Introduction

The use of photo-responsive nanoparticles for photothermal therapy (PTT) is becoming increasingly popular in cancer treatment, as it combines effectiveness, on-demand action, and remote stimulation. Among the various nano-heaters reported, Au-based nanoparticles are considered the gold standard due to their excellent stability and intense plasmonic resonance, which can be tuned to the near-infrared (NIR) biological window by adjusting the shape of the nanoparticles, such as nanorods1 or nanostars2. Other nanomaterials, such as iron oxides, have emerged as robust alternatives for inducing an anticancer photothermal effect3,4,5,6. Additionally, the field of nanomedicine is constantly incorporating other nanoparticle compositions7, including platinum (Pt), which offer a range of applications in imaging8,9,10, detection11,12, or therapy13,14,15. Pt nanoparticles are particularly attractive candidates for therapeutic action because of their high atomic number (78), which allows them to act as radiosensitizers and enhance the effectiveness of radiotherapy by improving the absorption of ionic radiation energies16,17,18.

It is worth noting that the cis-platinum complex [Pt(NH3)2Cl2], which is commonly used as chemotherapy for the treatment of various cancers (including lung, ovarian, testicular, and lymphoma), is based on Pt (II). However, the mechanism of action is linked to the complex, which acts by alkylation of DNA and transcription inhibition, rather than the metallic Pt itself. Nevertheless, Pt nanoparticles have been considered as potential initiators for a chemo-therapeutic effect19,20,21,22,23,24,25. The hypothesis of using Pt nanoparticles as initiators for a chemo-therapeutic action is double-edged, as it suggests an inherent cytotoxicity of Pt nanoparticles, which could be mediated by the release of reactive Pt-based species. However, recent studies have reported that Pt nanoparticles offer good cytocompatibility and in vivo tolerance, without causing any significant toxicity26,27,28,29,30.

It is also noteworthy to mention that emerging strategies in the field of nanomedicine are currently focused on intracellular nano-degradation of nanoparticles (for instance Fe-based31,32,33,34,35, but also FePt36,37,38,39) to elicit therapeutic action via the release of cytotoxic reactive species. Certain works are even considering the triggering or acceleration of nanoparticle degradation through external stimuli such as laser irradiation40,41,42,43. In light of these developments, Pt nanoparticles appear to be a potential candidate for merging PTT and chemotherapy, due to their potential for releasing reactive Pt-based species upon intracellular degradation, which could enhance the therapeutic effect.

However, the plasmonic resonance of Pt is primarily in the ultraviolet region44,45, which implies that Pt nanoparticles may not be the preferred choice for photothermal applications in the NIR window of nanomedicine. Despite this, a few studies have reported successful PTT applications of Pt nanoparticles under NIR irradiation26,46.

This study aims to integrate the recent field of Pt nanoparticles for nanomedicine, with a particular emphasis on designing nanoparticles for optimized photothermal conversion and evaluating their intracellular stability both with and without laser irradiation. The proposed Pt-based nanomaterials have a raspberry-like multi-core structure and are produced using a one-pot methodology that involves coating and assembling with alendronate. The significance of a multi-core structure has been previously demonstrated with Au-based NanoRaspberries47, which can shift their absorbance towards the NIR. Therefore, dendritic-shaped nanoparticles are considered to have promising characteristics for biomedical applications due to their distinct physicochemical, optical, and electronic properties48,49,50.

In this study, we put forward and demonstrate the hypothesis that confining a large number of Pt cores can also have a beneficial effect on PTT. We also systematically compare the effects of Pt multi-core NanoRaspberries (NR) and single-core ultrasmall Pt NanoSeeds (NS), on PTT for glioblastoma cancer spheroids. Additionally, we used X-ray absorption spectroscopy (XAS) at the Pt L3-edge (11,564 eV) to analyze the atomic and electronic structure of both nanoparticles, their behavior after being internalized in the cellular environment, as well as their stability after laser exposure. This assessment of nanomaterial stability is a crucial factor to consider for their potential therapeutic use.

