Abstract Purpose: Therapy resistance is a major clinical hurdle in bone cancer treatment and seems to be largely driven by poorly understood microenvironmental factors. Recent evidence suggests a critical role for a unique subpopulation of mesenchymal stem cells with inflammatory features (iMSC), though their origin and function remained unexplored. We demonstrate that cancer-secreted extracellular vesicles (EV) trigger the development of iMSCs, which hinder therapy response in vivo, and set out to identify strategies to counteract their function. Experimental Design: The role of iMSCs in therapy resistance was evaluated in an orthotopic xenograft mouse model of osteosarcoma. EV-induced alterations of the MSC transcriptome were analyzed and compared with single-cell RNA sequencing data of biopsies from patients with osteosarcoma and multiple myeloma. Functional assays identified EV components driving iMSC development. We assessed the efficacy of clinical drugs in blocking iMSC-induced resistance in vivo. Results: We found that iMSCs are induced by interaction with cancer EVs and completely abrogate the antimetastatic effect of TGFβ signaling inhibition. Importantly, EV-induced iMSCs faithfully recapitulate the inflammatory single-cell RNA signature of stromal cells enriched in biopsies from patients with multiple myeloma and osteosarcoma. Mechanistically, cancer EVs act through two distinct mechanisms. EV-associated TGFβ induces IL6 production, whereas the EV-RNA cargo enhances TLR3-mediated chemokine production. We reveal that simultaneous blockade of downstream EV-activated pathways with ladarixin and tocilizumab disrupts metastasis formation and overcomes iMSC-induced resistance. Conclusions: Our observations establish iMSCs as major contributors to drug resistance, reveal EVs as triggers of iMSC development, and highlight a promising combination strategy to improve therapy response in patients with bone cancer.

Suppl. Figure 8. a-c) RNA class distribution of osteosarcoma (n=10) (a) and multiple myeloma (n=5) (b) patient plasma EVs, and normalized counts (rpm) of known (TLR-3 and RIG-I activating) inflammatory RNAs in osteosarcoma plasma EVs, MM plasma EVs and 143B EVs (c). d) Expression levels of (endosomal and cytosolic) pattern recognition receptors in MSCs based on normalized RNA-seq counts. e-h) Knockdown confirmation of TLR3 (e), RIG-I (f), MDA5 (g) and LGP2 (h) in primary MSCs as compared to control (shGFP-transduced) cells. i) IL8 protein production in MSCs transduced with indicated shRNAs upon exposure to 143B EVs. Graph shows data from a replicate experiment performed using MSCs from donor #2.

Suppl. Figure 6. a-b) Expression of the most upregulated TGFβ-dependent (a) and independent/partially dependent (b) genes in bone marrow MSCs based on RNA-seq normalized counts. c-f) Relative expression levels of CXCL1, CXCL2, CXCL5, CXCL6 mRNAs in bone marrow MSCs (donor #1) exposed to 143B EVs in the presence or absence of the TGFBR1 inhibitor SB-431542 as assessed by RT-qPCR. g-m) Relative expression levels of IL6, IL8, CXCL1, CXCL2, CXCL3, CXCL5, CXCL6 mRNAs in bone marrow MSCs (donor #2) exposed to 143B EVs in the presence or absence of the TGFBR1 inhibitor EW-7197. n-o) Relative expression of IL6 and IL8 mRNAs in adipose-derived MSCs (donor #3) exposed to 143B EVs in the presence or absence of EW-7197.

Suppl. Figure 4. a-b) scRNA-seq mapping of human osteosarcoma tumors. MSC and other indicated clusters were identified through the UMAP (a) and marker gene expression analysis (b).

Suppl. Figure 3. a-e) Relative expression levels of IL6, IL8, CXCL2, CXCL3 and CXCL6 in MSCs from an independent donor (MSC#2) exposed to plasma-derived EVs from healthy donor #1 (H#1) and MM patient #1 (MM#1).

