3.1 Px1-Px3 effectively reduce the metabolic activity of Rb cells only after irradiation

The Y79-GFP-luc Rb and the RPE-D407 retinal cell lines were incubated for 24 h with high concentrations (5 μM and 10 μM) of each PS to assess the potential cytotoxic effects. Without the light activation step, the metabolic activity remained above 87.7 ± 3.0% (p = 0.0049, 5 μM, Px3, RPE-D407) (Fig. 1D). Also, RPE-D407 cells treated with 10 μM of Px1 showed a statistically significant decrease to 91.1 ± 4.0% (p = 0.0482). In the case of Y79-GFP-luc, 5 μM Px1 resulted in a significant reduction in metabolic activity to 87.0 ± 2.1% (p = 0.0017); however, this decrease was followed by an increase to 106.6 ± 1.8% (p = 0.0135) when the cells were incubated with 10 μM Px1. These results indicate no biologically relevant cytotoxic effect of these molecules per se.

Px1, Px2, and Px3-based PDT effects on the metabolic activity of Y79-GFP-Luc Rb cells were evaluated via MTT (Fig. 1E) and Alamar Blue (Fig. 1F) assays. Fitting the MTT assay data to a dose–response curve yielded the IC50 values of 57.2 nM for Px1, 262.8 nM for Px2, and 267.0 nM for Px3. Px1 demonstrated a significantly lower IC50 than Px2 and Px3 for Y79-GFP-luc. In the RPE-D407 cell line, however, Px2 did not impact the metabolic activity of the cells, not allowing IC50 calculation. Px1 had an IC50 of 39.3 nM, and Px3 had an IC50 of 134.4 nM. The Alamar Blue assay was performed as a confirmatory assay in Y79-GFP-Luc Rb cells (Fig. 1F), showing that, for Px1, metabolic activity significantly decreased to 23.9 ± 4.6% at 250 nM (p < 0.0001), and reached near-total inhibition at 1000 nM, with only 1.5 ± 0.6% activity remaining (p = 0.0005). For Px2, the treatment resulted in modest but statistically significant reductions in metabolic activity across all tested concentrations, with 1000 nM reducing activity to 74.1 ± 7.2% (p = 0.0042). For Px3, no significant effect was observed at 50 nM, but metabolic activity significantly decreased to 83.8 ± 5.4% (p = 0.0114) at 250 nM, with a further reduction at higher concentrations to 78.5 ± 8.8% (1000 nM, p = 0.0039). Overall, these data suggest that Px1 is more potent in reducing cell viability than Px2 and Px3.

3.2 Px1-based PDT is the most potent approach in inducing cell death and cell cycle arrest

The cell death pathways in Y79-GFP-Luc Rb cells were assessed following PDT with Px1, Px2, and Px3 (Fig. 2A). For Px1, cell viability dropped from 80.7 ± 8.7% to 1.4 ± 0.5% (p = 0.0075) at a concentration of 500 nM and to 0.7 ± 0.8% (p ≤ 0.0001) at a concentration of 1000 nM, indicating complete cell elimination. This was accompanied by a significant increase in late apoptosis/necrosis, reaching 75.8 ± 12.8% (p ≤ 0.0001) at 500 nM and 88.7 ± 7.0% (p ≤ 0.0001) at 1000 nM compared to 7.3 ± 3.2% in the controls. For Px2, cell death occurred through late apoptosis, rising to 77.5 ± 2.1% (p ≤ 0.0001) and 24.25 ± 8.4% (p ≤ 0.0001) at 500 nM and 1000 nM, respectively. Early apoptosis also increased to 13.0 ± 3.7% (p = 0.0312) and 47.3 ± 13.4% (p = 0.0016), while necrosis remained relatively unchanged. For Px3, cell viability decreased to 79.5 ± 3.8% (p = 0.0350), 1.3 ± 0.5% (p = 0.0021), and 1.3 ± 1.3% (p ≤ 0.0001) at increasing concentrations. Late apoptosis rose to 79.5 ± 5.1% (p ≤ 0.0001) at 500 nM and 81.3 ± 7.6% (p ≤ 0.0001) at 1000 nM, with a slight increase in early apoptosis. Overall, Px1 and Px3 demonstrated the most pronounced impact on cell viability, resulting in near-total cell death at higher concentrations. Morphological analysis, as shown in Fig. 2B-E, complemented these findings, demonstrating an increase in apoptotic cells, as well as cell membrane blebbing and the formation of apoptotic bodies, alongside cells undergoing necrosis following PDT. Apoptosis was most pronounced in cells treated with Px1 and Px3, whereas cells treated with Px2 displayed fewer apoptotic features and maintained a more rounded morphology than untreated cells. Necrotic cells, characterised by plasma membrane disruption, were observed in Px1- and Px3-treated cells. These results suggest that Px1 and Px3 are more potent in inducing apoptosis and necrosis, while Px2 has a less pronounced effect on cell death morphology.