Results and discussion

Synthesis and characterization of Pt nanoparticles

The synthesis of Pt nanoparticles involves the combination of an aqueous Pt(II) solution with sodium ascorbate, which acts as a reducing agent, along with a stabilizing solution of biphosphonate (alendronate) in either acidic (pH 3) or basic (pH 10) conditions. The formation of nanoparticles is facilitated by microwave treatment at 100 °C for 30 min. Under alkaline conditions, alendronate acts as a coating agent, resulting in the formation of ultra-small nanoparticles (referred to as NanoSeeds or NS) as shown in Fig. 1A. On the other hand, acidic pH causes alendronate to assemble the NanoSeeds together, leading to the formation of multi-core structures (referred to as NanoRaspberries or NR) as illustrated in Fig. 1B. NS have a mean diameter of 1.8 ± 0.3 nm, as determined from size measurements on transmission electron microscopy (TEM) images (Fig. 1C). In contrast, assembled NR have an average size of 69 ± 10 nm (Fig. 1D). Absorbance spectra of both types of nanoparticles show a decreasing trend from 300 to 800 nm (Fig. 1E). Notably, the absorbance of NR is higher than that of NS, with a 1.5-fold increase at 300 nm and a 3-fold increase in the NIR region at 800 nm. The structure of the nanoparticles was further analyzed using X-ray absorption spectroscopy (XAS), specifically focusing on the Pt L3-edge. The X-ray absorption near-edge structure (XANES) spectra of NS and NR at the Pt L3-edge (Fig. 1F, G) reveal that NS are partially oxidized compared to the Pt foil, while NR are more oxidized than NS. This is evident from the whiteline intensity and whiteline width in the normalized XANES spectra51 (Fig. 1F) as well as the shift of the first derivative maximum around 11568.5 eV in Fig. 1G, indicating that NR have the highest level of oxidation. This could be attributed to the synthesis of NR in an acidic environment. Moreover, a flattened post-edge resonance after the whiteline is observed for NS, and it is most prominent for NR, which suggests local disordering due to Pt oxidation52. Changes in the valence of the absorbing atom, as indicated by whiteline variations, are often accompanied by changes in structure, coordination, bond length, and disorder, as well as variations in particle size or induced strain53. These transformations are further investigated using extended X-ray absorption fine structure (EXAFS) analysis, as shown in Fig. 1H and Table 1. For NS, in addition to the Pt-Pt shell at a similar position as the Pt foil (around 2.7 Å), a new shell corresponding to Pt-O bonds at 1.9 Å, indicative of Pt oxide character, is observed. This Pt oxide character becomes more pronounced in NR. Furthermore, an increase in the Debye-Waller factor (related to local structural disorder) with the oxide fraction is observed, confirming the results obtained from XANES analysis. Figure 1I illustrates the heating cycles of both nanoparticles dispersed in water at a concentration of [Pt] = 6.7 mM when subjected to an 808 laser with an output power 1.65 W. This power corresponds to an average power density of 1 W cm−² at a distance of 4 cm from the laser fiber. The light-to-heat conversion efficiency was determined using the methodology described in the Methods section. The calculated values for the photothermal conversion efficiency were found to be η = (16 ± 3) % for NS and η = (25 ± 4) % for NR. Pt-based nanoparticles certainly do not perform better than gold-based ones, such as gold nanorods, that have already undergone extensive validation for photothermal therapy. Yet, while gold nanorods production54 generally involves a multistep protocol with the use of hexadecyltrimethylammonium bromide (CTAB) and AgNO3 that could pose biocompatibility issues, the Pt-based nanoparticles investigated here are synthesized in 30 min in water using alendronate only, a molecule already FDA approved for osteoporosis on g daily basis.

Fig. 1: Analysis of Pt-base nanoparticles.
figure 1

A Transmission electron microscopy images of NS, with larger view at the bottom; B Transmission electron microscopy images of NR, with larger view at the bottom; C NS size distribution analyzed on 200 independent nanoparticles on TEM images; D NR size distribution analyzed on 200 independent nanoparticles on TEM images; E Absorbance of NS and NR at the same Pt concentration; F XANES spectra at the Pt L3-edge for Pt NS and Pt NR compared with Pt foil. G First derivative of the XANES spectra to corroborate the changes in oxidation state. H Fourier transform (FT) of EXAFS signal for Pt foil, Pt NS and Pt NR. I Heating curves of NS and NR dispersed in water at [Pt]= 6.7 mM and irradiated with an 808 nm laser beam at an output power of 1.65 W (average power density of 1 W cm−²). Three 10 min laser on/off cycles are provided that prove the stability of the nanoparticles upon heating.

