Combination therapy targeting Erk1/2 and CDK4/6i in relapsed refractory multiple myeloma

Abstract

Oncogenic activated RAS mutations have been detected in 50% of de novo and 70% of relapsed multiple myeloma (MM) patients. Translocation t(11;14) involving IgH/CCDN1 and overexpression of cyclin-Ds are early events in MM pathogenesis, enhancing uncontrolled MM cell growth. We hypothesized that targeting both RAS/MAPK pathway molecules including Erk1/2 along with cyclin-Ds enhances MM cytotoxicity and minimizes side effects. Recent studies have demonstrated the high potency of Erk1/2 and CDK4/6 inhibitors in metastatic relapsed cancers, and here we tested anti-MM effects of the Erk1/2 + CDK4/6 inhibitor combination. Our studies showed strong synergistic (IC < 0.5) cytotoxicity of Erk1/2i + CDK4/6i in MM-cells. Erk1/2i + CDK4/6i treatment in a dose-dependent manner arrested MM-cells in the G0/G1 phase and activated mitochondrial apoptotic signaling. Our studies showed that Erk1/2i + CDK4/6i treatment-induced inhibition of key target molecules in Erk1/2 and CDK4/6 signaling, such as c-myc, p-RSK, p-S6, p-RB, and E2F1, suggesting on-target activity of these inhibitors. We identified Erk1/2i + CDK4/6i treatment associated five-gene signature which includes SNRPB and SLC25A5; these genes are involved in RNA processing and mitochondrial metabolism, respectively. Overall, our studies provide the preclinical framework for Erk1/2i + CDK4/6i combination clinical trials to target Ras+CDK pathways to improve patient outcome in MM.

Fig. 1. Effects of Erk1/2i + CDK4/6i on MM cell lines and HD-PBMC.

A Box plots showing ERK1/2 and CDK4/6 gene expression values in CD138 + PC from monoclonal gammopathy of undetermined significance (MGUS; n = 11), multiple myeloma (MM; n = 133), plasma cell leukemia (n = 9; PCL), and NPC (healthy donors), from dataset GSE13591. The error bars show the standard deviation. Horizontal bars indicate the mean value. The x-axis shows the samples analyzed, and the y-axis displays the expressions at log2 fold. These analyses confirmed significant (0.05 > P < 0.3E-0.07) overexpression of ERK1/2, CDK4 and CDK6 transcripts in clonal PCs. B Efficacy studies in MM cell lines (H929, MM.1Smch (mCherry (red fluorescent protein)), MM1R, AMO1, KMS12, RPMI8226, OPM2, and U266) and HD PBMCs at different time points after LY3214996 (Erk1/2i) and LY2835219 (CDK4/6i) treatment at different concentration ranges are shown on the x-axis; Cell viability (means and standard deviations) is shown as percentage of control untreated cells. C Determination inhibitory potency (IC50) of Erk1/2i for MM. MM (H929-NRAS, MM.1S, RPMI8226. OPM2 and U266) cell lines were treated with DMSO or ERK1/2i (0-15uM MM cell lines in RPMI medium with 0.1% DMSO and 10% FBS for 24–72 h. The concentration of drug that caused 50% inhibition of cell proliferation (IC50) relative to DMSO control was determined by non-linear regression using Prism, version 8.

Fig. 2. Synergistic cytotoxic effects of Erk1/2i + CDK4/6i on MM-cells.

A Schematic diagram of 9×9 checker-board matrix of ERK1/2i and CDK4/6i inhibitor concentrations in cell viability assays performed at 48 h (B) and 72 h (C) in MM1S, H929, and U266 cells co-cultured with BMSC-conditioned medium. CalcuSyn software was used to assess the synergistic effects of ERK1/2i + CDK4/6i. Calcusyn combination index (CI) is interpreted as follows: CI < 0.8 indicates synergism; CI < 0.55 indicates strong synergism; 0.6 < CI < 0.8 indicates moderate synergism; 0.9 < CI < 1.1 indicates additive effects; 1.2 < CI < 1.9 indicates moderate antagonistic effects. ED50, ED75, ED90 = effective dose at 50%, 75%, and 90%, respectively. H929-NRAS, MM1S-KRAS, and U266 WT cells were treated with ERK1/2i and CDK4/6i. These data demonstrate highly potent cytotoxicity of Erk1/2i + CDK4/6i combination 48 h after treatment. This highly synergistic effect of Erk1/2i + CDK4/6i was retained 72 h after treatment.

