Human bone marrow derived mesenchymal stem cells do not promote oral cancer cell growth in vitro and metastasis in vivo

Abstract

Mesenchymal stem cells (MSCs), particularly those derived from bone marrow (BMMSCs), hold substantial promise for bone regeneration in the maxillofacial region, especially after surgical resections with bone involvement. However, their use in patients with resections undergoing oral cancer treatment poses potential risks due to the effects of MSCs in modulating cancer cell behavior. This study aimed to assess the effect of circulatory BMMSCs on proliferation, migration, invasion, tumor growth, and metastasis of oral squamous cell carcinoma (OSCC) cells. Human BMMSCs were isolated, characterized, and their conditioned medium (CM) was tested on OSCC cell lines (Ca1 and OSCC1). Further, BMMSCs were transduced with lentiviral particles to express firefly luciferase for live cell tracking in vivo to study their biodistribution and homing capability to xenograft tongue tumors. In vitro assays revealed that BMMSC-CM did not significantly alter OSCC proliferation or invasion in 3D organotypic assays, while significantly reducing their migration in 2D scratch wound assay. In vivo, bioluminescent imaging and histological analyses indicated that human BMMSCs predominantly localized to the lungs without homing to other organs or to the human xenograft tongue tumors. Moreover, circulatory BMMSCs did not influence tumor size, nor did they promote lung metastasis in xenografted mice under these conditions. These findings suggest that circulating BMMSCs did not exacerbate OSCC progression, supporting their potential use in regenerative applications for patients post-OSCC resection.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-38370-5.

Introduction

Mesenchymal stem cells (MSCs) have shown substantial promise in the field of bone regeneration, offering new possibilities to treat a range of skeletal conditions and injuries¹. Although MSCs derived from different sources have different functions, they are characterized by their ability to self-renew and differentiate into various cell types, including osteoblasts, chondrocytes, adipocytes, and myocytes². In addition, MSCs possess immunomodulatory properties and can secrete various growth factors that promote tissue regeneration and repair³. Bone-marrow derived MSCs (BMMSCs) generally have a higher bone regeneration ability due to their capability of differentiating into bone-forming cells as well as their immune modulatory functions³. Moreover, BMMSCs have shown to migrate to injured tissues including cancer lesions thus emerging as a promising vehicle for drug delivery systems in cancer medicine⁴.

Oral squamous cell carcinoma (OSCC) is a common malignancy that arises in the epithelium of the oral mucosa⁵. Several established risk factors contribute to OSCC development, including tobacco use, excessive alcohol consumption, areca nut, and human papillomavirus (HPV) infection. The main therapeutic strategies involve surgery and radiochemotherapy⁶. Most often, OSCC is detected at late-stage and involves jaw resection as a surgical intervention creating a critical-sized defect that cannot heal without craniofacial reconstruction⁶. Unfortunately, these interventions are destructive to the hosts leading to severe physiological dysfunction, including issues with speech, chewing, swallowing, and breathing in addition to the detrimental psychosocial impact.

Newer regenerative approaches including the use of MSCs are being considered for reconstruction of major bone loss due to surgical resection^7,8. However, despite their immense potential, the use of MSCs in bone regeneration may pose certain challenges and risks, particularly when using them in patients undergoing OSCC treatment. It has been shown that MSCs have the ability to elicit both enhancing and restraining impacts on the growth, invasion, and spread of various cancer cells including OSCC⁹. Thus, there are justifiable concerns regarding their potential to fuel the growth and spread of some occult recurrences/metastases if used in cancer patients. With the increasing cancer incidence predicted to occur over the next decades¹⁰, most likely a significant part of the patients receiving MSC therapy in the future would have been surviving one or another type of cancer.

The efficacy and advantages of using BMMSCs for treatment of jaw defects in non-cancer related setting has been recently shown in clinical trials¹¹. The use of MSC cell therapy for bone reconstruction of jaw defects after surgical treatment in OSCC patients is thus an attractive avenue for managing this complex disease. However, before moving to clinical trials, preclinical safety data is necessary for the safe use of MSCs in OSCC patients. The aim of this study was thus to assess the effect of BMMSCs on proliferation, migration, tumor growth and metastasis of OSCC cells.

