For analyses of TLR expression in primary myeloma cells from patients, RNA sequencing data from CD138⁺ cells (MMRF_CoMMpass_IA14a_E74GTF_Salmon_Gene_Counts) were downloaded from the Multiple Myeloma Research Foundation CoMMpass IA14 release (https://research.themmrf.org/). The raw counts were cpm normalized and log transformed using EdgeR before TLR gene expression in diagnostic samples was analyzed (n = 772 patients). Spearman’s correlation test between TLR expression and chromosomal abnormalities was performed using the function cor.test in R, and significance was calculated using the function cor.mtest in R.

The MM cell line RPMI-8226 was obtained from the American Type Culture Collection (ATCC; Rockville, MD, USA). RPMI-8226 cells depleted for TLR4 (TLR4 KO), TLR9 (TLR9 KO), and control/mock (WT) cells were generated by gene editing using CRISPR/Cas9 as previously described (23). The cells were cultured in RPMI-1640 medium with 20% heat-inactivated fetal calf serum (FCS) during expansion and 10% FCS for experiments. Freshly isolated primary myeloma cells were obtained from the local hospital biobank (Biobank1). CD138⁺ cells were isolated from patient-derived bone marrow aspirates using the RoboSep automated cell separator and the Human CD138 Positive Selection Kit (StemCell Technologies, Grenoble, France). For patient characteristics, see Supplementary Table 1 . The primary cells were cultured in RPMI-1640 with 2% human serum, 1 ng/mL rhIL-6 (Gibco, Thermo Fisher Scientific, Waltham, MA, USA), and 20 µg/mL gentamicin. The study was approved by the regional ethics committee (REK #247909), and all patients gave informed consent before donation. The study was performed in accordance with the Declaration of Helsinki. Primary cells and cell lines were cultured at 37°C in a humidified atmosphere with 5% CO₂.

The following TLR agonists were used for stimulation of MM cells: TLR2/1, Pam3Cys (EMC Microcollections, Tübingen, Germany); TLR2/6, FSL-1 (EMC Microcollections); TLR3, Poly(I:C) HMW (Invivogen, San Diego, CA, USA); TLR4, Ultrapure LPS (E.coli 0111:B4, Invivogen); TLR5, Flagellin FLA-ST (Invivogen); TLR7 and TLR8, R-848 (Invivogen); and TLR9, CpG ODN 2006 (TIB Molbiol, Berlin, Germany). The following concentrations were used: Pam3Cys, 1 µg/mL; FSL-1, 1 µg/mL; Poly(I:C), 10 µg/mL; LPS, 0.1 µg/mL for RPMI-8226 cells and 1.0 µg/mL for primary MM cells; Flagellin, 1 µg/mL; R-848, 1 µg/mL; and CpG, 1 µM. For controls, cells were treated with equal volumes of cell culture medium. Patient samples with a limited cell number were treated with a cocktail containing all agonists.

For experiments assessing sensitivity to proteasome inhibitors, carfilzomib (#S2853) and bortezomib (#S1013) from Selleck Chemicals (Houston, TX, USA) were used at concentrations as indicated in the figure legends. For inhibition of the autophagy-lysosomal pathway, bafilomycin A1 (#sc-201550A) from Santa Cruz Biotechnology (Dallas, TX, USA) and hydroxychloroquine from Sigma-Aldrich (St. Louis, MO, USA) were used at concentrations as indicated in the figure legends.

