Evaluating the impact of CRBN mutations on response to immunomodulatory drugs and novel cereblon E3 ligase modulators in myeloma

Abstract

Immunomodulatory drug (IMiD) resistance is a key clinical challenge in myeloma treatment. Previous data suggest almost one-third of myeloma patients acquire genetic alteration of the key IMiD effector cereblon (CRBN) by the time they are pomalidomide refractory. Some events, including stop codons/frameshift mutations and copy loss, have clearly explicable effects on CRBN protein function. Missense mutations have also been reported throughout the length of CRBN but their functional impact has not been systematically studied. This study modeled selected missense mutations and examined their effect on CRBN function also analyzing whether any mutations deleterious to IMiD action could be overcome using the novel cereblon E3 ligase modulators (CELMoDs). Three patterns of response to missense mutations were apparent: mutations that led to complete loss of CRBN function for all agents, those that had no effect on CRBN function, and those with agent-dependent effect on CRBN function. The latter group of 4 mutations were profiled in more detail with confirmatory experiments demonstrating an ability of the more potent CELMoDs to lead to neosubstrate degradation and cell death even though IMiDs were not active. Dynamic modeling based on a newly generated crystal structure of the DDB1/CRBN/lenalidomide complex, with greater resolution than those published to date, helped to understand the impact of these mutations. These results have important implications for the interpretation of CRBN sequencing results from patients for future therapy decisions, particularly differentiating those who may, despite relapsing on IMiDs with CRBN mutations, have the potential to still benefit from the use of CELMoD agents.

Graphical abstract

Figure 1.

Overview of IMiD mechanism of action, the investigated compounds and reported CRBN mutations. (A) Schematic of the CRL4^CRBN E3 ubiquitin ligase complex and the IMiD mechanism of action that results in myeloma cell cytotoxicity. Figure created with BioRender.com. (B) Chemical structures for Len, Pom, iberdomide (CC-220), and Mezi (CC-92480). Len and Pom share a common glutarimide ring and differ only through the replacement of the isoindolinone in Len by a phthalimide in Pom. This difference leads to altered neosubstrate degradation, with Len being the only one able to potently degrade the kinase CK1α in cells. The chemical composition of the next-generation CELMoDs has led to higher CRBN-binding affinity. Iber achieves that through its additional phenyl and morpholino moieties, which extend the interactions with the β-hairpin sensor loop resulting in enhanced stabilization of the closed conformation of CRBN and ultimately increased neosubstrate ubiquitination and degradation.^, The design of Mezi further improves this by creating additional interactions with CRBN, which results in a further improvement in the degradation efficiency and kinetics.^, (C) Linear depiction of the CRBN protein; above are noted all CRBN mutations (n = 34) that have been reported in patients and below CRBNs are the 12 mutations investigated in this study. The color of the mark denotes the type of each mutation. Regarding mutation p,W415X, Gooding et al reported this mutation as W415X. Among the potential reported mutations for this region, it was selected to mutate Trp (W) into a Gly (G) as the most potentially harmful to proper folding and stability of CRBN due to the size difference of these residues. Figure generated using the ProteinPaint https://proteinpaint.stjude.org/ software by St. Jude Children’s Research Hospital. ∗Mutation not encountered in patients but added as an experimental positive control. One of the 3 Trp forming the IMiDs binding pocket, its mutation to an alanine has been previously demonstrated to functionally inactivate CRBN.^,

Figure 2.

The impact of CRBN mutations on the cell viability of IMiD-sensitive myeloma cell lines. (A) Growth inhibitory (GI50) cell viability measurements for the MM1.s^CRBNKO clone 1 cell lines each one expressing a differently mutated CRBN, CRBN^WT, or empty vector (EV) in comparison with parental MM1.s. The color of the title for each graph denotes the effect of each mutation on cell viability, where green is associated with reduction in cell viability comparable with parental MM1.s, red denotes no reduction in cell viability, and orange highlights the mutations whose presence leads to partial reduction in cell viability. Cells were treated with dimethyl sulfoxide (DMSO) or increasing concentrations of Len (maximum concentration 20 μM), Pom (maximum concentration 8 μM), Iber (maximum concentration 2 μM), and Mezi (maximum concentration 1 μM) for 5 days. Cell viability was determined using the CellTiter-Blue assay and expressed as a % of DMSO control. Results are the mean ± standard error of the mean (SEM) of n = 4 biological replicates. (B) Heat map summarizing the results shown in Figure 2A but with the drug concentration expressed as a percentage of the maximum for that cell line, Len 100% = 20 μM, Pom 100% = 8 μM, etc, and cell viability expressed as a percentage of DMSO control.

Figure 2.

