Converting lysosomes into photothermal organelles enables nanoparticle-free tumor ablation via intracellular vapor bubbles

Abstract

Photothermal nanomaterials enable precise tumor ablation but face limitations in biodistribution, tissue penetration, toxicity, and biodegradability. Here, we present a unique concept for nanoparticle-free photothermal therapy based on the lysosomal entrapment of cationic amphiphilic small molecular dyes for spatially controlled vapor bubble (VB)–mediated tumor cell ablation. This strategy, which exploits a universal biological and physical effect, uses intracellular pH gradients for extensive local dye enrichment in acidified organelles, transforming them into transient endogenous nanosized photothermal reactors for subsequent light activation. Using sunitinib, a clinically approved lysosomotropic anticancer drug, and the commercially available dye LysoTracker Deep Red, lacking intrinsic anticancer activity, we demonstrate pulsed laser-induced VB formation from dye-enriched lysosomal compartments, leading to selective photomechanical disruption of various ex vivo cancer cell models across two-dimensional (2D) cultures, 3D spheroids, patient-derived neuroblastoma tumoroids, and tumor fragments from a patient with ovarian carcinoma. This approach allows precise, low-fluence, and wavelength-tunable cancer tissue ablation without the need for synthetic photoresponsive nanoparticles.

INTRODUCTION

Nanoparticle-mediated photothermal therapy (PTT) has emerged as a promising strategy for killing accessible cancer cells using light (1–3). Compared to traditional anticancer treatments, phototherapies typically offer the advantage of high spatiotemporal precision (4–7). However, conventional PTT induces tumor cell death by heat induction, which may result in excessive heat dissipation and potential damage to healthy surrounding tissue (8). Laser-induced vapor bubble (VB) formation, mediated by pulsed laser irradiation of photothermal nanoparticles such as gold nanoparticles (AuNPs), iron oxide nanoparticles (IONPs), and polydopamine nanoparticles (PDNPs), has therefore attracted increasing attention (9). Upon exposure to femto-, pico-, or nanosecond laser pulses at specific wavelengths, such nanoparticles rapidly heat up, causing the surrounding water to vaporize and form VBs, thus reducing excessive heat accumulation and thermal diffusion (10–12). VBs can persist for tens to hundreds of nanoseconds before collapsing, thereby generating mechanical forces capable of locally disrupting targeted cells or tissues (13).

However, despite these advances, inefficient and heterogeneous intratumoral penetration of nanoparticles remains a central obstacle to effective tumor therapy (14–17). Following systemic administration, nanoparticle transport is strongly constrained by abnormal tumor vasculature, elevated interstitial fluid pressure, and dense tumor stroma, resulting in preferential perivascular accumulation and poor access to deeper tumor regions (18–20). These intrinsic transport barriers are highly variable across tumor types and patients, making it difficult to predict or control intratumoral nanoparticle penetration. This limits the translational impact of many nanoparticle-based strategies across both therapeutic and imaging modalities (21, 22). Even localized intratumoral injections into accessible tumors often results in shallow penetration depths and nonuniform tissue coverage, while off-target accumulation and poor biodegradability continue to challenge clinical translation and long-term safety (23–26).

To overcome these limitations, small molecules generally offer superior tissue distribution compared to nanoparticles. However, freely dissolved small molecular dyes typically fail to act as efficient photothermal reactors for VB induction due to suboptimal photophysical properties (27, 28). Cationic amphiphilic drugs (CADs) are a heterogeneous class of weakly basic molecules with high apparent volumes of distribution in vivo (29). Moreover, due to the pH partition effect and tumor acidosis, CADs are even more likely to accumulate in solid tumors (30). Intracellularly, the pH gradient between the near-neutral cytosol (pH ∼7.2) and the acidic lumen of lysosomes (pH ∼4.5) may lead to substantial drug enrichment within acidic intracellular compartments dependent on the CAD’s pK_a (31, 32). In this work, we propose to exploit this lysosomotropic behavior as a nanoparticle-free strategy for photothermal tumor cell ablation. More specifically, upon treating cancer cells with light-sensitive CADs, we hypothesize the spontaneous and transient conversion of dye-enriched lysosomal compartments into localized photothermal organelles for intracellular VB generation and tumor cell eradication. The latter builds on our recent work demonstrating melanoma cell ablation via VBs generated from intracellular melanosomes under nanosecond laser irradiation (33). The proposed strategy extends this concept to nonpigmented cancer cells, aiming to make the approach more broadly applicable.

To test this approach, we first repurposed the clinically approved cationic amphiphilic tyrosine kinase inhibitor sunitinib to induce intracellular laser-triggered VBs from lysosomal organelles. To demonstrate the broader applicability of the concept beyond repurposed drugs, lysosomal VB formation was also evaluated using LysoTracker Deep Red (LDR), a well-known cationic amphiphilic dye that accumulates in lysosomes via a similar mechanism. Exposure of sunitinib- or LDR-treated cancer cells to nanosecond laser pulses enabled effective and spatial-selective VB-mediated cell killing and tumor ablation in various two-dimensional (2D) and 3D tumor cell models, including patient-derived neuroblastoma tumoroids and tumor fragments from a patient with low-grade serous ovarian cancer (LGSOC).

Together, our findings provide ex vivo proof of concept of the potential of lysosomotropic cationic amphiphilic photosensitizers for lysosome-mediated intracellular VB formation, thus enabling nanoparticle-free tumor cell eradication with high spatial precision. Future work will focus on in vivo validation of this approach.

RESULTS

Intracellular accumulation and localization of sunitinib

Owing to its physicochemical properties as a hydrophobic weak base (logP: 5.2; pK_a: 9.0; DrugBank database), sunitinib is expected to predominantly accumulate in the lysosomes of cancer cells (Fig. 1A). Sunitinib exhibits strong absorbance in the 340–490 nm range and fluoresces around 540 nm, making it inherently suitable for light activation (fig. S1) (34, 35). Detection of sunitinib’s fluorescence by flow cytometry indicated that intracellular sunitinib levels in three different cancer models (i.e., HeLa cells, Jurkat cells, and HT1080 cells) rise in a concentration-dependent manner, with all cells having acquired the compound at concentrations ≥2.5 μM and 6 hours of incubation (fig. S2A). The relative mean fluorescence intensity (rMFI) also revealed a time-dependent intracellular accumulation of sunitinib, with 24-hour incubation showing the highest sunitinib fluorescence in the cells (Fig. 1B and fig. S3). Confocal microscopy imaging of the intracellular sunitinib distribution showed a clear colocalization of the drug with LDR-stained organelles (Fig. 1D and fig. S4), corroborating its lysosomotropic behavior (36). Moreover, sunitinib exposure increased the intracellular lysosomal density, indicated by both a heightened rMFI (Fig. 1C) and LDR fluorescence area per cell (Fig. 1E and fig. S5). The latter may result from both sunitinib-induced lysosomal swelling as well as de novo lysosomal biogenesis, as generally described for lysosomotropic CADs (37). Additionally, sunitinib exposure increased the side scatter (SSC-A) signal, a parameter indicative of lysosomal perturbation by CADs, in a time- and concentration-dependent manner (fig. S2B) (38, 39). Although absolute intralysosomal sunitinib concentrations were not quantified, taken together, our data indicate that sunitinib-treated cancer cells exhibit characteristic hallmarks of lysosomotropic accumulation.

Fig. 1. Characterization of sunitinib accumulation in HeLa cells.

(A) Schematic of intracellular sunitinib accumulation and lysosomal sequestration. (B) rMFI of cells accumulating sunitinib. (C) Relative increase in LDR staining of HeLa cells as a function of sunitinib concentration and incubation time. Data are presented as means ± SEM (N = 3, n = 3). Statistical analysis was performed using one-way analysis of variance (ANOVA) followed by Tukey’s multiple comparisons test. (D) Sunitinib accumulation (5 μM, 24 hours) and colocalization with LDR staining. (E) Quantification of LDR fluorescence area per cell (normalized). Data are presented as median ± interquartile range (N = 3, n = 10). Statistical analysis was performed using a two-tailed one-sample t test. h, hours.

