Ribosomal proteins as distinct “passengers” of microvesicles: new semantics in myeloma and mesenchymal stem cells’ communication

MAHMOUD DABBAH, MICHAEL LISHNER, OSNAT JARCHOWSKY-DOLBERG, SHELLY TARTAKOVER-MATALON, YARON S. BRIN, METSADA PASMANIK-CHOR, AVIVIT NEUMANN, and LIAT DRUCKER
KFAR SABA, TEL AVIV, AND HAIFA, ISRAEL

Aberrant mesenchymal stem cells (MSCs) in multiple myeloma (MM) bone marrows (BM) promote disease progression and drug resistance. Here, we assayed the pro- tein cargo transported from MM-MSCs to MM cells via microvesicles (MVs) with focus on ribosomal proteins (RPs) and assessment of their influence on translation initiation and design of MM phenotype. Proteomics analysis (mass spectrometry) demonstrated increased levels and repertoire of RPs in MM-MSCs MVs compared to normal donors (ND) counterparts (n = 3 8; P = 9.96E 08). We limited the RPs load in MM-MSCs MVs (starvation, RSK and XPO1 inhibitions), reapplied the modified MVs to MM cell lines (U266, MM1S), and demonstrated that the RPs are essential to the
proliferative effect of MM-MSCs MVs on MM cells (n = 3; P < 0.05). We also observed that inhibition with KPT-185 (XPO1 inhibitor) displayed the most extensive effect on
RPs delivery into the MVs ( 80%; P = 3.12E 05). Using flow cytometry we assessed the expression of select RPs (n = 10) in BM-MSCs cell populations (ND and MM; n
6 each). This demonstrated a heterogeneous expression of RPs in MM-MSCs with distinct subgroups, a phenomenon absent from ND-MSCs samples. These findings bring to light a new mechanism in which the tumor microenvironment participates in cancer promotion. MVs-mediated horizontal transfer of RPs between niche MSCs and myeloma cells is a systemic way to bestow pro-cancer advantages. This capacity also differentiates normal MSCs from the MM-modified MSCs and may mark their reprogramming. Future studies will be aimed at assessing the clinical and therapeutic potential of the increased RPs levels in MM-MSCs MVs. (Translational
Research 2021; 236:117—132)

From the Oncogenetic Laboratory, Meir Medical Center, Kfar Saba, Israel; Resaerch Institute, Meir Medical Center, Kfar saba, Israel; Hematology Unit, Meir Medical Center, Kfar saba, Israel; Autoimmunity laboratory, Meir Medical Center, Kfar saba, Israel; Orthopedics Department, Meir Medical Center, Kfar Saba, Israel; Bioinformatics Unit, G.S.W. Faculty of Life Sciences, Tel Aviv University, Tel Aviv, Israel; Sackler faculty of Medicine, Tel Aviv University, Tel Aviv, Israel; Oncology Department, Rambam Medical Center, Haifa, Israel.
Submitted for Publication January 27, 2021; received submitted March 23, 2021; accepted for publication April 12, 2021. Reprint requests: Liat Drucker, Oncogenetic Laboratory, Meir Medical Center, Kfar, Saba 44281, Israel. e-mail: [email protected]. 1931-5244/$ - see front matter
© 2021 Elsevier Inc. All rights reserved.117

Abbreviations: MM; multiple myeloma; MSCs; mesenchymal stem cells; BM-MSCs; bone mar- row mesenchymal stem cells; MM-MSCs; multiple myeloma mesenchymal stem cells; ND; nor- mal donor; MVs; microvesicles; RPs; ribosomal proteins; TI; translation initiation; eIF4E; eukaryotic translation initiation factor 4E; eIF4GI; eukaryotic translation initiation factor 4GI; XPO1; Exportin 1; RSK; ribosomal protein S6 kinases (p90RSK); MFI; mean fluorescence intensity

INTRODUCTION
Cancer cure is largely unattained due to develop- ment of drug resistant tumor clones that persist through allotted treatment and eventually thrive and expand.1 The selection of the drug resistant clones is applied to a mixed cancer cell population in diverse microenviron- ments.1 Understanding and clinically addressing this compound ecosystem of heterogeneous cancer popula- tion in its evolving niche is expected to improve our control and treatment efficacy.
Multiple myeloma (MM) is a malignancy of plasma cells that accumulate in the bone marrow (BM) and presents a good example for continuous clonal evolu- tion and development of critical and eventually fatal drug resistance.2,3 It is well recognized that the signifi- cant interaction of the myeloma cells with their sur- roundings is responsible for this course of events and the underlying cause for the fatality of MM despite many new treatments developed in the last 2 decades.2
In the past several years, others4,5 and we6-9 have delineated a significant role for BM mesenchymal stem cells (MSCs) in design of MM phenotype and drug response in a translation initiation (TI) dependent man- ner. Importantly, we have shown that BM-MSCs from MM patients promote the malignant cells’ prolifera- tion, migration and drug resistance, whereas BM- MSCs from normal donors (ND) do not and in some instances even inhibit them.9,10 These observations point to a fundamental difference between BM-MSCs according to their origin from healthy or pathological niches, present another layer of acquired alterations intertwined with disease progression, and may afford unique and selective opportunities for therapeutic intervention.8
A major contribution to cancer-niche plasticity and heterogeneity is increasingly attributed to circulating extracellular vesicles (EVs) shed from both cancer, immune and niche cells including MSCs.11 By transfer- ring cargoes from their origin cells to recipient cells as well as instigating membranal signaling, the various EVs shape the characteristics of their surroundings.
EVs consist of two major vesicle types based on their biogenesis and release: exosomes produced in the endosome and secreted via multivesicular bodies and microvesicles (MVs) all shed from the cell plasma membrane by budding and fission.12 All EVs carry pro- teins, nucleic acids and lipids and their biogenesis may be cell type and cargo dependent, affected by signaling and stimuli altogether leading to a highly diverse assemblage.12
In previous analyses of the BM-MSCs’ dialogue with
the MM cells we have applied fractionation studies to the BM-MSCs secretomes and found that the fraction that contains the MVs is involved.6 Further analyses showed that the MM-MSCs derived MVs promote MM whereas the ND-MSCs MVs do not.8 In an attempt to identify key effectors involved in advancing MM we assayed the cargoes of the MM-MSCs MVs and com- pared them to those carried by ND-MSCs MVs.8 Among other differences, we identified alterations in presence of multiple ribosomal proteins (RPs) in the MVs that origi- nated from the MM-MSCs, a phenomenon that is the focal point of the current study. Remarkably, previous observations depict the highest degree of RPs diversity in the hematopoietic system13 and several genetic hema- tological disorders have been classified as Ribosomopa- thies14 leading to the suggestion that RPs’ expression may regulate the hematopoietic developmental pro- gram.13 Despite the importance of such a function there are still only few reports regarding RPs in cancers, MM, its microenvironment or MVs.15,16

Traditionally, RPs were regarded as essential yet passive components of the translation machinery.17,18 Together with rRNAs the RPs assemble into stoichio- metric quantities of ribosomes that translate the cells’ proteome.19 Lately, this perception is challenged with new observations of unequal RPs’ expression under various conditions as well as evidence of ribosome- independent function in health and disease.17,20 Some of the data underscore an imbalance of RPs in cancer, which suggests a potential role in tumorigenicity and progression.19,21-23 The underlying mechanism of RPs’ altered expression and function in cancerous settings are topics of great interest yet much research is still needed, particularly if new anti-RPs therapeutics are to be introduced.19 The presence and role of elevated RPs levels in the cancer microenvironment, particularly BM-MSCs MVs, is novel. In the current study we aimed to further characterize this phenomenon and its significance to MM.