Suppl. Figure 7. a) Dynasore-mediated inhibition of PKH67-labeled EV uptake by MSCs as assessed by FACS. b-e) Relative expression levels of CXCL2, CXCL5, CXCL3, CXCL6 in MSCs treated with 143B EVs in the presence or absence of dynasore. Transcript levels are normalized to GAPDH and expressed as fold increase relative to the experimental controls (untreated or dynasore-treated MSCs). Graphs show the average of 3 (CXCL2 and CXCL6) or 2 (CXCL3 and CXCL5) experiments. Statistics was calculated on normalized expression data, *p < 0.05, two-tailed t test. f) Volcano plots depicting differentially expressed interferon stimulated genes (ISGs) in MSCs exposed to 143B EVs, and relative log2FC and FDR values. Human fibroblast (hF) EVs were used as control vesicles. g-j) Expression of the ISGs IFIT1, IFITM1, MX1 and OAS2 in BM-MSCs upon transfection with (143B) EV RNA pre-treated or not with RNase III or RNase A. k-n) Syntenin, IL8, CXCL2 and CXCL3 mRNA expression levels in 143B stably transduced with dox-inducible shRNAs against syntenin or GFP as non-targeting control, in the presence or absence of doxycycline. o-r) RAB35, IL8, CXCL2 and CXCL3 mRNA expression levels in 143B stably transduced with dox-inducible shRab35 or shGFP in the presence or absence of doxycycline.

Suppl. Figure 2. a-b) scRNA-seq mapping of the non-hematopoietic mononuclear cell fraction from bone marrow aspirates of MM patients and non-cancer control individuals. UMAP plot (a) and violin plots (b) of marker gene expression in the identified clusters. c) UMAP plot split into control and MM datasets. d-i) Density plots of differentially expressed cytokines and chemokine in control and MM samples.

Suppl. Figure 5. a-b) CD63 (a) and CD81 (b) protein quantification (western blot pixel density) in EVs released by engineered (shGFP, shSyntenin, shRab11b, shRab35) 143B cells. c) Growth curve of wt and engineered (shGFP, shRab11b, shRab35, shSyntenin) 143B showing no differences in cell growth over 72 hours). d-e) Relative mRNA expression levels of IL6 (d) and IL8 (e) in MSCs exposed to indicated volumes of wild type 143B EVs.

Suppl. Figure 1. a-b) Tumor growth measured by caliper (a) and BLI (b) in osteosarcoma-bearing mice receiving the TGFBR1 inhibitor EW-7197 and injected or not with 143B EV-educated MSCs. c) Western blot for CD63 and CD81 in EVs released by human fibroblasts (hF) and 143B, MDA-MB-231, PC3, HOS and MCF7 cell lines. d) GSEA enrichment plot for gene set ‘TNFα signaling via NFKB’ in 143B EV-exposed MSC compared to hF EV-exposed (ctrl) MSCs. e) List of gene sets enriched in 143B EV-exposed MSCs compared to hF EV-exposed (ctrl) MSCs.

Suppl. Figure 9. a) BLI images of tumor-bearing mice receiving Ladarixin and injected or not with human EV-induced iMSCs. b-c) Tumor growth measured by caliper (b) and lung nodules count (c) in tumor-bearing mice receiving ladarixin in combination with EW-7197 (TGFBR1 inhibitor) and injected or not with human EV-induced iMSCs.

Tissue-based biopsy is the present main tool to explore the molecular landscape of cancer, but it also has many limits to be frequently executed, being too invasive with the risk of side effects. These limits and the ability of cancer to constantly evolve its genomic profile, have recently led to the need of a less invasive and more accurate alternative, such as liquid biopsy. By searching Circulating Tumor Cells and residues of their nucleic acids or other tumor products in body fluids, especially in blood, but also in urine, stools and saliva, liquid biopsy is becoming the future of clinical oncology. Despite the current lack of a standardization for its workflows, that makes it hard to be reproduced, liquid biopsy has already obtained promising results for cancer screening, diagnosis, prognosis, and risk of recurrence.Through a more accessible molecular profiling of tumors, it could become easier to identify biomarkers predictive of response to treatment, such as EGFR mutations in non-small cell lung cancer and KRAS mutations in colorectal cancer, or Microsatellite Instability and Mismatch Repair as predictive markers of pembrolizumab response.By monitoring circulating tumor DNA in longitudinal repeated sampling of blood we could also predict Minimal Residual Disease and the risk of recurrence in already radically resected patients.In this review we will discuss about the current knowledge of limitations and strengths of the different forms of liquid biopsies for its inclusion in normal cancer management, with a brief nod to their newest biomarkers and its future implications.

