Integrin targeted photodynamic therapy in patient‐derived glioblastoma spheroids

Abstract

Glioblastoma multiforme (GBM) is the most aggressive primary brain tumor, with a median overall survival of 14.6 months. GBM is incurable because of its invasive growth. These local invasive cells, most significantly glioblastoma stem cells (GSCs), when left behind, resist standard treatment, and cause almost all recurrences. However, the treatment of these infiltrative margins remains a significant challenge, as there are currently no options to reach these margins safely. Photodynamic therapy (PDT) shows promise as localized treatment option using light‐activated compounds that target tumor cells and that generate reactive oxygen species (ROS) to destroy them. Far red light, combined with silicon phthalocyanines, could penetrate deeper making it more effective for reaching cancer cells in the tumor margin without compromise of healthy brain. In this study, we used patient‐derived GBM spheroids in vitro as a preclinical model to evaluate a new dual‐cRGDfK‐silicon phthalocyanine conjugate targeting integrin αvβ3, a protein expressed by GBM cells and vasculature. Targeted PDT was efficient in killing GSC spheroids, showing that the combination of far‐red light with more precise targeting can reach the type of cells found in the invasive margin, using silicon phthalocyanine as the photosensitizer.

A novel targeted photodynamic therapy (PDT) agent, SiPc‐2c(RGDfK), was used to treat patient‐derived glioma stem cell spheroids expressing varying levels of the V3 integrin. The quantitative measurement of cytotoxicity following treatment demonstrated the therapeutic efficacy of SiPc‐2c(RGDfK) on V3‐expressing single spheroids, effectively recapitulating the heterogeneity in patient responses.

Abbreviations

Δ RFU

Variation in Fluorescence Signal

φ Δ

Oxygen Quantum Yield

5‐ALA

5 Aminolevulinic acid

CSCs

Cancer Stem Cells

ECM

Extracellular Matrix

FBS

Fetal Bovine Serum

GBM

Glioblastoma Multiforme

GSCs

Glioblastoma Stem Cells

LDH

Lactate Dehydrogenase

NS

Neurosphere

PDT

Photodynamic Therapy

PpIX

Protoporphyrin IX

PS

Photosensitizer

ROS

Reactive oxygen species

SiPc

Silicon Phthalocyanine

SiPc‐2(RADfK)

Dual cRADfK Silicon Phthalocyanine Conjugate

SiPc‐2(RGDfK)

Dual cRGDfK Silicon Phthalocyanine Conjugate

SOSGR

Singlet Oxygen Sensor Green Reagent

INTRODUCTION

Glioblastoma multiforme (GBM) is the most aggressive primary brain tumor with a particularly poor prognosis. Despite progress in surgical techniques and localized therapies, the median survival of patients with GBM remains only 14–17 months. ¹ Maximizing the safe removal of GBM is a complex challenge because of the infiltrative nature of the disease. GBM is a heterogeneous disease, and the presence of cancer stem cells (CSCs) stimulates a rapid tumor growth which hampers an effective response to traditional treatments. ² The cancer stem cell theory suggests that small groups of CSCs are crucial in driving tumor expansion, exhibiting chemo/radioresistance and causing relapses. ³ When cancer cells are not fully removed during surgery, glioma stem‐like cells (GSCs), spread through the cerebrospinal fluid complicating treatment further. Indeed, their ability to self‐renew and transform into different cell types positions GSCs at the heart of resistance and recurrence. ⁴ , ⁵ Interestingly, GSCs, isolated from solid tumors, have been demonstrated to self‐renew and form three‐dimensional neurospheres when grown in vitro. Spheroids composed by such GSCs, which are known to functionally recapitulate the heterogeneity of the original tumor, can be used to study and develop novel treatments. ⁶