MATERIALS AND METHODS
Cell lines: Each experiment was conducted with U266 and MM1S purchased from the ATCC.23-25 U266 identity was authenticated by Genomics Core Facility of BioRap Technologies and the Rappaport Research Institute, Technion, Israel (STR of 15 loci). Authenticated MM1S were obtained from Karin Joehrer lab, Austria (STR of 15 loci). Fresh cell lines were repurchased from the ATCC recently and added to the study. All cell lines were propa- gated upon receipt/authentication and frozen ( 80˚ C) in aliquots.
BM-MSCs isolation and propagation: BM samples were obtained from femur head BM samples of con- secutive normal donors (ND), undergoing elective full hip replacement surgery (n = 15) and MM patients’ BM aspirates taken for medical purposes (n = 10) at Meir Medical Center, Israel. All partici- pants signed informed consent forms approved by Meir Medical Center Helsinki Committee. MSCs were isolated from BM samples on a Ficol gradient and seeded in flask at 40,000 cells/cm2 with RPMI 1640 supplemented with 10% FBS (Biological Industries). Nonadherent cells were removed with the medium within the first 10 days of culture, leav- ing the adhered MSCs in the culture dish. Media were replaced twice weekly until the culture was nearly confluent (2 3 weeks); at which time, the cells were harvested for identity validation (vimen- tin+, keratin-, CD271+, CD34-, CD45-; immunocy- tochemistry, flow cytometry [FACS]). The cells were also assayed for their ability to differentiate

into adipocytes and osteocytes as previously described by us.58-60
Microvesicles Isolation and application to MM cell lines: Microvesicles (MVs) were isolated from condi- tioned media collected from 80% confluent BM-MSCs cultures (2 6 weeks).60 Briefly, media was obtained after cell removal by centrifugation at 800 g for 5 min and then centrifuged at 4500 g for 5 min to discard large debris. After centrifugation twice at 20,000 g (Beckman Ti70 rotor; Beckman Coulter) for 60 minutes at 4˚C, the microvesicles were washed and re-suspended in PBS. Isolated BM-MSCs’ micro- vesicles were characterized by size and external expression of phosphatidylserine (electron microscope; Annexin V, FACS) as described by us previously8(sup- plementary Fig 2). Microvesicles’ total protein concen- tration was determined (Nanodrop). The dose of 50 mg/ml BM-MSCs MVs (ND, MM) per 100,000cells (MM) was chosen as the optimal concentration for the study according to calibrations performed by us previ- ously.8 This working concentration is also compatible with experimental conditions used by others.8
Flow cytometry (FACS): BM-MSCs MVs were identified by size and validated by Annexin binding to exposed phosphatidylserine. MVs uptake into the MM cells was assayed by staining them with the PKH67 dye, according to the manufacturer’s instructions (Sigma-Aldrich), and incubating them with MM cells (24 hours). Then, cells were washed (PBS) and fluores- cence was analyzed by (FACS; Navios Flow Cytome- tery, Beckman Coulter). In order to test the expression levels of the ribosomal proteins, BM-MSCs were fixed and permed prior incubation with RPs primary antibod- ies (RPL13a, RPS6, RPS27a, RPL19 and RPL37a, RPL28, RPL39, RPL10a, RPL4, RPL6, and RPS3; Pro-
teinTech 1:1000); secondary antibodies were added (Donkey anti rabbit Alexa Flour 647 and Donkey anti mouse Alexa Fluor 488, Jackson 1:1000) according to primary antibody source (Rabbit/Mouse) and the fluo- rescence was analyzed in (FACS).
Proteomics analysis: BM-MSCs (ND, MM) and their secreted microvesicles (n = 3 each) were collected, washed for FBS removal and assayed for protein con- tent (analytical proteomics) by mass spectrometry at the Smoler Protein Research Center (Technion). Initial analysis was performed using Perseus.61
Bioinformatics analysis: Mass spectrometry results were analyzed at the G.S.W Life Sciences Bioinformatics unit at Tel-Aviv University. Log LFQ intensities resulting from Perseus software were statistically analyzed for differentially expressed proteins (ND-MSCs vs MM- MSCs), (ND-MSCs MVs vs MM-MSCs MVs), (MM- MSCs/MVs CTRL vs KPT-185/BID-1870/1%FBS
treated MM-MSCs/MVs; cutoff P < 0.05 and FC >

1.25) using Partek Genomics suite (v 6.6; http://www.par tek.com/pgs). Commonly attenuated proteins in MM- MSCs (FC < 2, P < 0.05) and commonly enriched pro-
teins ( 2 < FC) in the MM-MSCs MVs by all three
treatments (KPT-185, BI-D1870 and 1%FBS) were iso-
lated (Venn diagram) and assayed for Gene Ontology enrichment (GO) with 3 different Bioinformatics tools (Webgestalt, ToppGene, Gorilla).62-64
Trypan blue: Total, viable, and dead cell counts were assayed by Trypan blue dye. Cells were automatically counted by Countess (Invitrogen).58-60
Cell viability assay: Cell viability was assayed with cell proliferation reagent WST-1 (Roche) as described before.8,10
Western blotting: Proteins lysates were immunoblot- ted as we described previously58-60 using rabbit/mouse anti-human: peIF4E (Ser209), peIF4GI (Ser1108), pmTOR (Ser2448), pERK1/2, pJNK, Histone H3, Smad5, XPO-1; (Cell Signaling Technology, Danvers, MA, USA); tubulin (Sigma); NFkB p50 (SC-8414); RPL13a, RPS6, RPL27a, RPL19, and RPL73a (Pro-
teinTech).
Inhibitors and drugs: KPT-185 (irreversible XPO-1 inhibitor; Cayman Chemical Company)65 and BI- D1870 (RSK inhibitor; MCE) were dissolved in DMSO and tested for dose response on BM-MSCs via- bility (50 200 nM and 5 20 mM, respectively). The lowest effective concentrations were selected for con- tinued studies (50 nm and 5 mM, respectively). All doses used in this study are clinically relevant accord- ing to previous publications.66,67
Statistical analysis: All experiments were conducted at least 3 separate times. Student’s paired t-tests were applied in the analyses of differences between 2 cohorts. An effect was considered significant when P value was less than 0.05. Mass spectrometry data sta- tistical analyses were done using SPSS-25 software.

RESULTS
RPs are overexpressed in MM-MSCs derived MVs. Hav- ing previously established that BM-MSCs MVs from normal and myelomatous settings differ in their effect on MM cells8 we set out to assay their protein cargoes. We focused on the MVs’ proteome because some of the effects we witnessed were measured within minutes thereby suggesting an immediate signaling most proba- bly attributed to proteins. Indeed, analytical mass spec- trometry assay demonstrated altered integrin expression and ensuing signaling discussed elsewhere.24 Pertaining to the current study, we also observed a significant increase in RPs levels and repertoire in MM-MSCs MVs compared to the protein cargoes of ND-MSCs
MVs ( 1). In detail, 72 RPs were detected in MM- MSCs MVs whereas only 57 of them were present also in ND-MSCs MVs ( 1A). Moreover, the RPs present in MVs of both sources were more abundant in the MM- MSCs MVs ( 1B, C; P = 9.96E 08). By comparing the expression levels of all proteins in the MM-MSCs MVs to the ND-MSCs MVs (without the RPs) we observed a 1.9 fold change (FC) increase whereas analy- sis of the RPs expression exclusively demonstrated an
8.4 FC increase. Furthermore, STRING protein-protein interaction (PPI) enrichment confirmed functional enrichment analysis of MM-MSCs MVs proteins cargo (P value of 1.0E 16). This indicates that increased bulk RPs is a unique systematic characteristic of the MM-MSCs MVs proteome . Principal Compo- nents Analysis (PCA) demonstrated that the RPs expres- sion pattern distinctly and differentially characterizes the BM-MSCs in accordance with their normal or mye- loma source (97.6%; 1E). These observations com- plemented our previous findings that MM-MSCs and their derived MVs promote recipient MM cells’ TI and proliferation. Moreover, the RPs stood out as an over- expressed family of proteins with distinct intercellular function being delivered by MVs.
The high load of RPs in MM-MSCs MVs function is essential for the MVs capacity to stimulate proliferation and TI in recipient MM cells. The increased RPs expres-
sion we report here is consistent with their role in ribo-
some composition17,18 and our previous published observations of increased TI and protein synthesis in MM cells with established uptake of MM-MSCs derived MVs.8,9 Our first aim in the current study was to confirm the mechanistic role of RPs’ transfer from the niche MSCs into the MM cells in promoting the malignant cells’ TI. Thus, we decided to lower the RPs levels in the MM-MSCs’ MVs and test the modified MVs’ effect on recipient MM cells. We made a point of analyzing the bulk-RPs mediated influence and importance rather than concentrate on specific RPs. We used three different strategies to achieve this goal:
(1) starvation by decreased serum concentration (1% FBS) that was previously reported to reduce RPs expression/protein synthesis25; (2) Inhibition of MAPK-activated p90 ribosomal S6 kinase (RSK) sig- naling known to control RPs production (BI-D1870);26 and (3) inhibition of RPs Exportin 1 (XPO-1) depen- dent export from the nucleus necessary for RPs deliv- ery into the MVs (KPT-185).27
MM-MSCs. In order to establish experimental condi- tions with these 3 approaches we tested their effects on BM-MSCs viability and secretion of MVs. Increasing doses of RSK and XPO1 inhibitors caused a dose dependent decrease in viability. Thus, we chose for fur- ther use the lowest effective BI-D1870 and KPT-185
1. RPs expression and repertoire are significantly increased in MM-MSCs MVs compared to ND-MSCs MVs. Venn diagram presents the mass spectrometry data analysis of 73 ribosomal proteins expressed in BM- MSCs (ND/MM) MVs (n = 8, each) (A). The top elevated proteins expressed in MM-MSCs MVs compared to ND-MSCs MVs are listed (fold-change and variance) (B), and their expression levels are also presented in box graph (n = (C). STRING protein-protein interactions and functional enrichment analysis of ND/MM MSCs
MVs elevated proteins (FC > 1.25) is presented and bulk RPs present only in MM-MSCs MVs are depicted (D). Principal Components Analysis (PCA) of RPs expression pattern in accordance with their normal or myeloma
source is presented (n = 3) (E).

