Background

Osteosarcoma (OS) is a malignant aggressive primary bone tumor that commonly arises in the long bones of children and young adults. Conventional clinical treatment consists of surgical resection of the tumor and adjuvant chemotherapy [1]. In spite of the effort made by clinicians in the last 30 years, the success of OS treatment is limited to a 70% 5-years survival rate with the remaining 30% of OS patients not responding to standard treatments [2] mainly because of the formation of lung metastases, which ultimately represent the primary cause of mortality [3]. Despite testing different drugs and regimens, a substantial improvement of survival rate has not been observed [4,5,6]. In addition to lung metastases, OS cells resistance to chemotherapeutics, such as doxorubicin [7], greatly hampers treatments’ efficacy, therefore the development of innovative and more selective strategies capable of improving the survival of OS patients is required.

Mesenchymal stromal cells (MSCs) have been proved to be powerful tools in cell therapy, being used for a wide array of clinical indications spanning from the treatment of graft-versus-host-disease to tissue engineering, and currently are tested in several hundreds of clinical trials [8]. Additionally, thanks to their proven ability to migrate and engraft in the stroma of several tumors [9], MSCs have been used in preclinical and clinical studies as carriers of antitumor drugs with the aim of enhancing their selective accumulation at the tumor site. In 2002, Studeny et al. firstly proposed MSCs as carrier cells for gene therapy [10]; at present, several studies have been published reporting MSCs as effective delivery vehicles of anticancer agents, such as pro-apoptotic molecules, chemotherapeutic drugs and oncolytic viruses [11,12,13,14,15,16,17,18,19,20]. In addition, it has been extensively demonstrated that MSCs can internalize and deliver nanoparticles loaded with therapeutic agents [21,22,23,24,25], including chemotherapeutic drugs and photosensitizers (PS) for photodynamic therapy (PDT) applications [26,27,28].

In PDT, light at a specific wavelength is applied to the tumor site where the PS localizes after administration; once irradiated, the PS enters an excited state that triggers the formation of various reactive oxygen species (ROS) responsible for killing cancer cells and the damage of tumor vasculature that in turn deprives the tumor of oxygen and nutrients [29]. In particular, the use of nanoparticles as PS delivery systems has been proposed as treatment for several tumors with the aim of bypassing biological barriers and cellular chemo-resistance [30]. PDT has proved to be a successful, clinically approved, and minimally invasive alternative/co-adjuvant therapeutic option to conventional therapies for the treatment of various tumors [31, 32]. In particular, PDT has been proved to be effective in decreasing tumor growth both in in vitro and in vivo OS models [33,34,35,36,37,38,39,40], and in a murine model of aggressive prostate tumor [41] as well as in clinical settings [42].

In order to establish whether the PDT driven MSCs strategy is an effective system for the treatment of OS, we designed a multistep process that could allow us to determine the best operating settings, such as: doses, radiant exposure and distance of light source, time after infusion of loaded MSCs, safety of the operating procedure etc., with in mind the possibility of clinical translation. With this in mind, in an earlier study we demonstrated that MSCs can be efficiently and safely loaded with fluorescently labelled poly-methyl methacrylate nanoparticles (FNPs) electrostatically decorated with the photosensitizer tetra-phenyl sulfonated porphyrin (TPPS), and that this system (TTPS@FNP@MSCs) exerts a ROS-mediated cytotoxic effect on surrounding OS cells upon irradiation with a 405 nm light in-vitro [26]. Based on these encouraging results, we improved our NPs system by loading a different PS, i.e. tetra-sulfonated aluminum phthalocyanine (AlPcS4), since it is well-established that the optimal light therapeutic window, ensuring the highest tissues penetration, falls in the near-infrared region [43]. In fact, unlike TPPS, AlPcS4 has a strong absorption peak in the near-infrared region of the spectrum; this upgraded system, i.e. AlPcS4@FNPs, was able to effectively kill human prostate cancer cells in in vitro 3D model as well as in in vivo mouse model [41].

Therefore, the objective of the current study is to prove whether our AlPcS4@FNPs particles are an effective PDT system against OS cells; more importantly, we aim to investigate in an in vivo OS ectopic model whether AlPcS4@FNPs loaded into MSCs have an improved tumor selectivity compared to AlPcS4@FNPs alone, while maintaining their cancer cells killing efficacy.

Material and methods

Reagents

Dulbecco’s Modified Eagle’s Medium-high glucose (DMEM-HG, glucose 4500 mg/L), McCoy’s medium, and Ficoll®-Paque PREMIUM 1.073 reagents were purchased by Sigma Aldrich (Saint Luis, Missouri, USA). α-Minimum Essential Medium Eagle (α-MEM) was purchased by Lonza (Verviers, Belgium). Fetal Bovine Serum (FBS), GlutaMAX, penicillin/streptomycin solution, Dulbecco’s phosphate buffered solution without calcium and magnesium (D-PBS), Puromycin, Alexa Fluor® 488 Annexin V/PI Dead Cell Apoptosis Kit, LIVE/DEAD® Viability/Cytotoxicity Kit (Calcein-AM and Ethidium homodimer-1), Alamar Blue, and WST-1 assay reagents were purchased from Thermo Fisher Scientific (Waltham, Massachusetts, USA). CellTiter-GLO® was purchased from Promega (Milano, Italy). Tetra-sulfonated aluminum phthalocyanine (AlPcS4) was purchased from LivChem Logistics GmbH (Frankfurt, Germany).