Pulsed laser irradiation of lysosomal accumulated sunitinib triggers VB formation

Considering the optical properties of sunitinib and its high degree of lysosomal accumulation, we hypothesized that sunitinib-loaded lysosomes could serve as endogenous intracellular photothermal organelles. Upon nanosecond pulsed laser irradiation, the photothermal conversion of the lysosomes may lead to rapid heating, potentially inducing the formation of intracellular VBs (Fig. 2A). To verify this, HeLa cells, A549 cells, HT1080 cells, and Jurkat cells were treated with 10 μM sunitinib during 24 hours and successively irradiated with a nanosecond pulsed laser (480 nm) with increasing laser fluence (i.e., energy applied per unit area, in joules per square centimeter). Dark-field microscopy clearly revealed the emergence of intracellular VBs via detection of the extensive light scattering produced during their lifetime (Fig. 2, B and D, and movie S1). In contrast, laser irradiation of sunitinib in solution (up to 2 mM) did not evoke VBs under identical conditions (movie S2), indicating that accumulation of sunitinib in the lysosomes is required. Quantification of the number of VBs formed within a defined laser-irradiated area as a function of the laser pulse fluence allowed us to calculate the VB generation threshold (Fig. 2, C and E), which is defined as the fluence of a single laser pulse at which VBs are formed with 90% probability. Based on the Boltzmann fitting parameters, the VB formation process can be categorized into three distinct regimes. The “Sparse” regime, spanning from the baseline to the inflection point, is characterized by a low and stochastic VB count. This is followed by the “Moderate” regime, defined between the inflection point and the 90% threshold, where the VB population increases markedly and nonlinearly per pulse before reaching the saturation plateau. The VB threshold exhibited minor variation among different adherent cell types. For instance, HeLa cells and A549 cells displayed a threshold of 1.0 J cm⁻², whereas Jurkat cells formed VBs at a substantially lower laser fluence of 0.8 J cm⁻² (fig. S6A and movie S3). A plausible explanation for the latter is that Jurkat cells tend to form clusters, allowing multiple sunitinib-loaded cells to contribute cooperatively to localized heat generation, facilitating VB formation at lower energy inputs. In support of this, resuspending Jurkat cells as a loose monolayer increased the VB threshold to 3.0 J cm⁻² (fig. S6B). Jurkat monolayers requiring substantially higher laser fluences than 2D cultures of either HeLa or A549 cells can be explained by the markedly lower sunitinib accumulation (figs. S2C and S3B). The latter likely limits localized intralysosomal heating upon laser irradiation, thereby raising the energy threshold for VB formation. Furthermore, according to the pH partition theory, a higher initial extracellular sunitinib dose is expected to lead to higher intralysosomal drug concentrations, thus increasing the photoresponsiveness of the organelles and lowering the VB threshold (fig. S7, A and B). Hence, in cells treated with 20 μM sunitinib, VB formation was already detectable at a laser fluence as low as 0.3 J cm⁻² (fig. S7D). HeLa cells incubated with 10 μM sunitinib undergo immediate and complete ablation upon a single laser pulse, while at a lower concentration of 5 μM, no visible change in cellular morphology was observed (movie S4). The applied laser fluence of 1.7 J cm⁻² is above the threshold for 10 μM sunitinib, while it fails to generate VBs in cells exposed to 5 μM sunitinib, supporting the fact that cell disruption is caused by VBs. At the highest sunitinib concentration of 20 μM, even the formation of large microbubbles was observed, presumably formed by coalescence of multiple VBs in the irradiated area (Fig. 2F and fig. S7D). After laser application, the cells in this illuminated spot were again clearly ablated (Fig. 2G and movie S5). Comparable results were also obtained for the A549 cell type (fig. S8). In summary, incubation of a variety of cancer cell lines with sunitinib can induce intracellular VBs from drug-loaded lysosomal compartments upon pulsed laser irradiation, dependent on the applied sunitinib concentration, incubation time and laser fluence.

Fig. 2. Detection of VBs in sunitinib-exposed cells upon nanosecond pulsed laser irradiation.

(A) Schematic of laser irradiation of cancer cells with sunitinib-loaded lysosomes to produce intracellular VBs, illustrating the VB formation and collapse from sunitinib-loaded lysosomes. (B) Dark-field microscopy images of HeLa cells before and during laser irradiation for visual detection of intracellular VBs and (C) the corresponding quantification of the intracellular VB formation threshold. (D) Dark-field microscopy images of Jurkat cells before and during laser irradiation and (E) the corresponding intracellular VB formation threshold. The thresholds were determined for HeLa and Jurkat cells treated with 10 μM sunitinib for 24 hours. (F) Dark-field microscopy images of HeLa cells incubated with different concentrations of sunitinib (2.5 to 20 μM) and exposed to pulsed laser irradiation at a fluence of 1.0 J cm⁻². (G) Time-resolved dark-field microscopy images capturing the formation and collapse of VBs in HeLa cells incubated with 20 μM sunitinib following a single nanosecond laser pulse. The sequence illustrates the full dynamic process of intracellular VB generation and subsequent collapse.

Cell killing induced by intracellular VB formation upon pulsed laser irradiation of sunitinib-treated cancer cells

Significant laser-induced cell death was observed across all sunitinib concentrations tested, in contrast to laser treatment alone (Fig. 3A). Near-complete HeLa cell eradication was achieved at 10 μM sunitinib and 1.8 J cm⁻² laser pulses, conditions leading to extensive VB formation and photomechanical cell ablation within the irradiated region. Higher laser fluences and sunitinib concentrations likely induce more damage to the cell membrane that cannot be easily repaired, thereby causing greater cytotoxicity (Fig. 3B) (40).

Fig. 3. Pulsed laser-induced cell killing of sunitinib-exposed cells.

(A) Decrease in cell viability after laser treatment (1.8 J cm⁻²) applied on HeLa cells incubated with sunitinib for 24 hours at concentrations of 0 to 10 μM. (B) Cell viability of HeLa cells treated with sunitinib (5 and 10 μM) decreased with increasing laser fluence. (C) TEM images of HeLa cells incubated with 10 μM sunitinib for 24 hours for sunitinib only control (no laser, sunitinib CTR), low-fluence irradiation (0.3 J cm⁻²), and VB-generating fluence (0.5 J cm⁻²). Low-magnification images (top) depict overall cellular morphology, while corresponding high-magnification views (bottom) focus on lysosomal structures. Yellow dashed boxes mark cytoplasmic voids observed after VB formation. (D) Quantification of average lysosome diameter in HeLa cells based on TEM analysis after different treatments. (E) Quantification of caspase-3/7 activation in HeLa cells incubated with 10 μM sunitinib for 24 hours, followed by laser irradiation at increasing fluences (0.3 to 1.0 J cm⁻²). Caspase activity was measured using the Caspase-Glo 3/7 assay and normalized to the nontreated control (NTC) group. (F) Assessment of intracellular ROS levels in Jurkat cells using the CellROX Deep Red assay following nanosecond pulsed laser irradiation. Tert-butyl hydroperoxide (TBHP) was included as a positive control. (G) Confocal microscopy images of HeLa cells treated with 10 μM sunitinib after scanning with different laser fluences in a predefined line pattern and (H) images of cells treated with different concentrations of sunitinib (5 to 20 μM) using a fixed laser fluence (1.1 J cm⁻²). (I) Confocal microscopy images of cells scanning in a predefined text pattern (10 μM, 1.1 J cm⁻²). Data are presented as means ± SEM (N = 3, n ≥ 3). Statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple comparisons test.