The impact of CRBN mutations on the cell viability of IMiD-sensitive myeloma cell lines. (A) Growth inhibitory (GI50) cell viability measurements for the MM1.s^CRBNKO clone 1 cell lines each one expressing a differently mutated CRBN, CRBN^WT, or empty vector (EV) in comparison with parental MM1.s. The color of the title for each graph denotes the effect of each mutation on cell viability, where green is associated with reduction in cell viability comparable with parental MM1.s, red denotes no reduction in cell viability, and orange highlights the mutations whose presence leads to partial reduction in cell viability. Cells were treated with dimethyl sulfoxide (DMSO) or increasing concentrations of Len (maximum concentration 20 μM), Pom (maximum concentration 8 μM), Iber (maximum concentration 2 μM), and Mezi (maximum concentration 1 μM) for 5 days. Cell viability was determined using the CellTiter-Blue assay and expressed as a % of DMSO control. Results are the mean ± standard error of the mean (SEM) of n = 4 biological replicates. (B) Heat map summarizing the results shown in Figure 2A but with the drug concentration expressed as a percentage of the maximum for that cell line, Len 100% = 20 μM, Pom 100% = 8 μM, etc, and cell viability expressed as a percentage of DMSO control.

Figure 3.

The impact of CRBN mutations on neosubstrate degradation after IMiD/CELMoD treatment. Approximately 1 × 10⁶ cells from each MM1.s^CRBNKO clone 1 cell line (expressing a differently mutated CRBN, CRBN^WT, or EV) and the parental MM1.s cell line were drug treated (10 μM Len, 1 μM Pom, 0.1 μM Iber, 0.01 μM Mezi, or DMSO) for 24 hours before harvesting for either western blotting or RNA extraction. (A) Immunoblotting results for the neosubstrate protein Aiolos in the investigated cell lines after IMiD/CELMoD treatment. For each cell line, protein-level measurements for all treatments are calculated as fold change normalized to the corresponding DMSO-treated control. The color of the title for each blot denotes the effect of corresponding mutation on protein levels, where green is associated with expression comparable with parental MM1.s (shown in lower right corner), red denotes no reduction in protein levels, and orange highlights the mutations whose presence leads to partial reduction in protein levels. Blots shown are representative of 3 biological replicates. (B) Optical densitometry quantification of the immunoblotting results shown in Figure 3A for the 3 biological replicates. Data are shown as mean ± SEM. (C) Quantitative reverse transcription polymerase chain reaction results for the messenger RNA expression levels of transcription factor IRF4 after IMiD/CELMoD treatment. Data are shown as mean ± SEM of n = 3 biological repeats.

Figure 4.

Compilation of the in vitro results and structural analysis of the CRBN mutations that demonstrated partial impact on CRBN function. (A) Immunoblotting results for CRBN neosubstrates Ikaros and ZFP91 and for transcription factor MYC for all CRBN mutations that demonstrated partial CRBN function. The blots for p.W386A and CRBN^WT have been included as negative and positive control, respectively. Cells were drug treated (10 μM Len, 1 μM Pom, 0.1 μM Iber, 0.01 μM Mezi, or DMSO) and harvested 24 hours after IMiD/CELMoD treatment for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)/western blotting. Left panel: data shown are representative of 3 biological repeats. Right panel: optical densitometry quantification of the 3 biological replicates. (B) GI50 cell viability results, displayed per compound, for all cell lines whose CRBN mutation demonstrated partial CRBN function in comparison with p.W386A (negative control) and CRBN^WT (positive control). All cell lines were treated with DMSO or increasing concentrations of Len (maximum concentration at 20 μM), Pom (maximum concentration 8 μM), Iber (maximum concentration 2 μM), and Mezi (maximum concentration 1 μM) for 5 days. Cell viability was determined using the CellTiter-Blue assay. Results are the average ± SEM of n = 3 biological replicates. (C-F) Spatial localization of the CRBN mutations that demonstrated an intermediate/variable effect on CRBN function with respect to the IMiDs and neosubstrate-binding sites. The presented figure is a composite figure generated by overlaying our high-resolution crystallographic CRBN/DDB1 structure bound to Len (where CRBN is present in its closed conformation, PDB code 9JFX), with the publicly available crystallographic structure of neosubstrate Ikaros bound to CRBN/DDB1/Pom complex (where CRBN is present in its open conformation, PDB code 6H0F¹²). CRBN is represented in gold, Ikaros in orange, DDB1 in gray, Len in cyan, the 3 Trp defining the IMiD’s binding pocket in light blue, Zn²⁺ atoms as gray spheres, and the position of the mutations in magenta. Figures were generated with Pymol. (C) Localization of CRBN Cys326, mutated to Gly in patients. (D) Localization of CRBN Pro352, mutated to Ser in patients. (E) Localization of CRBN Cys366, mutated to Tyr in patients. (F) Localization of CRBN Phe381, mutated to Ser in patients.

