Activation of Serum/Glucocorticoid Regulated Kinase 1/Nuclear Factor-κB Pathway Are Correlated with Low Sensitivity to Bortezomib and Ixazomib in Resistant Multiple Myeloma Cells

Multiple myeloma (MM) is an incurable malignancy often associated with primary and acquired resistance to therapeutic agents, such as proteasome inhibitors. However, the mechanisms underlying the proteasome inhibitor resistance are poorly understood. Here, we elucidate the mechanism of primary resistance to bortezomib and ixazomib in the MM cell lines, KMS-20, KMS-26, and KMS-28BM. We find that low bortezomib and ixazomib concentrations induce cell death in KMS-26 and KMS-28BM cells. However, high bortezomib and ixazomib concentrations induce cell death only in KMS-20 cells. During Gene Expression Omnibus analysis, KMS-20 cells exhibit high levels of expression of various genes, including anti-phospho-fibroblast growth factor receptor 1 (FGFR1), chemokine receptor type (CCR2), and serum and glucocorticoid regulated kinase (SGK)1. The SGK1 inhibitor enhances the cytotoxic effects of bortezomib and ixazomib; however, FGFR1 and CCR2 inhibitors do not show such effect in KMS-20 cells. Moreover, SGK1 activation induces the phosphorylation of NF-κB p65, and an NF-κB inhibitor enhances the sensitivity of KMS-20 cells to bortezomib and ixazomib. Additionally, high levels of expression of SGK1 and NF-κB p65 is associated with a low sensitivity to bortezomib and a poor prognosis in MM patients. These results indicate that the activation of the SGK1/NF-κB pathway correlates with a low sensitivity to bortezomib and ixazomib, and a combination of bortezomib and ixazomib with an SGK1 or NF-κB inhibitor may be involved in the treatment of MM via activation of the SGK1/NF-κB pathway.


Introduction
Multiple myeloma (MM), the second most common hematological cancer, is defined as the monoclonal proliferation of plasma cells in bone marrow and the discharge of monoclonal immunoglobulins [1]. The incidence of MM increased worldwide from 1990 to 2016. The 5-year survival rate for patients with MM is now 50%, owing to the development of novel therapeutic agents, such as immunomodulatory drugs, histone deacetylase inhibitors, proteasome inhibitors, and monoclonal antibodies [1,2]. However, in a fairly large proportion of cases, MM is an incurable malignancy because it often presents with primary and acquired resistance to therapeutic agents [3]. Thus, it is important to elucidate the primary and acquired resistance mechanisms of MM cells.

Proteasome Activity Assay
Proteasome β5 subunit activity was assessed as described previously [28].

Gene Expression Omnibus (GEO) Data Set
The gene expression profiles of the microarray datasets with the accession numbers GSE6205 and GSE9782 were obtained from the National Center of Biotechnology Information (NCBI) Gene Expression Omnibus (GEO) database (http://www.ncbi.nlm.nih.gov/ geo/). The expression of various genes in KMS-26, KMS-28BM, and KMS-20 cells was analyzed from GSE6205 and the expression of anti-phospho-serum and glucocorticoid regulated kinase (SGK)1 and nuclear factor (NF)-κB p65, and the overall survival rate of patients with multiple myeloma (MM) were analyzed from GSE9782.

Statistical Analysis
All results are represented as the means and standard deviations (SDs) of several independent experiments. All analyses were conducted using SPSS version 21.0 software (IBM Inc., Chicago, IL, USA), and Shapiro-Wilk analysis and one-way analysis of variance (ANOVA) were performed. When no differences in the Shapiro-Wilk test and satisfactory differences in ANOVA were confirmed, the data from the control group and various drugtreated groups were compared and analyzed using Dunnett's test. When our data did not show normal distribution, they were analyzed using the Kruskal-Wallis test followed by the Steel test. Survival rates were assessed using Kaplan-Meier curves and long-rank analysis. p values less than 0.05 were deemed significant. Drug interactions were analyzed using the combination index (CI) based on the method described by Chou and Talalay [29]. A CI value of less than 1.0 indicates synergy, while a CI value greater than 1 indicates antagonism.