Integrins, especially integrin αvβ3, are significant contributors to GBM progression, as overexpression of αvβ3 integrin is observed in tumor cells and their connected vasculature, facilitating tumorigenesis and angiogenesis. ⁷ The expression of the αvβ3 integrin is notably prevalent in astrocytic tumors, whereas a weaker expression has been observed in oligodendrogliomas. In general, the upregulation of this integrin in GBM, compared to the levels observed in healthy tissue, is well established, and its presence is closely associated with tumor malignancy. Notably, a pronounced overexpression of αvβ3 is observed at the tumor periphery, whereas healthy brain tissue does not express integrin αvβ3. ⁸ Targeting the integrin αvβ3 is therefore considered a promising strategy to enhance surgical outcomes by addressing residual cells in the tumor margin ⁹ thereby improving the specificity and effectiveness of cancer treatment. ¹⁰ , ¹¹

Photodynamic therapy (PDT) relies on induction of cell death by production of reactive oxygen species generated after excitation of the photosensitizer and subsequent transfer of energy to molecular oxygen. This process is initiated by a photosensitizer that specifically accumulates in tumor tissues and blood vessels. ¹² A well‐known class of photosensitizers is represented by phthalocyanines, including silicon phthalocyanine (SiPc), which share structural similarities with porphyrins, such as protoporphyrin IX (PpIX), but feature an expanded macrocyclic ring system, enabling absorption of longer wavelengths of light and thereby deeper tissue penetration. ¹³

To date, the only photosensitizer tested for glioblastoma treatment is protoporphyrin IX induced 5‐ALA which exploits the altered metabolism of cancer cells. ¹⁴ , ¹⁵ ALA is metabolized into protoporphyrin IX (PpIX) at a higher rate in tumors, facilitating its use in fluorescence‐guided surgery or photodynamic therapy to selectively destroy cancer cells. These complementary approaches aim to improve treatment efficacy while minimizing damage to healthy tissues. Unlike in dermatology, where laser light is applied locally and directly to the lesion, various approaches have been tested for brain treatments, including fiber optic diffusers, optical fibers, cavitation balloons, and interstitial applications. ¹⁶ However, 5‐ALA has limited targeting capacities and poor spectroscopic features (poor absorption/emission and activation by red‐shifted light). For these reasons, better photosensitizing molecules (PS) along with improved light‐delivery strategies need to be developed to treat GBM.

Silicon phthalocyanines have demonstrated potent photosensitizing activity in (pre)clinical studies. ¹⁷ In fact, photosensitizers that absorb light in the far red spectrum (up to 700 nm) offer significant advantages, as they can penetrate deeper into tissues while minimizing light scattering and avoiding major interference from hemoglobin and water, ¹⁸ allowing for a reach of approximately 1–1.5 cm in depth. ¹⁹ , ²⁰ Consequently, far red PDT shows great potential for targeting tumor cells beyond the resection cavity, effectively addressing the infiltrative margin.

In the case of GBM, around 80% of tumor recurrences happen within 2 cm of the boundary of the surgical resection. ²¹ This proximity to the resection cavity makes intra‐ or postoperative PDT a compelling option for targeting the residual cancer cells that prove a challenge to visualize and remove during surgery. The current study focuses on optimization of targeted PDT by developing a dual‐cRGDfK‐silicon phthalocyanine conjugate (SiPc‐2c(RGDfK)) and by testing its efficacy in patient‐derived GSC spheroids as representative models of GBM. ²²

MATERIALS AND METHODS

Characterization of the photosensitizer

Synthesis route and yield: see Supporting Information. Optical characterization: The absorbance spectra of the fluorophore SiPc‐2c(RGDfK) and of the fluorophore SiPc‐2c(RADfK) were acquired using a plate spectrofluorometer (Spectramax, molecular devices, San Jose, CA, USA) in a 100 μM concentration in either pure DMSO (100%), 5% DMSO in PBS or 5% in Medium (DMEM‐F12) at room temperature. Photophysical characterization: Photo‐oxidation of Singlet Oxygen Sensor Green Reagent (SOSGR) (Molecular Probes, NL) was used to determine the singlet oxygen formation by the PS conjugates SiPc‐2c(RGDfK) and SiPc‐2c(RADfK). This reagent was used because of its high selectivity for singlet oxygen and low sensitivity to hydroxyl radicals/superoxides. Stock solution of SOSGR (5 mM) was prepared in methanol as prescribed. The final concentration of SOSGR within the PS‐conjugate solutions were 10 mM. The stock solution of PS‐conjugate was dissolved in 100% DMSO. The final concentration of PS‐conjugate was 5 μM in 5% DMSO in PBS. Data were acquired in quartz cuvettes in aerobic conditions using stirring under uniform laser illumination at λirr = 662 nm (Spectra Physics, NL) at a measured fluence rate at 20 mW/cm². At intervals, the cuvette was removed from the magnetic stirrer and absorption and fluorescence emission spectrum (λexc = 475 nm) were acquired. The singlet oxygen quantum yields (ΦΔ) of conjugates in DMSO were determined in duplicate by a relative method with a ΦΔ = 0.52 of methylene blue in water.