Materials and methods

Cell culture

Previously isolated human BMMSCs from three different donors were used in this study¹². BMMSCs were cultured in Minimum Essential Medium (αMEM, Thermo Fisher Scientific, Waltham, USA) supplemented with 10% fetal bovine serum (FBS qualified, Gibco, Brazil) in tissue culture flasks (Sarstedt, Nümbrecht, Germany). The cells were sub-cultured and expanded in a humidified CO₂ chamber at 37 °C. All experiments were performed at passages less than 5. The use of these cells for research purposes from human donors was approved by the Regional Committee for Medical Research Ethics (REK), Norway (2013-1248/REK-sør-øst C and 2016-1266/REK-nord). All experiments were performed in accordance with the relevant guidelines and regulations and were approved by REK. Appropriate informed consent was obtained from all donors. The OSCC cell lines, Ca1 and OSCC1 are in-house derived cell lines previously characterized and used as representative OSCC in vitro and in vivo human OSCC cell models^13–15. Both Ca1 and OSCC1 were routinely maintained in FAD culture medium as previously described^13–15. Routine checks for mycoplasma contamination were performed using MycoAlert mycoplasma kit (LONZA, Rockland, USA) and all cells tested negative.

Characterization of BMMSC

The MSC phenotype of cells was analyzed using flow cytometry for cell surface markers, CD45-FITC, HLA DR- PE, CD90- PE Cy7 (all from eBioscience, Thermo Fisher Scientific), CD34-FITC, CD73- APC Cy7 and CD105- APC (all from BD Biosciences, San Jose, USA) using a well-established protocol in our laboratory¹². In addition, their multilineage differentiation capacity into osteogenic, adipogenic and chondrogenic lineages was evaluated based on Alizarin Red S, Oil O Red and Alcian Blue (all from Sigma-Aldrich, St- Louis, USA), respectively.

Conditioned medium (CM) and cell viability assay

BMMSCs were cultured either alone (cell density 10,000 cells/cm² in αMEM without FBS) or co-cultured with Ca1 or OSCC1 (1:1 ratio, 5000 cells/cm² of each cell type in αMEM + FAD without FBS) and allowed to adhere for 12 h. Subsequently, the cells were washed with phosphate-buffered saline (PBS, Thermo Fisher Scientific) and fresh medium without FBS was added. The cultures were then incubated for 24 h and the resulting medium was collected as conditioned medium (CM).

To study the effect of CM from monoculture or co-culture on cancer cell proliferation, Ca1 or OSSC1 were seeded at a density of 2000 cells/well in 96-well plate (Thermo Fisher Scientific) and allowed to attach. After 24 h, the culture medium was carefully removed, and the cells were washed with PBS and different CM were added. The experimental conditions tested were: (1) CM from monocultured BMMSC, (2) CM from cocultured BMMSC and Ca1, (3) CM from cocultured BMMSC and OSCC1, (4) FAD without serum serving as the control for cancer cells growing in monocultured medium, and (5) FAD + αMEM, serving as the control for cocultured medium. Cell viability was assessed using 0.01 mg/mL resazurin solution (Resazurin sodium salt, Sigma-Aldrich, St Louis USA) and the fluorescent readout was then measured at 560/590nm wavelength in a microplate reader (Varioskan™ LUX multimode microplate reader, Thermo Fisher Scientific).

Cell migration assay

To test the effect of CM on migration capability of cancer cells, a wound-healing assay was performed. 20,000 OSCC1 cells/well were seeded in 96 Imagelock microplates (Essen BioScience, Michigan, USA) and allowed to attach. After 24 h, a wound was created, and the cells were washed with PBS to remove floating cells and debris. Following this, CM either from monocultured BMMSCs or co-culture of BMMSCs + OSCC1 cells was added. Wells with culture medium without serum served as controls. The wounds were then assessed using IncuCyte Zoom timelapse imaging platform (Essen BioScience) for 48 h and analyzed using in-built software.

3D organotypic co-cultures and invasion assay

BMMSC cells were embedded in a type I collagen matrix (Corning, Bedford, USA) at a density of 0.5 × 10⁶ cells/mL, as previously described¹⁶ to form the stromal compartment. After 24 h, 0.5 × 10⁶ Ca1 or OSCC1 cells were then seeded on top of the collagen–BMMSC matrix to generate 3D organotypic co-cultures (3D OTs), as previously described¹⁶, or on top of collagen type I matrix without BMMSCs and allowed to grow for 48 h. The co-cultures were lifted to air–liquid interface on day 4 and maintained in total for 10 days in FAD serum-free medium, then fixed in 10% neutral-buffered formalin and paraffin-embedded for further histological evaluation.