To investigate the subcellular events associated with VB formation, we next performed transmission electron microscopy (TEM) on HeLa cells following sunitinib incubation and laser irradiation (Fig. 3C). In line with earlier observations, 10 μM sunitinib-treated cells exhibited pronounced lysosomal swelling. When cells were irradiated with laser fluences below the VB threshold (e.g., 0.3 J cm⁻²), serving as a heating-only control, no obvious morphological changes were observed. At a laser fluence of 0.5 J cm⁻², corresponding to the onset of VB formation at the given sunitinib concentration, distinct cytoplasmic voids emerged, suggestive of sites of VB formation and subsequent collapse. The latter coincides with a significant decrease in the number of intact lysosomes and a reduction in their size, indicative of selective rupture of the larger lysosomes or VB-induced lysosomal fragmentation (Fig. 3D). In some images, these voids appeared to impinge upon or even disrupt the nuclear envelope. In addition, cells irradiated at 0.5 J cm⁻² frequently exhibited elongated mitochondria, a well-established morphological hallmark of cellular stress responses (41–45). Additionally, a Caspase-Glo 3/7 assay revealed a clear activation of caspase 3/7, particularly when exceeding the VB threshold (Fig. 3E), indicating that apoptosis is triggered after VB formation. As intracellular reactive oxygen species (ROS) generation is a key cytotoxic mechanism in conventional PDT (46), the CellROX Deep Red Assay was used to detect ROS production following sunitinib exposure and laser treatment of Jurkat cells. However, both flow cytometry and confocal fluorescence microscopy revealed negligible ROS generation under the applied conditions (Fig. 3F and fig. S9G).

To explore the spatiotemporal nature of sunitinib-induced photoablation, a laser beam (spot size of ∼200 μm) was scanned across a monolayer of sunitinib-treated HeLa cells in a predefined pattern consisting of two vertical lines, each 1 mm wide and separated by a nonirradiated zone of 100 μm (fig. S9, A and B). Necrotic and late apoptotic cells, identified by positive TO-PRO-3 staining, were only found within the irradiated regions, whereas adjacent nonirradiated cells remained viable (i.e., calcein-AM positive), despite sunitinib exposure. These observations demonstrate precise spatial control over laser-induced intracellular VB formation and cell ablation (Fig. 3G). Confocal microscopy further confirmed that the extent of photomechanical cell killing was dependent on both sunitinib concentration and laser fluence. At the highest sunitinib concentration tested, TO-PRO-3–positive cells were no longer detectable within the irradiated zones, indicative of severe cell ablation (Fig. 3H). Comparable results were again obtained using A549 cells (fig. S9D). Cells located in nonirradiated areas retained their proliferation capacity after laser treatment, further verifying the spatial selectivity of the laser treatment (fig. S9E). Furthermore, we demonstrated highly selective cell killing through the creation of customized irradiation patterns, including defined letters and geometric shapes (Fig. 3I).

Last, the transient effect of drug exposure on subsequent laser-induced cell killing was assessed by replacing sunitinib-containing medium (after 24-hour incubation) with fresh culture medium for an additional 72 hours prior to laser treatment (fig. S9F). A marked reduction in the laser-induced cell killing effect was observed, likely explained by sunitinib washout. Furthermore, laser irradiation at suboptimal wavelength (e.g., 532 nm), which lies outside the absorption band of sunitinib, did neither induce any detectable VB formation nor compromise cell viability (fig. S9H). These findings highlight both the requirement for sufficient intralysosomal sunitinib concentration to enable VB-mediated cell ablation as well as the reversible nature of the lysosomal sunitinib sequestration.

Photoablation of sunitinib-treated tumor spheroids induced by VB formation

Spheroids more closely mimic the architecture, biological behavior, and microenvironmental characteristics of human tumors (47–49). HeLa cell spheroids were incubated with different sunitinib concentrations for 24 hours, resulting in a dose-dependent intracellular drug accumulation and a marked reduction in spheroid diameter (Fig. 4A and fig. S10A). The latter observation likely correlates with the reduced cell viability within the spheroids, approaching near-complete cell death at a concentration of 30 μM sunitinib (Fig. 4B). Due to the higher cell density, increased light scattering and absorption within 3D spheroids compared to 2D cultures, and the expected limited tissue penetration depth for laser light of the selected wavelength, photoablation was performed using multiple consecutive laser pulses (n = 10) at a fluence of 1.0 J cm⁻². To visualize VB formation and the resulting ablation effect more clearly, the laser was targeted at the periphery of the spheroid rather than the core (50, 51). Light microscopy enabled real-time imaging of VB formation at the illuminated area (Fig. 4D). Furthermore, edge-localized photoablation damage could be visualized under dark-field microscopy (Fig. 4E and movie S6).

Fig. 4. Photoablation of sunitinib-exposed tumor spheroids by pulsed laser irradiation.

(A) Diameter and (B) viability of HeLa spheroids incubated with different concentrations of sunitinib. (C) Viability of spheroids before and after laser treatment. Data are presented as means ± SEM (N ≥ 5, n = 3). Statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple comparisons test. (D) Optical microscopy images showing VBs at the spheroid’s edges as a function of localized laser irradiation. Yellow dashed circles indicate the irradiated region and highlight the structural disruption of the spheroid observed at the moment of VB collapse. Yellow arrows mark prominent VBs formed during laser exposure. (E) Dark-field microscopy images showing the photoablation effect of spheroids cultured with mounting concentrations of sunitinib. (F) Single time-point analysis of spheroids using bright-field microscopy for morphometric analysis. (G) Single time-point analysis of spheroids by confocal fluorescence microscopy. Fluorescence labeling with Hoechst (blue, indicating nuclei) and TO-PRO-3 (red, indicating dead cells) enabled evaluation of fluorescence intensity distribution within the spheroid area. Confocal fluorescence microscopy was performed using Z-stack imaging to further assess the spatial distribution of both markers following laser exposure, shown as maximum intensity projections. Sunitinib CTR, sunitinib only control.

At the lowest sunitinib concentration tested (5 μM), laser treatment resulted in a marked reduction in tumor cell viability from 79 to 25% (Fig. 4C), despite no overt morphological ablation being observed at this dose. By contrast, control spheroids showed no significant decrease in viability upon laser exposure, which confirms that sunitinib accumulation is required. At higher sunitinib concentrations, the effect of laser exposure was further amplified, with <5% viable cells remaining when combining 20 μM sunitinib with laser treatment. This level of cell killing was not reached even with high doses of photothermal nanoparticles (fig. S11, C and D), with spheroids demonstrating higher resistance to nanoparticle-induced cell death compared to 2D cultures.

Additional morphometric parameters, such as spheroid circularity and fragmentation, were evaluated by bright-field microscopy combined with automated image segmentation using AnaSP software (Fig. 4F) (52). Upon laser irradiation, especially at fluences of 1.6 J cm⁻² and above, clear edge defects became visible (white arrows), likely resulting from VB formation and collapse at the spheroid periphery. These structural alterations were further reflected in the quantification of circularity (fig. S10B). Although spheroids irradiated at 1.4 J cm⁻² did not exhibit visible fragmentation, viability analysis indicated over 70% cell death. This suggests that while the majority of cells were nonviable, the spheroid retained its physical integrity, possibly due to the presence of extracellular matrix (ECM) components (53). At fluences of 1.6 J cm⁻² and above, clear spheroid ablation and fragmentation was observed, correlating with increased TO-PRO-3 staining (Fig. 4G). Notably, in spheroids exhibiting fragmentation, the peripheral regions showed markedly elevated TO-PRO-3 fluorescence, indicating that the VB formation caused pronounced cell death at the spheroid edges. This observation highlights the role of photomechanical damage in compromising spheroid integrity and inducing localized ablation, which demonstrates the potent ability to kill sunitinib-loaded tumor cells, even within dense 3D structures.