In vitro experiments

Cell culture

Glioblastoma stem cells (GSCs) were cultured as described previously. ²³ In short, the GS359 and GS401 cells were cultured in DMEM‐F12 medium (Gibco™, Thermo Fisher Scientific, Catalog number 11320033) supplemented with penicillin/streptomycin (1%) (Gibco™, Thermo Fisher Scientific, Catalog number 15140122), fetal bovine serum (FBS, 10%) (Gibco™, Thermo Fisher Scientific, Catalog number 10270106), B27 supplement (2%) (Gibco™, Thermo Fisher Scientific, Catalog number 17504044), basic fibroblast growth factor (bFGF, 10–20 ng/mL) (Recombinant Human FGF‐basic, PeproTech, Catalog number 100‐18B), epidermal growth factor (EGF, 10–20 ng/mL) (Recombinant Human EGF, PeproTech, Catalog number AF‐100‐15), and heparin (5 μg/mL) (Heparin Sodium Salt, Sigma‐Aldrich, Catalog number H3149). Culture maintenance involved regular medium changes and incubation under optimal conditions of temperature and CO₂ concentration. After washing the cells with PBS, they were detached using Accutase (Sigma‐Aldrich, Catalog number A6964). To prepare adherent cells, the cell culture plates were coated with Cultrex ECM (Trevigen, MD, USA) allowing attachment and monolayer formation for in vitro analysis. A 1:100 dilution of the ECM solution in NS medium was then applied to the flask in the appropriate volume. The flask was then incubated at 37°C for 30 min. Coated flasks could be stored for up to 7 days at 37°C or for up to 3 weeks at 4°C. Regular monitoring of cell morphology and viability ensured key traits were preserved, maintaining experimental reproducibility. These cell lines are selected because they are a population of stem cells representative of the pathology studied, for their characteristics of autorenovation, tumorigenicity, differentiation capacity, and heterogeneity and microenvironmental interactions.

Determination of the target expression (integrin αVβ3)

In‐Cell Western™ assays were conducted according to the standard protocols provided by LI‐COR Bio, Lincoln, Nebraska, USA. In short, 1 × 10⁴ cells were seeded in ECM‐coated 96‐well plates. One day after seeding, cells were fixed, permeabilized, and exposed to primary (AbI LM609‐Abcam) and secondary (IRDye® 800CW Goat antimouse IgG) antibody. The fluorescence signals were measured at the 700 and 800 nm channel using the LICOR Odyssey® Imager and quantified using ImageStudio.

Staining of GSC cells in adhesion

2 × 10⁴ cells per well of cell line GS401 and GS359 were grown in adhesion in ECM‐coated chamber slides (IBIDI u‐slide eight well glass bottom slides), the next day the fluorophore SiPc‐2 c(RGDfK) was added to cells at a concentration of 500 nM (or control, regular medium) for 6 h. After washing with PBS, cells were fixed, stained with DAPI and imaged on a confocal microscope (LEICA SP5) using the 405 and 633 nm laser (40×).

Affinity assay of the photosensitizer with competitors

Cells (GS401) were seeded at a density of 1 × 10⁴ cells per well on day 0. On day 1, the cells were incubated with 100 μL of SiPc‐2 c(RGDfK) at various concentrations (1–500 nM), in the presence of the competitor peptide cRGDfK at 50 μM, for 2 h at 37°C. After the incubation, the treatment was removed, and the cells were washed once with PBS. Fluorescence was measured at 700 nm using an Odyssey scanner to assess drug uptake and competition.