Full size image
Table 1 Results of the EXAFS fittings of Pt foil together with NS and NF samples at Pt-L3 absorption edge.
Full size table

Cell internalization of Pt NS and Pt NR and photothermal heating effect

Glioblastoma cancer cells (U87) were then exposed to the Pt nanoparticles at varying concentrations ranging from [Pt] = 0.2 mM to 1.25 mM for 24 h. Afterward, the cells were fixed, processed for TEM, and imaged for different incubation conditions. For both nanoparticles NS and NR, images were taken for incubation conditions of [Pt] = 0.2 mM (Fig. 2A for NS, Fig. 2C for NR) and [Pt] = 0.5 mM (Fig. 2B for NS, Fig. 2D for NR). The images reveal that both types of Pt nanoparticles are internalized by the cells through endocytosis and consistently localize within endosomes. The number of endosomes filled with nanoparticles per cell increases with the administered dose. In the case of NR, several nanoparticles can be found within the same endosome, ranging from 1 to 5 nanoparticles. On the other hand, for NS, the nanoparticles appear to be highly aggregated within each endosome.

Fig. 2: Intracellular localization of the nanoparticles.
figure 2

Transmission electron microscopy images of cells having internalized Pt NS at [Pt] = 0.2 mM (A) or [Pt] = 0.5 mM (B) and Pt NR at [Pt] = 0.2 mM (C) and [Pt] = 0.5 mM (D).

Full size image

The amount of Pt internalized by the cells was quantified using atomic emission spectroscopy (ICP-AES) on cell lysates. The results showed that it ranged from 0.1 to 1 pg of Pt per cell, and was similar for both Pt NS (Fig. 3A) and Pt NR (Fig. 3B).

Fig. 3: Photothermal heating.
figure 3

A, B Internalization rate of both types of Pt nanoparticles, NS (A) and NR (B) in glioblastoma cells as a function of Pt concentration during the 24 h incubation, from [Pt] = 0.2–1.25 mm, measured by ICP-AES on 200,000 cells and renormalized per cell, finally expressed in pg of Pt per cell. C, D Typical photos of spheroids made of 200,000 cells loaded with NSs (C) or NRs (D) incubated at [Pt] = 0.7 mM (left image) or 1.25 mM (right image). E, F IR images upon laser irradiation (808 nm) of different spheroids with increasing Pt loads, indicated as the concentration within the tube in mM of Pt. G, H Temperature elevation reached for the different NS (E) and NR (F) loaded spheroids, dispersed in 5 µL, as a function of the total Pt concentration in this volume. Heating was also measured for 5 µL of aqueous dispersions (solutions) of NS (E) and NR (F) at different Pt concentration.

Full size image

To evaluate the photothermal heating effect of Pt-loaded cells, the cells were arranged in a 3D tumoroid-like configuration, which is closer to the solid tumors in vivo compared to 2D monolayers. To form the spheroids, 200,000 cells loaded with different doses of Pt NS and Pt NR were aggregated by centrifugation and allowed to mature in a 0.5 mL tube in complete medium for 2 days, forming compact spheroids resembling tumoroids (Fig. 3C, D). Photothermal measurements were then performed using an 808 nm laser at the single spheroid level, in the tube, with a total volume of 5 µL, which was sufficient to cover the spheroid. The Pt concentration in the samples was calculated by dividing the quantity of Pt in the spheroids (as measured by ICP) by the volume of 5 µL. Aqueous dispersions of nanoparticles were also prepared in the same 5 µL volumes at different concentrations to compare heating in cells and in water. The samples were irradiated with an 808 nm laser beam with an output power of 1.65 W, which corresponds to an average power density of 1 W cm-², but increases to 2.6 W cm-² at the center of the tube due to the Gaussian distribution of the beam. Infra-red camera was used for temperature mapping (Fig. 3E, F). Temperature elevations were averaged in Fig. 3G (NS) and 3H (NR) for the different concentrations tested. The results showed that NS are not efficient for photothermal heating in solution, but they start heating efficiently when confined in cells. In contrast, NR exhibit the same high photothermal conversion both in cells and in water. These measurements clearly demonstrate that NR are efficient photothermal agents due to their multi-core structure mediated by alendronate aggregation, and their heating capabilities remain unchanged in the intracellular environment, making them suitable for heating both the cancer cell environment (extracellular heating) and the cells themselves (intracellular heating). However, the situation is different for NS, which exhibit low heating in solution (and thus in the extracellular environment), but become almost as efficient as NR when located inside the cells due to massive aggregation within the endosomes. Of course, for this aggregation effect to yield beneficial results in vivo, a substantial quantity of nanoparticles must successfully penetrate the cancer cells, uneasily attainable after intravenous injection. However, this is precisely where NR exhibit their distinctive advantage. By controlling the aggregation process prior to in vivo administration, they enable optimization of photothermal conversion efficiency upstream, independent of achieving a specific dosage within the target cancer cells. This pre-administration control over aggregation thus holds potential for enhancing the effectiveness of the Pt-based nanoparticles in vivo.