doses that also instigated a significantly more powerful response in the MM-MSCs compared to ND-MSCs (5 mM and 50 nM, respectively; P < 0.01; supplemen- tary 1). The decreased serum level was chosen based on previous publications28,29 and lack of induced
death (data not shown). A mild yet significant attenuat- ing effect of 1%FBS on MM-MSCs viability was also demonstrated (supplementary Fig 1A, right panel).
All chosen treatment doses diminished the ND- MSCs’ MVs secretion ( 20% 40%; P < 0.05; 2A). Yet, the effect of the three treatments on the secretion of MVs from MM-MSCs varied. While BI-
D1870 and KPT-185 decreased MVs secretion from MM-MSCs at similar rates as when applied to ND- MSCs ( 20% 25%; P < 0.05) 1%FBS triggered an increase in MVs secretion from MM-MSCs ( 27%; P < 0.01; Fig 2A). This observation corresponds with
previous publications.28
Next we wanted to determine that the chosen treat- ments indeed diminished the RPs levels in treated BM-

MSCs and attenuated key proteins in protein synthesis. For this purpose we selected the representative RPs: RPL37a, RPS27a, RPL19, and RPS6. Their selection was based on previous publication regarding their extra ribosomal function in cancerous settings indicating their potential importance in the MM niche (detailed in supplementary Table 1). We also assayed by immuno- blotting the treatments’ effects on mammalian target of rapamycin (mTOR) and eukaryotic initiation factors E and GI (eIF4E and eIF4GI, respectively) since they are critically involved in protein synthesis dependent on RPs in their classical role (Supplementary Fig 1B, Fig 2B). All assayed proteins displayed dose dependent reductions in expression upon BI-D1870 or KPT-185 administration (supplementary Fig 1B) with no signifi- cant reduction to MSCs viability upon starvation (10% and 5% FBS) except for the most extreme conditions
of 1% FBS ( 20%, P < 0.01; 2B). Specifically, chosen treatments’ doses (BI-D1870 5 mM; KPT-185
50 nM; 1%FBS) significantly attenuated the expression

2. The effect of KPT-185/BI-D1870/1%FBS on BM-MSCs (ND/MM) MVs’ secretion and RPs expression. BM-MSCs (ND/MM; n = 4) were treated for 72 h with KPT-185 (50 nm), BI-D1870 (5 mm), and decreased lev- els of FBS in media (1%). Their secreted MVs were enumerated by FACS (A). Treated BM-MSCs proteins lysates were harvested and assayed for RPs expression levels (RPL37a, RPS27a, RPL19, RPS6, and RPL13a), TI factors (peIF4E, peIF4GI) and p/t mTOR (immunoblotting). Results were quantified by Multigauge, and pre-
sented bar graph as percent (mean § SE, n ≥ 4) of respective protein expression in control BM-MSCs (ND, MM) not treated with RPs inhibitors (dotted line) (B). Bulk RPs expression levels after treatment were analyzed,
normalized to RPL13a (housekeeping; ref) and presented in box graph (n = 82) (C). All RPs detected by mass spectrometry in MM-MSCs treated with KPT-185/BI-D1870/1%FBS were analyzed for unique and common proteins by Venn diagram (D). Asterisks depict statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001).

of our RPs panel ( 20% 85%, P < 0.05; Fig 2B, Table 1) yet did not affect housekeeping RPL13a (sup-
plementary 1C). We also tested the selected doses effects on mammalian target of rapamycin (mTOR),

Table 1. List of depleted RPs in MM-MSCs MVs following treatment with KPT-185/BI-D1870/1%FBS compared to untreated MM-MSCs MVs

KPT-185 BI-D1870 1% FBS

RPL11 RPL7A RPS9 RPS15A RPL26L1 RPL12 RPL9 RPL26L1 RPL26L1 RPL35A RPL21 RPS13 RPL35A RPL35A RPL37 RPL22 RPS14 RPL37 RPL37 RPL39P5 RPL23 RPS15 RPL39P5 RPL39P5 RPL3L RPL23A RPS15A RPL3L RPL3L RPL7L1 RPL27A RPS19 RPL7L1 RPL7L1 RPS29 RPL28 RPS21 RPS29 RPS29 RPS4Y1 RPL31 RPS24 RPS4Y1 RPS4Y1 RPS6KA3 RPL36 RPS27 RPS6KA3 RPS6KA3
RPL36A RPS27L
RPL37A RPS28

Bold font highlights depleted RPs common to all treatments.

translation initiation factors E and G previously dem- onstrated by us to be affected by the MM-MSCs MVs cargoes8 and expected to respond to RPs availability. Indeed, we determined a dose dependent decrease in TI factors in KPT/Bi-D1870/FBS treated MM-MSCs. These results supported the biological relevance of our treatments. Since the strategies we applied are not spe- cific to RPs biogenesis or trafficking we wanted to vali- date that the effect was more prominent in their RPs inhibition than with all other proteins detected and again used mass spectrometry to analyze the BM- MSCs with and without inhibitors. We compared the rates of attenuated proteins (excluding the RPs) with each treatment versus the general 3289 proteins anno-
tated in the MM-MSCs (>1.25FC, P < 0.05). To our satisfaction, while most RPs were reduced by KPT-185
97% (71/73), BI-D1870 90% (66/73) and 1%FBS 84%
(61/73), the general proteome was only mildly affected with a decrease of 5% (177/3289) with KPT-185, 6% (187/3289) with BI-D1870 and 7% (220/3289) with 1%FBS. A proteomic analysis (mass spectrometry) of