Animal experiments

Experimental design

A total of seventeen four to six-week-old NOD SCID IL2R gamma chain deficient mice (NSG, The Jackson Laboratory, Bar Harbor, USA) weighing between 25 and 30 g were used for this study. Of these, 15 received one million Ca1 cells resuspended in 30 µL of Matrigel (Corning, Bedford, USA) in the tongue of mice. After the tumors had developed and reached measurable sizes, 1 × 10⁶ BMMSCs cells (in PBS, passages 3 or 4) were injected intravenously (iv) in the tail vein of 12 mice and 3 mice received PBS in the tail vein to serve as cancer controls. The mice (n = 12) were randomly divided into four groups based on study time-points: 2 h (n = 1), 2 days (n = 1), 14 days (n = 4) and 28 days (n = 5) post-injection. One mouse died after tail vein injection, likely a reason of technical failure. Two mice, without Ca1 xenografts were kept as controls to study biodistribution of BMMSCs and were euthanized via cervical dislocation at 2 days and 28 days post iv injection. The mice were subjected to bioluminescent imaging (BLI) once every week. Tongue tumors, lungs, liver, spleen, and kidneys were harvested for immunohistochemistry (IHC) and qPCR. All animal experiments were performed in accordance with the relevant guidelines and regulations. The study was approved by the local ethical board (FOTS ID: 202022627) at the Animal Research Facility, University of Bergen, Norway. All methods are reported in accordance with ARRIVE 2.0 guidelines.

Tumor growth measurements

Tumor volumes and weight were monitored every second day since tongue injections until mice reached the planned endpoint or euthanized due to ethical reasons. The effect of BMMSCs on tumor growth was studied by recording the tumor size before iv injection and every second day after iv injection of BMMSCs. Tumor volume was calculated as: volume, V (in mm3) = (L × W2)/2, where L is tumor length and W is tumor width.

Redifect transduction and BLI

The BMMSCs were transduced using RediFect™ Red-FLuc Lentiviral Particles (PerkinElmer, Waltham, USA), to express a red-shifted Luciola italica luciferase transgene according to the manufacturer’s instructions. The lentiviral particles contain a luciferase transgene which is fused to a green fluorescent protein (GFP) gene, thus the transduced cells co-express luciferase and GFP. A multiplicity of infection (MOI = 25) was used to obtain optimal signal. The survival and distribution of transduced BMMSCs were tracked before each euthanasia time point for all mice. Photon emissions from the luciferase expressing BMMSCs were measured using the IVIS Lumina (In Vivo Imaging System, Lumina XRMS, PerkinElmer) with Living Image software (version 4.3.1, PerkinElmer), which was used to estimate the level of photon emission per mice.

Quantitative reverse transcriptase polymerase chain reaction (qRT-PCR)

Tongue tumor, lung, liver, spleen and kidney tissues were stored in RNA-later and homogenized using RLT-buffer (RNA easy kit, Qiagen, Venlo, Netherlands). RNA was extracted following the manufacturer’s instructions by RNA easy kit (RNeasy Kit, Qiagen). RNA purity and quantity was measured by Nanodrop 2000 (Thermo Fisher Scientific). cDNA was synthesized from 200ng of isolated RNA using manufacturer’s protocol (high-capacity cDNA reverse transcription kit, Thermo Fisher Scientific). Further, qPCR was employed to trace luciferase (TaqMan Mr03987587_mr, Thermo Fisher Scientific) gene in all the mice tissues collected using a protocol mentioned elsewhere¹⁷. The expression levels of luciferase were normalized to human GAPDH (TaqMan Hs02786624_g1, Thermo Fisher Scientific), and the fold changes were reported using the 2 − ΔΔCt method.