Proliferation was assessed by quantification of cells in cell cycle S-phase, measured as 5-bromo-2′-deoxyuridine (BrdU)-positive live cells using flow cytometry. Cells were seeded at a concentration of 250,000 cells/mL and treated with LPS or CpG. The following day, additional media containing TLR agonists was added to ensure exponential growth of the cells. Following 48 hours of TLR stimulation, the cells were incubated with 10 µM BrdU (#ab142567, Abcam, Cambridge, UK) in cell culture media for 4 hours. Cells were washed in phosphate-buffered saline (PBS) and further stained with Fixable Viability Dye eFluor™ 450 (#65–0863-14, eBioscience, San Diego, CA, USA) for 30 minutes on ice. The cells were washed with PBS and fixed with ice-cold methanol. Fixed cells were kept in the fridge in methanol overnight for further analysis. The following day, cells were washed with PBS before DNA was hydrolyzed by slowly adding 1 mL of 2N HCl/0.05% Triton X-100 upon vortexing. Following 30 minutes of incubation at room temperature, 1 mL of 0.1 M Na₂B₄O₇ (borax) was added to neutralize the pH. Cells were spun down and resuspended in PBS with 1% bovine serum albumin (BSA)/0.1% Tween. Next, cells were stained with Alexa Fluor 488 Anti-BrdU (#ab220074, Abcam) for 1 hour at room temperature. Cells were washed before analysis on the LSR II instrument (BD Biosciences, San Jose, CA, USA) with FACS Diva software (BD Biosciences). FlowJo version 10 (TreeStar, Ashland, OR, USA) was used to gate BrdU positive cell fractions of live cells.

Cell viability measurements were performed using CellTiter Glo 2.0 Cell Viability Assay from Promega (Madison, WI, USA) according to the manufacturer’s protocol. Ten thousand cells were seeded per well before treatment with LPS or CpG. For experiments investigating the effect of TLR activation on sensitivity to proteasome inhibitors, cells were pre-treated with LPS or CpG for 2–2.5 hours before treatment with bortezomib/carfilzomib or dimethyl sulfoxide (DMSO) control. For experiments investigating autophagy, cells were pre-treated with LPS or CpG for 4 hours before treatment with bortezomib or carfilzomib and HCQ. Cells were incubated for 24–48 hours before viability was assessed. Luminescence was recorded using a Victor3 plate reader and Wallac 1420 Work Station software (PerkinElmer Inc., Waltham, MA, USA) or the PolarSTAR OMEGA plate reader with MARS Data Analysis Software version 2.10 R3 (BMG Labtech GmbH, Ortenberg, Germany).

mRNA transcript counts in RPMI-8226 cells challenged with LPS for 48 hours were performed using the nCounter analysis system (NanoString Technologies, Seattle, WA, USA) and the gene expression panels Human Metabolic pathways panel (XT-CSO-HMP1–12) and the PanCancer pathways Panel (XT-CSO-PATH1–1) from NanoString Technologies. The assay was performed according to the manufacturer’s protocol using 100 ng total RNA input. nSolver Analysis Software (NanoString Technologies) was used for the calculation of transcript numbers, applying normalization against housekeeping genes. Using normalized count matrices, differentially expressed genes between control and LPS-stimulated RPMI-8226 cells were analyzed in R using linear models for microarray and RNA-seq data (limma-voom) with TMM normalization. p-Values were adjusted using the Benjamini–Hochberg correction. Analyses were run using R version 4.3.2 (2023–10-31). The packages used with version number include the following: packageVersion (“org.Hs.eg.db”) ‘3.14.0’, packageVersion (“limma”)’3.50.3’, packageVersion (“edgeR”) ‘3.36.0’, and packageVersion (“ggplot2”) ‘3.4.4’. To identify which biological pathways were most affected by LPS stimulation, the significantly differentially expressed genes (adjusted p-value <0.05, N = 112) were used as input in a gene set enrichment analysis (Enrichr) (32) using the Panther metabolic and cell signaling database.

Oxygen consumption rate (OCR) and extracellular acidification rate (ECAR) were measured using the Seahorse XF96 bioscience extracellular flux analyzer (Agilent, Santa Clara, CA, USA). Seahorse XF cell culture microplates were coated with Cell-Tak (#354240, Corning, New York, NY, USA) following the manufacturer’s protocol, and cells were seeded at a concentration of 25,000 cells/well. Glycolysis and oxidative respiration were measured by combining the Glycolysis stress test and Mito stress test assays (Agilent) to allow for assessment of glycolysis and oxidative respiration in the cells at the same time, as previously described (33). Seahorse XF base medium was supplemented with 10 mM sodium pyruvate and 2 mM glutamine (both from Sigma-Aldrich). Glucose (Merck) at mass of 10 mM was added in the first injection, followed by 1 µM oligomycin in the second injection, 0.8 µM carbonyl cyanide p-trifluoro-methoxyphenyl hydrazone (FCCP) (#C2920) in the third injection, and finally, a combination of 2 µM rotenone (#R8875) + 2 µM antimycin (#A0149) + 50 mM 2-deoxyglucose (2-DG; #D8375), all from Sigma-Aldrich. Basal respiration was determined using the following formula: (last rate measurement before oligomycin injection) − (non-mitochondrial respiration rate). Glycolysis was determined using the following formula: (maximum rate measurement before oligomycin injection) − (last rate measurement before glucose injection).