Immuno-oncology demonstrated substantial efficacy in cancer treatment. Immune-related adverse events (irAEs) can virtually involve every organ, with different incidence depending on the different immune-checkpoint inhibitor. irAEs consequences can range from quality of life worsening and therapy discontinuation to death, if not recognized promptly. However, patients interrupting therapy due to irAEs in absence of progressive disease can benefit from immuno-oncology over time after discontinuation. We present the case of a man affected by metastatic renal cell carcinoma (mRCC) that experienced a long-term response to programmed cell death-1 inhibitor, nivolumab, after interruption due to immune-related pnenumonia. IrAEs can be associated to efficacy and very long-term response in mRCC patients treated with immuno-oncology.

Immune checkpoint inhibitors (ICIs) are largely used in the treatment of patients with advanced non-small-cell lung cancer (NSCLC). Novel biomarkers that provide biological information that could be useful for clinical management are needed. In this respect, extracellular vesicles (EV)-associated microRNAs (miRNAs) that are the principal vehicle of intercellular communication may be important sources of biomarkers. We analyzed the levels of 799 EV-miRNAs in the pretreatment plasma of 88 advanced NSCLC patients who received anti-PD-1 therapy as single agent. After data normalization, we used a two-step approach to identify candidate biomarkers associated to both objective response (OR) by RECIST and longer overall survival (OS). Univariate and multivariate analyses including known clinicopathologic variables and new findings were performed. In our cohort, 24/88 (27.3%) patients showed OR by RECIST. Median OS in the whole cohort was 11.5 months. In total, 196 EV-miRNAs out 799 were selected as expressed above background. After multiplicity adjustment, abundance of EV-miR-625-5p was found to be correlated with PD-L1 expression and significantly associated to OR by RECIST (p = 0.0366) and OS (p = 0.0031). In multivariate analysis, PD-L1 staining and EV-miR-625-5p levels were constantly associated to OR and OS. Finally, we showed that EV-miR-625-5p levels could discriminate patients with longer survival, in particular in the class expressing PD-L1 ≥50%. EV-miRNAs represent a source of relevant biomarkers. EV-miR-625-5p is an independent biomarker of response and survival in ICI-treated NSCLC patients, in particular in patients with PD-L1 expression ≥50%.

Bone metastases are common in genitourinary cancers, but they are underreported and not well researched. Synchronous bone metastases occur in 1.39-5.5% of bladder cancer patients, while 30-40% of cases are metachronous. Bone morphogenetic proteins (BMPs) play a key role in regulating proliferation, migration and invasion of tumor cells in bone microenvironment of bone metastases from metastatic urothelial carcinoma (mUC). Bone metastases represent a poor prognostic factor due to high morbidity and mortality correlated to skeletal-related events (SREs). The incidence rate of SREs in bladder, renal pelvis, and ureteral cancer varies from 39 to 68%. Radiotherapy is the most frequent treatment for SREs. The early use of bone targeted therapies (BTT), zoledronic acid and denosumab, improves SREs incidence and morbidity and it seems to improve overall survival (OS). To date, several new agents (immunotherapy and targeted drugs) demonstrated efficacy in mUC. However, subgroup analysis for bone metastases is often not available, due to difficulties in analysing bone samples, non-RECIST lesions and delay in systemic treatment due to SREs that limit the enrolment of bone mUC patients in clinical trials. Larger solid tumor studies that included UC patients are the main source of data for the management of mUC patients with bone metastases. For these patients, multidisciplinary approach should be preferred, involving orthopaedics, radiotherapists and rehabilitation to improve outcome and quality of life. New prospective trials should characterize clinical and molecular features of patients with bone metastases and the impact of new drugs on this poor prognostic metastatic site.