Table 2. Gene ontology (GO) analyses
A. Altered proteins in MM-MSCs following treatment with KPT-185/BI-D1870/1%FBS (-2>FC>2, p<0.05)
Translation ToppGene Gorilla WebGestalt
Eukaryotic Translation 3.51E-72 1.94E-05 0.00E+00
Eukaryotic Translation Initiation 3.80E-59 4.86E-06
cotranslational protein targeting to membrane
translational initiation 8.37E-73
6.08E-55 2.71E-07
4.86E-06
SRP-dependent cotranslational protein targeting to membrane 1.26E-74 3.26E-07
mRNA metabolic process 1.76E-88 0.00E+00
Metabolism of amino acids and derivatives
Biosynthesis of amino acids 0.00E+00
0.00E+00
Amino Acid metabolism 0.00E+00
ribosome
RNA binding
1.22E-106
structural constituent of ribosome 7.24E-54 5.17E-06
mRNA metabolic process 1.76E-88 0.00E+00
RNA splicing
ribosome biogenesis 1.03E-36
8.14E-31 0.00E+00
0.00E+00
Formation of a pool of free 40S subunits 8.86E-64 4.04E-08
rRNA processing 5.74E-49 2.10E-02
Cytoplasmic Ribosomal Proteins
B. KPT-185/BI-D1870/1%FBS modified MM-MSCs MVs proteome (FC>-2, p<0.05)
ECM and adhesion 0.00E+00
cell adhesion molecule binding 1.77E-26 2.21E-11
cytoskeletal protein binding
integrin binding 1.63E-21
5.72E-17 1.50E-05
3.07E-05
actin binding 2.83E-11 1.50E-05
cadherin binding
fibronectin binding 3.82E-11
1.35E-08
proteoglycan binding 1.61E-08 4.26E-14
extracellular matrix organization 1.24E-25 0.00E+00
cell adhesion
Focal adhesion 1.92E-23
1.36E-15 2.21E-11
Integrin signalling pathway 1.03E-13 2.21E-11
Ensemble of genes encoding extracellular matrix and extracellular 2.18E-06 0.00E+00
matrix-associated proteins Metabolism
ATPase activity 2.01E-08 3.77E-01
Glycolysis and Gluconeogenesis
nucleotide metabolic process
3.96E-01 5.93E-02
nucleobase-containing small molecule metabolic process 1.00E+00
Immunity
myeloid leukocyte mediated immunity
Innate Immune System 7.98E-37
4.08E-20 6.62E-05
0.00E+00
human complement system 6.62E-05
Signaling
calcium ion binding
nucleoside-triphosphatase activity 3.45E-18
1.15E-13
3.90E-01 0.00E+00
purine nucleoside binding 2.74E-10 3.96E-01
protease binding
transferase activity, transferring amino-acyl groups 3.61E-07
1.82E-05
1.27E-01
S100 protein binding 2.25E-05
wound healing 9.53E-22
response to cytokine
RHO GTPases activate PAKs 3.85E-15
2.17E-06
3.14E-03
Urokinase-type plasminogen activator (uPA) and uPAR-mediated signaling 1.29E-05
VEGFA-VEGFR2 Pathway 3.10E-05 2.36E-13
Signalling to RAS
Ion transport by P-type ATPases 6.24E-05
7.48E-05
Vesicles
membrane fusion 1.36E-51

(continued)vesicle organization 2.33E-49
exocytosis
vesicle fusion to plasma membrane 1.63E-46
1.63E-46
1.90E-01
Phagosome 9.09E-20 0.00E+00
Vesicle-mediated transport 8.76E-07 3.63E-01
the treated cells demonstrated that the inhibitors (KPT- 185, BI-D1870 and to a lesser extent 1% FBS (NS) extensively attenuated RPs expression levels (n = 73; 28% 55%; P < 0.05; Fig 2C). Using the Venn dia- gram analysis we isolated the proteins commonly
attenuated by all three treatments compared to control (FC < 2, P < 0.05; Fig 2D) and studied their Gene Ontology (GO) enrichment with three different Bioin- formatics tools (Webgestalt, ToppGene, Gorilla; Table 2A). The assays demonstrated a major effect of
the inhibitors on pathways associated with protein syn- thesis. Taken together, our results demonstrate the effectiveness of our approach to downregulate RPs as a bulk in MM-MSCs. Our results also demonstrated the biological significance of reduced RPs expression to MM-MSCs’ TI status. Finally and pertinent to our study model the inhibitors displayed minimal effect on general cellular proteome.
MM-MSCs MVs. Next, we set out to test whether the reduced RPs in KPT-185/BI-D1870/1%FBS treated MM-MSCs translated into reduced RPs’ load in their MVs. We collected the treated and untreated control MM-MSCs’ conditioned media after 3 days in culture, isolated MVs and extracted their proteins. Immuno-
blotting of MVs lysates demonstrated yet again an attenuation of the RPs panel (Fig 3A; P < 0.05). Encouraged by these results we tested RPs levels in the MVs using mass spectrometry. This analysis demon- strated unequivocally that KPT-185 and BI-D1870 but
not 1%FBS diminished bulk presence of RPs in MM- MSCs MVs (Fig 3B, P < 0.01). Of note, some RPs (n = 34 and n = 10, respectively; Table 1) were completely eliminated following KPT-185 and BI- D1870 treatments. Using the Venn diagram analysis
we isolated the proteins that were commonly enriched (n = 519) in the MVs after all 3 treatments ( 2 < FC; Fig 3C) and assayed their functional Gene Ontology (GO) with 3 different Bioinformatics tools (Webges- talt, ToppGene, Gorilla). Our results demonstrated that
the modified MVs cargo was enriched in adhesion, metabolism, immune and vesicles related proteins. Importantly they lacked proteins implicated and essen- tial for translation related process (Table 2b).
MM cell lines treated with modified MM-MSCs MVs. Having established that KPT-185, BI-D1870 and to a lesser extent 1%FBS are capable of reducing the pres- ence of RPs in MM-MSCs MVs we decided to test

whether this strategy curtailed the pro-MM effect of these treated MVs compared to untreated MM-MSCs MVs. In order to resolve this we applied MVs from KPT-185, BI-D1870 or 1%FBS treated/untreated MM- MSCs to MM cell lines (U266 and MM1S) and assayed their viability and cell count after 3 days (Fig 3D, E). As we have described previously, the untreated MM- MSCs MVs elevated the recipient MM cell lines’ via-
bility and proliferation ( 50% 80%; P < 0.01). In comparison to MM cells treated with unconditioned
MM-MSCs MVs, MM cells (U266/MM1S) treated with KPT-185/BI-D1870 MM-MSCs MVs and in some instance 1%FBS showed significant decrease of live cells count, total cell count and viability, with no significant effect on dead cell count (Fig 3D, E)
( 10% 30%; P < 0.05). Our results demonstrate that the MM-MSCs’ MVs with depleted RPs had less pro-
tumorigenic effect on the recipient MM cells viability/ proliferation. These findings are also supported by the lack of modified MM-MSCs MVs induction of TI (sup- plementary Fig 1D). Specifically, while the un-treated MM-MSCs MVs increased TI status (eg, factors [eIF4E, eIF4GI; their regulators {mTOR, ERK, JNK}]; and targets [SMAD5, NFkB]) MVs from KPT-185/BI- D1870/1%FBS treated MM-MSCs did not activate the TI process. These findings substantiate that increased RPs presence in the MM-MSCs MVs is critical to their pro-MM effect.
RPs are exported into BM-MSCs MVs unequally in health and myeloma. Our observations so far have substanti- ated that source dependent (ND, MM) differential expression of RPs in BM-MSCs and their MVs is involved in modulation of recipient MM cells’ pheno- type. Next, we wanted to examine whether the alloca- tion of the RPs to MVs is proportional to their cellular expression or controlled as well. For this we compared the respective distribution of RPs in origin cells and their derived MVs in ND and MM-MSCs (mass spec- trometry; Fig 4A). Heatmap unsupervised hierarchical clustering of the RPs levels in MSCs supported our findings so far by indicating higher level of RPs in MM-MSCs (depicted in red) compared to the lower levels of RPs in ND-MSCs (relatively presented in blue). Interestingly, when assessing the RPs distribu- tion into the MVs it was readily observed that it is sig- nificantly higher in MM-MSCs MVs than in their respective normal counterparts (ie, RPL3). This

Fig 3. The effect of KPT-185/BI-D1870/1%FBS on MM-MSCs MVs RPs load and recipient MM cells viability and proliferation.MM-MSCs (n = 4) were treated with BI-D1870 (5 mm), KPT-185 (50 nm) and decreased level of FBS in media (1%) for 72h. Conditioned media were collected from the treated cells and MVs were sepa- rated. MVs’ proteins were extracted and assayed for RPs expression levels: RPL37a, RPS27a, RPL19, and RPS6 (immunoblotting). All immunoblots were quantified and normalized for Histone H3 (housekeeping). Results are presented in bar graph and expressed as percent (mean § SE, n ≥ 4) of respective protein expression in
control MM-MSCs MVs not treated with RPs inhibitors (A). Bulk RPs expression levels upon treatment were analyzed, normalized to RPL13a and presented in box
graph (B). The mass spectrometry unique and common altered RPs in MM-MSCs MVs after treatments were presented by Venn diagram (C). U266 and MM1S (n = 3) were incubated with MVs from MM-MSCs treated with RPs inhibitors (BI-D1870, KPT-185, 1% FBS) for 72 h, and assayed for live/dead/total cell count (try- pan blue) and viability (WST1) (D, E). Results are presented in bar graphs and expressed as percent (mean § SE, n ≥ 3) of respective parameter measured in control
U266/MM1S treated with unconditioned (naive) MM-MSCs MVs (*). Asterisks depict statistical significance (*P < 0.05, **P < 0.01, ***P < 0.001) (D, E).