Repurposing lysosomotropic dyes for lysosome-assisted intracellular VB formation

Next, we sought to investigate whether other photoresponsive lysosome-targeting dyes might elicit a similar photomechanical response. To validate this, LDR, a commercially available probe used for visualizing acidic organelles, was selected (Fig. 5A). Unlike sunitinib, LDR lacks intrinsic anticancer activity and exhibits strong absorption in the 500–600 nm range (fig. S1D). Notably, LDR exhibits a relatively low fluorescence quantum yield (reported to be <0.4) (54, 55), indicating that a substantial fraction of the absorbed photon energy is dissipated through nonradiative decay pathways. This enables efficient photothermal conversion required for VB formation. First, HeLa cells were incubated with 0.5 or 1 μM LDR for 24 hours and subsequently irradiated with a 532-nm nanosecond pulsed laser, using a commercially available benchtop photoporation device (LumiPore, Trince, Ghent, Belgium) (9, 10, 56). Clear VB formation was observed, with lower LDR concentrations requiring higher laser fluences (Fig. 5, B and C, and fig. S12, A and B). Similar results were obtained in HT1080 cells (Fig. 5, D and E, and movie S7). Time-resolved dark-field imaging in both HeLa and HT1080 cells incubated with 1 μM LDR and irradiated at 1.2 J cm⁻² clearly illustrates the VB lifecycle, demonstrating a VB lifetime within 10 ms (fig. S13A). Cells irradiated at 561 nm, a wavelength closer to the absorption maximum of LDR, generated more VBs compared to 532 nm at a slightly reduced VB threshold (fig. S14). Together, these data confirm VB generation in LDR-treated cells following pulsed laser irradiation at a wavelength matching the dye’s absorption spectrum.

Fig. 5. Detection of VBs in LDR-treated cells upon nanosecond pulsed laser irradiation.

(A) Schematic of pulsed laser irradiation of lysosome-stained cells with LDR to produce VBs, illustrating their formation and collapse. (B to E) Dark-field microscopy of HeLa and HT1080 cells before and during laser application and visual detection of VBs and VB threshold quantification for cells treated with 1 μM LDR for 24 hours. (F) Confocal fluorescence microscopy of HeLa cells treated with 1 μM LDR before and after laser irradiation at 1.2 J cm⁻², illustrating intracellular morphological deformations following VB generation and collapse. (G) Cell viability of HT1080 cells treated with LDR (1 μM) with increasing laser fluence. (H) Quantification of HT1080 cell viability after incubation with different concentrations of LDR (0.1 to 1 μM), followed by laser irradiation at 0.5 and 1.2 J cm⁻². (I) Quantification of caspase-3/7 activation in HT1080 cells incubated with 1 μM LDR for 24 hours, followed by laser irradiation at increasing fluences (0.3 to 1.0 J cm⁻²). Caspase activity was normalized to the nontreated control group. Data are presented as means ± SEM (N = 3, n ≥ 3). Statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple comparisons test. (J) Confocal fluorescence microscopy analysis of MC38-GFP-Gal8 cells incubated with 0.5 μM LDR for 24 hours and subsequently irradiated with nanosecond pulsed laser at varying fluences. Representative images show Gal8-GFP puncta formation, indicative of lysosomal membrane permeabilization. GFP channel images were converted to grayscale using ImageJ for enhanced contrast. (K) Time-resolved dark-field microscopy images showing sequential frames of HeLa cell spheroids incubated with 1 μM LDR before and after exposure to laser at a fluence of 1.2 J cm⁻². (L) Single time point analysis of spheroids using bright-field microscopy.

We next assessed whether VB formation and collapse translated into effective cell killing. HT1080 cells were incubated with either 0.5 or 1 μM LDR for 24 hours, followed by irradiation with increasing laser fluences ranging from 0.08 to 1.6 J cm⁻². Control groups without laser exposure (LDR only) and cells treated with low laser fluences (≤0.3 J cm⁻²) did not induce significant cytotoxicity. Starting at 0.5 J cm⁻², corresponding to the previously identified VB threshold for 1 μM LDR, cell viability dropped substantially, inducing ∼70% cell killing at the highest fluence tested (1.6 J cm⁻²) (Fig. 5G). At LDR concentrations <0.5 μM, no significant cell killing was noted (Fig. 5H). Tests performed on HT1080 and CT5.3 cell lines showed consistent trends (fig. S13, B and C). Furthermore, we also demonstrated that LDR-assisted cell killing was spatially confined (fig. S13E). Together, these findings support a dose- and fluence-dependent cell killing effect mediated by LDR-assisted VB generation.

Following laser irradiation at 1.0 J cm⁻², in contrast to sunitinib-treated cells, LDR-treated cells exhibited less violent photomechanical effects. Nevertheless, real-time confocal microscopy revealed the emergence of transient intracellular cavities consistent with subcellular VB formation (Fig. 5F). Notably, using MC38–enhanced green fluorescent protein (eGFP)–galectin 8 reporter cells, 1 μM LDR exhibited a progressive increase in galectin puncta formation with mounting laser fluences, indicative of endolysosomal membrane permeabilization (Fig. 5J and fig. S13F). In line with sunitinib, significant caspase activation was only observed at laser fluences above LDR’s VB threshold, reinforcing that VB-induced mechanical damage serves as the primary trigger for subsequent apoptotic cell death (Fig. 5I).

Following 2D cell culture, LDR-induced VB formation was also confirmed in HeLa tumor spheroids (Fig. 5K and movie S8). Given the negligible cytotoxicity of LDR, the spheroid diameter remained unchanged after 24-hour incubation with 1 μM LDR (Fig. 5L). Minor morphological changes were detected upon laser treatment, again supporting less violent photomechanical ablation compared to sunitinib. At higher fluences, peripheral cells appeared more dispersed and the spheroid boundary less defined, indicating ablation effects at the spheroid surface. LDR-treated spheroids irradiated with 1.8 J cm⁻² exhibited a clear loss of calcein-AM signal and disrupted morphology in the laser-exposed region, confirming localized cell death (fig. S13G). Although no macroscopic fragmentation or collapse of the spheroids was observed, a significant reduction in cell viability occurred at fluences above 1.6 J cm⁻² (fig. S13D), confirming that VB generation in LDR-loaded spheroids can induce substantial cell death.

Lysosomal VB-mediated tumor ablation in patient-derived tumor models

To further validate the concept of lysosome-mediated VB formation in a complex 3D tumor context, patient-derived neuroblastoma AMC691B-GFP tumoroids were exposed to 1 μM LDR (Fig. 6A). Following a 532-nm wavelength single nanosecond laser pulse, the central region of the tumoroid appeared darker with bright-field microscopy imaging, likely due to extensive light scattering caused by intracellular VBs (fig. S15A). This reduced optical transmission gradually returned to baseline levels within 0.5 s, corresponding to VB collapse. These observations were corroborated by dark-field microscopy, clearly showing the emergence of numerous VBs (Fig. 6B and fig. S15B). In line with the data obtained on tumor spheroids, no clear macroscopic disruption of the tumoroid structure was noted. Nevertheless, tumoroid viability was significantly reduced at a laser fluence of 0.9 J cm⁻² (>VB threshold; Fig. 6C). Notably, upon applying a second laser pulse, neither additional VBs nor further viability reduction could be observed (fig. S15C). This result supports the notion that (i) once photosensitized lysosomes are actuated to generate VBs, they are structurally compromised, and (ii) LDR-mediated intracellular VB formation and lysosomal disruption is driving cell death induction.

Fig. 6. Lysosomal VB-mediated tumor ablation in patient-derived tumoroids and tumor fragments.

(A) Schematic illustration of the workflow for LDR-treated AMC691B-GFP neuroblastoma tumoroids preparation and culture. (B) Dark-field microscopy images of VBs in LDR-incubated tumoroids upon 532-nm laser exposure. (C) Quantitative viability analysis of tumoroids treated with varying laser fluences (0.3 to 2.6 J cm⁻²). Data are presented as means ± SEM (N = 3, n = 6). Statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple comparisons test. (D) Schematic illustration of the workflow for sunitinib-exposed patient-derived tumor fragments preparation and culture. (E) Light microscopy images of tumor fragments from a patient with LGSOC before and after nanosecond pulsed laser irradiation at fluences of 2.3 and 4.6 J cm⁻². Nontreated control is included for comparison. (F) Corresponding dark-field microscopy images revealing the generation of VBs upon laser exposure, with more prominent scattering observed at higher fluence. (G) Confocal fluorescence microscopy images of TO-PRO-3–stained tumor fragments showing the extent of cell death in the sunitinib only control (sunitinib CTR, no laser) and after irradiation at 2.3 and 4.6 J cm⁻². (H) Representative H&E-stained sections of patient-derived tumor fragments following VB-mediated treatment. Top: Untreated tumor fragments incubated with sunitinib only and two groups exposed to a 2.3 J cm⁻² pulsed laser fluence. Bottom: Corresponding zoom-in views highlighting local tissue morphology. The images highlight the preservation of the overall ECM architecture surrounding the treated regions, supporting the spatially confined nature of the photomechanical effect. (I) Quantification of tumor fragment’s viability across conditions, confirming increased cell death with mounting laser fluence. Data are presented as median (n = 7). Statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple comparisons test.