Preparation of AlPcS4@NPs/FNPs

Poly-methyl methacrylate (PMMA) core-shell fluorescent nanoparticles (FNPs) were obtained by an emulsion-co-polymerization reaction as previously described [26, 41]. Briefly, 2-(dimethyloctyl)-ammonium ethyl-methacrylate bromide (0.52 g, 1.5 mmol) in water (50 mL) was placed into a 250 mL three-necked reactor equipped with a mechanical stirrer, a condenser, a thermometer and a nitrogen inlet. The mixture was heated to 80 °C under stirring (300 rpm), and 2-aminoethyl methacrylate hydrochloride (AEMA, 0.25 g, 1.48 mmol) was added to the solution. Afterward, a mixture of allyl 2-(3-allyloxy-6-oxo-6H-xanthen-9-yl) benzoate [44] (0.003 g, 0.007 mmol) and methyl methacrylate (0.93 mL, 9.35 mmol) was added to the previously obtained solution. After 10 min, 15 mg (0.05 mmol) of 2,2′-azobis (2-methylpropionamidine) dihydrochloride (AIBA), dissolved in 0.5 mL of mQ water was added to the mixture, which was then allowed to react for 4 h. The reaction product was purified by dialysis (against water) to remove residual monomer and stabilizer. Whenever needed, and to avoid interference with fluorescent staining, non-fluorescent PMMA nanoparticles (NPs) were prepared by the same procedure without the addition of the fluorescent comonomer, i.e. allyl 2-(3-allyloxy-6-oxo-6H-xanthen-9-yl) benzoate AlPcS4@FNPs or AlPcS4@NPs stock solution was prepared by adding 50 μl of AlPcS4 (1 mg/mL in milliQ water) to 50 μl of FNPs or NPs (10 mg/mL) and milliQ water to a final 1 mL volume. The stock solution was diluted in complete cell culture medium to the desired concentration. Where not explicitly stated, the indicated concentrations refer to the amount of FNPs/NPs per volume unit, resulting in an equivalent AlPcS4 concentration of 1/10 (e.g. 90 μg/ml AlPcS4@FNPs are equivalent to 9 μg/mL AlPcS4).

Human osteosarcoma cell lines

Human osteosarcoma cell lines, MG-63 (CRL-1427), Saos-2 (HTB-85) and U-2 OS (HTB-96) were purchased from ATCC (Manassas, Virginia, USA). Briefly, cells were cultured respectively in DMEM-HG (MG-63) or McCoy’s medium (Saos-2, U-2 OS) containing 10% of FBS, 1% GlutaMAX and 50 U/ml penicillin/streptomycin at 37 °C in a humidified atmosphere with 5% CO2.

Saos-2-Luc/GFP cell line was generated by transduction with Lentivirus particles containing the CMV promoter for the expression of humanized firefly luciferase (hLUC) and SV40 promoter for the expression of GFP protein according to manufacturer’s protocol (GeneCopoeia). Three days after infection, cells with high levels of GFP expression were selected by Cell Sorter NIR Aria II (BD Bioscience) and expanded for a week in culture medium supplied with Puromycin to generate a stable cell line.

Isolation and culture of human mesenchymal stromal cells (MSCs)

MSCs were obtained from bone marrow samples of five patients undergoing surgery at Rizzoli Orthopedic Institute (Bologna, Italy). Isolation and culture expansion of human MSCs was performed as previously described in Pierini et al. [45] with minor modifications. Briefly, mononucleated cells were isolated from bone marrow through gradient separation with Ficoll®-Paque PREMIUM 1.073, then placed in 150 cm2 culture flasks in complete growth medium (αMEM+ 20% FBS) at a density of 4 × 105 cells/cm2 and incubated at 37 °C in 5% CO2 atmosphere. Medium was changed every 3–4 days; after the first passage sub-culturing was performed at 2 × 103 cell/cm2 every time the cells reached a 70–80% confluence. Complete characterization in terms of fibroblast-colony forming unit (CFU-F) efficiency, immunophenotypic profile, proliferation rate, and trilineage-differentiation potential of each MSC line was performed. Since ex-vivo expanded MSCs are a heterogeneous population and are known to be highly sensitive to the protocols used to isolate and expand the cells in culture [46, 47], detailed protocols are provided as Supplementary Methods (Additional file 1), and all MSCs’ characterization and in-process data are provided for review in Table 1S (Additional file 2), as suggested by Reger and Prockop [48]. In order to account for unpredictable variations on test results due to the well-known donor-to-donor variability of MSCs [49], at least 3 different MSCs lines were tested in independent experiments. When stated, all the 5 MSCs lines were tested, in order to strengthen the reproducibility of the results. Only cells from the third to the sixth passage were used in all experiments.