Primary cells were stimulated with a cocktail of TLR agonists covering TLR1–TLR9 (concentrations as described above) for 6 hours. RNeasy mini kit was used to isolate total RNA (Qiagen, Hilden, Germany) before cDNA was synthesized with a High Capacity RNA-to-cDNA kit (Applied Biosystems, Carlsbad, CA, USA). Real-time quantitative PCR was performed using TaqMan Gene Expression Assays (Thermo Fisher, Waltham, MA, USA) and the StepOnePlus Real-Time PCR system (Applied Biosystems) using standard settings (2 minutes 50°C, 10 minutes 95°C, 40 cycles at 95°C for 15 sec, and 1 minute 60°C). The comparative Ct-method was used to estimate relative changes in gene expression, and ACTB (for CD138⁺ primary cells) or TBP (for RPMI-8226) were used as housekeeping genes. The following gene expression assays were used: ACTB (Hs01060665_g1), TBP (Hs00427620_m1), IRF4 (Hs01056533_m1), MYC (Hs00153408), BCL2 (Hs00608023_m1), BCL2L1 (Hs00236329_m1), MCL1 (Hs01050896_m1), and TNFRSF17 (Hs03045080_m1).

Cells were lysed in lysis buffer containing 50 mM Tris–HCl, 1% NP40, 150 mM NaCl, 10% glycerol, 1 mM Na₃VO₄, 50 mM NaF, and cOmplete Protease Inhibitor (Roche Diagnostics, Mannheim, Germany). To assess the protein expression of SQSTM1, cells were lysed in 8 M urea 0.5% Triton X-100, and lysates were sonicated to shear DNA before electrophoresis. Lysates were mixed with 1× NUPage LDS sample buffer supplemented with 0.1 mM DTT and heated at 70°C for 10 minutes for protein denaturation. Samples were next separated on NuPAGE™ 4–12% Bis-Tris polyacrylamide gel before the proteins were transferred to a nitrocellulose membrane using the iBlot Dry Blotting System (Invitrogen, Thermo Fisher) or by wet blotting using the Criterion™ Blotter (Bio-Rad, Hercules, CA, USA). The membranes were blocked with 5% bovine serum albumin in 0.1% Tris-buffered saline with 0.01% Tween (TBS-T) or 5% dry milk in TBS-T, followed by incubation with primary antibodies for 1–3 days. Horseradish peroxidase (HRP)-conjugated antibodies (DAKO, Glostrup, Denmark) and Super Signal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific) were used for detection, and images were obtained using Odyssey Fc Imager (LI-COR Biosciences, Cambridge, UK). Images were analyzed using Image Studio Software (LI-COR Biosciences). The following primary antibodies were used, all from Cell Signaling Technology (Danvers, MA, USA): anti-β-actin (#4967), anti-p-IκBα (#2859), anti-IκBα (#4812), anti-c-Myc (#5605), anti-IRF4 (#4964), anti-Bcl-2 (#15071), anti-Bcl-xl (#2762), and anti-SQSTM1/p62 (#39749).

Chymotrypsin-like proteasome activity was measured using Proteasome-Glo™ Chymotrypsin-Like Cell-Based Assay (#G8660, Promega) according to the manufacturer’s protocol. RPMI-8226 cells were seeded at a concentration of 10,000 cells/well. Cells were pre-treated with LPS or CpG for 2 hours before treatment with bortezomib. Following 24 hours, equal numbers of the treated cells were transferred to separate wells for analysis of chymotrypsin-like proteasome activity and cell viability. Cell viability was measured using CellTiter Glo 2.0 Cell viability assay (Promega) according to the manufacturer’s protocol. Luminescence was measured using the PolarSTAR OMEGA plate reader with MARS Data Analysis Software version 2.10 R3 (BMG Labtech GmbH, Ortenberg, Germany). Data are presented as chymotrypsin-like proteasome activity normalized against cell numbers.