Spheroid formation

GSCs were plated (day 0) in 96‐well transparent low attachment round bottom plates (Bio Float F202003) at a density of 1 × 10³ cells/well. Spheroid growth was monitored under a microscope until spheroids reached a largely homogeneous size with diameters of an average of 200 μm by day 2 for each cell line.

Uptake evaluation in spheroids

Spheroids were formed as described above. Spheroids were then treated with a 5% DMSO solution of SiPc‐2(cRGDfK) or SiPc‐2(cRADfK) prepared in NS medium (500 nM), with 70 μL added per well. Treatments were administered for 30 min, 2 h, and 6 h. After the incubation period, the cells were washed with PBS to remove any unbound SiPc‐2(cRGDfK). Fluorescence intensity in PBS was measured using an Odyssey scanner, utilizing the 700 nm channel. Fluorescence data were subsequently analyzed to quantify uptake across different concentrations and cell lines.

Photodynamic therapy on GS Spheroids

In this study, 1 × 10³ GS359 and GS401 cells, with passage numbers between 8 and 12, were seeded in a low attachment clear round bottom plate (BioFloat F202003). After 48 h incubation, spheroids were treated with either 500 nM of targeted photosensitizer, SiPc‐2c(RGDfK), or SiPc‐2c(RADfK) (each 1% DMSO in NS medium). After 2 h and 6 h of incubation with the photosensitizer at 37°C, the medium was replaced and photodynamic therapy was performed at 20 mW × cm⁻² to a dose of 20 J × cm⁻² using a 690 nm wavelength laser on an orbiting plate shaker. After PDT, positive control wells were incubated with Triton 0.2% (for 24 h) in medium, whereas negative control wells were kept in the same medium. A control plate, under the same conditions but not exposed to the laser, was included in the experiments. At 24 h after photodynamic therapy, the LDH assay (LDH Glo Promega J2380) was performed to quantify the cytotoxicity of the treatment. ²⁴ Bioluminescence measurements of collected medium were taken using the GloMax Discovery (Promega). The treated spheroids were concomitantly imaged under the microscope to study changes in morphology.

RESULTS

Synthesis of drugs and optical and photophysical profiles

The chemical structure of the molecule studied in this work consists of a planar aromatic macrocyclic structure made up of four isoindole units, with a central silicon atom (Si) coordinated at the core. The silicon atom is bonded to axial ligands, which contribute to the stability of the structure. Here we incorporate an integrin targeting vector in both axial positions of SiPc (Figure 1).

FIGURE 1.

Chemical structure of compound SiPc‐2c(RGDfK). The silicon phthalocyanine was linked, through axial ligands, with two cyclic peptides, cRGDfK (Arg–Gly–Asp–D–Phe–Lys).

The synthetic cyclic RGDfK (Arg–Gly–Asp–D–Phe–Lys) peptide and cyclic RADfK (Arg–Ala–Asp–D–Phe–Lys) negative control peptide have been used as targeting vectors; their conjugation to SiPc is described in Supporting Information (Figures S1 and S2). The cyclic nature of the RGD peptide is crucial for binding integrin receptors, particularly αvβ3, which are overexpressed in tumor and angiogenic cells.

UV–VIS absorption spectrum of SiPc‐2c(RGDfK) in DMSO shows a characteristic peak at 676 nm, typical of phthalocyanines (Figure 2). This peak corresponds to the Q‐band absorption, reflecting the extended π‐conjugation in its macrocyclic structure. ²⁵ The Q‐band peak shifts to 720 nm when J‐aggregates of the molecule form, predominantly in aqueous solutions. However, the fluorescent spectrum collected in cells after SiPc‐2(cRGDfK) uptake suggests that the monomeric form is bound to cells (see Figure S3).

FIGURE 2.

Absorption spectroscopy profiles of SiPc‐2c(RGDfK) and SiPc‐2c(RADfK) in pure DMSO at the final concentration of 100 μM.