Cytotoxicity assay

MSCs were seeded in 96 well plates and incubated for 1 h with AlPcS4@FNPs, FNPs or AlPcS4 at increasing doses (45, 90, 180 μg/mL for AlPcS4@FNPs or 4.5, 9, 18 μg/mL for AlPcS4). Cells were then washed twice with D-PBS and new complete medium was added to each well. WST-1 assay was performed 1, 2 and 6 days after loading following the manufacturer’s instructions. The optical density of each well was measured by a microplate reader (Synergy HT, BioTek Winooski, VT, USA) set at 450 nm with the correction wavelength set at 690 nm.

MSCs loading with nanoparticles

MSCs were seeded at 104 cells/cm2 in complete medium and allowed to adhere to the plates overnight before loading. MSCs were exposed to AlPcS4@FNPs or AlPcS4@NPs diluted in complete medium for 1 h, then washed two times with D-PBS. AlPcS4@FNPs or AlPcS4@NPs loaded MSCs were allowed to recover for a time varying from 2 h to overnight (o/n) in complete medium, before being detached from culture flasks, for subsequent experiments. Before each experiment, AlPcS4@FNPs loaded MSCs were checked through automated cell counter Countess II® FL (Thermo Scientific, Waltham, Massachusetts, USA) in order to verify the loading efficiency.

Cellular uptake and accumulation

The percentage of loaded MSCs were determined by BD FACScanto II cytometer (Becton-Dickinson, Franklin Lakes, NJ, USA) and by the automated cell counter Countess II® FL (Thermo Scientific, Waltham, Massachusetts, USA), taking advantage of FITC fluorescence of FNPs. For confocal microscopy (Nikon, Amsterdam, Netherlands) analysis, MSCs were seeded onto glass coverslips, loaded with FNPs and at the indicated time points fixed for 10 min in 10% Neutral Buffered Formalin at room temperature, thoroughly washed with D-PBS, stained with Hoechst and imaged.

In vitro migration study

Cell migration was assessed by the Boyden chamber technique. Cell culture inserts for 24-well plate with 8 μm pore diameter were used (Millipore, Darmstadt, Germany). After a recovery period of 2 h in complete medium, AlPcS4@FNPs loaded MSCs, along with unloaded MSCs used as control, were exposed to o/n starvation, switching the complete medium to αMEM+ 0.2%FBS. Then 104 cells were placed in the upper chamber in 200 μL of αMEM+ 0.2%BSA. Six hundred microliters of αMEM supplemented with 20% FBS (chemo-attractant) or 0.2% BSA (neutral) was added into the lower chamber. After overnight incubation, cells on the upper face of the membrane were removed with a cotton swab while the ones on the bottom surface were fixed in 100% methanol and stained with Hema-stain kit (Fischer Scientific, Hampton, New Hampshire, USA). Cells migrated through the microporous membrane were counted in 10 randomly chosen fields under an inverted Nikon Eclipse TE2000-U microscope (Nikon, Amsterdam, The Netherlands).

2D co-culture

MSCs were loaded with 90 μg/ml AlPcS4@FNPs and left o/n to recover in complete medium. AlPcS4@FNPs loaded MSCs were then trypsinized, counted with Countess II® FL and 5 × 103 cells were seeded into 24-well plate mixed with 5 × 103 or 15 × 103 Saos-2 cells, i.e 1:1 or 1:3 ratio respectively. Photoirradiation was delivered after overnight cell adhesion.

3D co-culture

MSCs were loaded with 90 μg/ml AlPcS4@FNPs and left for a recovery period of 4 h in complete medium. AlPcS4@FNPs loaded MSCs were then trypsinized, counted and mixed with MG-63 in different ratios (1:1, 1:3 and 1:7) to a final concentration of 105 mixed cells/mL in DMEM-HG + 10%FBS. One hundred microliters aliquots of the suspension were dispensed in an ultra-low attachment U-bottom 96-well plate (Corning Costar, Amsterdam, The Nederlands) and allowed to aggregate for 4 days to form regularly shaped spheroids.

Photodynamic therapy parameters

In in vitro experiments, AlPcS4@NPs loaded MSCs were photoactivated using a LED light source (λmax = 668 ± 3 nm) at room temperature, with the light-emitting unit placed directly under the tissue culture plates (radiant power: 140 mW). Monolayer cultures (2D) received photoactivation for 5 min, while spheroids (3D) for 10 min. Viability assays were performed, in all experiments, 24 h after PDT treatment.

In in vivo model, the tumor bearing area was irradiated for 20 min using the same LED source but with the addition of a focusing device (i.e. a cylinder of 0.6 cm diameter and 2 cm length, with a light-reflecting internal surface). The end of the focusing device was placed in close proximity to the mouse skin (Radiant power: 130 mW). Treatment was repeated twice, once a week.

Cell viability assays

In 2D co-culture, cell death was evaluated by Alexa Fluor® 488 Annexin V/Propidium Iodide Dead Cell Apoptosis Kit according to the manufacturer’s protocol and analyzed with BD FACScanto II cytometer (Becton-Dickinson, Franklin Lakes, NJ, USA). Cell survival rate was determined by Alamar blue assay following the manufacturer’s instructions. The fluorescence of each well was measured by a microplate reader (Synergy HT, BioTek Winooski, VT, USA) with excitation/emission wavelengths of 530/590 nm. The fluorescence intensity from the samples was corrected using a cell-free control as blank.