High-throughput photothermal treatment on spheroids in microwells

Next step was to evaluate the photothermal therapy of glioblastoma cells with the two types of nanoparticles. To do so, the live/dead assay was used to quantify both the viability and the death of spheroids after laser treatment. In order to have a significant number of spheroids, and the possibility for fluorescence imaging (live cells in green/dead cells in red) at spheroid level, spheroids were grown in 200 µm diameter microwells fabricated in agarose inside each well of a 96-well plate. 500 cells loaded with the nanoparticles were deposited in each microwell by centrifugation. After two days of maturation, all microwells were containing 100 µm diameter live spheroids (Fig. 4A for NS-loaded cells and Fig. 4B for NR-loaded cells). The entire wells of the plate could then be irradiated sequentially with the laser, for 5 min. To do so, the laser beam was adjusted the well diameter. Figure 4C, D show typical fluorescence imaging of the spheroids in microwells for NS-loaded cells (Fig. 4C) and NR-loaded cells (Fig. 4D), without laser treatment (0 W cm−²) and for the maximal laser power density (2.6 W cm−²) showing the two extreme cases of live only spheroids or dead only ones. Figure 4E–H finally present the live and dead fluorescence measured for all spheroids (n > 100) irradiated with different laser power density. It shows similar efficiency in terms of cell death induction for both NS and NR, in line with their similar heating once in the cells. For both, cell death is almost total for laser power density of 2.1 and 2.6 W cm-². Importantly, for spheroids made of unlabeled (control) cells, no increase of cell death is detected, whatever the laser power density (Fig. 4I–K).

Fig. 4: High-throughput evaluation of cell death at spheroid level after photothermal treatment.
figure 4

A, B Growth of spheroids in microwells starting from 250 cells loaded with Pt NS (A) or Pt NR (B). Images are shown at day 0 (seeding), and day 1 and day 2 of maturation. At day 2, live staining (green) is included. C, D Live (green) and dead (red) images at day 2 of the microwell array containing the NS (C) and NR (D) loaded spheroids, without (0 W cm-²) and with (2.6 W cm-²) laser exposure. E–H Quantification on single spheroids of the live and dead signal as a function of the laser power density: (E) and (F) correspond to the live fluorescent signal of NS and NR loaded spheroids, respectively; (G) and (H) correspond to the dead fluorescent signal of NS and NR loaded spheroids, respectively. I–K Quantification of the live and dead signal of spheroids formed of unlabeled (control) cells; (I) live (green) and dead (red) staining for spheroids exposed to the highest laser power density of 2.6 W cm−². J, K correspond to respectively the live and dead fluorescent signal for all measured spheroids, as a function of the laser power density.

Full size image

The comparison between control samples lacking internalized nanoparticles (Figs. 4I–K) and the control samples depicted in Fig. 5E–H, which do contain internalized nanoparticles but were not subjected to laser application (0 W cm−²), also serves as evidence that the Pt-based nanoparticles by themselves do not induce any cell death. Indeed, the occurrence of cell death could have been expected due to the enzymatic catalytic activity of Pt that could have catalyzed toxic superoxide radicals from hydrogen peroxide present in tumor cells.