The determined ΦΔ for SiPc‐2c(RGDfK) and SiPc‐2c(RADfK) were 0.54 and 0.60, respectively. The free SiPc in DMSO has a quantum yield between 0.6 and 0.7. However, the conjugates show a slightly lower yield, likely due to self‐quenching of singlet oxygen by the conjugate itself.

Specificity of targeted compound

For the GSC cell lines, integrin αVβ3 expression was retrieved from previous RNA sequencing data. GS401 was determined to have a low gene expression of integrin αVβ3, whereas GS359 was found to have a high expression. ²⁶ In cell western of the GSC cell lines revealed that GS359 have higher expression compared to GS401 (Figure 3A).

FIGURE 3.

Characterization of integrin expression in GSCs, visualization of drug internalization, and competition between drug and peptide alone. (A) Expression of integrin αVβ3 evaluated in GSCs through In Cell Western Assay shows significantly higher integrin expression in GS359. (B) Confocal microscopy images of cell lines GS359 and GS401 at 40× (405, 633 nm laser). DAPI and far‐red fluorescence imaging demonstrate bright signal in the cytosol surroundings of GS359 cells after 6 h incubation of targeted compound, SiPc‐2c(RGDfK). The signal is considerably brighter in the GS359 cell line than in the GS401 cell line. (C) ΔRFU in GS359 adherent cells measured based on the signal of the photosensitizer incubated alone and in competition with the cRGDfK peptide at 50 μM concentration.

GS359 cells showed higher fluorescent signal compared to GS401 cells after specific internalization of the targeted drug SiPc‐2c(RGDfK) (Figure 3B) confirming difference in relative expression of the target.

In the competitive binding assay performed on GSC cell lines, the introduction of an unconjugated cRGDfK peptide as a competitor, results in receptor blockade, markedly decreasing the internalization of the probe in adherent cells (Figure 3C). Variation in fluorescence signal (ΔRFU) observed between the PS alone and PS in competition assay confirms specific binding of the peptide.

Once the cells are grown, they spontaneously form spheroids, which show different sizes according to plate density and specific cell line. The results demonstrate that spheroids reach an ideal size at day 2 after seeding, when the spheroid is approximately 200 μm in diameter (Figure 4A). The internalization of the selective probe is appreciated over time, both in spheroids (Figure 4B) and in the adherent cells, following coating of the plate (see Figure S4). Our internalization studies on spheroids revealed higher specific drug uptake in GS359 cells, which overexpress the target integrin, when compared to GS401 cells. Internalization of the compound is also appreciable in fluorescence confocal images (see Figure 4C and Movie S1).

FIGURE 4.

Spheroid characterization and drug uptake. (A) Diameter of the spheroids shows growth from the day of seeding (i.e., day 0) up until day 2 to obtain approximately 200 μm in diameter. (B) Normalized fluorescence signals, captured with a LICOR Odyssey system using a 700 nm emission filter, show changes over time following the incubation of spheroids with SiPc‐2c(RGDfK) or SiPc‐2c(RADfK). The two‐way ANOVA revealed significant differences between groups and conditions. Notably, at 30 min incubation, significant differences were observed only between cells treated with SiPc‐2(cRGDfK): GS359 vs. GS401 (p < 0.0001). After 2 h and 6 h of incubation with SiPc‐2c(RGDfK), a significant difference is observed between the two cell lines (p < 0.0001). (C) Confocal imaging of spheroids, performed using the Opera Phenix Plus system after a 2 h incubation with SiPc‐2 c(RGDfK), shows where the probe is located (red signal), while the nuclei are stained with a blue dye.

Photodynamic therapy on spheroids

After photodynamic treatment of the integrin expressing spheroids, cells showed morphological changes under the microscope (Figure 5A). Spheroids derived from the GS 359 cell line that were treated with the SiPc‐2c(RGDfK) appeared visibly disintegrated, with photosensitizer aggregates around the spheroid, whereas the SiPc‐2c(RADfK)‐treated spheroids as well as the negative control spheroids remained largely similar in morphology, appearing vital and compact. This spheroid disintegration is seen to a considerably lesser extent in the GS401 cell line. The response to the PDT treatment was quantified through LDH assay, which indicated increased bioluminescence signals, reflecting LDH release (and thereby cellular damage) (see Figure S5). At 2 h, the targeted molecule, SiPc‐2c(RGDfK), showed a 34% and 41% higher cytotoxicity than the SiPc‐2c(RADfK), in the GS359 and GS401 cell lines, respectively. At 6 h, however, the targeting effect was less pronounced, here the specific killing of targeted PDT was maintained only in GS359 spheroids derived from the high integrin expressing cell line (Figure 5B).