For 3D co-culture system, cell death was evaluated through the ATP content–based assay CellTiter-Glo® 3D following the manufacturer’s protocol. Additionally, a LIVE/DEAD® staining was performed. Spheroids were incubated with 2.5 μM Calcein-AM in DMEM Phenol Red-free for 2 h, then Ethidium homodimer-1 (EthD-1) was added to a 5 μM final concentration for 10 min. Z-stacks images, for a total depth of 100-120 μm, were acquired with an A1R confocal laser scanner (Nikon, Amsterdam, The Netherlands) using Nikon Plan Apo VC 20x/0.75 NA DIC N2 objective lens and 3D rendering was performed with NIS elements software using the Alpha-blending algorithm.

Transmission electron microscopy (TEM)

Spheroids were fixed with 2.5% glutaraldehyde in 0.1 M cacodylate pH 7.6 buffer for 1 h at room temperature. After post-fixation with 1% OsO4 in cacodylate buffer for 1 h, cells were dehydrated in an ethanol series and embedded in Epon resin. Semithin sections of 0.8 μm were cut using an ultramicrotome and stained with toluidine blue. Ultrathin sections (70 nm) were contrasted with uranyl acetate and lead citrate and observed with a Jeol Jem-1011 transmission electron microscope (Jeol Jem, USA).

Animal study

Eighteen female Athymic-nude mice, aged 6–8 weeks, were subcutaneously injected into the left flank with a mixture of Saos-2/Luc cells (1 × 106) and MSCs (1 × 106) in 50 μL of PBS/Matrigel. When tumors reached 100–150 mm3, approximately 2 weeks post-injection, the mice were divided into four groups: two control groups (group I and II respectively PBS and AlPcS4), group III AlPcS4@FNPs alone and group IV AlPcS4@FNPs loaded into MSCs. Fifty microliters of PBS, AlPcS4 (9 μg/mL), AlPcS4@FNPs (90 μg/mL) and AlPcS4@FNPs loaded-MSCs (1 × 106) were intra-tumorally injected. The next day, the mice were exposed for 20 min to PDT. Intra-tumor injection and PDT treatment were performed weekly for 2 weeks. All animals were euthanized 1 week after the last treatment. After intra-tumor administration of test substances, the whole animal fluorescent imaging (excitation/emission wavelengths: 640/680 nm) was performed using the IVIS Lumina II (PerkinElmer, Waltham, MA) to observe AlPcS4@FNP nanoparticles biodistribution. The same instrumentation was used to monitor tumor growth through bioluminescence imaging (BLI). D-luciferin (GolBio, St Louis, MO) dissolved in PBS (1.5 mg luciferin/100 μL PBS) was injected intraperitoneally at a dose of 150 mg D-luciferin/kg. The BLI imaging was performed prior NPs/NPs loaded MSCs injections and after PDT treatment. Regions of interest (ROIs) were drawn within the tumor to measure average radiance (expressed as photons/s/cm2/sr) using Living Image® 4.2 software (Caliper Life Sciences, Hopkinton, MA).

Histology

Tumors were collected, fixed in 4% paraformaldehyde solution, and embedded in paraffin. Samples were sectioned at a thickness of 4 μm and hematoxylin and eosin (H&E) staining was performed for a general inspection of the pathologic specimens. To evaluate the extent of tumor apoptosis and validate the BLI results, a terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay was performed with a commercial kit (Roche, Mannheim, Germany). Ki-67 staining for cell proliferation was also performed. Images of tumor tissue were taken by a NIKON Upright BF&Fluorescent light microscope.

Statistical analysis

All results were obtained from at least three independent experiments and expressed as the mean ± SD. Two-way ANOVA followed by Dunnett’s multiple comparisons test was used to determine statistical probabilities in the in vivo results. Results were considered to be statistically significant at P-values < 0.05. The statistical analysis was processed with GraphPad Prism 6 Software (GraphPad; San Diego, CA, USA).

Results

AlPcS4@FNPs internalization does not affect MSCs viability and migratory capacity

Fluorescent core-shell PMMA nanoparticles, namely FNPs, were characterized in terms of size, zeta potential, and morphology. In particular, FNPs were obtained with an average hydrodynamic diameter of 75 ± 0.92 nm (five measurements, PDI = 0.16 ± 0.01; Table 2S, Additional file 3), and a zeta-potential of 54 ± 2 mV (five measurements; Table 3S, Additional file 3). The number of ammonium group available for AlPcS4 loading was determined by potentiometric titration of the bromide ions obtained after complete ion exchange and was found to be 571 μmol per gram of nanospheres. The morphological analysis, performed through scanning electron microscopy (SEM), confirmed the spherical shape of FNPs (Figure 1S, Additional file 3).

In order to determine the AlPcS4@FNPs concentration to guarantee a 90% uptake into MSCs without altering their viability, we optimized the NPs loading parameters. Several MSC lines were incubated for 1 h with different concentrations of AlPcS4@FNPs (45, 90 and 180 μg/mL) and 24 h after the loading, the FITC fluorescence intensity was quantified by flow cytometry. Unloaded cells were used as control. As shown in the representative histograms of Fig. 1a, 98 to 100% of the MSCs internalized the NPs at all tested concentrations. In addition, for all concentrations, 1 h loading was sufficient for AlPcS4@FNPs internalization, therefore this incubation time was used in all the experiments.