Fig. 3. Erk1/2i + CDK4/6i treatment triggers cell-cycle arrest and induces apoptosis.

Erk1/2i + CDK4/6i induced cellular effects were evaluated in H929, MM 1S, and U266 cells co-cultured with BMSC-conditioned medium after 48 h treatment with ERK1/2i (0; 0.9; 3 and 6 μM) and CDK4/6i (0; 0.1; 1 and 3uM), alone or in combination. A1 Cell cycle profiles were analyzed using standard DAPI staining. MM-cells in G0/G1, S, and G2/M phases were measured on the BD Fortessa X-20, followed by analysis using ModFit LT software. B1 ERK1/2i + CDK4/6i effects on apoptosis were quantified using annexin V-fluorescein isothiocyanate and 7-AAD staining and flow cytometry; percentages of early apoptotic (early A; Annexin V+/7-AAD−) and late apoptotic (late A, Annexin V+/7-AAD+) events were identified using FlowJo V10 software. A2 and B2 Protein lysates were obtained after treatment of MM-cells with Erk1/2i or CDK4/6i or Erk1/2i + CDK4/6i. Cell lysates were subjected to immunoblotting using the antibodies indicated; GAPDH served as a loading control. A3 Densitometry analysis of protein bands was performed using the Image J software. The color key next to each heatmap shows the protein expression level compared to DMSO. On the heatmaps, protein expression levels are indicated with intensity shades of black color. These analyses showed a dose-dependent overexpression of a cell cycle inhibitor p27 protein (A2), and increased PARP cleavage (B2), the signature of cell death. B3 Caspase (CAS) 3/7 activation in MM cell lines was measured by flow cytometry in response to Erk1/2i, CDK4/6i, and Erk1/2i + CDK4/6i treatment. Relative CAS3/7 activation is summarized as heatmaps; mean fluorescent intensities (MFI) are calculated in comparison with MFI of controls. Each treatment with a specific concentration of Erk1/2i and/or CDK4/6i was done in duplicate, plus/minus the standard error of the mean has been hidden for simplicity. C BH3 profiling to measure early changes in net pro-apoptotic signaling of mitochondria in response to ERK1/2i and CDK4/6i, alone or in combination. C1 Interaction map for BH3 peptides and BH3 mimetics with BCL-2 family proteins. Darker green color indicates Kd<100 nM determined by fluorescence polarization. C2, C3 Heatmap of delta priming response, assessed by dynamic BH3 profiling, in MM cell lines plated and treated for 16 h with ERK1/2i and CDK4/6i, alone or combination. Delta priming = % cytochrome c loss_(drug) − % cytochrome c loss _(DMSO). A, B, C Results are summarized as heatmaps; color-keys are shown next to the each heatmap for data interpretation. Expression levels are indicated with intensity shades of green color.

Fig. 4. Effects of ERK1/2i + CDK4/6i on gene and protein expression in MM cell lines.

H929, MM 1S, and U266 cells co-cultured with BMSCs were treated for 48 h with ERK1/2i (Ei) (0; 0.9; 3 and 6 μM) and CDK4/6i (Ci) (0; 0.1, 1, and 3 μM), alone or combination (Ei + Ci). MM-cells were then separated from BMSC using EasySep magnetic bead cell separation method. A Total mRNAs were isolated from MM1S and H929 cells and transcribed into cDNAs. Gene expression profiling was performed using custom TaqMan Gene Expression Array Plates containing primers and probes for detection of RAS and cell cycle signaling pathway genes. Results were analyzed using the relative standard curve method, and final data analyses and heatmaps were generated based on dCt values using Partek Genomic Suite. Expression levels are indicated with intensity shades of green/red colors. GAPDH and TBP genes were used as housekeeping genes. B Protein lysates (20–40 μg of protein/lane) of H929 and MM1S cells were separated on SDS-PAGE, and after blotting onto nitrocellulose membranes were probed with anti- Erk1/2, -p-ERK1/2, -c-myc, -p-S6, p-RSK, -p-RB, and -E2F1 antibodies; GAPDH served as a loading control. C Densitometry analysis of protein bands was performed using the Image J software. The color key next to each heatmap shows the protein expression level compared to DMSO. On the heatmaps, protein expression levels are indicated with intensity shades of black color.

Fig. 5. Erk1/2i + CDK4/6i reduces tumor burden in an MM1.S-Luc⁺ xenograft in vivo model.