BCMA surface expression in RPMI-8226 cells treated with LPS or CpG for 48 hours was analyzed by flow cytometry. TLR-stimulated cells were washed with PBS before being stained with eBioscience™ Fixable Viability Dye eFluor™ 450 (#65–0863-14, Invitrogen) for 30 minutes according to the manufacturer’s protocol. Cells were washed twice in flow buffer (0.1% BSA in PBS) before being stained with APC anti-human CD269 (BCMA) Antibody (#357506, BioLegend, San Diego, CA, USA) for 30 minutes on ice. Cells were washed twice in flow buffer and analyzed using LSR II (BD Biosciences) and FACS Diva software (BD Biosciences). Samples were analyzed using FlowJo V.10.4 (TreeStar, Ashland, Oregon, USA). Debris and dead cells were gated out before the median fluorescence intensity (MFI) of BCMA expression was analyzed. Fluorescence minus one (FMO) controls not stained with APC anti-human CD269 (BCMA) antibody were used as a negative staining control.

Statistical analyses were performed using GraphPad Prism Version 10.2.0 for MacOS (GraphPad Software). For comparison of more than two groups, one-way or two-way ANOVA followed by Sidak’s multiple comparisons test was used as indicated. Differences were considered significant when p < 0.05.

To obtain an overview of TLR expression in primary MM cells, we examined TLR mRNA expression in CD138⁺ cells obtained from 772 patients at diagnosis (CoMMpass IA14). We found that overall, TLR1, TLR4, TLR6, TLR9, and TLR10 were the most highly expressed (median cpm > 2) in this cohort. Importantly, however, the expression levels varied considerably between patients ( Figure 1A ). Using data from CoMMpass, we investigated the correlation between the presence of chromosomal abnormalities and gene expression of TLR genes. We found that four TLR genes correlated significantly with chromosomal abnormalities, of which TLR2 correlated with t(14:20) and t(14:16), TLR1 correlated with t(12:14), and TLR8 and TLR10 both correlated with t(4:14) ( Figure 1B ).

The MM cell line RPMI-8226 expresses a broad range of TLRs (23) including TLR4 and TLR9, which were also expressed in a high proportion of the primary MM cells. In line with previously published data (22, 25, 27–29), we found that both cell proliferation ( Figure 1C ) and cell viability ( Figures 1D, E ) were increased following LPS and CpG treatment.

To characterize how TLR4 activation altered the cells’ gene expression, we examined mRNA expression in RPMI-8226 cells after LPS treatment for 48 hours using NanoString gene expression assays. For genes present in the Human Metabolic pathways panel, 57 were significantly increased, and 34 were downregulated upon LPS stimulation ( Figure 2A ; Supplementary Table 2 ). For the PanCancer pathways panel, a total of 16 genes were significantly upregulated with LPS treatment, and six genes were downregulated ( Figure 2B ; Supplementary Table 3 ). Genes that were significantly differently expressed between LPS-stimulated and unstimulated cells were enriched in pathways such as glycolysis, apoptosis, and oxidative stress response ( Figure 2C ). In support of these findings, we found that TLR signaling increased glycolysis and basal OXPHOS ( Supplementary Figures 1A, B ). Importantly, among the upregulated genes, there were several genes encoding established pro-survival and anti-apoptotic factors in myeloma, including MYC, IRF4, NFKB1/2, BCL2, and BCL2L1 ( Supplementary Tables 2, 3 ). Gene expression analyses by qPCR demonstrated that also TLR9 signaling in RPMI-8226 cells induced similar changes in terms of anti-apoptotic/pro-survival genes ( Supplementary Figure 1C ). MYC, IRF4, NFKB, and BCL2 were further upregulated at the protein level in response to LPS and CpG ( Figure 2D ). Moreover, cells depleted of TLR4 (TLR4 KO) and TLR9 (TLR9 KO) did not change their protein expression of IRF4 and BCL2 in response to TLR stimulation, supporting that the effect is dependent on TLR signaling ( Figure 2E ). Another striking effect of TLR stimulation was the increase in several proteasomal components as well as an increase in the expression of the SQSTM1, an autophagosomal cargo receptor ( Supplementary Table 2 ). Thus, TLR stimulation may promote an increased capacity of the two major protein degradation systems in the cells, i.e., both enhanced degradation through the proteasome and the autophagosome–lysosome pathway.