Overexpression of Anti-Phospho-Serum and Glucocorticoid-Regulated Kinase (SGK1) Correlated with Low Sensitivity of Bortezomib and Ixazomib in Multiple Myeloma (MM) Cells
To reveal differences in gene expression between KMS-20 cells and KMS-26 or KMS-28BM cells, we used the publicly available gene expression profiling (GEP) dataset, GSE6205. These analyses revealed that the expression of several genes was elevated only in KMS-20 cells. Above all, we focused on three genes; anti-phospho-fibroblast growth factor receptor 1 (FGFR1), a member of the growth factor receptor tyrosine kinase family;

Overexpression of Anti-Phospho-Serum and Glucocorticoid-Regulated Kinase (SGK1) Correlated with Low Sensitivity of Bortezomib and Ixazomib in Multiple Myeloma (MM) Cells
To reveal differences in gene expression between KMS-20 cells and KMS-26 or KMS-28BM cells, we used the publicly available gene expression profiling (GEP) dataset, GSE6205. These analyses revealed that the expression of several genes was elevated only in KMS-20 cells. Above all, we focused on three genes; anti-phospho-fibroblast growth factor receptor 1 (FGFR1), a member of the growth factor receptor tyrosine kinase family; C-C chemokine receptor type 2 (CCR2); and anti-phospho-serum and glucocorticoid regulated kinase (SGK1), a member of the serine/threonine kinase family (Table 1). The FGFR family includes FGFR1, FGFR2, FGFR3, and FGFR4, and binding of FGFR to fibroblast growth factor (FGF) induces the activation of signaling molecules, such as RAS/ERK, phosphoinositide 3-kinase (PI3K)/Akt, Janus kinase (JAK)/signal transducer and activator of transcription (STAT), and the phospho-anti-phospho-c-Jun N-terminal kinase (JNK) pathway [30]. Additionally, activation of FGFR by FGF2 stimulation was involved in prednisolone resistance in B cell precursor acute lymphoblastic leukemia cells [31]. It has been reported that FGFR3 overexpression is less sensitive to bortezomib in U266 cells [32]. The CC chemokine ligand (CCL2)-CCR2 axis is involved in multi-tyrosine kinase inhibition and microtubule inhibition resistance in the colon and prostate cancer cells [33,34], and CCR2 promotes cell growth and cell cycle progression via Src and Akt activation in breast cancer cells [35]. Activation of SGK1 by tongue cancer resistance-related protein 1 induces tamoxifen resistance [36], and a high expression of SGK1 correlates with a poor prognosis in patients treated with neoadjuvant chemotherapy and esophageal squamous cell carcinoma [37]. These findings suggest that overexpression of FGFR1, CCR2, or SGK1 is involved in resistance to various anti-cancer drugs. Thus, the overexpression of FGFR1, CCR2, and SGK1 may explain the low sensitivity of multiple myeloma (MM) cells to bortezomib and ixazomib. We investigated the expression of SGK1, CCR2, and FGFR1 mRNA and protein in KMS-26, KMS-28BM, and KMS-20 cells. The expression of SGK1, CCR2, and FGFR1 mRNA and total protein levels were elevated in KMS-20 cells compared to those in KMS-26 and KMS-28BM cells ( Figure 3A,B). Additionally, phosphorylated SGK1 and FGFR1 levels were increased in KMS-20 cells with an increase in total SGK1 and FGFR1 protein levels ( Figure 3B). Next, we examined whether the inhibition of these signal molecules by selective inhibitors enhanced the cytotoxicity of bortezomib and ixazomib in KMS-20 cells.
PD166866, an FGFR1 inhibitor, and a CCR2 antagonist did not affect cell death induced by bortezomib and ixazomib but GSK650394, an SGK1 inhibitor, enhanced the cytotoxicity of bortezomib and ixazomib to KMS-20 cells (p < 0.05) ( Figure 3C-E). Additionally, combined treatment with GSK650394 and bortezomib or ixazomib enhanced the Annexin V-positive cells compared to treatment with bortezomib, ixazomib, and GSK650394 alone (Supplementary Materials, Figure S1). Moreover, bortezomib and ixazomib enhanced the SGK1 total or phosphorylated protein in KMS-20 cells (Supplementary Materials, Figure S2). These results indicate that bortezomib and ixazomib enhance SGK1 overexpression/activation and that the SGK1 inhibitor overcame bortezomib-and ixazomib-resistance via suppression of SGK1 activation in KMS-20 cells.
To analyze whether the expression of SGK1 is correlated with the response to bortezomib and overall survival in patients with MM, we used the publicly available GEP dataset, GSE9782. Patients with high SGK1 expression had a lower sensitivity to bortezomib and a shorter overall survival than patients with low SGK1 expression (p < 0.01) ( Figure 3F,G). Thus, the overexpression and activation of SGK1 were associated with the low sensitivity of MM cells to bortezomib and ixazomib. by bortezomib and ixazomib but GSK650394, an SGK1 inhibitor, enhanced the cytotoxicity of bortezomib and ixazomib to KMS-20 cells (p < 0.05) ( Figure 3C-E). Additionally, combined treatment with GSK650394 and bortezomib or ixazomib enhanced the Annexin V-positive cells compared to treatment with bortezomib, ixazomib, and GSK650394 alone (Supplementary Materials, Figure S1). Moreover, bortezomib and ixazomib enhanced the SGK1 total or phosphorylated protein in KMS-20 cells (Supplementary Materials, Figure  S2). These results indicate that bortezomib and ixazomib enhance SGK1 overexpression/activation and that the SGK1 inhibitor overcame bortezomib-and ixazomib-resistance via suppression of SGK1 activation in KMS-20 cells.
To analyze whether the expression of SGK1 is correlated with the response to bortezomib and overall survival in patients with MM, we used the publicly available GEP dataset, GSE9782. Patients with high SGK1 expression had a lower sensitivity to bortezomib and a shorter overall survival than patients with low SGK1 expression (p < 0.01) ( Figure  3F,G). Thus, the overexpression and activation of SGK1 were associated with the low sensitivity of MM cells to bortezomib and ixazomib.
Moreover, we analyzed whether the expression of NF-κB p65 is associated with the response to bortezomib and the overall survival in patients with multiple myeloma (MM) using the GSE9782 dataset. High expression of NF-κB p65 in patients with MM led to a low sensitivity to bortezomib and a shorter overall survival than low expression of NF-κB p65 in patients (p < 0.01) ( Figure 4E,F). Thus, both the overexpression and activation of NF-κB p65 were associated with the low sensitivity of MM cells to bortezomib and ixazomib.