4. RPs are exported into BM-MSCs MVs unequally (ND, MM) and enriched according to their function. Mass spectrometry of BM-MSCs (ND, MM) revealed 82 different RPs. Their expression levels in cells were analyzed and compared to respective expression levels in their secreted MVs. Expression levels are relatively presented in heatmap (A). MM-MSCs RPs expression levels were also compared to their expression levels in their MVs with/without treatments (KPT-185 and BI-D1870) and presented in heatmap (n = 3) (B). BM-MSCs MVs (ND/MM) enriched RPs were defined and divided into subgroups according to published literature: Tumor promotors (n = 40), suppressors (n = 23), and RPs with extra-ribosomal tumor promotor functions (n = 24) and with extra-ribosomal tumor suppressor functions (n = 14). RPs with no relevant function were excluded (n = 18). The tumor promotor/suppressors RPs expression levels were compared to the levels of the tumor promotor/sup- pressor RPs with extraribosomal functions and presented in bar graph (C). Asterisks depict statistical signifi-
cance (*P < 0.05).

phenomenon is even more profound when addressing RPs with equally moderate levels in both ND and MM- MSCs source cells such as RPL27a that present with extremely high levels in the MM-MSCs MVs only. These findings suggest that not only are RPs overex- pressed in MM-MSCs but that they are actively and increasingly transported to the MVs. Indeed, in assess- ment of the XPO-1 inhibitor compared to the RSK inhibitor significant changes arise (Fig 4B). Both KPT- 185 and Bi-D1870 attenuate the expression of the 73 RPs in the treated MM-MSCs, yet only the XPO-1 inhi- bition caused a profound ablation of RPs presence in MVs (Fig 4B, Table 1). Unsurprisingly, the number of RPs detected in the mass spectrometry analysis with known function as tumor promoters is higher in MM-

MSCs compared to RPs with established tumor sup- pressor activity (n = 40 and n = 14, respectively; data not detailed). This observation is in concordance with the RPs indispensable role in ribosome construction and the importance of protein synthesis to malignant progression. Remarkably, the distribution of RPs with extra-ribosomal tumor promoting function demon- strated additional enrichment in the MM-MSCs MVs whereas their presence in ND-MSCs MVs was reduced
( 112% and 65%, respectively; P < 0.05; Fig 4C). In contrast, RPs with known extra-ribosomal tumor sup-
pressing function were equally distributed to MM- MSCs MVs and ND-MSCs MVs (Fig 4D). These results suggest an active and organized delivery of RPs into MM-MSCs MVs. It remains to be determined

whether export from the nucleus is the key to this trans- port or additional steps are necessary.
Heterogeneity of RPs expression in MM-MSCs defines distinct sub-populations. With the growing understanding that tumor heterogeneity is critical to clinical manage- ment and outcome and the accumulating evidence that the cancer niche co-evolves with the malignancy,30-32 we wondered whether elevated RPs characterize all MM-MSCs or specific cell sub-groups. Moreover, PCA presented in Fig 1E depicts a higher rate of heterogene- ity in the MM-MSCs group than in the ND-MSCs sam- ples which are relatively similar to each other.
To address this question we assayed by Flow Cytom-
etry the expression of 10 select RPs in samples of ND- MSCs (n = 6) and MM-MSCs (n = 7). The RPs for this

assay were chosen based on their overexpression and/ or literature associating them with cancer and extra- ribosomal functions (supplementary Table 1). As we suspected, the increased expression of the RPs was not similar in all MM-MSCs’ samples or for all RPs. Spe- cifically, ND-MSCs displayed different distribution of cells averaging around a certain expression level for each RP. In general, it is visibly evident that ND- MSCs display a dot plot of RP expression in cells cen- tered on a single focus whereas the MM-MSCs display dot plots with two average expression levels, indicating a split population of 2 subgroups for the assayed RPs (Fig 5). On the contrary, all MM-MSCs displayed 2 subpopulations of cells averaging around different RPs’ expression levels for several of the assayed RPs

5. Analysis of RPs expression in MM-MSCs defines distinct subpopulations. BM-MSCs (ND, MM; n ≥ 6 each) were fixed, permeabilized and stained (single staining) for 10 different RPs with primary antibodies indi- cated at the top of each graph (detailed in materials and methods). Fluorescent secondary antibodies were added according to primary antibodies source (Rabbit, Mouse) and analyzed with flow cytometry (>10,000 cells/sam-
ple). Unstained and isotype control were used as negative controls (not shown). MFI of the positively stained
cells were determined (low and high) and relative distribution into the groups registered (percentage of total cells assayed). Relative distribution in assayed MM-MSCs samples (n = 7) is presented in bar graphs with color- ing indicating expression levels of respective RPs (yellow-negative, blue- low, red-high). Representative FACS plots of ND and MM are depicted beneath bar graphs.
. Furthermore, all RPs assayed had a varied expression in at least several of the MM-MSCs sam- ples. The 2 populations of MM-MSCs expressed all RPs at equal or higher levels than the ND-MSCs. Alto- gether, we can conclude that the increase in RPs expression in MM-MSCs is not a homogenous process but with different distributions in 2 distinct subgroups of the cells and a diverse panel of RPs. These findings may indicate a categorical shift in expression of RPs in a subgroup of MM-MSCs. Additional studies are war- ranted to determine the co-expression of the elevated RPs in the same cells. Finally, we wanted to determine if this heterogeneity is indeed transferred to the recipi- ent MM cells. For this purpose we chose MVs from two varied MM-MSCs samples with distinct differen- ces in RPs expression patterns. Specifically, we col- lected MVs from MM1 that displayed high expressions of RPL4, RPS27a, and RPS6 in addition to low expres- sion of RPL6. We also collected MVs from MM7 that displayed high expressions of RPL6, RPS27a and RPS6 along with low expression of RPL4. MVs from each sample (50 mg) were applied to MM1S cells for 24 hours after which the myeloma cells were assayed for RPs expression by FACS. Untreated MM1S cells were used to determine the base line expression of the RPs prior MVs’ mediated transfer (MFI). As expected we witnessed an increase in MFI of RPS27a and RPS6
in MM cells treated with MVs from both sources of MM-MSCs (MM1, MM7; 12% 20%, P < 0.05).
Yet, increased MFI of RPL4 ( 16%, P < 0.05) was
observed in MM1S treated with MM1 MVs only and
increased levels of RPL6 was witnessed in MM1S treated with MM7 MVs only ("20%, P < 0.01).

DISCUSSION
Our current study delineates a differential presence of RPs in MVs derived from BM-MSCs under normal condi- tions and myelomatous settings with increased levels in MM-MSCs MVs. We also depicted a mixed expression of the RPs in the MM-MSCs themselves creating two distinct sub-groups one with normal levels of RPs and another with increased RPs expression. As we have previously shown, the RPs-enriched-MVs are internalized by neigh- boring MM cells prompting increased TI dependent prolif- eration,8,24 By depleting the RPs content in the MM-MSCs MVs and applying them to MM cells we have demon- strated here that the TI dependent pro-MM effect was abro- gated indicating that the RPs load is essential. We also showed that by exposing MM cells to MM-MSCs with dif- ferent patterns of RPs expression we altered the internal MM RPs content suggesting that the tumor microenviron- ments’ heterogeneity promotes diversity in the malignant