Fig. 1
figure 1

MSCs internalize AlPcS4@FNPs nanoparticles without cytotoxic effect. Representative flow cytometry analysis of FNPs uptake at increasing doses (45, 90, 180 μg/mL) 24 h after 1 h-loading in MSCs (a) and cell cytotoxicity analysis (WST-1 assay) of MSCs exposed for 1 h to increasing concentrations of FNPs, AlPcS4, or AlPcS4@FNPs, at the end of incubation (day 0) and after 1, 2 and 6 days (b). All data are expressed as mean ± SD (n = 3)

Full size image

To evaluate the potential cytotoxicity of AlPcS4@FNPs on MSCs without light irradiation, i.e. dark toxicity, cells were incubated with 45, 90 and 180 μg/mL of AlPcS4@FNPs for 1 h, as well as with each of the NP components alone (FNPs and AlPcS4). WST-1 assay was performed at 1, 2 and 6 days after loading. As shown in Fig. 1b, the viability of MSCs was not affected by the internalization of AlPcS4@FNPs or the single components. In fact, WST-1 values in MSCs control cells were comparable to those of MSCs exposed to AlPcS4@FNPs, FNPs or AlPcS4 at different concentrations and at all tested times (Fig. 1b).

The loading efficiency was further investigated using an automated cell counter to measure the percentage of AlPcS4@FNPs positive cell and the average fluorescence intensity of 5 different MSCs lines, incubated with 45, 90 and 180 μg/mL of AlPcS4@FNPs. Results shown in Table 1 demonstrate that, regardless of the MSC lines used, the 90 μg/mL concentration ensures the highest and most uniform internalization rate, therefore this concentration was selected for all subsequent experiments.

Table 1 Nanoparticles uptake tested in MSCs lines isolated from five patients
Full size table

The retention of AlPcS4@FNPs (90 μg/mL) in MSCs over time (1, 2 and 3 days) was determined by flow cytometry (Fig. 2a) and by fluorescence microscopy (Fig. 2b). Both assays demonstrated that the fluorescence intensity remains constant for 3 days. In particular, the internalization was close to 100% at all tested time points, and the FNPs localization was intracellular (Fig. 2b).

Fig. 2
figure 2

AlPcS4@FNPs internalization and retention analysis. Representative flow cytometry analysis of AlPcS4@FNPs (90 μg/mL) loaded MSCs over time (0, 24, 48, 72 h) (a). Representative images of FNPs internalization in MSCs after 1 h loading (0 h) over the time (up to 72 h) by confocal microscope (merge images of green (FITC of FNP) and blue (Hoechst, nuclei) channels are shown) (scale bar = 200 μm) (b). Representative images and quantification of cells migrated through the porous membrane of a Boyden chamber, in absence (0.2% BSA) or presence (20%FBS) of chemotactic stimuli; MSCs loaded with 90 μg/mL AlPcS4@FNPs were compared to unloaded MSCs (c). All data are expressed as mean ± SD (n = 3)

Full size image

Since one of the major aims of this study was to exploit MSCs as delivery vehicles of AlPcS4@FNPs, the migration of NPs loaded MSCs was investigated using the Boyden chamber assay. Results in Fig. 2c showed that nanoparticles internalization does not affect MSCs migration profile in both neutral and chemoattractant conditions (0.2% BSA or 20% FBS respectively), suggesting that AlPcS4@FNPs do not modify the migratory potential of MSCs.

AlPcS4@NPs loaded MSCs induce OS cell death upon photostimulation

PDT efficacy in 2D co-culture system

To determine the effect of photoirradiation on OS cells, MSCs loaded with 90 μg/mL AlPcS4@NPs were co-cultured with Saos-2 cell line (ratios 1:1 and 1:3) and PDT applied for 5 min. Following irradiation, cells were kept in the dark and cultured under standard conditions for 24 h. Cell death and survival rate were evaluated by Annexin/PI labelling and Alamar Blue assay respectively; as shown in Fig. 3, nearly 80% of cell death was observed in the co-culture with Saos-2 (Fig. 3a), which was confirmed also by Alamar Blue assay that showed approximately a 20% survival rate for both the 1:1 and the 1:3 ratios (Fig. 3b). In order to distinguish the death ratio between MSCs and Saos-2 cells, AlPcS4@NPs loaded MSCs were co-cultured with GFP-tagged Saos-2 cells at a ratio of 1:1 and 1:3; Annexin V/PI Dead Cell Apoptosis Kit was performed. The percentage of surviving cells for each cell type was measured by flow cytometry 24 h upon photoirradiation. In particular, at the 1:1 ratio, 9% of MSCs and 12% of Saos-2 cells survived, while at the 1:3 ratio, 4% of MSCs and 28% of Saos-2 cells survived (Fig. 3c). These results demonstrate that photoirradiation is effective in killing OS cells, although the percentage of surviving OS cells is greater when the number of OS cells is increased.