Fig. 5: Intracellular stability of the local electronic and atomic structure investigated with XANES and EXAFS.
figure 5

Normalized XANES spectra of solution and cells without and with laser application, at day 1 or day 12 of spheroids maturation, are shown for NS (A) and NR (B). Insets show a detailed view of whiteline changes. FTs of the EXAFS signals are shown for spheroids at day 1 and day 12 without or after laser exposure, for NS (C) and NR (D).

Full size image

Biostability evaluation

The final step was to determine the biostability of the Pt nanostructures once in the intracellular environment. XAS at the XANES and EXAFS regimes can provide unique in situ description of the local electronic and atomic structure of nanomaterials in their environment, possibly biological. XAS measurements were thus performed on spheroids either at day 1 after their formation, or 12 days later (day 12), and compared with the spectra in aqueous dispersion. Besides, other spheroids that had been laser irradiated were also measured, at both day 1 and day 12 to assess the effect of laser irradiation at structural level.

XANES spectra for all conditions are shown for NS and NR in Fig. 5A, B, respectively. NS in the intracellular environment after day 1 shows a change in the XANES response with respect to the solution, both in the position of the whiteline towards that of the Pt foil and in the resonances that follow, indicating a change in the environment of the Pt absorbing atoms. After day 12 no XANES variations are identified compared with day 1 while a whiteline decrease is identified for the irradiated nanostructures. For the NR system, XANES results show a reduction of the Pt atoms as NR are internalized after day 1 and day 12 (similar absorption signal), with the whiteline at positions of the Pt foil but a lower intensity and a similar XANES spectra. After laser irradiation, an intensity increase of whiteline is obtained for both cases with a higher intensity after day 12. In any situation, the whiteline intensity are lower than that of the Pt foil. Those modifications of whiteline intensity could be attributed to some electron transfer or/and coordination modifications55. The Fourier transform (FT) of the EXAFS signal are then shown in Fig. 5C, D, to investigate the distributions of atoms within the nanoparticles, without and with laser irradiation and compared with the aqueous dispersion, at day 1 and day 12 for NS (Fig. 5C) and NR (Fig. 5D), respectively). For NS system for all conditions, a shift of the first shell at lower position with respect to the solution is obtained falling around 2.4 Å, confirming the change in the environment around the Pt atoms after the internalizations into cells. Different scenarios are considered and the EXAFS signal is fitted considering the Pt atoms coordinated with P and/or S atoms. This structural variation could be related with the heating efficacy of nanostructures in the cell environment and not in aqueous dispersion. For NR into the cell for all conditions, just one shell is obtained at positions close to that of the Pt shell around 2.8 Å with lowest intensity related to the coordination. After the laser irradiation, an increase of coordination is obtained for both day 1 and day 12, being higher for the latter. Remarkably, not only these measures evidence no degradation of the nanoparticles, but it shows on the contrary a biotransformation towards metallic Pt for NR and some compound with Pt-S/P coordination at short order for NS inside the cells. And last but not least, no significant structural change is identified. Just a change in the coordination number increasing it is detected, which could be related to a marked increase of the crystallinity.

At this stage, we cannot definitively determine the long-term in vivo stability. However, we have observed that there is no degradation at the spheroid-tissue level over a period of approximately two weeks. In contrast, significant degradation was observed when using a similar spheroid model with nanoparticles made of iron oxide56, silver57, or copper sulfide58. Interestingly, gold nanoparticles also did not degrade at all57,59, suggesting that Pt behaves similarly to Au based on these initial in vitro observations. However, it will be essential in the future to investigate this similarity in terms of both long-term in vivo stability and efficiency. Additionally, it is yet to be explored whether it would be advantageous to utilize nanomaterials that can degrade over time and be cleared out after accomplishing their therapeutic mission, despite the risk of triggering a biological response through the release of ionic species. Alternatively, nanoparticles that remain intact inside the cells for an extended period of time may also be considered.