cells. Taken together, these observations elaborate on our previous understanding that elevated TI in MM is instru- mental to disease progression and drug resistance6,9,10,33 and substantiate a role for RPs delivery from reprog- rammed neighboring BM-MSCs.
MM clonal heterogeneity is a major obstacle to dis- ease cure and at the basis of minimal residual disease (MRD) and relapse.34 Both genetic and nongenetic fac- tors underlie the constant appearance of new subclones and intratumoral variety.35,36 Interestingly, the diver- sity of MM clones is clearly witnessed at the spatial pathological level with different mutational character- istics of MM cells at different BM locations in the same patient as well as the coexistence of more than one clone at a specific location.35 Current understand- ing of the malignant process and progression acknowl- edges the primary role of external evolutionary forces such as cellular and secreted stroma as well as thera- peutics.35 Our observations regarding the delivery of BM-MSCs MVs with heterogeneous cargoes to the malignant MM cells are consistent with this view.
The discovery that RPs are more than passive constit- uents of ribosomes is recent and still largely disre- garded.17,18 Our initial observation of RPs’ overexpression in MM-MSCs and their secreted MVs suggested a deliberate alteration in the MM microenvi- ronment with possible relevance to the carcinogenic pro- cess. This is also reflected in the existence of two distinct MM-MSCs’ subgroups expressing different lev- els of RPs rather than a shift of the average expression in all the MM-MSCs, which would reflect a progressive and common increase in RPs levels. By depleting the RPs cargo in the MM-MSCs MVs and cancelling their pro-MM effect we reinforced this understanding. It is readily understood how increased ribosome biogenesis contributes to intensive protein synthesis needed by can- cer cells yet accumulating data indicate that RPs may introduce competitive advantages by modifying cellular proteome as well, particularly in cancer.37 Acquired aberrations in RPs have been described in glioma, colo- rectal cancers, chronic lymphocytic leukemia, and T- cell acute lymphoblastic leukemia (mutations and/or biogenesis).38 There have also been several reports of specific low occurring mutations in RPL5, RPL10 and over expression of RPS9 in MM that affected cell phe- notype and clinical response.15,39-41 Pathologies stem- ming from alterations in ribosome biogenesis and haploinsufficiency have been coined ribosomopathies38 and the best described disease in this category is Dia- mond-Blackfan anemia (DBA).42 Ribosomopathies are tissue specific and most often manifest in cells with high rates of protein synthesis and proliferation. Consistently, hematological systems are more prone to present with

symptoms upon inheritance of genetic mutations in ribo- somal proteins.38
Patients with ribosomopathies suffer from BM fail- ure on the one hand and higher risk to develop cancer on the other.38 There are several theories that address the tissue specificity of ribosomopathies and the patients’ propensity for malignant transformation. It is suggested that RPs may direct the specificity or transla- tional efficiency of ribosomes by creating “specialized ribosomes” and in the context of cancer “onco- ribosomes.”43,44 Another approach suggests that the mere concentration of ribosomal proteins may drive the favored translation of specific mRNAs at the expense on other less ideal candidates.43,44 A third the- ory considers the site of RPs biogenesis, the nucleolus, as the signaling hub that translates RPs imbalance into altered homeostasis and neoplastic promoters.43,44 Spe- cifically, RPs imbalance causes local stress and meta- bolic shifts that underlie aging, neurodegeneration and cancer and possibly provide a prolific setting for accu- mulation of malignant clones.45 Finally, chronic inflammation is characteristic of patients with riboso- mopathies and may also explain elevated cancer risk.38 Taken together, it is unsurprising that MM known for its elevated protein production as well as activated stress and inflammatory responses also presents with changes in ribosomal proteins.
Published data demonstrated changes in specific RPs expression in cancer cells. In contrast, we have shown alterations in the MM niche. We also depicted a profound shift in bulk RPs expression rather than isolated changes. Most importantly, we observe these changes in a sub- group of MM-MSCs whereas the remaining MM-MSCs resemble the ND-MSCs in their RPs expression. These data may suggest a reprogramming and/or expansion of a select group of BM-MSCs in the myelomatous niche. Several of our observations support the former option since we have previously shown that exposure ND-MSCs to MM cells alters their phenotype7 and repetitive expo- sure alters the ND-MSCs into MM supporters in vitro (unpublished data). Ongoing studies by our group are underway to address this very interesting possibility that MM-MSCs originate from ND-MSCs upon continuous exposure to ever expanding MM clones.
There is paucity of data regarding the malignant niche with respect to RPs expression and even less descriptions of RPs expressed externally to the cell.46-
48 Interestingly, a recent publication described the hori-
zontal transfer of ribosomal duplicated coding genes in bacterial evolution,49 which may enable the attainment of RPs functions that are distinct from ribosome struc- ture and translation. To the best of our knowledge, our report is the first to describe RPs transfer between human cells and in the neoplastic context.

In addressing the RPs as bulk and not the quantity of specific proteins we needed to devise a general method of inhibition to test the relevance of their delivery into MM cells. For this purpose we used three strategies: the first, highly unspecific, starvation by depleted serum known to restrain the energy consuming process of translation; the second, targeted p90 ribosomal S6 kinase RSK essential for RPs biogenesis. We used BI- D1870, an ATP-competitive inhibitor of the N-terminal AGC kinase domain, that targets RSK1-4 only and pre- vents the phosphorylation of GSK3a, GSK3b, and LKB150,51; the final approach included the inhibition of RPs export from the nucleus to the cytoplasm. We used KPT-185 known for its antiproliferative effects in cancers such as Mantle Cell Lymphoma.27 XPO1, its target, is a recognized nuclear exporter of multiple oncogenes as well as RPs. Interestingly, this approach yielded profound insight into the RPs delivery into MVs. In fact, while both BI-D1870 and KPT-185 diminished the expression of RPs in the MM-MSCs in a similar fashion they had profoundly different effects on MVs contents. The targeting of RPs export from the nucleus into the cytoplasm was sufficient to limit their delivery into the membranal fission sights of the MVs. These observations suggest that regulation of the cyto- plasmic levels of RPs is sufficient to control its hori- zontal delivery to other cells yet additional studies are required to substantiate this discernment.
The targeting of XPO1 was already flagged in MM therapy52,53 and our findings provide insight into a new underlying mechanism that may prompt novel and more effective drug combinations in the future that combine the targeting of the stroma on top of the MM cells.
Data accumulated in the past several years indicates that translation is driven by ribosome heterogeneity and that different ribosomal components lead to diverse translational signatures.54 Our findings previously6,8-10 and in the current study depict a differential role for BM-MSCs in proteome design of co-cultured MM cells according to their normal or MM source. Yet, the transi- tion from ND-MSCs to MM-MSCs has not been addressed. We conjectured that this reprogramming of the MM niche is a process and that not all cells undergo changes at the same pace or direction. Taking it a step forward, we speculated that the robust expression of RPs in the MM-MSCs may afford an ideal means to generate heterogeneity in the BM niche and the MM cells residing in it. In order to test our hypothesis we altered our approach from analysis of bulk RPs expres- sion to the assay of the cellular expression of select RPs in BM-MSCs of ND and MM subjects by flow cytome- try, which allowed us a population based approach. We concentrated on RPs highly overexpressed in the MM- MSCs MVs and/or with published reports associating

them with cancer. Consistent with our preface, we observed that ND-MSCs were relatively homogeneous in RPs expressions whereas MM-MSCs were composed of distinct sub-populations differing in RPs expression levels and diversity. Specifically we determined all 10 assayed RPs were heterogeneously expressed in at least several of the MM-MSCs samples, always more prominently than in ND-MSCs and most often in 2 distinct sub-groups. Fur- thermore, all RPs assayed displayed heterogeneity in some of the myeloma samples. It remains to be determined if the phenomenon is progressive with an end stage of high expressing cells only and whether the expression pattern may correspond to any clinical characteristics in the MM patients or treatment protocols.
Scientific reports have begun depicting roles for spe- cific RPs in malignancy with underlying participation in unique functions separate from ribosome configuration (supplementary Table 1). Proteomics studies have dem- onstrated that the stoichiometric relationships within translating polysomes promote preferential translation of specific mRNAs.54,55 This was specifically demonstrated with RPL10a.54 The mechanistic explanation for RPL10a selectivity was suggested to be its binding to IRES (Inter- nal Ribosome Entry Site).55 Another study reported that leukemia cells display a mutation in RPL10 that allows increased IRES dependent BCL2 expression and survival under hypoxic conditions.45 This is particularly intriguing in light of the increased eIF4GI expression in MM cells also assigned with IRES binding capabilities.56,57 In our study RPL10a was the most elevated among all RPs in the MM-MSCs MVs (FC 166). Moreover, analyses of its expression in MM-MSCs demonstrated a unanimous het- erogeneous representation.
At the launch of this study we performed high throughput mass spectrometry proteomics analysis that provided insight into the aberrant expression of RPs in the MM microenvironment. The biological importance of this observation was further corroborated in func- tional in vitro studies while favoring an approach that addressed the RPs as a bulk. In continued work we determined that the RPs imbalance is unequal in MM- MSCs and can define subgroups in these cells.
Taken together, our observations highlight a new mech- anism for reprogramming the tumor microenvironment (BM-MSCs) and translating these changes into direct design of the cancer cells (via MVs). Further studies are planned to determine the consequence of the RPs expres- sion profiles in MM-MSCs sub-populations in defining the MSCs themselves and in manipulating adjacent MM cells’ proteome and phenotype. Positive results are expected to yield new markers for MM progression, improve progno- sis and perhaps drug response, and most importantly afford new therapeutic targets in the cancer niche and means to monitor and optimize treatment timing.