A We specifically used this model in which MM1S cells home to the mouse BM, mimicking the tumor microenvironment of human MM. BLI images of three representative mice bearing MM1.S-Luc⁺ tumors from each indicated group. BLI images showing localization MM1.S-Luc⁺ tumor cells to hind limb BM. The scale represents luminescence signal from MM1.S-Luc⁺ cells. B BLI imaging quantification of tumor burden for monitoring tumor progression; data are presented as mean values ± SD (n = 7-8 animal/group); P values were calculated using Log-rank (Mantel-Cox) test in Prism V9. C Total mRNAs were isolated from bone marrow cells obtained from flushed femurs of animals treated with Erk1/2i (Ei) or CDK4/6i (Ci), alone or combination (Ei + Ci); then these samples were transcribed into cDNAs and subjected to the RAS and cell cycle signaling pathway gene-expression profiling using custom TaqMan Gene Expression Arrays. Each of the TaqMan assays was individually evaluated as a single reaction. The TaqMan assays that specifically recognize human mRNA (polyadenylated transcripts encoding proteins) were included in the final screening arrays. Results were analyzed using the relative standard curve method and presented as heatmaps (Partek Genomic Suite). Gene-expression levels (dCt values) are indicated with green/red color intensity. GAPDH/TBP was used as a housekeeping gene. These analyses showed significant downregulation of RAS and CDK4/6 signaling pathway genes in samples treated with Erk1/2i + CDK4/6i.

Fig. 6. Effects of Erk1/2i and CDK4/6i on primary MM-cells (A 1, B1, B3).

Dose-response effects of Erk1/2i (Ei), CDK4/6i (Ci), or Erk1/2i + CDK4/6i (Ei + Ci) treatment on MM-cells freshly obtained from MM BM aspirates of patients are summarized as heatmaps. CD138 + PCs selected from MM patient BM using EasySep cell separation, with purity >95% (confirmed using flow cytometry), were treated with Erk1/2i, CDK4/6i, or Erk1/2i + CDK4/6i at the concentrations indicated. On the figures A1, B1 the results are shown as percent (%) of cell death relative to DMSO controls. The % cell death is indicated with intensity shades of green color scale shown at the right edge of each heatmap. Each concentration of Erk1/2i and/or CDK4/6i was tested in triplicate. The results, plus or minus the standard error of the mean, ranged from +0.01 to +0.7, and have been hidden for simplicity. A2 CD138 + PC were collected 48 h after treatment and cell lysates were prepared to measure phosphorylation/total levels of Erk1/2 using InstantOne enzyme-linked immunosorbent assay kits, according to the manufacturer’s protocol. Absorbance was measured at 450 nm. The results are presented as bar graphs. The x-axis shows the samples analyzed, and the y-axis displays the phosphorylation/total Erk1/2 levels as an absorbance. Results obtained from positive and negative control samples are not displayed on the graph. The mean ± S.D. is shown for two independent experiments. A3 Caspase (CAS) 3/7 activation in MM CD138 + PCs from Pt 1 and Pt 2 were measured by flow cytometry in response to ERK1/2i and CDK4/6i treatment, alone or in combination. In this experiment, unfractionated BM cells from MM patients were used, and CAS3/7 activation was measured in MM CD138 gated PCs. The results are summarized as heatmaps; the percentage of mean fluorescent intensity (MFI) was calculated and compared with the MFI of the controls, and presented in shades of red scale shown on the right side of each heatmap. Each treatment with a specific concentration of Erk1/2i and/or CDK4/6i was done in duplicate, plus/minus standard error (+0.01 to +0.75) of the mean MFI have been hidden for simplicity. B2 Functional effects of ERK1/2i and CDK4/6i or ERK1/2i + CDK4/6i treatments were validated at the protein level by immunoblotting. Cell lysates were prepared 2D after treatment from CD138 + PCs selected using a CD138 positive magnetic bead selection method (Stem Cell Technology).

Fig. 7. Effects of ERK1/2i + CDK4/6i on primary MM-cells co-cultured with or without BMSC-CM.