Histology and IHC

For evaluation of OSCC cell growth and invasion in 3D models, 3–5 µm sections of the 3D OTs co-cultures were stained with hematoxylin and eosin (H&E) and scanned using a whole slide scanner (Hamamatsu NanoZoomer-XR whole slide scanner, Hamamatsu Photonics) at 40× magnification. Epithelial thickness and depth of OSCC invasion were quantified using NDP.view2 software. Each section was divided into five equal parts; the central and two outer fifths were excluded from measurement to minimize edge and central effects, where variations in oxygen and nutrient diffusion are highest. Epithelial thickness (ET) and depth of invasion (DOI) were assessed in the remaining two fifths at six measurement points per 3D organotypic co-culture. A reference line corresponding to the uppermost intact collagen surface (basement membrane zone) was drawn, and ET was measured as the vertical distance from this reference line to the epithelial surface while DOI was determined as the vertical distance from this reference line to the deepest invading epithelial cell.

The harvested mice tissues were fixed in formalin and embedded in paraffin for histological and immunohistochemical staining as per protocol well established in our laboratory¹⁷. Paraffin-embedded tissue sections were incubated twice in xylene for 5 min each and then in 100%, 96% and 70% ethanol, 5 min each. For antigen retrieval, the sections were microwaved in retrieval buffer pH 9.0 for 20 min. After cooling, peroxidase block (Agilent Dako, Golstrup, Denmark) was applied for 20 min. Following washes with tris-buffered saline with tween-20 (TBST), nonspecific binding was blocked with 10% goat serum (Agilent Dako) in 3% bovine serum albumin (Sigma, St Louis, USA). For single IHC staining, the sections were incubated overnight at 4°C with the primary antibody (anti-GFP (1:200), sc-9996 Santa Cruz biotechnology, USA). The reaction was then detected using the DAB chromogen, EnVision kit (Agilent Dako). For double staining, the sections were incubated at room temperature for 1 h with the first primary antibody (anti-GFP, 1:200) and following the above protocol, the reaction was visualized using a DAB chromogen, EnVision kit (Agilent Dako). Subsequently, an additional denaturation step was carried out to prevent antibody cross-reactivity. This involved heat-induced antigen retrieval treatment for 8 min using a pH 6.0 retrieval solution in the microwave. After cooling, the sections were incubated with the second primary antibody (anti-p53, 1:1000, clone DO-7, Agilent Dako) overnight at 4 °C. The second primary antibody was visualized using magenta chromogen (Agilent Dako). Slides were then counterstained using hematoxylin, washed, covered, and mounted with pertex (Histolab, Askim, Sweden). Stained sections were digitally scanned at ×40 magnification using the Hamamatsu NanoZoomer-XR whole slide scanner and analyzed using the open-source software QuPath (version 0.4.0; Queen’s University).

Statistical analysis

Two-way anova test was used when experiments involved multi-factor comparisons. In this study, it was used to determine significance between the controls and treated in the cell viability assays and migration assays. Students t-test was used strictly for two factor comparison to compare two independent means under the same assumptions and was used to determine significance between the number of GFP positive cells in the lungs of mice at different time points after tail vein injections. Students t-test was also used to verify significance between the number of OSCC cells metastasized to lungs in Ca1 only versus Ca1 + BMMSCs. All statistical analyses were performed using GraphPad Prism 9 and p < 0.05 was considered significant.

Results

BMMSCs displayed typical mesenchymal stem cell phenotype and BMMSC-CM did not increase proliferation or migration of oral cancer cells

BMMSCs exhibited typical mesenchymal-like morphology and showed plastic adherence in vitro (Fig. 1A). The immunophenotype of BMMSCs was characterized by flow cytometry and the cells exhibited a characteristic MSCs immunophenotype with more than 99% cells being positive for CD73, CD90 and CD105 while being negative for CD34, CD45 and HLA-DR (Fig. 1B). These cells also showed the capacity to differentiate into osteogenic, adipogenic and chondrogenic lineages that was confirmed by formation of mineralized matrix staining positive to Alizarin Red S after 21 days, intracellular lipid vesicles staining positive to Oil O Red after 14 days, and cartilaginous proteoglycan matrix staining positive to Alcian Blue after 28 days, respectively (Fig. 1C).

Fig. 1.

