 p = 0.0004, ***; STAT1 versus CREB1 + STAT1 p = 0.0003, ***. L Western blot analysis for phospho-STAT1, STAT1, HLA-E, and GAPDH in OPM-2 and H929 cells treated with DMSO, IFN-γ 1 ng/mL, 666-15 (CREBi) 1 μM, and the combination of IFN-γ + CREBi for 24 h.

Fig. 4. HLA-E expression is modulated by anti-MM drugs.

A HLA-E mRNA fold change in OPM-2 and H929 cells treated with DMSO, lenalidomide (LEN) 1 μM for 72 h, pomalidomide (POM) 1 μM for 72 h, and panobinostat (PANO) 100 nM for 24 h. n = 2, t test, two-tailed. OPM-2: DMSO versus LEN p = 0.0149, *; DMSO versus POM p = 0.0293, *; DMSO versus PANO p = 0.05, *. H929: DMSO versus LEN p = 0.0349, *; DMSO versus POM p = 0.0063, **; DMSO versus PANO p = 0.0002, ***. B HLA-E MFI fold change in OPM-2 cells and H929 cells treated with DMSO, LEN 1 μM for 72 h, POM 1 μM for 72 h, and PANO 100 nM for 24 h. n = 3, t test, two-tailed. OPM-2: DMSO versus LEN p < 0.0001, ****; DMSO versus POM p = 0.0021, **; DMSO versus PANO p = 0.0105, *. H929: DMSO versus LEN p = 0.0004, ***; DMSO versus POM p = 0.0075, **; PANO p = 0.97, ns. C HLA-E MFI fold change in OPM-2 cells treated with DMSO, LEN 0.1 μM, or POM 0.1 μM for 4 and 7 days. n = 2, t test, two-tailed. Day 4: DMSO versus LEN p = 0.05, *; DMSO versus POM p = 0.0115, *; day 7: DMSO versus LEN p = 0.0255, *; DMSO versus POM p = 0.0313, *. D HLA-E MFI fold change in OPM-2 cells and MM patient samples treated with DMSO, POM 1 μM, 666-15 (CREBi) 0.3 μM, or POM + CREBi. OPM-2: OPM-2 cells were treated for 72 h. n = 2, t test, two-tailed; POM versus POM + CREBi p = 0.0018, **. MM patients: patient samples were treated for 48 h. n = 3, t test, two-tailed; POM versus POM + CREBi p = 0.0059, **. E Gene-expression profiling for HLA-E in n = 16 relapsed refractory patients with MM with matched pre- and post-lenalidomide (LEN) data from dataset GSE8546. F Western blot analysis for p-CREB1, STAT1, and GAPDH in OPM-2, H929, and MM.1S cells treated with DMSO and POM 1 μM for 6, 24, and 48 h.

Fig. 5. NK cell-mediated cytotoxicity is increased by CREB1 inhibition.

A Cellular interactions between HLA-E⁺ MM cells and CD94⁺ NKG2A⁺ NK cells (either CD56^bright or CD56^dim). B Percentages of CD3⁻ CD56^bright CD16⁻ CD94⁺ NKG2A⁺ NK cells (t test, two-tailed; p = 0.021, *) or CD3⁻ CD56^dim CD16⁺ CD94⁺ NKG2A⁺ NK cells (t test, two-tailed; p = 0.004, **) in patients with <10% of CD56-expressing clonal MM cells (n = 27) or >10% of CD56-expressing clonal MM cells (n = 26). C t-SNE analysis combining CD94⁺ NKG2A⁺ cells in the two conditions. D Schema of co-culture experiments using gain-of- or loss-of-function cells. Freshly isolated NK cells were co-cultured with MM cells for 4 h. SYTOX staining was used to distinguish viable from dead MM cells. E U266 control cells (CNT) or U266 cells overexpressing CD56 (CD56) were incubated with NK cells derived from three healthy donors (E:T ratio 5:1 p = 0.11, ns; E:T ratio 1:10 p = 0.014, *). F H929 scrambled cells (scr) or H929 cells silenced for CD56 (shCD56) were incubated with NK cells derived from two healthy donors (E:T ratio 5:1 p = 0.08, ns; E:T ratio 10:1, p = 0.009, **). G Schema of co-culture experiments in cells pretreated for 48 h with DMSO, 666-15 (CREBi) 0.3 μM, pomalidomide (POM) 1 μM, or combination of CREBi with POM (CREBi + POM). Freshly isolated NK cells were co-cultured with MM cells for 4 h. SYTOX staining was used to distinguish viable from dead MM cells. H NK cell-mediated cytotoxicity by SYTOX staining in H929, OPM-2, and RPMI-8226 cells treated with DMSO, CREBi 0.3 μM, POM 1 μM, or combination of CREBi with POM (CREBi + POM) with or without NK cells. NK cells were isolated from n = 6 different healthy donors. E:T ratio was 5:1. % of lysis is calculated as described in the “method” section. H929: ANOVA p < 0.0001; “DMSO + NK cells” versus “CREBi + NK cells” p < 0.0001, ****; “DMSO + NK cells” versus “CREBi + POM + NK cells” p < 0.0001, ****; “CREBi + NK cells” versus “CREBi + POM + NK cells” p = 0.0041, **; “POM + NK cells” versus “CREBi + POM + NK cells” p < 0.0001, ****. OPM-2: ANOVA p = 0.0041; “DMSO + NK cells” versus “CREBi + NK cells” p = 0.0149, *; “DMSO + NK cells” versus “POM + NK cells” p = 0.0124, *; “DMSO + NK cells” versus “CREBi + POM + NK cells” p = 0.0005, ***; RPMI-8226: ANOVA p < 0.0001; “DMSO + NK cells” versus “CREBi + NK cells” p = 0.0261, *; “DMSO + NK cells” versus “CREBi + POM + NK cells” p < 0.0001, ****; “CREBi + NK cells” versus “CREBi + POM + NK cells” p = 0.001, **; “POM + NK cells” CREBi + POM “combo + NK cells” p < 0.0001, ****. I NK cell-mediated cytotoxicity by SYTOX staining in MM patient samples treated with DMSO, CREBi 0.3 μM, POM 1 μM, or combination of CREBi with POM (CREBi + POM) with or without NK cells. NK cells were isolated from n = 2 different healthy donors. ANOVA p = 0.0050; “DMSO + NK cells” versus “CREBi + POM + NK cells” p = 0.0015, **; “CREBi + NK cells” versus “CREBi + POM + NK cells” p = 0.0071, **; “POM + NK cells” versus “CREBi + POM + NK cells” p = 0.0022, **.