Dimethyl Fumarate (DMF), a Nuclear Factor (NF)-κB Inhibitor, Enhanced the Sensitivity of KMS-20 Cells to Bortezomib and Ixazomib
Our results suggest that a low sensitivity to bortezomib and ixazomib is correlated with the activation of nuclear factor (NF)-κB by serum/glucocorticoid regulated kinase 1 (SGK1). We examined whether dimethyl fumarate (DMF), an NF-κB inhibitor, enhances the sensitivity of KMS-20 cells to bortezomib and ixazomib. The combination of DMF and bortezomib or ixazomib enhanced the sensitivity of KMS-20 cells to bortezomib and ixazomib (p < 0.05) ( Figure 5A,B and Supplementary Materials, Figure S6). Additionally, DMF suppressed B-cell lymphoma-2 (Bcl-2), Bcl-xL, and survivin expression and enhanced Bcl-2-like protein 11 (Bim) expression via the inhibition of NF-κB nuclear localization in KMS-20 cells ( Figure 5).
(SGK1). We examined whether dimethyl fumarate (DMF), an NF-κB inhibitor, enhances the sensitivity of KMS-20 cells to bortezomib and ixazomib. The combination of DMF and bortezomib or ixazomib enhanced the sensitivity of KMS-20 cells to bortezomib and ixazomib (p < 0.05) ( Figure 5A,B and Supplementary Materials, Figure S6). Additionally, DMF suppressed B-cell lymphoma-2 (Bcl-2), Bcl-xL, and survivin expression and enhanced Bcl-2-like protein 11 (Bim) expression via the inhibition of NF-κB nuclear localization in KMS-20 cells ( Figure 5). To validate these observations, we confirmed the sensitivity to bortezomib and ixazomib in other multiple myeloma (MM) cells. We found that L363 had a lower sensitivity to bortezomib and ixazomib, similar to that of KMS-20 cells, but ARH-77 and RPMI8226 cells showed a high sensitivity to bortezomib and ixazomib, which was similar to that of the KMS-26 and KMS-28BM cells (Supplementary Materials, Figure S7). Additionally, activation of SGK1 and NF-κB p65 was higher in the L363 cells than in the ARH-77 and RPMI8226 cells, and GSK650394 and DMF enhanced the sensitivity to bortezomib and ixazomib in L363 cells (Supplementary Materials, Figure S7). These results indicate that the SGK/NF-κB pathway may play a role in the underlying mechanism of the low sensitivity of KMS-20 and L363 cells to bortezomib and ixazomib.