Characterization of BMMSCs. (A) Fibroblast-like morphology exhibited by BMMSCs at low and high confluency. Scale bar- 50 µm. (B) Graphs showing negative expression of the cell surface markers CD34, CD45 and HLA-DR and positive expression of the cell surface markers CD73, CD90 and CD105 in BMMSCs as analyzed by flow cytometry. (C) Representative images showing the multi-differentiation capacity into osteogenic, adipogenic, and chondrogenic lineages confirmed by formation of mineralized matrix staining positive to Alizarin Red S after 21 days, intracellular lipid vesicles staining positive to Oil Red O after 14 days, and cartilaginous proteoglycan matrix in a section of cultured cell pellet staining positive to Alcian Blue after 28 days, respectively. All experiments were performed at least twice with three different human donors.

To study the gross effect of growth factors secreted by BMMSCs on oral cancer cells, CM was collected from BMMSCs in monoculture or in co-culture with cancer cells (OSCC1 or Ca1, Fig. 2A). This CM was used to study proliferation of cancer cells. Although a trend for increased proliferation of OSCC1 and Ca1 cancer cells upon treatment with CM from monocultured BMMSCs was observed when compared to media control containing no serum, these differences were not statistically significant (Fig. 2B). In addition, CM from co-culture of cancer cells and BMMSCs did not increase the proliferation of cancer cells (Fig. 2C). This may indicate that the crosstalk between BMMSCs and cancer cells did not alter the secretory profile of either BMMSCs or cancer cells towards a stimulatory one able to induce in vitro proliferation of cancer cells. Moreover, the wound-healing assay revealed a slower migration rate of cancer cells when grown in CM from all BMMSC donors as compared to media control and this was statistically significant between treated and non-treated OSCC1 cells (p = 0.045) (Fig. 2D,E).

Fig. 2.

Growth kinetics of cancer cells when exposed to CM from BMMSCs. (A) Schematic diagram showing the in vitro experimental design. Effect of CM obtained from BMMSCs from three human donors (D1, D2 and D3) on proliferation of oral cancer cell lines OSCC1 and Ca1 in (B) monoculture setting (n = 3), and (C) co-culture setting (n = 3). Proliferation assay was repeated thrice with each replicate having five technical replicates. Two-way anova test was used to establish significance and values p < 0.05 were considered significant. (D) Representative images showing the effect of CM on migration of OSCC1 cells in a wound-healing assay at different time points. The assay was performed thrice with each replicate performed further in triplicates. Only representative images are shown here. (E) Graphical representation of % wound closure over time of OSCC1 in media control versus OSCC1 in CM from BMMSCs isolated from different donors. Two-way anova test was used at 24 h, 28 h and 32 h timepoint to determine significance. *p = 0.045.

BMMSCs did not increase growth or invasion of oral cancer cells in 3D organotypic models

When cultured on type I collagen matrices alone, both Ca1 and OSCC1 cells formed cohesive, relatively thick epithelial layers on top of the collagen matrix, with epithelial thickness of 51.8 ± 11.2 μm and 53.8 ± 8.27 μm, respectively. Invasion into the collagen was completely absent for Ca1 cells and only minimal for OSCC1 cells with DOI of 20.8 ± 9.41 μm (Fig. 3A,C). No significant changes in either ET or DOI were observed when the carcinoma cells were grown on matrices containing BMMSCs embedded within the collagen (Fig. 3E,F). The epithelial-stromal interface remained well-organized and continuous for both cell lines, without evidence of increased invasive protrusions or dissociation from the epithelial layer into the matrix populated with BMMSCs (Fig. 3B,D). Of note, increased differentiation of OSCC was observed when grown on matrix with BMMSC suggestive of relatively normal differentiation profile.

Fig. 3.

Growth and invasion of oral cancer cells in 3D organotypic cultures. Ca1 (A, B) and OSCC1 (C, D) cells were grown on top of type I collagen gels without (A, C) or with (B, D) BMMSCs. Representative H&E-stained sections show epithelial morphology and growth patterns after 10 days of culture. Scale bar 50 µm. Quantification of epithelial thickness (ET, E) and invasion (F) is illustrated in the accompanying bar graph (n = 4 independent experiments per cell line, with 2–3 technical replicates per condition). Students t-test was used to determine significance.

