Validation of myeloma survival kinases. (A) Top-ranked kinases targeted by lethal siRNA during primary high-throughput screens were tested in KMS11 cells using 4 unique siRNAs per kinase, assayed individually in separate wells, to verify survival kinase identification and to exclude kinases falsely identified during primary screening by chance or by off-target siRNA effects. As not all siRNAs efficiently silence their target, only the top 3 of 4 siRNAs tested are shown. The effects on viability of siRNAs are shown in arbitrary relative units, with UBB siRNA inducing approximately 100% cell death. To exclude false positives, only kinases whose targeting yielded concordant lethal results from ≥ 3 of 4 unique siRNAs were considered validated survival kinases. For validated survival kinases, the median viability loss induced by all siRNAs tested is shown as a bar. (B) XY plot of kinase gene mRNA expression in KMS11 and INA6 versus kinase siRNA lethality during primary kinome screening. Kinase mRNA levels were determined by Affymetrix (U133_2+ chip) expression profiling and represent the mean signal intensity (MAS5) of all probe sets per gene. Validated survival kinases are labeled.

Expression of selected myeloma survival kinases in primary multiple myeloma versus an atlas of nonhematopoietic human tissues. Histograms show the differential patterns of expression of myeloma survival kinases (GRK6, HK1, and FGFR3) and of control genes (CD138 and β-actin) in primary human somatic tissues (103 arrays), primary multiple myeloma tumor cells (115 arrays), and in early stage plasma cell disorders: monoclonal gammopathy of uncertain significant (MGUS) and smoldering multiple myeloma (SMM; data from SymAtlas, Novartis Research Foundation and from Mayo Clinic; GEO accession no. GSE 6477). Array-based gene expression data were normalized per chip (sample) to the median signal intensity of expressed genes (genes with present detection call and/or MAS5 signal intensity > 200); a value of 1.0 thus represents the median expression intensity of genes whose expression can be reliably detected within each sample type.

Cytotoxic effects of GF109203X, a dual inhibitor of protein kinase C (IC₅₀, ∼ 0.1μM) and of GRK6 (IC₅₀, ∼ 1-10μM), on myeloma cells versus nonmyeloma cells. (A) Cytotoxicity of 20μM GFX109203X at 72 hours is summarized for a spectrum of human myeloma and nonmyeloma cells, with statistical comparison by Mann-Whitney test. (B) An illustrative example of selective cytotoxic effect of GF109203X versus primary myeloma cells in mixed culture. Primary bone marrow cells from myeloma patients were treated in culture with 1.25 to 20μM GF109203X, or with dimethyl sulfoxide (DMSO) vehicle control. After 72 hours, samples were stained with CD138–fluorescein isothiocyanate to identify myeloma cells (bottom right population) versus other marrow cells (bottom left population) and propidium iodide to distinguish late-stage apoptotic cells. At high concentrations (reflective of GRK6 rather than protein kinase C effect), GF109203X induced selective cytotoxicity that was confined to the tumor cell population, evidenced by CD138 shedding and propidium iodide uptake.

Kinome-wide siRNA lethality screening in myeloma cells. Prevalidated siRNA directed against 639 human kinases and associated subunits were transfected into (A) KMS11 and (B) IL6-dependent INA6 human multiple myeloma cells in high-throughput screening experiments using 2 siRNAs per target tested in a single-siRNA-per-well format. Cell viability was determined at 96 hours by adenosine triphosphate concentration. siRNA-induced changes in viability are shown, measured in units of standard deviation (B-score) from the median of noninhibitory siRNA, plotted in the order in which siRNAs were screened. Cells treated with transfection reagent alone, GFP siRNA, or 1 of 2 unique nonsilencing (NS) scrambled siRNAs represent negative controls. Positive controls for transfection efficiency were included at regular intervals throughout the screen and are cells treated with an siRNA targeting ubiquitin B. (C-D) Results of kinome siRNA studies in myeloma cells were highly reproducible, with significant correlation between duplicate studies performed at separate times on KMS11 and on INA6, respectively. (E-F) Operative transfection efficiency during primary kinome screening studies in (E) KMS11 and (F) INA6, measured by modulation of viability by control siRNA. The median result for each condition is shown. (G) Kinases in KMS11 and INA6 myeloma cells associated with at least 1 lethal siRNA (with B-score < −3) during kinome-scale screening; top-ranked kinases were forwarded to validation studies.

Selective vulnerability of myeloma survival kinases. (A-D) Effect on viability of siRNA directed against top-ranked myeloma survival kinases in (A) KMS11 and (B) JJN3 myeloma cells, and in (C) 293 embryonic kidney and (D) A549 lung carcinoma cells. Cells were transfected with 2 unique siRNAs per kinase in separate wells, using conditions optimized for each cell line each associated with > 90% transfection efficiency. Viability was assessed at 96 hours and is plotted in units of standard deviations from control siRNA. (E-F) Western blot analysis of kinase knockdown after siRNA targeting of (E) AURKA and (F) PKN1, confirming equivalent or greater transfection and protein knockdown in 293 and A549 cells, compared with KMS11 myeloma cells. Cells were transfected using the same conditions as panels A through D optimized for each line, and lysates were prepared at 48 hours; 10 μg of total protein was loaded per lane.