This cohort includes two patients with NRAS mutations, three patients with KRAS mutations, five patients with RAS-WT, and two patients with the t(1;14). A CD138 + PCs were cultured with or without BMSC-CM and treated for 3D and 5D. MM PCs were maintained in 15% fetal calf serum spiked with the serum obtained from autologous peripheral blood of MM patients. Cells were treated with Erk1/2i (Ei), CDK4/6i (Ci), or Erk1/2i + CDK4/6i (Ei + Ci) at the concentrations indicated. The % cell death is indicated with intensity shades of green colors, and the color scale is shown at the right edge of each heatmap. Each concentration of Erk1/2i and/or CDK4/6i was tested in triplicate. The results, plus or minus the standard error of the mean, ranged from +0.01 to +0.7. B Erk1/2i, CDK4/6i, or Erk1/2i + CDK4/6i effects on apoptosis in MM BM PCs were quantified using Annexin V + 7-AAD staining. Unfractionated BM cells were treated at the indicated concentrations, and cell death was monitored 3D after treatment; % of apoptotic events were detected in CD138 gated PCs using FlowJo V10 software. Flow cytometry data that illustrates the gating strategy for the apoptosis assay is shown in supplementary Fig. S5.

Fig. 8. Gene signature-based approach identified targets associated with Erk1/2i + CDK4/6i treatment.

A H929 MM-cells were transfected with either scrambled control (SC) or Erk1/2-siRNA (Erk1/2-siR); Erk1/2 knockdown efficiency was evaluated at mRNA (A1) and at protein (A2) levels. Erk1/2 knockdown H929 cells were treated with Erk1/2i, CDK4/6i, or Erk1/2i + CDK4/6i; after 48 h of treatment cell lysates were collected (A3). Protein lysates A2, A3 were subjected to immunoblotting using Erk1/2, p-ERK1/2, pRSK, and E2F1 antibodies; GAPDH served as a loading control. Bar graphs in A2 and A3 show a densitometric analysis of the protein bands measured by ImageJ software. Fold expressions on the Y-axis show the protein expression level compared to loading controls. B A heatmap shows unsupervised clustering of deregulated genes detected in H929 cells, with or without Erk1/2 knockdown or treated with Erk1/2i and CDK4/6i, alone or in combinations. The color scale for log expression values is shown at the bottom of the heatmaps, while sample clustering is presented as a dendrogram on the top. C Transcriptome changes in these cells are shown by principal component analysis (PCA). On PCA plot C1, different groups are presented in different colors; on PCA plot C2, cells treated with different concentrations of the drugs or drug combinations are presented with different colors. D Gene and functional enrichment analyses of Erk1/2i + CDK4/6i signature gene-sets (G-set-1 (D1) and G-set-2 (D2)) were performed using DAVID v6.8. G-set-1 comprises the genes that are upregulated in response to Erk1/2 knockdown (G-set-1); G-set-2 includes genes that were downregulated in response to Erk1/2i + CDK4/6i treatment. Enrichment analyses considered p-value enrichment with p < 0.05, and fold enrichment with a p > 5. Bar plots display the −log10 p value enrichments. E1 The Venn method was used to identify commonly deregulated signature genes associated with Erk1/2i + CDK4/6i treatment. The left Venn diagram–intersecting deregulated G-set-1 with deregulated genes in MM Cohort-1(GSD4968), and Cohort-2(GSE5900, GSE2658); the right Venn diagram-intersecting deregulated G-set-2 with deregulated genes in MM Cohort-1/2. E2 Gene and functional enrichment analyses of commonly deregulated Erk1/2i + CDK4/6i signature genes identified in MM cohort-1/2. Enrichment analyses considered p value enrichment with p < 0.05, and fold enrichment with a p > 5. Bar plots display the −log10 p value enrichments. F Identification five gene-signature associated with Erk1/2i + CDK4/6i. The Venn diagram–intersecting deregulated G-set-1and G-set-2 with deregulated genes in MM Cohort-1/2. G Violin plots showing gene expression values in CD138 + PC from MGUS (n = 44), sMM (n = 12), MM (n = 559), and HD (healthy donor, n = 22), from GSE5900 and GSE2658. The x-axis shows the samples analyzed; violin plots are colored by sample type. The y-axis displays the expressions at log2 fold. Significance between groups was evaluated using a Wilcoxon Rank-Sum test; the type I error cut off was 5%. Multiple testing adjustments between groups were then made using the Bonferroni correction. These analyses show significant overexpression of SNRPB and SLC25A5 transcripts in clonal PCs. H The relevance of SNRPB and SLC25A5 expression to clinical outcome was examined in 559 MM patient samples. The samples were classified based on SNRPB and SLC25A5 expression levels. Survival curves were estimated using the Kaplan-Meier method; the type I error cut off was 5%. differences in survival were assessed using the log-rank test and Cox regression models. For analyses R version 4.0.0, along with the survival and survminer packages, were used. >quantile = >75th percentile; <quantile = <25th percentile.