Biodistribution of systemic BMMSCs after iv injection in mice models

Biodistribution of BMMSCs was assessed by BLI at 2 h, and on days 2, 6, 14, 21 and 28 after iv injection (Fig. 4A). Bioluminescence signal was observed primarily in the lungs area 2 h after iv injection but subsequently decreased and was absent at day 6 and onwards, suggesting massive stem cell death and clearing of BMMSCs (Fig. 4B,C). These results were also confirmed by IHC, wherein GFP positive cells were stained and quantified. Corroborating with BLI results, the maximum number of GFP + cells were seen at day 2 with a decrease observed overtime (Fig. 4D–G). Although the observations at 2 h and day 2 are constrained by limited sample size, the findings support expected localization of BMMSCs to the lungs following iv injections. Some GFP+ cells were also detected on day 14, again supporting observations from BLI (Fig. 4H). Interestingly, no signal was observed in liver or spleen areas in any of the mice, and this was later confirmed by IHC (Figure S1). Owing to the sensitivity of PCR, quantitative real-time PCR was performed on lungs, liver, spleen, and kidney samples at all the time points to detect BMMSCs. The fluorescent signal above cutoff was observed only in lung tissues on day 2 and 14. No signal was observed from any other mice tissues at all time points (Figure S2A).

Fig. 4.

Biodistribution of BMMSCs in vivo. (A) Schematics of the experimental design of the animal experiment. (B) In vivo bioluminescent imaging showing signal observed in the lung area after iv injection at day 0 (n = 7) but no signal was observed at day 6. (C) The bioluminescent signal was quantified by measuring the total flux per unit area and represented as a graph. Immunohistochemical staining of BMMSCs at (D) 2 h (n = 1), (E) 2 days (n = 1), (F) 2 weeks (n = 4) and (G) 4 weeks (n = 5) after iv injection. (H) Graph showing percentage of GFP positive cells/mm² of lung tissue quantified at different timepoints. Statistical analysis using Students t-test was performed only on timepoints where n > 1, at 2 weeks (n = 4) and 4 weeks (n = 5).

Systemic BMMSCs did not affect tongue tumor size or lung metastasis in xenografted mice

To assess the tropism of human BMMSCs towards human oral cancer, a human OSCC cell line (Ca1) was xenografted in the tongue of the immunocompromised mice and human BMMSCs were injected intravenously. Throughout the observation period, no bioluminescence signal was observed in the tongue of the animals. To validate these findings, we performed qPCR and IHC of the tongue tissue of mice at all time points. No fluorescent signal was observed in qPCR and no GFP+ cells could be detected by IHC in the tumor or peritumoral region of mice tongue tumors (Fig. 5A,B and Figure S2B). Tongue tumor size was comparable between the mice xenografted with Ca1 only as compared to Ca1 + BMMSC iv injection (tumor size 77.87 ± 3.107 and 78.37 ± 1.678 mm³ respectively, p = 0.999) (Fig. 5C). No significant differences in mice body weight were found between BMMSCs iv injected mice and Ca1 controls over the study period (Figure S2C).

Fig. 5.

Effect of BMMSCs on tongue tumors. Double immunostained mice tongue tumors showing only p53 + cells (in magenta) in both (A) Ca1 xenografted only (n = 3) and (B) Ca1 xenografted + BMMSCs (n = 5) iv injection. Scale bar- 1mm. Inset showing zoomed image of boxed area. Scale bar- 50 µm. (C) Graphical analysis of tumor volume measured using vernier calipers over observation period. Students t-test was used to determine statistical significance.

Mesenchymal origin cells like fibroblasts and MSCs are known to prepare the ‘metastatic soil’. Since cells injected via the tail vein often home to the lungs (from tail vein to vena cava, then to the right heart, and finally to the lung capillary system), we assessed lung metastasis of Ca1 cells. This was done by performing double IHC for anti-p53 and anti-GFP, with Ca1 cells positive only for p53 whereas BMMSCs positive only for GFP. Lung metastasis was observed in both groups of mice with Ca1 xenografts with BMMSCs (n = 5/5, 100%) and without BMMSCs (n = 3/3, 100%) iv injection (Fig. 6A,B). When the number of p53+ Ca1 cells was quantified, a trend towards less p53+ Ca1 cancer cells in the lungs of Ca1 xenografted with BMMSCs iv injection mice was observed, although no significant statistical difference was observed between the two groups (p = 0.1429) (Fig. 6C).

Fig. 6.