Discussion
Here, KMS-20 cells presented a lower sensitivity to bortezomib and ixazomib than KMS-26 and KMS-28BM cells. Although proteasome β5 subunit overexpression and the induction of autophagy by a proteasome inhibitor have been associated with resistance in multiple myeloma (MM) cells [45][46][47], our results showed that there was no change in the expression level of the proteasome β5 subunit, inhibition level of proteasome β5 subunit activity, or degree of autophagy induction by bortezomib and ixazomib. These results suggest that a low sensitivity to bortezomib and ixazomib relies on other factors.
To ascertain the mechanism underlying the low sensitivity to bortezomib and ixazo- To validate these observations, we confirmed the sensitivity to bortezomib and ixazomib in other multiple myeloma (MM) cells. We found that L363 had a lower sensitivity to bortezomib and ixazomib, similar to that of KMS-20 cells, but ARH-77 and RPMI8226 cells showed a high sensitivity to bortezomib and ixazomib, which was similar to that of the KMS-26 and KMS-28BM cells (Supplementary Materials, Figure S7). Additionally, activation of SGK1 and NF-κB p65 was higher in the L363 cells than in the ARH-77 and RPMI8226 cells, and GSK650394 and DMF enhanced the sensitivity to bortezomib and ixazomib in L363 cells (Supplementary Materials, Figure S7). These results indicate that the SGK/NF-κB pathway may play a role in the underlying mechanism of the low sensitivity of KMS-20 and L363 cells to bortezomib and ixazomib.

