Identification of cereblon-binding proteins and relationship with response and survival after IMiDs in multiple myeloma

Abstract

Cereblon (CRBN) mediates immunomodulatory drug (IMiD) action in multiple myeloma (MM). Using 2 different methodologies, we identified 244 CRBN binding proteins and established relevance to MM biology by changes in their abundance after exposure to lenalidomide. Proteins most reproducibly binding CRBN (>fourfold vs controls) included DDB1, CUL4A, IKZF1, KPNA2, LTF, PFKL, PRKAR2A, RANGAP1, and SHMT2. After lenalidomide treatment, the abundance of 46 CRBN binding proteins decreased. We focused attention on 2 of these-IKZF1 and IKZF3. IZKF expression is similar across all MM stages or subtypes; however, IKZF1 is substantially lower in 3 of 5 IMiD-resistant MM cell lines. The cell line (FR4) with the lowest IKZF1 levels also harbors a damaging mutation and a translocation that upregulates IRF4, an IKZF target. Clinical relevance of CRBN-binding proteins was demonstrated in 44 refractory MM patients treated with pomalidomide and dexamethasone therapy in whom low IKZF1 gene expression predicted lack of response (0/11 responses in the lowest expression quartile). CRBN, IKZF1, and KPNA2 levels also correlate with significant differences in overall survival. Our study identifies CRBN-binding proteins and demonstrates that in addition to CRBN, IKZF1, and KPNA2, expression can predict survival outcomes.

Figure 1

Identification of proteins in CRBN protein-protein interaction. Schematic diagrams show the workflow designed for identifying CRBN-binding partners via Co-IP or Ni⁺ beads pull-down and proteomic-based analysis. The first approach includes Co-IP, separation by SDS-PAGE, in-gel digestion and peptide extraction, and HPLC-ESI-MS/MS quantitative analysis. The second approach included using Ni⁺ beads to pull down the proteins from control cells (OCI-MY5 vector) and CRBN-overexpressing cells (OCI-MY5 CRBN-His) untreated or pre-treated with lenalidomide for 48 hours, separation by SDS-PAGE, in-gel digestion and peptide extraction, and HPLC-ESI-MS/MS quantitative analysis.

Figure 2

In silico network and pathway enrichment analysis. Lists of the proteins in the CRBN complex that upregulated (A) or downregulated (B) in response to lenolidomide were evaluated against the latest versions of the following databases (covering >50% of human proteins): CellMap, Reactome, KEGG, NCI-PID, and BioCartel. Linker genes were included in the network modeling and pathway enrichment predictions. Results were filtered for a P value < .05 and a false discovery rate < 0.05. The separate networks modeled from downregulation vs upregulation lists are presented in (A) and (B), and the association or connection between IKZF1 and CRBN was a custom annotation added to the figures.

Figure 3

Validation of IKZF1 and IKZF3 as CRBN-binding proteins. (A) Immunoblotting of the Ni⁺ beads pull-down proteins from OCI-MY5 vector (control) and OCI-MY5 CRBN-His with (+) without (–) lenalidomide treatment of 48 hours (2 independent experiments). CRBN, DDB1, CUL4A, IKZF1, IKZF3, and KPNA2 showed similar changes detected by MS analysis. (B-C) OCI-MY5 CRBN-His and MM1.S cells were treated with lenalidomide (10 μM) in a time course, and the changes of some top proteins identified from MS analysis or related with lenalidomide-induced cytotoxicity were measured by western blot. IKZF1 and IKZF3 downregulation was identified as the earliest change after lenalidomide treatment. (D) OCI-MY5 CRBN-His cells were treated with dimethyl sulfoxide or lenalidomide for 3 hours, and Co-IP was performed with anti-CRBN antibody to immunoprecipitate CRBN and CRBN-associated proteins, using a normal mouse IgG as a negative control. IKZF1 and IKZF3 were greatly increased in CRBN antibody–immunoprecipitated samples after lenalidomide treatment (left panels), but they are decreased in cell lysate after lenalidomide treatment (right panels). CRBN increased in both immunoprecipitated and nonimmunoprecipitated lysate.

Figure 4

Gene expression analysis of IKZF1 and IKZF3 across cell lines originated from multiple hematological and solid tumors. The number of cases of each tumor entity is indicated in parenthesis. Red boxes indicate the set of MM cell lines. Results show IKZF1 (A) and IKZF3 (B) expression in most cell lines established from hematological tumors. The findings in hematological neoplasms contrast with the lack of expression of both genes found in solid-tumor cell lines. AML, acute myeloid leukemia; B-ALL, B-cell acute lymphoblastic leukemia; CML, chronic myeloid leukemia; T-ALL, T-cell acute lymphoblastic leukemia; MCL, mantle cell lymphoma; DLBCL, diffuse large B-cell lymphoma; MM, multiple myeloma.^,

Figure 5

Expression of Ikaros proteins in HMCLs and induction of cytotoxicity after knockdown of IKZF1 and IKZF3. (A) Expression of IKZF1 and IKZF3 in HMCLs. Nine HMLCs either sensitive or resistant to lenalidomide were analyzed for the expression of Ikaros by western blot. (B) Expression of IKZF1 in HMCLs was detected by flow cytometry analysis. (C-D) Knockdown of IKZF1 and IKZF3 induced cytotoxicity. MM1.S cells were infected with lentivirus-expressing nontargeting control or IKZF1 or IKZF3 shRNAs. MTT assays were performed at day 6 after infection and western blot analysis was performed at day 3 after infection. (E-G) Overexpression of IKZF1 and IKZF3 in a lenalidomide-sensitive cell line, MM1.S, did not substantially change lenalidomide sensitivity. MM1.S cells were infected with control virus (expressing vector alone) or lentivirus-expressing IKZF1 and IKZF3 after 72 hours (E), or after sorting GFP⁺ cells in MM1.S cells infected by IKZF1 lentivirus at 14 days after infection (F), cells were seeded in a 96-well plate and were treated with lenalidomide. (G) MTT assay was performed at day 6 after treatment and western blot analysis was performed at day 3 to confirm overexpression of IKZF1 and IKZF3.

Figure 6

Correlation of IMiD responsiveness and OS with IKZF1 and KPNA2 gene expression level in pomalidomide and dexamethasone therapy. A cohort of 44 MM patients treated with pomalidomide with GEP measured before initiation of chemotherapy was used in this study. (A) Responsiveness to pomalidomide in quartiles of IKZF1 expression. No patient in the lowest IKZF1 quartile expression responded to pomalidomide treatment. (B)The gene expression of the 11 top lenalidomide-regulated putative CRBN-interacting proteins was analyzed to examine the association between drug response and survival by using COX modeling. IKZF1 was identified as the top gene, which expression is associated with OS (P = .002). (C-D) Low IKZF1 and high KPNA2 expression levels were correlated with shorter OS.