Conclusion

The new class of Pt multi-core raspberries-like nanoparticles alendronate-assembled were efficiently internalized in cancer cells and provided sufficient heating to kill cancer spheroids upon laser exposure. These Pt NanoRasbperries (NR) were compared to ultrasmall Pt nanoparticles, NanoSeeds (NS), similar to the cores composing the Pt NR. While NR photothermal conversion was much higher that NS one for nanoparticles in aqueous dispersion, once confined within the endosomes of the cancer cells, NS were almost as efficient nano-heaters than NR, confirming the importance of aggregation in photothermal conversion. The Pt biostability of both nanoparticles was excellent, as revealed by XAS, after 1 day or 12 days of cellular processing. Remarkably, even after laser application, none were degrading, but conversely, were reshaped into more crystalline structure. To conclude, Pt nanoparticles with multi-core design appear as promising photothermal therapeutic agents both in the extracellular and intracellular environments, without triggering any degradation, even by laser irradiation.

Materials and methods

Synthesis of Pt nanoparticles

The multi-core synthesis is similar to previously reported ones based on Au47 or Pd60. 80 µL of 20 mM potassium tetrachloroplatinate at 20 mM and 180 µL of alendronate at 40 mM were mixed in 2 mL ultrapure water, at pH 3 or pH 10. 40 µL of a sodium ascorbate (17.6 mg/mL) were then added and the mix was microwave-heated (Monowave 300, Anton Paar GmbH) for 30 min at 80 °C. The solutions containing the nanoparticles were finally dialyzed (Spectra.Por Float-A-Lyzer G2) and were ready for use.

Cell culture and nanoparticles incubation

U-87 MG human glioblastoma cells (ATCC) were grown in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin in a humidified incubator with 5% CO2 at 37 °C. After reaching 80% confluence, the cell culture medium was replaced by a solution of nanoparticles dispersed in complete culture medium at concentrations [Pt] = 0.2, 0.5, 0.7, and to 1.25 mM. Cells were left to incubate for 24 h. They were then rinsed twice with complete medium and detached using 0.05% trypsin-EDTA solution. After cell counting, 200,000 cells were centrifuged at 1100 rpm for 5 min in 1 mL of culture medium in a 15 mL Falcon tube. The pellet is left in the incubator in tubes with partially open cap to mature for one or two days and consistently formed a cohesive spheroid in 1 day.

Transmission electron microscopy

For nanoparticles imaging, 10 μl of diluted nanoparticles suspension were seeded on 200 mesh copper, left for 1 min, and then dried with absorbing paper. For imaging inside the spheroids, they were processed as follows. Spheroids were washed with cacodylate buffer (0.2 M), then fixed with 5% glutaraldehyde in 0.1 M cacodylate buffer for 1 h min at room temperature and rinsed again in the same cacodylate buffer (0.2 M). The samples were then contrasted with Oolong Tea Extract (OTE) diluted to 0.5% in cacodylate buffer, and post-fixed in 1% osmium tetroxide containing 1.5% potassium cyanoferrate and then dehydrated in increasing ethanol baths. Samples were then embedded in Epon epoxy resin, sectioned (70 nm) and placed on a copper grid for observation. All nanoparticles and cell samples were observed with a HITACHI HT 7700 electron microscope operating at 120 kV (INRA, France).

Pt quantification by ICP-AES

The total amount of platinum in the cell samples was determined by ICP-AES (inductively coupled plasma atomic emission spectroscopy - Spectrogreen, SPECTRO). For each measurement, the number of cells was counted using a Malassez chamber, and the samples were digested with aqua regia consisting of a mixture of hydrochloric acid (2/3) and nitric acid (1/3). The mixture was then placed on a hot plate for evaporation, then re-suspended in 200 µL of hydrochloric acid and left at room temperature for two days. This final solution was then diluted with ultra-pure filtered water to reach 2% HCl solution, before measuring the total amount of Pt by ICP-AES.

Photothermal heating by 808 nm laser application

The heating profile was systematically acquired in 5 µL volume at the bottom of 0.5 mL Eppendorf tubes.