Effect of BMMSCs on lung metastasis of oral cancer cells. Immunostained lung sections showing p53 + cells (in magenta) in both (A) Ca1 xenografted only (n = 3) and (B) Ca1 xenografted + BMMSCs (n = 5) iv injection. Scale bar- 5mm. Inset showing zoomed image of boxed area. Scale bar- 50 µm. (C) Graphical representation of % p53+ cells counted digitally in Ca1 xenografted only and Ca1 xenografted + BMMSCs iv injection. Students t-test was used to determine statistical significance. ns, not significant.

Discussion

BMMSCs have gained attention for their therapeutic potential in various clinical applications, including anti-inflammatory therapy, tissue regeneration, graft-versus-host response, and treatment of autoimmune diseases^1,3,9. Additionally, BMMSCs hold promise in tumor therapy due to their unique property of homing to tumor sites. As such, BMMSCs can be used as an excellent drug delivery vehicle. However, the role of BMMSCs in cancer progression is controversial, with studies showing both tumor promoting and inhibiting effects^18–21. These effects were either related to their homing to the tumor sites^22,23 and direct effect on cancer cells or the tumor microenvironment²⁴, or indirectly, due to their systemic immunomodulatory effects³. In this study we have investigated the direct effect of BMMSCs on cancer cell proliferation, migration and invasion in vitro and systemic effects of BMMSCs on cancer cell metastasis in vivo. The results show that the BMMSCs did not promote cancer cell proliferation in 2D in vitro models. The presence of BMMSCs in the stromal compartment did not enhance the growth or invasive behavior of OSCC cells in 3D organotypic co-cultures either. Moreover, conditioned medium from BMMSC decreased cancer cell migration significantly in vitro. In vivo, live cell imaging revealed that the biodistribution of BMMSCs was limited to the lungs of the animals with no BMMSCs found ‘homing’ to the tongue tumors. Although lung metastasis was observed in BMMSCs-injected and control mice, iv injection of BMMSCs resulted in with a trend towards less Ca1 cancer cells present in the lungs of BMMSCs injected mice. These findings indicate that circulatory BMMSCs do not stimulate oral cancer progression and metastasis.

Previous studies have shown that BMMSC can promote or inhibit oral cancer cell proliferation, migration and invasion in vitro^24–26. Our results on proliferation, but not on migration/invasion of OSCC cells corroborate with previous results. In contrast to Salo et al., who had showed that presence of BMMSCs repressed proliferation but increased the invasion of oral tongue cancer cells in a 3D myoma organotypic invasion model²⁴, our results from the 3D organotypic models show that BMMSCs did not induce any significant difference in the growth (ET) or invasion (DOI) of OSCC cells. These contrasts could be due to the differences in extracellular matrix (ECM) between our rat tail collagen-I based model and the 3D myoma model of Salo et al. The myoma model is known to provide a very complex extracellular human tumor microenvironment, whereas the rat tail collagen model is limited to predominant collagens and can markedly influence cell–matrix interactions, invasion patterns and stromal behavior¹⁵.

Our findings demonstrate that BMMSC-CM exerts an inhibitory effect on OSCC cell migration in vitro, while in vivo we observed a non-significant trend toward fewer lung metastatic foci following BMMSC administration under these experimental conditions. The in vitro antimigratory effect of BMMSC-CM suggests that BMMSCs secrete soluble factors capable of modulating tumor cell motility²⁴. Such paracrine interactions potentially involving chemokines, matrix-modifying enzymes, or anti-inflammatory mediators may reduce the migratory or invasive phenotype of OSCC cells under controlled culture conditions³. Although in vitro environments cannot fully reproduce the complexity of the tumor microenvironment, these findings provide a biologically plausible framework for understanding the directionality of the in vivo trend. In the in vivo model, BMMSCs administered systemically did not exacerbate lung metastatic burden and instead showed a mild reduction in metastatic foci that, while not statistically significant, aligns directionally with the in vitro antimigratory phenotype. This consistency between datasets suggests a potential biological relationship, however, warrants further mechanistic investigation. Mesenchymal cells such as fibroblasts and MSCs are known to prepare a pre-metastatic niche by interacting with other cell components and proving onco-promoting signals²⁷. Systematically administered BMMSCs trapped in the capillary bed of the lung may present an ideal ‘soil’ for seeding metastatic cells. However, our study shows that even though the BMMSCs were lodged in the lungs, they did not promote lung metastases of oral cancer xenografts. In fact, they did not influence tongue tumor size in the iv injected mice when compared to controls. In addition, the most relevant outcome from the safety perspective is that the systemically administered BMMSCs do not worsen the progression of oral cancer in a xenograft mice model. These observations are without precedent to the best of our knowledge. Nonetheless, our study did not assess the viability, functional state, or secretory profile of BMMSCs after lodging in the lungs, nor did it examine whether direct interactions between BMMSCs and circulating tumor cells occur. Therefore, the current findings cannot delineate a causal chain from in vitro inhibition of migration to in vivo modulation of metastatic behavior and further studies are required to deep dive into the mechanistic interactions.