Although such tumoroids capture the 3D cellular architecture of tumors, they still lack the ECM density and heterogeneity of genuine patient-derived tumor tissues, where the dense ECM can hinder both drug diffusion and laser penetration (57). Hence, we next investigated ex vivo cultured patient-derived tumor fragments, which preserve the native ECM and cellular composition of clinical tumor specimens. Here, tumor fragments were obtained from a low-grade serous ovarian carcinoma (LGSOC), a tumor type characterized by abundant collagen-rich stroma with small nests of cancer cells. Following a 24-hour incubation with 10 μM sunitinib, fragments were irradiated with nanosecond laser pulses at fluences of 2.3 or 4.6 J cm⁻² (Fig. 6D). Light microscopy imaging revealed increased optical transparency of the tumor fragments postirradiation compared to the untreated control, suggestive of structural disruption and decreased tumor cell density (Fig. 6E). Real-time dark-field microscopy further confirmed the formation of VBs exclusively in sunitinib-incubated samples. Moreover, VBs were only observed in tumor cell nodules and not in the ECM, further supporting the need for intracellular sunitinib accumulation (Fig. 6F and movies S9 and S10). Confocal microscopy combined with TO-PRO-3 staining revealed markedly higher cell death in the laser-treated group compared to the sunitinib-only control, while preserving the fragment’s structural integrity (Fig. 6G). Consistent with this, hematoxylin and eosin (H&E) staining (Fig. 6H) provides histological evidence that the surrounding ECM architecture was preserved following VB-mediated treatment. Last, the CellTiter-Glo 3D cell viability assay quantitatively demonstrated a near-complete loss of ATP-dependent viability following sunitinib exposure and laser-induced VB generation within the ECM-rich tumor fragments (Fig. 6I). These results collectively confirm that lysosomal accumulation of light-sensitive CADs can be exploited to induce localized tumor ablation, even in clinically relevant, ECM-rich tumor tissues.

DISCUSSION

The application of photothermal nanoparticles for cancer tissue ablation often faces several limitations, including poor biodegradability, potential long-term retention in organs such as the liver and spleen, and concerns regarding their immunogenicity and systemic toxicity (58, 59). In addition, their limited tumor penetration and uneven intratumoral distribution can compromise treatment uniformity and efficacy (60). In this study, we therefore propose a nanoparticle-free strategy for photothermal tumor cell killing by exploiting the lysosomotropic properties of cationic amphiphilic small molecular dyes, as demonstrated ex vivo for the clinically approved light-sensitive tyrosine kinase inhibitor sunitinib and the commercially available dye LDR (61).

While the lysosomal accumulation of sunitinib has been described to contribute to drug resistance (36, 37, 62), it also enables its repurposing as an intracellular photosensitizer. Exposure of a variety of sunitinib-treated cancer cell models to pulsed laser light led to VB formation, depending on the drug concentration and the applied laser fluence. In contrast, laser treatment of free sunitinib in solution (up to 2 mM) does not generate VBs. The concentrations of sunitinib used throughout this study mostly remain below the acute 24 hours IC₅₀ in HeLa cells (∼16.6 μM; fig. S16). Based on the pH partition theory for weakly basic drugs, under equilibrium conditions up to 800-fold higher concentrations of sunitinib can be expected in the lysosomal compartment at pH 4.5, compared to the pH neutral cytosol (note S1). Of note, this value represents a theoretical equilibrium estimate; the absolute sunitinib concentrations in different organelles under the given conditions have not been experimentally verified in this work. It has been demonstrated in literature that CADs can also insert into the membrane of anionic intraluminal vesicles in lysosomal compartments. Moreover, it is known from our earlier work that localized accumulation of dyes may promote their propensity toward VB formation by reducing intermolecular distance. Hence, it is hypothesized that the combination of sunitinib’s optical and lysosomotropic properties converts lysosomes into endogenous photoresponsive substrates for intracellular VB formation, leading to violent photomechanical damage to organelles and/or cells upon their collapse. At the same time, lysosomotropic agents may also distribute within late endosomes (pH ∼5.0 to 6.0) (63, 64). Nevertheless, based on the substantially higher drug accumulation anticipated in the more acidic lysosomes based on the pH-partition theory, we believe that the latter organelles are the main contributors to the observed VB-induced photomechanical effects.

Despite a clear correlation between laser application, VB induction, and cell killing, other mechanisms cannot be excluded. Both photomechanical, photothermal, and photochemical effects can contribute to cell damage evoked by photosensitizers (65). In particular, our experiments with sunitinib revealed a certain degree of cell killing at laser fluences below the VB threshold. The latter could in part be explained by localized production of heat, thus increasing the permeability of the endolysosomal limiting membrane and therefore promoting cytosolic release of sunitinib from the lysosomes or evoking lysosomal cell death (10). Of note, the latter also presents a potential strategy to overcoming the challenge of lysosomal entrapment, a major obstacle that may contribute to drug resistance to cationic amphiphilic anticancer drugs. In addition, previous studies demonstrated generation of ROS upon continuous laser irradiation of intracellular sunitinib (34, 66). Nevertheless, our results suggest that VB-mediated cytotoxicity induced by pulsed laser irradiation is not driven by such photochemical mechanisms. Opposed to continuous laser illumination, a single laser pulse delivers energy in a short burst, thus preventing ROS accumulation over time.

Next to sunitinib repurposing, the concept of using lysosomotropic dyes to convert lysosomes into photothermal organelles was additionally validated with the commercial dye LDR, inducing intracellular VBs at a higher wavelength (>500 nm). Notably, while LDR lacks intrinsic anticancer activity and is not clinically used, it serves here as a mechanistic probe to demonstrate that VB-mediated cytotoxicity is primarily governed by photomechanical effects rather than drug-specific biochemical activity. Our results again demonstrated a clear correlation between intracellular VB formation and cell death, albeit without the extensive photomechanical damage to the treated cells typically observed for sunitinib exposure. Most likely, the subcellular VB-mediated disruption of the lysosomes itself is driving cytotoxicity instead of macroscopic cell ablation. More dedicated experiments would be required to shed light on the exact cell death modalities involved under different experimental conditions.

Notably, we were able to show ex vivo VB-mediated cell killing with both sunitinib and LDR in tumor spheroids, as well as human patient-derived tumoroids and tumor fragments. Such tumor models retain key features of the tumor microenvironment and provide a relevant platform for evaluating localized photomechanical effects. Multiple reports have demonstrated limited penetration of nanoparticles of varying sizes in tumor tissue. In contrast, our results reveal that small molecule photosensitizers such as sunitinib and LDR can effectively reach the tumor cells in more complex 3D tumor models and accumulate into their lysosomal compartments, thereby rendering them amenable to VB-mediated cell killing upon pulsed laser application.

To benchmark our nanoparticle-free strategy against conventional photothermal agents, we evaluated the therapeutic efficacy of synthetic nanomaterials in both 2D and 3D HeLa cell models (note S2). Both 80-nm AuNPs and 150-nm PDNPs induced higher cytotoxicity in monolayer cultures upon pulsed laser irradiation compared with 3D tumor spheroids. Of note, sunitinib induces comparable dose-dependent reduction in cell viability in both 2D and 3D models, with a near-complete loss of viability observed at the highest dose tested. These findings highlight the inherent transport barriers and restricted interstitial diffusion of larger synthetic particles within dense tumor architectures, whereas our small molecule approach may enable improved tissue access and effective ablation, even in stroma-rich tumors. Nevertheless, the latter statement still requires further in vivo validation.