Fig. 3
figure 3

Cell death evaluation after PDT of AlPcS4@NPs loaded MSCs in co-culture with Saos-2 cells. Graph representing quantification of total cell death (a) and survival rate (b), 24 h after PDT by Annexin V/PI and by Alamar Blue assay respectively, 5 × 103 MSCs loaded with 90 μg/ml AlPcS4@NPs were seeded into 24-well plate alone (grey bar) and in co-culture with 5 × 103 or 15 × 103 with Saos-2 cells (AlPcS4@NPs@MSC:Saos-2; black bars) at different ratios (1:1 and 1:3 respectively). Graph representing quantification by flow cytometry of the percentage of live or dead cells for Saos-2 (dark grey bars) and AlPcS4@NPs loaded MSCs (light gray bars) 24 h upon photoirradiation, 5 × 103 MSCs loaded with 90 μg/ml AlPcS4@NPs were seeded into 24-well plate in co-culture with Saos-2 cells at 1:1 and 1:3 ratios (5 × 103 or 15 × 103 cells respectively) (c). All data are expressed as mean ± SD (n = 3)

Full size image

PDT efficacy in 3D co-culture system

To further confirm the efficacy of the photoactivation of MSCs loaded with AlPcS4@NPs against OS cells, we developed a 3D spheroid model. Preliminary experiments were conducted using three different human OS cell lines (Saos-2, U-2 OS, MG-63), to identify the most reliable cell lines for the 3D model. MG-63 cells showed the ability to rapidly form more compact and homogeneous (in size and shape) spheroids, as respect to Saos-2 and U-2 OS (Figure 2S and Table 4S in Additional file 4) and were therefore selected for the task. Furthermore, we verified that FNPs loaded MSCs can be suitably combined with MG-63 to form multicellular spheroids, and that the FNPs were retained inside the spheroids for several days, as shown in the representative images of Figure 3Sa (Additional file 5). Indeed, from image analysis, no reduction of the fluorescence intensity due to FNPs, was observed inside the spheroids during 5 days of culture (Figure 3Sb, Additional file 5).

To determine whether the efficacy of the photoactivation decreases along with the increase of cancer cells proportion, MSCs loaded with 90 μg/mL AlPcS4@NPs were co-cultured with MG-63 cells at different ratios, i.e. 1:1, 1:3 and 1:7 MSCs:MG-63. After 4 days from generation, spheroids were irradiated for 10 min and viability was tested after 24 h as summarized in Fig. 4a. ATP measurement showed a dramatic decrease of the cell viability at the 1:1 MSCs:MG-63 ratio (survival rate lower than 5%) in this 3D settings; additionally, all the 5 MSCs lines tested exhibited similar results as detailed in Figure 4S (Additional file 6). Moreover, when the number of MSCs in the spheroid was decreased respect to OS cells (ratios 1:3 and 1:7), a higher percentage of cell survival after 10 min of photoirradiation was observed (Fig. 4b).

Fig. 4
figure 4

Cell death evaluation after PDT in a 3D co-culture system. Schematic summary of 3D in-vitro testing (a). Quantification of survival rates observed in multicellular spheroids composed by different ratios of AlPcS4@NPs loaded MSCs and MG-63 after 10 min irradiation. Data are expressed as mean ± SD (ratio 1:1 n = 5, ratio 1:3 n = 4, ratio 1:7 n = 3) (b). Representative confocal images (scale bar = 100 μm) of Live&Dead staining (green Calcein AM staining of live cells and red EthD-1 staining of dead cells’ nuclei) (c) and representative TEM images (scale bar = 5 μm) (d) of control (−PDT) and irradiated (+PDT) spheroids at 1:1, 1:3 and 1:7 ratios

Full size image

Confocal images of Calcein AM/Ethidium homodimer staining showed a significant cellular death due to PDT (red cells) in the 1:1 and 1:3 ratios, while the effect was less evident in the 1:7 ratio where only few damaged cells were visible. In control spheroids (no PDT) any significant alteration of cellular viability was observed, as confirmed by the small number of dead cells found (Fig. 4c).

In order to further investigate the effect of PDT on cells, TEM analysis was performed; in particular, a high degree of cell necrosis in the whole spheroid mass of the 1:1 ratio sample was found (Fig. 4d). Similarly, the samples produced with a lower MSCs proportion showed considerable damage of the cellular structures and, even though some cellular compartments were still recognizable (Fig. 4d, black arrow), the morphology of the cytoplasm and its related organelles appeared altered. The effect of PDT treatment was less evident in the 1:7 ratio sample; in this case, besides the presence of large necrotic and altered areas, few viable cells were observed with perfectly conserved mitochondria (Fig. 4d, highlighted frame), accounting for a higher survival rate as also indicated by the ATP measurement.

The distribution and frequency of cells that survived the PDT treatment in the 1:7 ratio spheroids were further evaluated by toluidine blue staining on semi-thin slices taken across the entire spheroid at different levels (Figure 5S, Additional file 7). Interestingly, we observed that viable cells were predominantly located at the external rim or in the polar region of the spheroid, while internal areas showed necrosis or morphological alterations in the majority of cells. In the spheroids composed by unloaded MSCs and MG-63, no difference in term of viability was observed between irradiated or not irradiated samples (data not shown).