FIGURE 5.

Microscopic imaging (pre‐ and post‐PDT) and LDH release of photodynamic therapy on patient‐derived gliomaspheres. (A) Microscopy of GS359 and GS401 spheroids before PDT (first column of each cell line) and 24 h after PDT. Spheroids underwent PDT after 2 h and 6 h of drug incubation with SiPc‐2c(RGDfK) or SiPc‐2c(RADfK). (B) Results of the PDT treatments show the cytotoxic effect quantified with LDH assay in GS359 and GS401 after 2 h and 6 h of incubation with SiPc‐2c(RGDfK) and SiPc‐2c(RADfK) compared to maximum LDH release after triton treatment (***p < 0.001, *p < 0.05).

DISCUSSION

In this study, we demonstrated the treatment efficacy of PDT in three‐dimensional patient‐derived GBM models using a novel photosensitizer that specifically targets integrin αVβ3.

Unlike monolayer cell cultures, which do not mimic the complexity of in vivo conditions, GSC spheroids offer a more accurate three‐dimensional representation of the tumor biology. This three‐dimensional setup enhances the relevance and reliability of experimental results by better reflecting cell‐to‐cell interactions. GSC spheroids have proven to be valuable as in vitro cancer models, particularly because of their ability to reveal differential integrin expression, which correlates with how a probe is internalized in vivo. ²⁷

We found the target expression of GS359 spheroids to be three times higher than the measured expression in spheroids of GS401. According to this, the PS internalization over time in the spheroids varied depending on the expression of the target. The photodynamic activity observed at 2 h mainly in GBM cell spheroids expressing the target confirms the relationship between target, internalization, and treatment efficacy.

The elevated expression of αVβ3 in GSCs promotes tumor angiogenesis and contributes to the aggressive behavior of GBM. Both αVβ3 and αVβ5 integrins are highly expressed not only in glioma cells but also in the surrounding tumor vasculature, playing key roles in GBM progression and (neo)angiogenesis. Targeting the tumor vasculature with PDT deprives tumor cells of necessary nutrients and oxygen while also increasing vascular permeability, which improves the delivery of subsequent therapeutic agents. ²⁸ This dual effect, direct cytotoxicity and vascular disruption, make PDT a promising treatment. ²⁹ To better study this effect, future research may benefit from the development of vascularized organoid models, incorporating endothelial cells to replicate blood vessels. Organoids can grow endothelial cells and form blood vessels using co‐culture techniques that include both endothelial cells and angiogenic factors like VEGF. These methods mimic the (micro)‐environment of the tumor, allowing vascular networks to form. Other approaches, such us implanting organoids in tissues with existing blood vessels or using microfluidic systems, can encourage integration of blood vessels from the host or stimulate the growth of new vessels within the organoid itself. These vascularized organoids are particularly useful for studying (peri)tumoral blood vessel formation and testing treatments that target tumor blood supply. Such models would enable a more precise investigation of how treatments such as PDT impact both tumor cells and vasculature, offering critical insights for improving therapeutic outcomes. ³⁰ , ³¹

By combining a silicon phthalocyanine with cRGDfK, we aimed to improve effectiveness of PDT through enabling the targeted delivery of the photosensitizer to the tumor. The bis‐axial SiPc‐RGD conjugate was chosen to minimize aggregation through steric hindrance, enhancing water solubility and overcoming a key limitation of phthalocyanines for in vivo applications. This improvement is attributed to functionalization rather than increased cellular uptake, as the distance between the two RGD peptides likely prevents simultaneous binding to dual receptor sites. Although not directly demonstrated, the conjugate is expected to primarily interact with a single receptor, ensuring better solubility under physiological conditions. SiPcs have demonstrated potent photosensitizing activity in both preclinical and clinical studies. ³²