A majority of studies looking into biodistribution of BMMSCs after iv injection report initial retention of BMMSCs in the lungs and later redistribution of cells to liver, spleen and kidney^28,29. In our study, we observed bioluminescent signal originating from lungs of injected mice upto 6 days post iv injection. However, we did not see the redistribution of cells either to liver, spleen or kidney, suggesting massive cell death and clearance in the recipient animals. Moreover, in our study we did not see any BLI signal from the tongue tumor area or detected BMMSCs via IHC or qPCR in the tongue tumors. The typical diameter of human BMMSCs is 15–25 µm whereas in comparison, lymphocytes are 5–12 µm³⁰. The relatively large size of BMMSCs could explain entrapment in the lung capillaries of mice and may impede their migration to other organs or tongue tumors. Our results are in line with study of Eggenhofer et al.²⁹, in which they suggest that MSCs are short-lived cells and do not migrate beyond the lungs. Further, the authors indicate that live MSCs do not pass the lungs after iv injection and suggests that other studies describing MSC in tissues like liver or spleen could be MSC debris or phagocytosed MSC and not live MSC^29,30. Our study points to the fact that other routes of BMMSC administration must be investigated if BMMSCs need to be delivered to other tissues than the lungs.

Although the limitation of our work is that the results were obtained using xenogeneic BMMSCs on oral cancer cells in a xenograft mice model, the results represent a new biosafety insight for the use of MSC-based cell therapy for reconstruction of jaw defects after surgical treatment in OSCC patients. The primary MSCs used in our study as is inherent to all primary MSC cultures, have limited expansion potential and could not be passaged extensively without altering their phenotype, thus limiting the number of replicates. This limitation reduced the number of replicates feasible in some experiments and should be considered when interpreting these observations. Our manuscript focuses on observing the biosafety of BMMSC implantation and detailed mechanistic pathway studies, although warranted, are beyond the scope of the current study.

Conclusion

Our findings indicate that BMMSCs do not enhance proliferation and inhibit the migration of oral cancer cells in vitro. Furthermore, BMMSCs did not exacerbate the progression of oral cancer in vivo.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Conceptualization, DS, HP, KM, DEC and HND; data curation, DS, HP, IEH, DK, LL, SMA and HND.; formal analysis, DS, HP and HND; funding acquisition, HR, KM, DEC and HND; investigation, DS, HP, IEH, DK, LL, SMA, DEC and HND.; methodology, DS, HP, DEC and HND.; project administration, HR, KM, DEC and HND.; supervision, HP, DEC and HND.; validation, DS, HP and HND; visualization, DS, HP and DEC; writing—original draft, DS and HND; writing—review & editing, DS, HP, HR, SMA, KM, DEC and HND. All authors have read and agreed to the published version of the manuscript.

Funding

Open access funding provided by University of Bergen. The authors declare that financial support was received from the Research Council of Norway through its Centers of Excellence funding scheme, (DEC and HND, Grant No. 22325 to Center of Excellency for Cancer Biomarkers CCBIO) and The Western Norway Regional Health Authority (DEC, Helse Vest Grant Nos 912260/2019 and F-13105/2024). The authors also acknowledge the Trond Mohn Foundation (KBEM, OsteoStem Grant No. TMS2021TMT08) for financial support.

Data availability

All data generated or analysed during this study are included in this published article [and its supplementary information files].

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

Appropriate informed consent was obtained from all donors and the study was approved by the Regional Committee for Medical Research Ethics (REK), Norway (2013–1248/REK-sør-øst C and 2016-1266/REK-nord).