Figure 4.

Compilation of the in vitro results and structural analysis of the CRBN mutations that demonstrated partial impact on CRBN function. (A) Immunoblotting results for CRBN neosubstrates Ikaros and ZFP91 and for transcription factor MYC for all CRBN mutations that demonstrated partial CRBN function. The blots for p.W386A and CRBN^WT have been included as negative and positive control, respectively. Cells were drug treated (10 μM Len, 1 μM Pom, 0.1 μM Iber, 0.01 μM Mezi, or DMSO) and harvested 24 hours after IMiD/CELMoD treatment for sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE)/western blotting. Left panel: data shown are representative of 3 biological repeats. Right panel: optical densitometry quantification of the 3 biological replicates. (B) GI50 cell viability results, displayed per compound, for all cell lines whose CRBN mutation demonstrated partial CRBN function in comparison with p.W386A (negative control) and CRBN^WT (positive control). All cell lines were treated with DMSO or increasing concentrations of Len (maximum concentration at 20 μM), Pom (maximum concentration 8 μM), Iber (maximum concentration 2 μM), and Mezi (maximum concentration 1 μM) for 5 days. Cell viability was determined using the CellTiter-Blue assay. Results are the average ± SEM of n = 3 biological replicates. (C-F) Spatial localization of the CRBN mutations that demonstrated an intermediate/variable effect on CRBN function with respect to the IMiDs and neosubstrate-binding sites. The presented figure is a composite figure generated by overlaying our high-resolution crystallographic CRBN/DDB1 structure bound to Len (where CRBN is present in its closed conformation, PDB code 9JFX), with the publicly available crystallographic structure of neosubstrate Ikaros bound to CRBN/DDB1/Pom complex (where CRBN is present in its open conformation, PDB code 6H0F¹²). CRBN is represented in gold, Ikaros in orange, DDB1 in gray, Len in cyan, the 3 Trp defining the IMiD’s binding pocket in light blue, Zn²⁺ atoms as gray spheres, and the position of the mutations in magenta. Figures were generated with Pymol. (C) Localization of CRBN Cys326, mutated to Gly in patients. (D) Localization of CRBN Pro352, mutated to Ser in patients. (E) Localization of CRBN Cys366, mutated to Tyr in patients. (F) Localization of CRBN Phe381, mutated to Ser in patients.

Figure 5.

The effect of CRBN mutations in a heterozygous model. (A) GI50 cell viability measurements for the CRBN^WT/WT cell line (WT plasmid overexpression) and each cell line with heterozygous CRBN (mutated CRBN overexpression). The corresponding CRBN^KO clone 1 homozygous mutation GI50 curve from Figure 2A has been added to each graph (gray) only to aid in the visual comparison between heterozygous and homozygous mutations. The color of the title for each graph denotes the effect on cell viability associated with each mutation in its homozygous state, where green is associated with reduction in cell viability comparable with parental MM1.s, red denotes no reduction in cell viability, and orange highlights the mutations whose presence leads to partial reduction in cell viability. Cells were treated with DMSO or increasing concentrations of Pom (maximum concentration 8 μM) and Mezi (maximum concentration 1 μM) for 5 days. Cell viability was determined using the CellTiter-Blue assay and expressed as a % of DMSO control. Results are the mean ± SEM of n = 3 biological replicates.

Figure 6.

Clashes predicted between the IKZF1 degron and Len bound to WT and p.P352S CRBN constructs. (A) Unfavorable contacts shown as orange dashes between Len bound to WT CRBN (C atoms shown in light blue with proline 352 highlighted) and IKZF1 degron atoms from PDB 6h0f (C atoms in yellow with glycine 151 highlighted) after structural alignment on CRBN-binding site residues. (B) A larger number of clashes is predicted between the IKZF1 degron and Len bound to the p.P352S CRBN construct (representative conformation from MD; C atoms in orange with serine 352 highlighted). (C) Distribution of clash volumes calculated between heavy atoms of the IKZF1 degron and Len in MD trajectory frames of the WT (blue box), p.P352S (orange), and p.F381S (green) CRBN constructs.

Figure 7.

Analysis of conformational changes predicted by MD simulations of the p.F381S CRBN construct. (A) Conformations of residues located in the surroundings of the p.F381S mutation are shown in the representative conformation from MD simulations of the mutant construct (C atoms in green) compared with the crystal structure of WT CRBN (C atoms in light blue). (B) A similar comparison focused on Len binding site shows smaller deviations mainly for residues exposed to the solvent (eg, H357, H397, W400). (C) Distribution of RMSD values calculated on MD simulations of the p.F381S construct for residues located in both binding site and mutation site (5 Å from Len and the mutated residue, respectively).