The singlet oxygen quantum yields (ΦΔ) of the SiPc conjugates were very similar and slightly lower than the unconjugated SiPc. Previous studies have also documented variations in singlet oxygen quantum yields between free PS and conjugates. ³³ The differences in ΦΔ observed for the SiPc conjugates may be related to variations in amino acids to which the PS is conjugated, thus affecting the environment surrounding PS molecules. Although the efficient generation of singlet oxygen is an important factor in the induction of phototoxic effects, the localization of the PS, and variables affecting this localization, also significantly influence the overall therapeutic effect. Among the factors contributing to the therapeutic shortcomings of PDT in GBM is the concentration of oxygen in the brain. In fact, oxygenation levels drop from 5%–15% in healthy brain tissue to as low as 0.1% in necrotic brain tumors. ³⁴ In our research, the emphasis shifts from addressing the entire tumor mass to specifically targeting the remaining positive infiltrating tumor cells. This approach is particularly targeted at regions with more physiological concentrations of oxygen, which are not the bulk of the tumor but the surrounding area. The elevated oxygen levels render these areas susceptible to the effective application of photodynamic therapy, leaving surgery as the treatment of choice for eradicating the primary tumor mass. Treatment of a patient population positive for the integrin biomarker in the cells that remain in the positive margins, represents an emerging strategy. ³⁵ The combination of these two elements; maintaining normal oxygen levels in the treated area and ensuring that treatment selectively spares healthy tissues, forms the foundation for the therapeutic efficacy of this approach in defined conditions. ³⁶

In summary, targeted PDT represents a promising therapeutic strategy for local treatment of the GBM resection cavity. ³⁷ This is underlined by the targeting mechanisms that can be combined with local light delivery methods, which collectively render this therapeutic approach a prospective option to delay GBM recurrence. ³⁸ Future investigations into NIR‐PDT up to 750 nm could enhance tissue penetration even more and should therefore be assessed in representative three‐dimensional models to study to what extent deeper light penetration enhances tumor damage. Additionally, another challenge that should be addressed before moving toward clinical application should involve investigating heterogeneity of integrin expression at the infiltrative edge of primary and recurrent GBM. ³⁹ This approach could improve the precision and effectiveness of targeted PDT, ultimately enhancing therapeutic outcomes for these challenging tumor regions. Further research is needed to highlight the potential of the targeted therapy to improve treatment of GBM patients. Our study indicates that PDT is feasible and effective in patient‐derived GSCs when using an integrin targeted PDT agent.

AUTHOR CONTRIBUTIONS

M.R., L.M., and E.T. conceived and designed the study. M.R. and M.A. carried out the main experiments and contributed to data analysis and manuscript drafting. I.Q. was responsible for image processing starting from the 3D samples. R.S., E.T., F.B., and L.D.A. designed the drug and contributed to its chemical synthesis. H.S.B. and D.J.R. assisted with experimentation and data analysis in spectroscopy and PDT. E.T. and L.M. supervised the study. All authors engaged in discussions of the results and contributed to the final revision of the manuscript.

FUNDING INFORMATION

This work was partially funded by Dutch Cancer Society (KWF) (Grant No. 11924) and by the Italian Ministry of University and Research through the PNRR project SEELIFE (call M4/C2/L3.1.1). Additional funding was provided by the “Call for the award of mobility grants to Ph.D. candidates for research periods abroad – Year 2023” from the University of Torino, which supported Miriam Roberto's exchange at the Erasmus Medical Center in Rotterdam.

Supporting information

Appendix S1.

Movie S1.

Roberto M, Ali M, Que I, et al. Integrin targeted photodynamic therapy in patient‐derived glioblastoma spheroids. Photochem Photobiol. 2025;101:1241‐1250. doi: 10.1111/php.14097

DATA AVAILABILITY STATEMENT

All the data are included in the article and Supporting Information. Raw data are available at Zenodo repository. DOI 10.5281/zenodo.15112992