Last, several translational considerations for future development of this platform can be highlighted. Compared to conventional continuous-wave photothermal approaches that primarily rely on gradual bulk heating and thermal diffusion, the present strategy operates under conditions of thermal confinement, enabling localized intracellular VB generation and photomechanical disruption with reduced heat diffusion to surrounding tissue. Such spatial confinement may be advantageous for applications requiring high precision and minimized collateral tissue damage. While visible-wavelength excitation, as required for both sunitinib and LDR, inherently limits tissue penetration depth in larger tissue volumes, the proposed strategy may be particularly suitable for applications in which direct optical access to tumor tissue is feasible, including intraoperative treatment of tumor margins to eliminate microscopic residual disease, as well as endoscopic or laparoscopic interventions. Nanosecond pulsed laser systems are clinically available and are already widely implemented in ophthalmology and dermatology for highly localized tissue disruption.

Of note, although clear ex vivo proof of concept could be obtained, both sunitinib and LDR have not been developed and optimized for photothermal applications. Hence, future research will be focused on developing next-generation lysosomotropic cationic amphiphilic dyes with optimized optical properties, particularly within the NIR-I and NIR-II spectral windows. We anticipate that these optimized agents could improve tissue penetration and photothermal performance, thereby facilitating clinical translation of this platform. Moreover, as in vivo data are currently lacking, it will be of interest to evaluate in vivo delivery, tissue selectivity, therapeutic efficacy, and systemic safety of novel lead photoresponsive lysosomotropic compounds in future work.

In conclusion, we have identified a nanoparticle-free strategy to enable spatial-selective VB-mediated photoablation of cancer cells using lysosomotropic cationic amphiphilic photosensitizers. As it is based on a universal cellular and physicochemical mechanism, we anticipate that this concept can be used across multiple cancer types, photosensitizers, and wavelengths, including applications with NIR dyes.

MATERIALS AND METHODS

Materials

Dulbecco’s modified Eagle’s medium/Nutrient Mixture F-12 (DMEM/F-12), DMEM (high glucose), Roswell Park Memorial Institute (RPMI) 1640 medium, Ham’s F-12 Nutrient Mix (catalog no. 11765054), B-27 supplement (50×, serum-free, catalog no. 17504001), recombinant human FGF-basic (amino acids 1 to 155, catalog no. PHG0266), animal-free recombinant human EGF (PeproTech, catalog no. AF-100-15), recombinant human IGF-I (PeproTech, catalog no. 100-11), recombinant human PDGF-AA (PeproTech, catalog no. 100-13A), recombinant human PDGF-BB (PeproTech, catalog no. 100-14B), penicillin/streptomycin solution (100 IU ml⁻¹ penicillin, 100 μg ml⁻¹ streptomycin), l-glutamine (2 mM), and 0.25% Trypsin-EDTA and Collagen I, rat tail (catalog no. A1048301) were purchased from Gibco-Life Technologies (New York, USA). Dulbecco’s phosphate-buffered saline without Ca²⁺/Mg²⁺ (DPBS⁻), bovine serum albumin (BSA), and sodium azide (NaN₃) were obtained from Sigma-Aldrich (St. Louis, USA). Fetal bovine serum (FBS) was from Hyclone (GE Healthcare, Machelen, Belgium). Sunitinib malate (CAS no. 341031-54-7) was purchased from Cayman Chemical (Michigan, USA). LDR (catalog no. L12492), CellROX Deep Red (catalog no. C10491), and CellROX Green (catalog no. C10492) Flow Cytometry Assay Kits, TO-PRO-3 iodide (catalog no. T3605), and calcein-AM (catalog no. C3099) were obtained from Invitrogen (Eugene, USA). The CellTiter-Glo Luminescent Cell Viability Assay (catalog no. G7571), CellTiter-Glo 3D Cell Viability Assay (catalog no. G9682), and Caspase-Glo 3/7 Assay (catalog no. G8091) were obtained from Promega (Belgium).

Cell lines and culture conditions

Human cervical epithelial adenocarcinoma HeLa cells, lung epithelial adenocarcinoma A549 cells, and T lymphoblast Jurkat E6-1 cells were obtained from American Type Culture Collection (Manassas, USA). Fibrosarcoma HT1080 cells and hTERT-immortalized colon cancer–associated fibroblasts (CT5.3hTERT CAFs) were provided by O. De Wever (Laboratory of Experimental Cancer Research, Department of Human Structure and Repair, Ghent University). MC38-GFP-Gal8 cells was provided by B. De Geest (Biopharmaceutical Technology Unit, Department of Pharmaceutics, Ghent University). HeLa cells were cultured in DMEM/F-12, and A549, HT1080, and CT5.3hTERT cells were cultured in DMEM with high glucose. Jurkat E6-1 cells were cultured in RPMI 1640. All culture media were supplemented with 10% FBS, 2 mM l-glutamine, and 100 U ml⁻¹ penicillin/streptomycin (i.e., complete cell culture medium). All cell lines were maintained in a humidified atmosphere containing 5% CO₂ at 37°C, and culture medium was renewed every other day. Cell lines are regularly tested and found negative for mycoplasma.

Tumor spheroids

Spheroids were produced via seeding cells in ultralow attachment (ULA) round-bottom 96-well plates (PrimeSurface, S-BIO, Japan) with a density of 4 × 10³ cells per well. The spheroids were incubated in a humidified atmosphere containing 5% CO₂ at 37°C, and culture medium was renewed every day. Spheroid morphology was monitored daily using bright-field microscopy. Acquired images were subjected to automated segmentation and morphometric analysis, including diameter and circularity, using AnaSP software (52). On day 3, the spheroids were incubated with sunitinib at concentrations ranging from 0 to 40 μM for 24 hours. Following incubation, the spheroids were collected, washed three times with PBS, and transferred for subsequent sample preparation and analysis.

The AMC691B-GFP tumoroids were provided by K. Durinck (Pediatric Precision Oncology Lab, Department of Biomolecular Medicine, Ghent University) (67). Tumoroids were maintained in a specialized tumoroid medium consisting of DMEM (low glucose, GlutaMAX) and Ham’s F-12 Nutrient Mix (4:1 v/v), supplemented with 2% B-27 (minus vitamin A), 1% N-2 supplement, 100 U ml⁻¹ penicillin, and 100 μg ml⁻¹ streptomycin. This medium was further enriched with recombinant human growth factors, including EGF (20 ng ml⁻¹), FGF-basic (40 ng ml⁻¹), IGF-I (200 ng ml⁻¹), PDGF-AA (10 ng ml⁻¹), and PDGF-BB (10 ng ml⁻¹).

Preparation and ex vivo treatment of patient-derived tumor fragments

Patient-derived tumor material was applied consistent with our previously reported protocols (68). Collection of tissue samples from patients was conducted in accordance with relevant guidelines and approved by the Ethics Committee of Ghent University Hospital (reference BC-06978). Patient-derived LGSOC tumor fragments were obtained at the time of surgery following informed written consent. Solid tumor lesions were macroscopically selected by a pathologist and immediately placed in ice-cold DMEM LG (catalog no. 31885023, Thermo Fisher Scientific) supplemented with 10% FBS (Sigma-Aldrich), 100 U ml⁻¹ penicillin, and 100 μg ml⁻¹ streptomycin (catalog no. 15070063, Life Technologies). All tumor material was immediately processed by automated cutting into small tumor fragments of 0.125 mm³ and placed on ice. After processing, the tumor fragments from different regions within a tumor were mixed to ensure uniform representation of the tumor. Thereafter, individual tumor fragments were seeded in prewarmed supplemented DMEM LG. To prevent evaporation, ULA round-bottom 96-well plates (PrimeSurface, S-BIO, Japan) were sealed with Breathe-Easy semipermeable membranes (catalog no. Z380059, Merck Life Science). Tumor fragments were cultivated for up to 5 days (37°C, 5% CO₂) under normoxic conditions.