For aqueous dispersions, the solutions were diluted at different concentrations ranging from [Pt] = 0.125 to 6.7 mM. For spheroids, they were placed in tubes, and total volume was adjusted to 5 µL. A 808 nm laser was used (Laser Components). The laser beam was positioned at a distance of 4 cm from the liquid surface, and the output laser was set at 1.65 W. This corresponded to a mean power density of 1 W cm-² in the sample plane. However, the laser beam has a gaussian profile and at the center of it, where the sample is located, the power density is higher, measured with mask and powermeter at 2.6 W cm-². The temperature elevation was measured in real time using an infrared thermal camera (FLIR A615) and data were processed with the FLIR ResearchIR software. The light-to-heat conversion coefficient is calculated from heating curves61, as \({{{{{\rm{\eta }}}}}}={{{{{\rm{B}}}}}}.\triangle {{{{{\rm{T}}}}}}.\frac{{{{{{{\rm{m}}}}}}}_{{{{{{\rm{sample}}}}}}}{{{{{{\rm{C}}}}}}}_{{{{{{\rm{w}}}}}}}}{{{{{{{\rm{P}}}}}}}_{0}(1-{10}^{-{{{{{\rm{A}}}}}}})}\) with the sample mass msample approximated to the mass of water, Cw the specific heat capacity (4.18 J g−1 K−1), P0 the incident laser power and A the absorbance of the sample. B, the rate of heat dissipation, is calculated from the decreasing temperature curves when the laser is turned off, \({{{{{\rm{B}}}}}}=\frac{-1}{{{{{{\rm{t}}}}}}}.\, {{{{\mathrm{ln}}}}}\left(\frac{{{{{{\rm{T}}}}}}\left({{{{{\rm{t}}}}}}\right)-{{{{{{\rm{T}}}}}}}_{0}}{{{{{{{\rm{T}}}}}}}_{{{{{{\rm{m}}}}}}}-{{{{{{\rm{T}}}}}}}_{0}}\right)=0.025{s}^{-1}\).

Photothermal therapy on spheroids in microwells

U87 cells incubated with [Pt] = 0.7 mM of NS and NR for 24 h in complete DMEM medium were detached and deposited in wells of 96-well plate (20,000 cells per well) containing agarose molded microwells. Those microwells were microfabricated thanks to 3D printed stamps containing a network of micro-pillars (37) of 200 µm diameter and 200 µm height. In all, 50 µL of liquid 2% agarose in PBS was poured in each well, the stamps were placed, and the agarose was left to solidify (15 min) before removing the stamps. Cells deposited in well were centrifuged (1200 rpm, 3 min) to fall within the microwells (500 cells per microwell). They were left to mature into cohesive spheroids for 2 days before laser application, performed in situ of the 96-well plate containing the microwells. The laser was adjusted at 3 cm above the plate surface with a laser holder centered on top of each well. Laser power density as calibrated with a mask of the well surface, positioned at 3 cm. 4 power densities of 1.1, 1.6, 2.1, and 2.6 W cm−² were selected, and applied well per well, for 5 min in a 37 °C - 5% C02 environment. Live/Dead staining was performed after laser irradiation according to supplier instructions (LIVE/DEAD™ Cell Imaging Kit R37601, Invitrogen™). Multi-well images were acquired in fluorescence mode with EnSight Microplate Reader (Perkin Elmer). Images were processed with ImageJ.

X-ray absorption spectroscopy

Both the X-ray absorption near-edge structure (XANES) and X-ray absorption fine structure (EXAFS) experiments were performed on samples at the BL22 CLÆSS beamline of the ALBA synchrotron facility in Cerdanyola del Vallès (Spain)62. Measurements were achieved at room temperature in fluorescence mode at the Pt L3-edge (11564 eV). The monochromator used in the experiments was a double Si crystal oriented in the (311) direction. Pt metal foil was measured and used to calibrate the energy. Solutions and cells were measured in transmission or fluorescence configuration, according to convenience, in order to have an absorption jump of 1. The XAS analysis was performed using the Athena and Artemis softwares63.

The modulus of the Fourier transform (FT) of the EXAFS signal at the Pt L3-edge is performed in the k2χ(k) weighted EXAFS signal between 2.4 and 9.0 Å-1. Experimental EXAFS results are fitted in R-space in the range 1.4–3.5 Å using the FEFFIT code64. The fitting was performed by fixing the shift at the edge energy E0 and using the number coordination N, the interatomic distance R and the Debye-Waller (DW) factors as free parameters.