The increase in pro-survival genes combined with a potentially enhanced protein degradation capacity supports reduced efficacy toward proteasomal inhibitors upon TLR signaling. Indeed, both TLR4 and TLR9 activation significantly reduced MM cell sensitivity to both bortezomib and carfilzomib ( Figures 3A, B ). The reduction in drug sensitivity was dependent on TLR signaling since the protective effect of TLR activation was lost in TLR4 and TLR9 KO cells ( Figures 3C, D ; Supplementary Figures 2A, B ). Importantly, while bortezomib and carfilzomib reduced IRF4, MYC, and BCL2 levels, the expression of these proteins was maintained in cells co-treated with LPS or CpG except for the highest drug concentrations ( Figures 3E, F ), further supporting that TLR-induced expression of these pro-survival factors could be implicated in the reduced drug sensitivity in the presence of TLR agonists.

Proteasome subunits were among the upregulated genes in LPS-stimulated MM cells ( Supplementary Table 2 ), but TLR activation did not counteract the effect of bortezomib on proteasomal degradation efficacy ( Supplementary Figure 3A ). However, targeting autophagy by adding HCQ slightly but significantly reduced the protective effect of TLR signaling on PI-induced cytotoxicity ( Figure 4 ). HCQ has previously been shown to increase the efficacy of carfilzomib, but not bortezomib (34). Indeed, HCQ had no effect on cell survival in cells that were not exposed to PI (untreated) or that were exposed to bortezomib alone (bortezomib, ctrl) but inhibited the effect of carfilzomib (carfilzomib, ctrl). Importantly, HCQ reduced the protective effect of LPS and CpG on both bortezomib-induced cytotoxicity (bortezomib and CpG) and carfilzomib-induced cytotoxicity (carfilzomib, LPS, and CpG). Cells resistant to carfilzomib showed elevated levels of SQSTM1 (34, 35). Here, we also observed increased mRNA levels of SQSTM1 in cells treated with LPS ( Supplementary Table 2 ). Moreover, treating cells with LPS or CpG led to the accumulation of SQSTM1 protein, and the increase was also evident upon inhibition of autophagy ( Supplementary Figure 3B ), supporting that TLR activation can increase flux through the lysosomal protein degradation pathway.

As shown in Figure 1A , plasma cells from patients expressed a broad range of TLRs, but the expression levels varied greatly between patients. Thus, to examine how TLR activation affected pro-survival genes in freshly obtained CD138⁺ cells from MM patients, we treated the cells with a cocktail of TLR agonists covering TLR1–TLR9 for 6 hours. As expected, there were donor variations. However, TLR stimulation increased BCL2, BCL2L1, and MYC in a large proportion of the patients, while the effect on IRF4 and MCL1 gene expression was less prominent ( Table 1 ; Figure 5 ).

BCMA is a target for CAR T cells and bispecific antibodies approved to treat MM patients (36). From our NanoString analyses, we observed a significant reduction in TNFRSF17, the gene encoding BCMA, upon LPS stimulation ( Figure 2A ; Supplementary Table 2 ). CpG stimulation similarly reduced TNFRSF17 gene expression ( Figure 6A ), and importantly, cell surface expression of BCMA was also reduced upon TLR4 or TLR9 activation ( Figure 6B ). We also observed the downregulation of TNFRSF17 in primary cells obtained from 6/14 patients. Interestingly, however, in an equal fraction of the patient samples, TNFRSF17 was upregulated ( Figure 6C ; Table 1 ). Thus, TLR signaling in MM cells can theoretically both enhance and reduce the efficacy of BCMA-targeting drugs.