The tumor fragments were first transferred using sterile 200-μl pipette tips into a new ULA round-bottom 96-well plates containing prewarmed culture medium and treated with sunitinib at a final concentration of 10 μM. Following a 24-hour incubation, the tumor fragments were carefully transferred to flat-bottom 96-well plates (VWR, Avantor, USA). The drug-containing medium was replaced with 200 μl of fresh culture medium per well to remove extracellular sunitinib. Subsequently, the fragments were subjected to laser treatment.

Quantification of cellular accumulation of sunitinib by flow cytometry

Cells were seeded at a cell density of 8 × 10³ cells per well in flat-bottom 96-well plates (VWR, Avantor, USA) and allowed to adhere overnight. The next day, serial dilutions of sunitinib in cell culture medium were applied at final concentrations ranging from 0 to 10 μM. After incubation for 6, 12, or 24 hours at 37°C, lysosomes were labeled by incubation with 75 nM LDR (Thermo Fisher Scientific, Waltham, MA, USA) for 15 min at 37°C. Sunitinib fluorescence, LDR intensity, and SSC signal were measured by flow cytometry, using a CytoFLEX flow cytometer equipped with a 96-well plate reader (Beckman Coulter, Krefeld, Germany) and CytExpert software. Data analysis was performed using FlowJo_v10.6.2 (Treestar, Costa Mesa, CA, USA).

Colocalization of sunitinib with lysosomes observed by confocal microscopy

Cells were seeded in sterilized and tissue culture–treated 96-well microplates (PhenoPlate, Revvit, The Netherlands) with black well walls and an optically clear cyclic olefin bottom at a density of 1.2 × 10⁴ cells per well and were allowed to adhere overnight. The following day, cells were incubated with different concentrations of sunitinib for 3, 6, 12, 24, 48, or 72 hours. Before imaging, sunitinib-containing cell culture medium was removed, cells were stained with 75 nM LDR for 15 min at 37°C, and the nuclei were labeled with Hoechst 33342 (Molecular Probes, Belgium) in cell culture medium (1/2000 dilution) for 15 min at 37°C. Microscopy imaging was performed with a laser scanning confocal microscope (Nikon A1R HD confocal, Nikon, Japan), equipped with a 60× water objective lens (60× WI Plan Apo, NA 1.27, WD 180 μm, Nikon, Japan) with a laser box (LU-N4 LASER UNIT 405/488/561/640, Nikon Benelux, Brussels Belgium) and detector box (A1-DUG-2 GaAsP Multidetector Unit, GaAsp PMT for 488 and 561 nm and Multi-Alkali PMT for 640 and 405 nm). The 405-, 488-, and 561-nm laser were applied to excite the Hoechst-labeled nuclei, sunitinib, and LDR, respectively. Fluorescence emission was detected through 450/50 nm (MHE57010), 525/50 nm (MHE57030), and 595/50 nm (MHE57050) filter cubes, respectively. A Galvano scanner was used for unidirectional scanning to acquire the sequential channels with two times line averaging and scan speed of 0.042 FPS. The pinhole was set to 20.43 μm, and the pixel size was 100 nm pixel⁻¹. NIS Elements software (Nikon, Japan) was applied for imaging. Images were analyzed with ImageJ (FIJI, https://fiji.sc/) (69).

Laser irradiation and VB formation

VB detection was performed using a previously reported in-house developed setup including an optical system and an electric timing system (56, 70). The system includes a pulsed laser source (Opolette HE 355 LD, OPOTEK Inc., CA, USA) with a pulse duration of ∼7 ns, a repetition rate of 20 Hz, and a beam diameter of 222 μm. The excitation wavelength was set to 480 nm. Cells were seeded at a density of 1 × 10⁵ cells per dish onto 50-mm glass-bottom petri dishes (MatTek, P50G-1.5-30-F) that had been precoated with Gibco Collagen I, rat tail (Thermo Fisher Scientific) for 1 hour at room temperature. After 24 hours of incubation to allow cell attachment, cells were treated with varying concentrations of sunitinib (0 to 20 μM) and further incubated for an additional 24 hours prior to laser irradiation. VBs were visualized via dark-field microscopy using a dark-field condenser, capitalizing on the strong light scattering properties of nano/microbubbles. An EMCCD camera (Cascade II: 512, Photometrics, Tucson, AZ, USA) was synchronized with the laser pulses through an electronic pulse generator (BNC575, Berkeley Nucleonics Corporation, CA, USA), enabling the acquisition of images before, during, and after laser exposure. Laser pulse energy was monitored in real time using an energy meter (J-25MB-HE&LE, EnergyMax-USB/RS sensors, Coherent), and laser fluence was calculated as the average energy per pulse divided by the beam area (J cm⁻²). By counting the number of VBs within the laser irradiation area for increasing laser fluence, one can determine the VB threshold, defined as the laser pulse fluence at which 90% of the plateau of producing VB (71). To precisely determine this threshold and characterize the VB generation dynamics, the experimental data were fitted using a Boltzmann sigmoidal model

$$y=A_2+\frac{A_1-A_2}{1+\text{exp}\left[(x-x_0)/d_x\right]}$$

In this equation, $A_0$ and $A_1$ represent the lower and upper plateaus of the VB count, respectively, while $x_0$ denotes the inflection point (midpoint) of the transition and $d_x$ signifies the width of the response curve. The VB generation threshold $x_{90}$ was then calculated as $x_{90}=x_0+d_x·\text{ln}\left(9\right)$, representing the laser fluence required to reach 90% of the saturation plateau. Furthermore, based on these fitting parameters, the VB induction process was categorized into three distinct regimes: (i) the “Sparse” regime, spanning from the baseline $A_1$ to the inflection point $x_0$, characterized by a low and stochastic VB count; (ii) the “Moderate” regime, defined from $x_0$ to the $x_{90}$ threshold, where the VB number increases markedly and nonlinearly; and (iii) the “Threshold” regime (above $x_{90}$), where frequent and stable VB formation occurs, leading to consistent photomechanical effects.

To facilitate the identification and quantification of individual VBs against the background, frame-by-frame image analysis was performed using ImageJ (FIJI, https://fiji.sc/), allowing precise visualization of VB formation and collapse dynamics. In addition to EMCCD-based detection, an sCMOS camera was also used for real-time monitoring of VB formation and threshold measurements. To support reproducibility and visualization, the operation window of the Micro-Manager software was screen-recorded using Free Cam (Digital Thoughts Software), and representative videos have been provided as supplementary videos.

Programmable area-defined laser scanning

For quantification of cell viability as a function of laser irradiation, HeLa cells were seeded at a cell density of 8 × 10³ cells per well in flat-bottom 96-well plates (VWR, Avantor, USA) and were allowed to adhere overnight. The following day, sunitinib dilutions in cell culture medium at final concentrations ranging from 0 to 10 μM were applied, and cells were incubated for 24 hours at 37°C. For laser treatment, the plates were placed on an electronic motorized translation stage (H117, Prior, UK), enabling automated laser scanning across each well. The laser was scanned in a “snake”-type pattern (fig. S9A), with a scanning speed of 7 × 10³ μm s⁻¹. The beam was moved line by line across the well with a fixed step width of 0.2 mm in the Y direction, matching the laser spot diameter to ensure continuous, nonoverlapping coverage of the entire well. X-field and Y-field parameters were user-defined to control the number of scans in each direction, and the edge of the 96-well plate was used as a reference point to initialize each scanning sequence. This setup ensured that the laser beam uniformly covered the entire well surface without redundancy or omission.

Evaluation of cellular health

Cell viability was then measured via detection of cell’s metabolic activity using a CellTiter-Glo assay (Promega, Belgium) following the manufacturer’s instructions. Correspondingly, the viability of spheroids was measured using the CellTiter-Glo 3D Cell Viability Assay. Data are presented as a percentage of viable cells calculated from the luminescence signal of each condition relatively to nontreated cells and by taking into account the background fluorescence of the medium.