Discussion
Here, KMS-20 cells presented a lower sensitivity to bortezomib and ixazomib than KMS-26 and KMS-28BM cells. Although proteasome β5 subunit overexpression and the induction of autophagy by a proteasome inhibitor have been associated with resistance in multiple myeloma (MM) cells [45][46][47], our results showed that there was no change in the expression level of the proteasome β5 subunit, inhibition level of proteasome β5 subunit activity, or degree of autophagy induction by bortezomib and ixazomib. These results suggest that a low sensitivity to bortezomib and ixazomib relies on other factors.
To ascertain the mechanism underlying the low sensitivity to bortezomib and ixazomib, we analyzed the gene expression profiling (GEP) database to reveal the genes expressed in KMS-26 or KMS-28BM cells, compared to those in KMS-20 cells. Considering the possible resistance genes revealed in this study, fibroblast growth factor receptor 1 (FGFR1), C-C chemokine receptor type 2 (CCR2), and serum/glucocorticoid regulated kinase 1 (SGK1) were overexpressed in KMS-20 cells. FGFR1 contributes to tyrosine kinase inhibitor resistance and chemoresistance in lung, breast, and urothelial cancers [48][49][50]. The activation of CCR2 is involved in resistance to regorafenib, a multikinase inhibitor, and cabazitaxel in colon and prostate cancer cells [33,34]. SGK1 is associated with Akt and phosphoinositide 3-kinase inhibitor and paclitaxel resistance in breast and ovarian cancer cells [51][52][53]. Additionally, SGK1, CCR2, and SGK1 mRNA and protein expression, and the activation of SGK1 and FGFR1, were higher in KMS-20 cells than in KMS-26 and KMS-28BM cells. Furthermore, the inhibition of SGK1 by an SGK1 inhibitor, GSK650394, enhanced the cytotoxic effect of bortezomib and ixazomib, whereas the combined administration of bortezomib or ixazomib with an FGFR1 inhibitor, PD166866, or a CCR2 antagonist did not. Moreover, patients with MM who were non-responsive to bortezomib had a high expression of SGK1 and those with a high expression of SGK1 had a significantly lower overall survival than patients with a low expression of SGK1. Thus, the overexpression and activation of SGK1 have important roles in a low sensitivity to bortezomib and ixazomib.
Although it is known that the activation of SGK1 accelerates the phosphorylation of N-myc downstream regulated gene 1 and activation of murine double minute 2, the effects of signaling crosstalk is unknown [54]. The expression of extracellular regulated protein kinase 1/2 (ERK1/2), nuclear factor (NF)-κB, Akt, and c-Jun N-terminal kinase (JNK) was higher in KMS-20 cells than in KMS-26 and KMS-28BM cells. GSK650394 suppressed phosphorylated NF-κB expression but did not affect ERK1/2, Akt, and JNK phosphorylation. Moreover, GSK650394 inhibited B-cell lymphoma-2 (Bcl-2), Bcl-xL, and survivin expression, increased Bcl-2-like protein 11 (Bim) expression and did not affect Bcl-2-associated X (Bax), X-linked inhibitor of apoptosis protein (XIAP), phorbol-12-myristate-13-acetate-induced protein 1 (Noxa), and p53 upregulated modulator of apoptosis (Puma) expression. The combination of dimethyl fumarate (DMF), an NF-κB inhibitor, and bortezomib or ixazomib strongly induced cell death in KMS-20 cells via the suppression of NF-κB activation, which inhibited the expression of Bcl-2, Bcl-xL, and survivin and enhanced the expression of Bim. NF-κB p65 phosphorylation at Ser536 is involved in transcriptional activity, and NF-κB p65 activation modulates the expression of several apoptosis-regulating factors, such as those in the Bcl-2 and IAP families [55][56][57][58][59][60]. Activation of NF-κB promotes the expression of Bc-2, Bcl-xL, survivin, and XIAP, and downregulated Puma, Noxa, Bim, and Bax expression via suppression of p53 function [43,44,[61][62][63]. Bcl-2 and Bcl-xL suppress the apoptosis-inducing function of proapoptotic proteins such as Bax, Bim, Noxa, and Puma [64]. Survivin and XIAP inhibit the function of caspases, which are apoptosis-leading proteinases, and suppress the induction of apoptosis [65]. The Bcl-2 overexpression was associated with a low clinical response to bortezomib in MM patients [66]. Moreover, it indicated that upregulation of Bcl-xL by NF-κB signaling activation induced proteasome inhibitor resistance in MM cells [67]. Additionally, low levels of Bim expression were associated with bortezomib resistance in bortezomib-resistant U266PS-R cells and primary MM cells, whereas ABT-737, a Bcl-2 homology 3 (BH3) mimetic, overcame the bortezomib resistance in U266PS-R cells [68]. Furthermore, Survivin was overexpressed in the bortezomib-resistant NCI-H929-R20.1 cells compared to the parent cells, and BAY-11-7082, an IκB kinase inhibitor, suppressed the Survivin expression via inhibition of NF-κB p65 activation in bortezomib-resistant NCI-H929-R20.1 cells [60]. Moreover, NF-κB p65 activation induced bortezomib resistance, but the inhibition of NF-κB signaling reduced bortezomib resistance in MM cells [69,70]. Moreover, we found that high expression of NF-κB p65 in MM patients was associated with a low sensitivity to bortezomib and a poor prognosis compared to that of patients with a low NF-κB p65 expression. These results suggest that the SGK/NF-κB signaling pathway contributes to a low sensitivity to bortezomib and ixazomib and an SGK or NF-κB inhibitor may enhance the cytotoxic effect of bortezomib and ixazomib in MM cells.
It has been demonstrated previously that SGK1 activation is involved in acquired bortezomib resistance in MM cells. Also, SGK1 inhibition by bortezomib or short hairpin RNA (shRNA) has been shown to enhance its cytotoxicity via activation of the JNK pathway and induction of endoplasmic reticulum (ER) stress [20]. During the present study, we showed that the overexpression of SGK1 is correlated with primary resistance to bortezomib and ixazomib via activation of the NF-κB pathway. Moreover, the inhibition of SGK1 and NF-κB by the inhibitors, bortezomib and ixazomib, aided in repressing primary resistance. Additionally, we found that SGK1 inhibitors did not induce CHOP expression and activation of JNK. These findings indicate that the activation of the SGK1/NF-κB pathway plays an important role in developing bortezomib and ixazomib resistance in MM cells.
It has previously been indicated that induction of SGK1 expression was regulated via activation of the JAK/STAT3, MEK/ERK, PI3K/Akt, or p38MAPK pathways [19,41,42]. However, in this study we found that the extent of activation of the STAT3 and p38MAPK pathways was comparable among the KMS-20, KMS-26, and KMS-28 cells. Additionally, U0126, a MEK inhibitor, and LY294002, a PI3K inhibitor, did not affect the expression of the SGK1 protein, and did not enhance the cytotoxicity of bortezomib and ixazomib in KMS-20 cells. These results indicate that SGK1 transcription is not involved in the activation of the JAK/STAT3, MEK/ERK, PI3K/Akt, or p38MAPK pathways in KMS-20 cells. However, the mechanism underlying SGK1 induction via these signaling pathways is not yet clear. Therefore, future studies are required to investigate their effects on SGK1 transcription.