GRK6 kinase silencing induces selective apoptosis of myeloma cells. (A) KMS11 myeloma cells were infected with lentivirus driving the expression of 3xFLAG-tagged GRK6 (+), or with control lentivirus (−), and were assessed by anti-FLAG immunoblot (left panel) and immunoprecipitation (right panel), verifying abundant tagged GRK6 protein production. (B) KMS11 cells stably infected with 3xFLAG-GRK6 expressing LV, or with control LV, were treated with Lipofectamine 2000 (L2K) transfection reagent and control or GRK6 3′UTR siRNA. GRK6 3′UTR siRNA induced native GRK6 mRNA suppression in control cells, but did not suppress 3xFLAG-GRK6 (which lacks GRK6 3′ UTR sequence), as shown by QRT-PCR at 48 hours (right panel). The effect of GRK6 3′ siRNA on KMS11 viability at 96 hours was assessed in parallel by MTT assay (left panel). Ectopic 3xFLAG-GRK6 rescues KMS11 from GRK6 3′UTR siRNA-induced lethality, verifying that the siRNA's lethality is due to on-target GRK6 silencing. Both KMS11 lentiviral lines were tested with NS and UBB siRNA, revealing equivalent transfection efficiency and minimal L2K toxicity. (C) The lethality of GRK6 silencing in KMS11 myeloma cells was further verified by 4 GRK6 siRNAs (GRK6-2, -9, -3, -C) that target GRK6 exons: after transfection of KMS11, GRK6 mRNA knockdown was assessed by QRT-PCR at 48 hours and viability was assessed by annexin V + propidium iodide binding and by MTT assay (at 96 hours). Cellular inhibition was due largely to apoptosis. (D) Lentiviruses nos. 66 to 68 expressing distinct short hairpin RNA (shRNA) targeting GRK6 were screened for induction of GRK6 knockdown in KMS11 by immunoblot at 48 hours. LV no. 66 and no. 68 both induced GRK6 knockdown in contrast to no. 67 and control LV expressing nonsilencing (NS) shRNA. LV no. 66 induced similar suppression of GRK6 in MCF7 and SF767 cells. (E) OPM1 myeloma cells infected with GFP-expressing LV no. 66 were mixed 1:1 with OPM1 cells infected with control LV. Cells in which GRK6 was targeted by shRNA no. 66 were progressively eliminated from culture, as determined by serial flow cytometry for the GFP marker, despite the presence of healthy control OPM1, which became enriched. (F) Comparison of the effect of GRK6-shRNA no. 66 on the viability of myeloma (KMS11 and OPM1) and nonmyeloma (MCF7 and SF767) cell lines. Cells were infected in parallel with equal titer LV producing NS-shRNA, GRK6-shRNA no. 66, or GFP cDNA; percentage cellular infection was determined by GFP expression. Specific apoptosis induction by GRK6 silencing at 96 hours was determined by annexin V and 7-amino-actinomycin D binding, comparing control LV (producing NS-shRNA) with LV GRK6-shRNA no. 66. (G) The effect of GRK6 inhibition on cellular viability is summarized for a spectrum of human myeloma and nonmyeloma cells. Results were obtained as shown in panel F and are plotted as the GRK6-shRNA attributable apoptosis normalized to the percentage infection obtained with each cell line. Error bars are calculated from 5% absolute error in assessments of infection and apoptosis.

GRK6 binds to and is regulated by the chaperone HSP90 in human myeloma cells and is required for phospho-STAT3 pathway signaling. (A) 3xFlag-tagged GRK6 protein complexes were affinity purified from 10⁸ KMS11 myeloma cells using M2 antibody and separated by SDS-PAGE. A single GRK6 copurified protein (*) was detected by Coomassie staining, migrating at 90 kDa, and was subsequently identified by nanoLC-ESI-MS/MS as HSP90A/B. Additional immunoprecipitation (IP) Coomassie staining bands reflect M2 antibody fragments and “sticky” lysate proteins retained nonspecifically by sepharose. The identity of isolated 3xFLAG-GRK6 was confirmed by nanoLC-ESI-MS/MS. (B-C) HSP90 potently and rapidly (< 24 hours) regulates GRK6 protein levels in KMS11 (B) and OPM1 (C) myeloma cells. Both myeloma tumor lines were treated with the HSP90 inhibitor, geldanamycin, 1μM, for the specified times, and were then assayed for ensuing effects on GRK6 protein expression by immunoblot. (D) To assess the effect and kinetics of GRK6 inhibition on the activity of myeloma survival pathways, FLAG-GRK6–expressing OPM1 cells were infected with control lentivirus expressing nontargeting shRNA, or with lentivirus expressing GRK6 shRNA, and the phospho-activation status of key intracellular signaling mediators AKT, MAPK p42 and p44 (ERK1/2), JNK, and STAT3 was determined by immunoassay sequentially from 24 to 96 hours after viral shRNA infection. MCL1 phosphorylation and levels were examined in parallel.