Apoptosis was assessed via detection of the cell’s caspase activity using the Caspase-Glo 3/7 assay (Promega, Belgium) following the manufacturer’s instructions. Briefly, cells were seeded in 96-well plates (VWR, Avantor, USA) and incubated under standard conditions. Prior to laser irradiation, the culture medium containing sunitinib was removed and replaced with 100 μl of fresh culture medium. Immediately after laser treatment, the plates were returned to the incubator and cells were allowed to recover for an additional 6 hours. Subsequently, 100 μl of Caspase-Glo 3/7 reagent was added directly to each well, bringing the total volume to 200 μl. Plates were gently shaken at 100 rpm for 10 min and then incubated at room temperature for 20 min to allow luminescence development. Following incubation, 150 μl of the supernatant was transferred from each well into a new white-walled 96-well microplate (PS, F-Bottom, Chimney Well, Greiner, Germany), and luminescence was measured using a GloMax Microplate Reader. Caspase activity was expressed as relative caspase-3/7 activity (fold of control).

Confocal microscopy imaging of spatial-selective cell killing

HeLa cells and A549 cells were irradiated according to a predefined pattern. Two vertical lines with a width of 1 mm, separated by a nonirradiated zone of 100 μm (fig. S9B), were scanned and cells were subsequently incubated for 4 hours at 37°C, 5% CO₂. After incubation, cells were stained with calcein-AM green cell-permeant dye (0.5 μg ml⁻¹; Invitrogen) and 1 μM TO-PRO-3 iodide (Invitrogen) nucleic acid stain for 15 min at 37°C, 5% CO₂, to identify viable and necrotic cells, respectively, and afterward visualized using confocal microscopy.

Wide-field fluorescence and spinning disk confocal images were recorded with a Nikon Ti microscope using an Andor DU-897 X-8026 camera with a 20× objective lens (Plan Apo λ 20× 0.45 NA, Nikon), numerical aperture of 0.75, and a refractive index of 1. Two planes were imaged at 488/640 nm. The camera settings included a 1 × 1 binning, 100 ms exposure, and a multiplier of 300. The readout speed was set to 17 MHz with a conversion gain of Gain 1, and the vertical shift speed was 3.3 μs. The PFS-S system was active with an offset of 8094 and the mirror inserted. The illumination was controlled via the NIDAQ system. Imaging was performed at a scan speed of 4900 rpm using the CSU system, with the CSU shutter open and fiber output set to EPI. Image acquisition was carried out in multiexcitation mode, ensuring precise control of the laser lines and their corresponding wavelengths. NIS Elements was used as acquisition software. A large field-of-view image was composed of 9 by 9 images (512 × 512 image size) with 15% overlapping and stitching by a blending algorithm. Images analysis was performed using ImageJ (FIJI, https://fiji.sc/), including correction of field nonuniformity.

ROS detection

Intracellular ROS generation was assessed using the CellROX Deep Red Reagent (Thermo Fisher Scientific), a cell-permeant fluorogenic probe that becomes highly fluorescent upon oxidation by ROS. The dye exhibits excitation/emission maxima at ∼640/665 nm, allowing for reliable detection of oxidative stress in live cells via flow cytometry and confocal fluorescence microscopy.

To avoid interference from trypsinization-induced oxidative stress, suspension Jurkat cells were used. Cells were first incubated with either sunitinib or LDR at indicated concentrations for 24 hours. After incubation, cells were pelleted by centrifugation, washed once with PBS, and resuspended in fresh medium containing 5 μM CellROX Deep Red. Laser treatment was immediately applied following dye addition. Subsequently, 1 mM SYTOX Blue Dead Cell Stain (Thermo Fisher Scientific) was added to each sample and incubated for 30 min at 37°C to distinguish live and dead cells.

Samples were then analyzed by flow cytometry (CytoFLEX, Beckman Coulter) or imaged by confocal microscopy. ROS generation was evaluated based on the intensity of deep red fluorescence in CellROX-positive cells, while SYTOX Blue was used concurrently to assess membrane integrity. No-laser control was included to determine baseline ROS levels in the absence of photothermal stimulation.

TEM sample preparation and imaging

HeLa cells were cultured on acid-cleaned glass coverslips (BRAND, 20 mm × 20 mm square) for TEM analysis. Coverslips were pretreated by overnight immersion in concentrated sulfuric acid (ACS reagent, 95.0 to 98.0%), followed by extensive rinsing with RNase-free water and thorough drying. To ensure stable adhesion of the coverslip during cell seeding, 10 μl of PBS was first applied to the center of a 12-well plate (VWR, Avantor, USA), and the coverslip was placed onto the droplet. This improved surface wetting and prevented displacement during medium addition. A 100-μl cell suspension containing 4 × 10⁴ HeLa cells was then gently dispensed onto the coverslip. After a 30-min settling period to allow cell sedimentation onto the coverslip, 1 ml of fresh culture medium was slowly added along the inner wall of the well. Cells were incubated for 24 hours before treatment. Following initial attachment, the culture medium was replaced with medium containing 10 μM sunitinib for 24 hours. After incubation, the medium was removed and replaced with fresh medium prior to laser treatment. Using a microscope integrated with the nanosecond pulsed laser system, the coverslip region was identified and subsequently irradiated at specified fluences (0.3 or 0.5 J cm⁻²). The cells on coverslips were fixed using a two-step primary fixation protocol. First, culture medium was gently replaced with a 1:1 mixture of fresh medium and primary fixative solution [4% formaldehyde (EM-grade), 2.5% glutaraldehyde (EM-grade) in 0.1 M cacodylate buffer], followed by gentle shaking for 10 min at room temperature. This was then replaced by 100% fixative solution, and cells were further fixed for 1 hour at room temperature. After fixation, samples were washed three times for 10 min each at 4°C with 0.1 M cacodylate buffer. Postfixation was performed in reduced 1% osmium tetroxide solution {0.132 g of K₃[Fe(CN)₆] + 2 ml of 4% OsO₄ + 6 ml of 0.1 M cacodylate buffer} for 1 hour on ice with gentle shaking. Samples were subsequently washed three times in distilled water (3 × 10 min, 4°C, shaking) and then stained en bloc in 1% uranyl acetate at 4°C under light-protected conditions. Dehydration was carried out at room temperature with sequential ethanol solutions (15, 30, 50, 70, 95, and two changes of 100% ethanol), each for 15 min with gentle shaking. Resin infiltration was performed at room temperature using Spurr’s resin mixed with 100% ethanol in increasing ratios (1:3, 2:3, and 3× 3:3), with each step lasting at least 8 hours under continuous shaking. For embedding, coverslips were gently blotted to remove excess resin and inverted onto freshly filled embedding molds containing Spurr, with the cell side facing downward. Samples were polymerized at 69°C for ∼16 hours. Ultrathin sections (∼70 nm) were obtained using a Leica EM UC6 ultramicrotome. Subsequently, the sections were poststained with 1% uranyl acetate for 40 min and 3% lead citrate for 10 min using a Leica EM AC20 automatic stainer. Images were acquired using a JEOL JEM-1400Plus TEM operating at an acceleration voltage of 80 kV. Morphological changes were assessed with a focus on lysosomal structure and subcellular damage induced by laser treatment. Lysosome diameters were quantified across different treatment groups.

Statistics and reproducibility

Experiments were performed in biological triplicate (N = 3), each including at least three technical replicates (n = 3), and results are presented as means ± standard error of the mean (SEM), unless stated otherwise. Statistical analysis was performed using the ninth version of the GraphPad Prism software. One-way analysis of variance (ANOVA) with Tukey correction was applied to compare multiple conditions.

Supplementary Materials

The PDF file includes:

Supplementary Text

Legends for movies S1 to S10

Figs. S1 to S16

Other Supplementary Material for this manuscript includes the following:

Movies S1 to S10