DECLARATIONS
Ethics approval and consent to participate: All par- ticipants signed informed consent forms approved by Meir Medical Center Helsinki Committee.

AVAILABILITY OF DATA AND MATERIAL
Datasets of proteomics analyses and other material related to the current study are readily available upon request.

AUTHOR CONTRIBUTIONS
Mahmoud Dabbah conceived, carried out experi- ments, analyzed data and wrote the paper, Michael Lish- ner, Shelly Tartakover Matalon and Avivit Neumann advised on experimental design and data analysis, Osnat Jarchowsky collected MM patients for the study and analyzed data, Yaron S. Brin supplied the femur heads for the study, Metsada Pasmanik-Chor analyzed the mass spectrometry results and performed bioinformatics analysis, Liat Drucker conceived experiments, analyzed data and wrote the paper. All authors had final approval of the submitted and published versions.

ACKNOWLEDGMENTS
Conflicts of interest: There is no conflict of interest. This work was supported by the Israel Cancer Asso-
ciation (ICA) #KM281100124, and the Dotan Hemato- Oncology Center, Tel Aviv University # 0601243683.
The study was performed in partial fulfillment of the requirements for a Ph.D. degree by Dabbah Mahmoud, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel. We are grateful to the staff of the Hemato- cytological Laboratory for its dedicated technical sup- port, the orthopedics and hematologists for their willing participation in collection of samples at Meir Medical Center, Kfar Saba, Israel. We are also most grateful to Adam Neumann for his generous support and faith in our work. All authors have read the jour- nal’s policy on conflicts of interest and journal’s authorship agreement. There is no conflict of interests.

SUPPLEMENTARY MATERIALS
Supplementary material associated with this article can be found in the online version

REFERENCES

1. Ramon YCS, Castellvi J, Hummer S, Peg V, Pelletier J, Sonen- berg N. Beyond molecular tumor heterogeneity: protein synthe- sis takes control. Oncogene 2018;37:2490–501.
2. Ria R, Vacca A. Bone marrow stromal cells-induced drug resis- tance in multiple myeloma. Int J Mol Sci 2020;21.
3. Schurch CM, Rasche L, Frauenfeld L, Weinhold N, Fend F. A review on tumor heterogeneity and evolution in multiple mye- loma: pathological, radiological, molecular genetics, and clinical integration. Virchows Arch 2020;476:337–51.
4. Xu S, De Veirman K, De Becker A, Vanderkerken K, Van Riet I. Mesenchymal stem cells in multiple myeloma: a therapeutical tool or target? Leukemia 2018;32:1500–14.
5. Roccaro AM, Sacco A, Maiso P, et al. BM mesenchymal stromal cell-derived exosomes facilitate multiple myeloma progression. J Clin Invest 2013;123:1542–55.
6. Marcus H, Attar-Schneider O, Dabbah M, et al. Mesenchymal stem cells secretomes’ affect multiple myeloma translation initi- ation. Cell Signal 2016;28:620–30.
7. Dabbah M, Attar-Schneider O, Zismanov V, Tartakover Matalon S, Lishner M, Drucker L. Multiple myeloma cells promote migration of bone marrow mesenchymal stem cells by altering their translation initiation. J Leukoc Biol 2016;100:761–70.
8. Dabbah M, Attar-Schneider O, Tartakover Matalon S, et al. Microvesicles derived from normal and multiple myeloma bone marrow mesenchymal stem cells differentially modulate mye- loma cells’ phenotype and translation initiation. Carcinogenesis 2017;38:708–16.
9. Attar-Schneider O, Zismanov V, Dabbah M, Tartakover-Mata- lon S, Drucker L, Lishner M. Multiple myeloma and bone mar- row mesenchymal stem cells’ crosstalk: effect on translation initiation. Mol Carcinog 2016;55:1343–54.
10. Ibraheem A, Attar-Schneider O, Dabbah M, et al. BM-MSCs- derived ECM modifies multiple myeloma phenotype and drug response in a source-dependent manner. Transl Res 2019;207:83–95.
11. Wang SE. Extracellular vesicles and metastasis. Cold Spring Harbor perspectives in medicine; 2019.
12. Cufaro MC, Pieragostino D. Extracellular vesicles and their potential use in monitoring cancer progression and therapy: the contribution of proteomics. J Oncol 2019;2019:1639854.
13. Guimaraes JC, Zavolan M. Patterns of ribosomal protein expres- sion specify normal and malignant human cells. Genome Biol 2016;17:236.
14. Raiser DM, Narla A, Ebert BL. The emerging importance of ribosomal dysfunction in the pathogenesis of hematologic disor- ders. Leuk Lymphoma 2014;55:491–500.
15. Hofman IJF, Patchett S, van Duin M, et al. Low frequency muta- tions in ribosomal proteins RPL10 and RPL5 in multiple mye- loma. Haematologica 2017;102:e317–20.
16. Hofman IJF, van Duin M, De Bruyne E, et al. RPL5 on 1p22.1 is recurrently deleted in multiple myeloma and its expression is linked to bortezomib response. Leukemia 2017;31:1706–14.
17. Ferretti MB, Karbstein K. Does functional specialization of ribosomes really exist? RNA (New York, NY). 2019;25:521–38.
18. Penzo M, Montanaro L, D T, M D. The ribosome biogenesis- cancer connection. Cells 2019;8.
19. Lessard F, Brakier-Gingras L, Ferbeyre G. Ribosomal proteins control tumor suppressor pathways in response to nucleolar

stress, 41. BioEssays: news and reviews in molecular, cellular and developmental biology; 2019:e1800183.
20. Pelletier J, Thomas G, Volarevic S. Ribosome biogenesis in can- cer: new players and therapeutic avenues. Nat Rev Cancer 2018;18:51–63.
21. Zhou X, Liao WJ, Liao JM, Liao P, Lu H. Ribosomal proteins: functions beyond the ribosome. J Molecul Cell Biol 2015;7: 92–104.
22. de Las Heras-Rubio A, Perucho L, Paciucci R, Vilardell J, LL ME. Ribosomal proteins as novel players in tumorigenesis, 33. Cancer metastasis reviews; 2014:115–41.
23. Stepinski D. The nucleolus, an ally, and an enemy of cancer cells. Histochem Cell Biol 2018;150:607–29.
24. Dabbah M, Jarchowsky-Dolberg O, Attar-Schneider O, et al. Multiple myeloma BM-MSCs increase the tumorigenicity of MM cells via transfer of VLA4 enriched microvesicles. Carcino- genesis 2020;41:100–10.
25. Wang J, Zhang J, Lee YM, et al. Quantitative chemical proteo- mics profiling of de novo protein synthesis during starvation- mediated autophagy. Autophagy 2016;12:1931–44.
26. Bain J, Plater L, Elliott M, et al. The selectivity of protein kinase inhibitors: a further update. Biochem J 2007;408:297–315.
27. Tabe Y, Kojima K, Yamamoto S, et al. Ribosomal Biogenesis and Translational Flux Inhibition by the Selective Inhibitor of Nuclear Export (SINE) XPO1 antagonist KPT-185. PloS One 2015;10:e0137210.
28. Sun L, Wang HX, Zhu XJ, et al. Serum deprivation elevates the levels of microvesicles with different size distributions and selectively enriched proteins in human myeloma cells in vitro. Acta Pharmacol Sin 2014;35:381–93.
29. Haraszti RA, Miller R, Dubuke ML, et al. Serum deprivation of mesenchymal stem cells improves exosome activity and alters lipid and protein composition. iScience 2019;16:230–41.
30. Morgan GJ, Walker BA, Davies FE. The genetic architecture of multiple myeloma. Nat Rev Cancer 2012;12:335–48.
31. Rasche L, Kort€um KM. The impact of tumor heterogeneity on diagnostics and novel therapeutic strategies in multiple mye- loma. Int J Mol Sci 2019;20.
32. Peinado H, Lavotshkin S, Lyden D. The secreted factors respon- sible for pre-metastatic niche formation: old sayings and new thoughts. Semin Cancer Biol 2011;21:139–46.
33. Attar-Schneider O, Drucker L, Zismanov V, Tartakover-Matalon S, Rashid G, Lishner M. Bevacizumab attenuates major signal- ing cascades and eIF4E translation initiation factor in multiple myeloma cells. Lab Invest 2012;92:178–90.
34. Ledergor G, Weiner A, Zada M, et al. Single cell dissection of plasma cell heterogeneity in symptomatic and asymptomatic myeloma. Nat Med 2018;24:1867–76.
35. Sch€urch CM, Rasche L, Frauenfeld L, Weinhold N, Fend F. A review on tumor heterogeneity and evolution in multiple mye- loma: pathological, radiological, molecular genetics, and clinical integration. Virchows Arch 2020;476:337–51.
36. Heuck CJ, Mehta J, Bhagat T, et al. Myeloma is characterized by stage-specific alterations in DNA methylation that occur early during myelomagenesis. J Immunol 2013;190:2966–75.
37. Bustelo XR, Dosil M. Ribosome biogenesis and cancer: basic and translational challenges. Curr Opin Genet Dev 2018;48:22–9.
38. Piazzi M, Bavelloni A, Gallo A, Faenza I, Blalock WL. Signal transduction in ribosome biogenesis: a recipe to avoid disaster. Int J Mol Sci 2019;20.
39. Song KP, Shu P, Ma JL, Wang B, Chen B. Ribosomal protein S9 expression in multiple myeloma and its effect on cell biological