PDT efficacy in an ectopic murine osteosarcoma model

The effect of irradiation of AlPcS4@FNPs loaded MSCs was further investigated in an ectopic OS mouse model.

After 2 weeks from Saos-2-Luc/MSCs co-injection, subcutaneous tumors were visible. A volume of 50 μl of AlPcS4@FNPs alone or loaded in MSCs were administrated intra-tumorally, as well as PBS or AlPcS4 alone (control groups) (timeline in Fig. 5a). The localization of the nanoparticles (NPs) within the tumor was monitored by taking advantage of AlPcS4 fluorescence, using the IVIS Lumina II System. As shown in Fig. 5b the AlPcS4@FNP distribution was more localized around the site of injection when delivered by MSCs, than when NPs were injected alone. Twenty-four hours after the intra-tumor injection, the tumor was exposed to irradiation for 20 min. This cycle of treatment was repeated once a week for 2 weeks. As shown in Fig. 5c and d, after the first irradiation (day 21), no significant difference was observed among all groups. The second photoirradiation, instead, delayed the tumor growth in mice treated with AlPcS4@FNPs alone (− 65 ± 14%, p < 0.001) or with AlPcS4@FNPs loaded MSCs (− 62 ± 13%, p < 0.01) (day 28). As side effect of the PDT, a circumscribed dark eschar on the mouse skin was observed a few days after photoirradiation on the tumor surface only in mice injected with AlPcS4@FNPs alone but not in any of the other groups (Figure 6S, Additional file 8). One week after the last treatment, the reduction on the tumor growth was further enhanced for both AlPcS4@FNPs (− 72 ± 10%, p < 0.0001) and AlPcS4@FNPs loaded MSCs (− 68 ± 10%, p < 0.001).

Fig. 5
figure 5

In vivo photodynamic therapy of OS tumors. Schematic representation of the in vivo treatments (a). Representative fluorescent luminescent imaging of AlPcS4 alone or loaded in NPs (AlPcS4@FNPs) and AlPcS4@FNPs loaded MSCs (AlPcS4@FNPs@MSCs) localization after intra-tumor injection (b). Representative BLI images showing the evolution of luciferase-expressing tumor cells treated (c). Quantification of luminescence intensity of regions-of-interest (ROI) (tumor) (the light events recorded in the acquired images expressed in mean ± SD vs time) ** p < 0.001 (AlPcS4@FNPs alone), * p < 0.01 (AlPcS4@FNPs loaded MSCs) at day 28 (d). Histological analysis of tumor tissues after treatments: H&E, Ki-67 and TUNEL staining (scale bar = 100 μm, black arrow = necrotic areas) (e). For this study a total of 18 mice were used, mice were divided in 4 group as followed: mice treated with PBS (n = 3), with AlPcS4 alone (n = 3), with AlPcS4@FNPs NPs alone (n = 6) and with AlPcS4@FNPs loaded MSCs (n = 6)

Full size image

Following the last BLI acquisition, mice were euthanized, and tumors were explanted and processed for histology. The anti-tumor effect of the PDT treatment in combination with AlPcS4@FNPs and AlPcS4@FNPs loaded MSCs injections were assessed with Hematoxylin and Eosin (H&E), Ki-67 and TUNEL staining. H&E and TUNEL staining on tumor histological sections revealed areas of necrosis (highlighted in Fig. 5e by black arrow). In particular, we identified degraded areas composed predominantly of necrotic/apoptotic cells in tissue sections from AlPcS4@FNPs and AlPcS4@FNPs loaded MSCs photoirradiated samples. Instead, mice treated with PBS or free AlPcS4 had an extensive Ki-67 staining, in fact almost all the cells showed uniform brown color, suggesting that the photoirradiation did not affect OS cell proliferation. However, cell proliferation decreased in both AlPcS4@FNPs and AlPcS4@FNPs loaded MSCs treated groups as shown in Fig. 5e, where the number of cells expressing Ki-67 is greatly reduced.

In summary, we showed that under these experimental conditions, photostimulation treatments of intra-tumorally injected AlPcS4@FNP loaded in MSCs were able to kill OS cells and decrease tumor burden compared to both control groups (PBS and free AlPcS4). Moreover, the internalization of the AlPcS4@FNPs by the MSCs improved the local concentration of NPs at the injection site, eventually reducing side effects when compared to AlPcS4@FNP alone.

Discussion

Despite the effort made by clinicians in the last 30 years, 30% of osteosarcoma (OS) patients still do not respond to standard treatments, succumbing to the disease [50]. Major issues for poor OS survival rate include the insurgence of distal metastases [51, 52], mainly localized in the lungs, and the development of multi drug resistance (MDR) [53, 54]. Possible strategies for improving OS patients’ survival rate include methods for selectively targeting the therapeutic agent to the tumor stroma, as well as the use of alternative therapeutic approaches able to either circumvent MDR or kill chemoresistant cells.