Conclusions
To conclude, both the overexpression and activation of the serum/glucocorticoid regulated kinase 1 (SGK1)/Nuclear Factor (NF)-κB pathway are involved with the low sensitivity of multiple myeloma (MM) cells to bortezomib and ixazomib, and SGK1 or NF-κB inhibitors can be used to increase the cytotoxic effects of bortezomib and ixazomib. These results indicate that the combination of an SGK1 or NF-κB inhibitor and bortezomib or ixazomib may be a potential therapy for MM harboring the overexpression and activation of the SGK1/NF-κB pathway.
Supplementary Materials: The following are available online at https://www.mdpi.com/2227-905 9/9/1/33/s1, Figure S1: Combined effect of GSK650394 and bortezomib or ixazomib on Annexin-V positive cells in KMS-20 cells, evaluated by Muse™ Annexin V and Dead Cell kit, Figure S2: Effect of bortezomib and ixazomib on phosphorylated SGK1 and total SGK1 expression in KMS-20 cells, Figure S3: Effect of U0126, perifosine, and SP600125 on bortezomib or ixazomib resistance, Figure S4: Effect of GSK650394 on CHOP expression in KMS-20 cells, Figure S5: Effect of SGK1 siRNA on bortezomib or ixazomib resistance, Figure S6: Combined effect of dimethyl fumarate (DMF) and bortezomib or ixazomib on Annexin-V positive cells in KMS-20 cells, evaluated by Muse™ Annexin V and Dead Cell kit, Figure S7: Effect of bortezomib and ixazomib on L363, ARH-77, and RPMI8226 cell viability.

Data Availability Statement:
The data presented in this study are availabl on request from the corresponding author.