function. Zhongguo yi xue ke xue yuan xue bao Acta Academiae Medicinae Sinicae 2019;41:175–82.
40. Ali M, Ajore R, Wihlborg AK, et al. The multiple myeloma risk allele at 5q15 lowers ELL2 expression and increases ribosomal gene expression. Nat Commun 2018;9:1649.
41. Hofman IJF, van Duin M, De Bruyne E, et al. RPL5 on 1p22.1 is recurrently deleted in multiple myeloma and its expression is linked to bortezomib response. Leukemia 2017;31:1706–14.
42. Aspesi A, SR E. Rare ribosomopathies: insights into mecha- nisms of cancer. Nat Rev Cancer 2019;19:228–38.
43. Sulima SO, Kampen KR, De Keersmaecker K. Cancer biogene- sis in ribosomopathies. Cells 2019;8.
44. Farley-Barnes KI, Ogawa LM, Baserga SJ. Ribosomopathies: old concepts, new controversies. Trends Genet 2019;35:754–67.
45. Kampen KR, Sulima SO, Verbelen B, et al. The ribosomal RPL10 R98S mutation drives IRES-dependent BCL-2 transla- tion in T-ALL. Leukemia 2019;33:319–32.
46. Baroncelli M, van der Eerden BC, Kan YY, et al. Comparative proteomic profiling of human osteoblast-derived extracellular matrices identifies proteins involved in mesenchymal stromal cell osteogenic differentiation and mineralization. J Cell Physiol 2018;233:387–95.
47. Prisic S, Hwang H, Dow A, et al. Zinc regulates a switch between primary and alternative S18 ribosomal proteins in Mycobacterium tuberculosis. Mol Microbiol 2015;97:263–80.
48. Wei X, Cao S, Zhang L, et al. Comparative proteome analysis of the extracellular proteins of two Haemophilus parasuis strains Nagasaki and SW114. Biochem Biophys Res Commun 2014;446:997–1001.
49. Makarova KS, Ponomarev VA, Koonin EV. Two C or not two C: recurrent disruption of Zn-ribbons, gene duplication, lineage-spe- cific gene loss, and horizontal gene transfer in evolution of bacte- rial ribosomal proteins. Genome Biol 2001;2.Research 0033.
50. Sapkota GP, Cummings L, Newell FS, et al. BI-D1870 is a spe- cific inhibitor of the p90 RSK (ribosomal S6 kinase) isoforms in vitro and in vivo. Biochem J 2007;401:29–38.
51. Zhang Z, Liu R, Townsend PA. Proud CG. p90(RSK)s mediate the activation of ribosomal RNA synthesis by the hypertrophic agonist phenylephrine in adult cardiomyocytes. J Mol Cellular Cardiol 2013;59:139–47.
52. Gandhi UH, Senapedis W, Baloglu E, et al. Clinical impli- cations of targeting XPO1-mediated nuclear export in multi- ple myeloma. Clin Lymphoma Myeloma Leuk 2018;18: 335–45.
53. Azizian NG, Li Y. XPO1-dependent nuclear export as a target for cancer therapy. J Hematol Oncol 2020;13:61.
54.
Calamita P, Gatti G, Miluzio A, Scagliola A, Biffo S. Translat- ing the game: ribosomes as active players. Front Genet 2018;9:533.
55. Shi Z, Fujii K, Kovary KM, et al. Heterogeneous ribosomes pref- erentially translate distinct subpools of mRNAs genome-wide. Mol Cell 2017;67:71–83.
56. Attar-Schneider O, Drucker L, Zismanov V, Tartakover-Matalon S, Lishner M. Targeting eIF4GI translation initiation factor affords an attractive therapeutic strategy in multiple myeloma. Cell Signal 2014;26:1878–87.
57. Attar-Schneider O, Pasmanik-Chor M, Tartakover-Matalon S, Drucker L, Lishner M. eIF4E and eIF4GI have distinct and dif- ferential imprints on multiple myeloma’s proteome and signal- ing. Oncotarget 2015;6:4315–29.
58. Dabbah M, Attar-Schneider O, Zismanov V, Tartakover Matalon S, Lishner M, Drucker L. Multiple myeloma cells promote migration of bone marrow mesenchymal stem cells by altering their translation initiation. J Leukoc Biol 2016;100:761–70.
59. Attar-Schneider O, Zismanov V, Dabbah M, Tartakover-Mata- lon S, Drucker L, Lishner M. Multiple myeloma and bone mar- row mesenchymal stem cells’ crosstalk: effect on translation initiation. Mol Carcinog 2016;55:1343–54.
60. Attar-Schneider O, Zismanov V, Drucker L, Gottfried M. Secre- tome of human bone marrow mesenchymal stem cells: an emerg- ing player in lung cancer progression and mechanisms of translation initiation. Tumour Biol 2016;34:4755–65.
61. Tyanova S, Temu T, Sinitcyn P. The Perseus computational plat- form for comprehensive analysis of (prote)omics data. Nature Methods 2016;13:731–40.
62. Eden E, Navon R, Steinfeld I, Lipson D, Yakhini Z. GOrilla: a tool for discovery and visualization of enriched GO terms in ranked gene lists. BMC Bioinform 2009;10:48.
63. Chen J, Bardes EE, Aronow BJ, Jegga AG. ToppGene Suite for gene list enrichment analysis and candidate gene prioritization. Nucleic Acids Res 2009:W305–11, Jul;37(Web Server issue).
64. Wang J, Duncan D, Shi Z, Zhang B. WEB-based GEne SeT AnaLysis Toolkit (WebGestalt): update 2013. Nucleic Acids Res 2013;41(Web Server issue):W77–83.
65. Tai YT, Landesman Y, Acharya C, et al. CRM1 inhibition indu- ces tumor cell cytotoxicity and impairs osteoclastogenesis in multiple myeloma: molecular mechanisms and therapeutic implications. Leukemia 2014;28:155–65.
66. Ranganathan P, Yu X, Na C, et al. Preclinical activity of a novel CRM1 inhibitor in acute myeloid leukemia. Blood 2012;120:1765–73.
67. Roff´e M, Lupinacci FC, Soares LC, Hajj GN, Martins VR. Two widely KPT-185 used RSK inhibitors, BI-D1870 and SL0101, alter mTORC1 signaling in a RSK-independent manner. Cell Signal 2015;27:1630–42.