In this context, mesenchymal stromal cells (MSCs) are increasingly considered as an ideal vector for delivering antineoplastic drugs, because of their well-established ability to home towards the stroma of several primary and metastatic tumors [55, 56]. Indeed, MSCs have been used for the in vitro and in vivo delivery of, among others, diagnostic and therapeutic agents, small interfering RNA and nanoparticles [57]. In particular, several authors have shown that MSCs easily internalize different types of nanoparticles [27] and can reach the tumor, showing limited or no toxicity effect of NPs to MSCs [58, 59]. Additionally, other authors have investigated whether MSCs can transport nanoparticles for therapeutic purposes [22, 60, 61].

Among alternative cancer treatments, photodynamic therapy (PDT) has been successfully used to kill OS cells in vitro [35, 37, 62,63,64,65]; in particular Kusuzaki et al. has shown that PDT is also able to kill multidrug resistant OS cells [66]. In vivo the effectiveness of PDT has been demonstrated in OS animal models [34, 67, 68]. Moreover, PDT has been previously successfully used to treat sarcomas in a group of patients in which PDT inhibits the local recurrence after intralesional tumor resection [38, 69,70,71].

Our results successfully demonstrate that the internalization of AlPcS4@FNPs into MSCs takes place in 1 h, and that particles are retained in the cells for at least 3 days. This result is in line with Roger et al. work, where they demonstrated that PLA nanoparticles are internalized by MSCs up to 100% within 1 h, and that the particles are retained for at least 3 days [23, 24]. This aspect is particularly significant in view of clinical application since a 3 days interval is compatible with the migration of MSCs to the tumor from the site of the injection [72].

The awareness recently acquired that a monolayer culture is not predictive of in vivo results, prompted us to find an ex vivo model able to simulate in vivo physiology [73]. Tumor spheroids are an established model to investigate novel cancer treatments, as they provide a better recapitulation of tumor pathophysiological aspects, such as the in vivo-like differentiation pattern due to the appropriate 3D extracellular matrix (ECM) assembly and the complex cell–matrix and cell–cell interactions [74, 75]. Particularly, in our case the thickness of the cell aggregate (~ 400 μm of diameter), together with the presence of extracellular matrix and the inevitable oxygen gradient, provided a more challenging model for the PDT treatment. In this model, we clearly demonstrated that the efficacy of the photoactivation of AlPcS4@FNPs loaded MSCs depends on the ratio between MSCs and OS cells. As expected, this result suggested that the in vivo efficacy of this system will strongly depend on both the tumor’s dimensions and on the number of loaded MSCs that will reach the neoplastic region.

Additionally, by combining the results obtained from the APT content analysis, the Live&Dead staining assay and the TEM microscopy studies, we were able to establish that by decreasing the MSCs:OS ratio from 1:1 to 1:7, cell death is far higher in the center of the spheroid as respect to the outer region. This observation could be explained by the outgrowth of OS cells in the 4 days of spheroids formation that would ultimately confine MSCs in the more internal part of the spheroid. A similar distribution of MSCs in a spheroid model has been observed by Zhang et al. in melanoma cell spheroids [25].

As a starting point to test the efficacy in vivo of AlPcS4@FNPs loaded MSCs, an ectopic OS model was developed. An arbitrary dose of MSCs loaded with AlPcS4@FNPs was selected and injected intra-tumorally and the efficacy of two photoactivation cycles was tested. Compared to the control groups (PBS and AlPcS4 alone), after the second photoirradiation, tumor growth was reduced in both groups (AlPcS4@FNPs and AlPcS4@FNPs@MSCs, 72 and 68% respectively). Results observed by luminescence analysis were confirmed by histological changes in the treated tumor sections. Clear evidence of apoptosis, observed by H&E and TUNEL staining, were associated with PDT treatment in combination with AlPcS4@FNPs injection (alone or loaded in MSCs), supporting the role of our system in the cell-killing process after photoirradiation.

The results described herein demonstrate that the AlPcS4@FNPs@MSCs system is very promising for treating OS tumors. However, treatment outcome could be improved by increasing PDT efficacy, by either performing more irradiation cycles or/and by optimizing the AlPcS4@FNPs loaded MSCs dose as well the irradiation conditions. Importantly, even if a reduction of tumor growth was demonstrated in both groups of mice, i.e. AlPcS4@FNPs alone and AlPcS4@FNPs@MSCs, in the mice injected with AlPcS4@FNPs alone, the NPs had a larger distribution as shown in Fig. 5a, thus supporting the tumor targeting effect of MSCs. Furthermore, it is interesting to notice that only mice injected with AlPcS4@FNPs alone, displayed skin superficial burns, probably generated by an excessive local heating due to an over-concentration of the particles in the skin and their subsequent photoactivation, as already observed by others [76]. This data further supports the potential advantage of using MSCs as a delivery system in terms of selective localization at the target tissue, which in turn allows the control of unwanted side effects.

Conclusions

In summary, our results obtained both in vitro and in vivo show that the use of MSCs to deliver photoactivatable NPs together with PDT can be a promising treatment for OS. However, our findings indicate that MSCs loaded with AlPcS4@FNPs could be especially promising when used either in patients that have developed chemoresistance or when the tumor is small and located in inoperable sites. To conclude, we believe that MSCs based PDT technology will impact the design of clinical trial for personalized treatment.