Fig. 2

Px1 is the most potent PS in inducing cell death and cell cycle arrest, emphasising apoptosis. A Cell viability and cell death pathways in Y79-GFP-luc cell line 24 h after PDT with Px1, Px2, and Px3; B–E Morphologic features of Y79-GFP-luc cells assessed by May-Grünwald-Giemsa staining in control conditions (B), and 24 h after PDT with Px1 (C), Px2 (D), and Px3 (E). The black arrows represent necrotic cells, characterised by the breakdown of the plasma membrane. The red arrows show blebbing (a characteristic process of apoptotic cells), and the green arrow shows the cells’ disassembly into apoptotic bodies; F Cell cycle of Y79-GFP-luc cells, 24 h after PDT with Px1, Px2, and Px3. Results are presented as the mean and standard error of the mean of n = 5 to 17 assays. Statistical significance is represented with * for p < 0.5, ** for p < 0.1, and **** for p < 0.0001

Cell cycle arrest was analysed using flow cytometry after PDT with Px1, Px2, and Px3, as shown in Fig. 2F. Overall, PDT did not significantly alter the distribution of cells in the Sub-G1, G0/G1, S, or G2/M phases compared to untreated controls, indicating that none of the PSs had a significant impact on the cell cycle progression. However, treatment with 500 nM and 1000 nM Px1 led to a significant increase in the Sub-G1 population, indicative of apoptosis, in Y79-GFP-Luc Rb cells. Specifically, 3.8 ± 2.2% of total cells were in the sub-G1 phase after treatment with 500 nM Px1, and 5.5 ± 3.3% after treatment with 1000 nM Px1, compared to 0.5 ± 1.5% in the control group.

3.3 PX1-based PDT preserves retinal structure and function in eyes with small Rb tumours

In vivo studies were conducted to assess the efficacy of Px1, selected for its performance in the in vitro studies, including a significantly lower IC50 associated with no cytotoxicity in the absence of irradiation. OCT was used to evaluate the retinal structure in eyes with or without Rb, treated or not, at two timepoints (pre-PDT and post-PDT).

Tumour growth in the eyes with Rb (Group 1—Left eye) was observed in the vitreous cavity without affecting the retinal structure at the initial stages (pre-PDT), as seen in Fig. 3. Nevertheless, tumour size increased over time, precluding OCT imaging at the final examination (post-PDT). In the LT-PDT subgroup (Group 2—Left eye) at the pre-PDT stage, tumour cells were visible as small light blueish spots in the vitreous, with slight growth near the nerve fibre layer (NFL). At the post-PDT stage, tumour growth obstructed OCT imaging, meaning that the retina could not be assessed in vivo. By contrast, the ST-PDT subgroup (Group 2—Left eye) exhibited small tumour clusters at the pre-PDT stage with no significant impact on the retina structure. At the post-PDT stage, slight tumour growth was observed, but it still allowed the light to pass through, allowing OCT imaging and suggesting no major alterations in retinal structure. Moreover, to exclude the possibility of retinal alterations caused by PDT per se, Supplementary Fig. 3 shows the eye fundus and retinal architecture of the animals’ left eyes without Rb that were treated with PDT (Group 3). This evidence indicates that PDT per se does not affect the retinal architecture. Additionally, the absence of alterations in retinal thickness, as shown in Supplementary Fig. 4, corroborates that Px1-based PDT is safe and that it does not induce changes in retinal structure.

Fig. 3

Impact of PDT on retinal structure in eyes with Rb. OCT was performed in both eyes, in which the left eye was inoculated with Rb, and the right eye was not. Representative images of the experimental groups 1 and 2 (including LT and ST) are presented. The qualitative analysis was performed both pre- and post-PDT. Both images of the eye fundus and OCT images are represented. Complex formed by nerve fibre layer, ganglion cell layer and inner plexiform layer (NFL + GCL + IPL), inner nuclear layer (INL), outer nuclear layer (ONL), inner and outer photoreceptor segments (IS/OS) layer. Scale bar: 50 μm

ERG was used to evaluate the retinal function and involved four tests: scotopic threshold response (STR), scotopic ERG, photopic ERG, and the Flicker test (Fig. 4). STR assesses retinal ganglion cell (RGC) function under dim light in a dark-adapted retina. The pSTR test primarily reflects RGC activity, while the nSTR also has contributions from amacrine cells. STR pre-inoculation responses revealed that each animal presented a distinct pre-PDT response. Hence, inter-subject comparisons were avoided, and the assessments at each time point were made using data from the right eye, which was not subjected to any intervention.

Fig. 4

Impact of PDT on the retinal function of eyes with Rb. ERGs were performed at pre-inoculation (Day 0), pre-PDT (Day 23), and post-PDT (Day 29). Analysis was performed in both eyes, in which the left eye (in red) was inoculated with Y79-GFP-luc cells, and the contralateral eye (in black) was considered the control. Retinal ganglion cells’ function was assessed by STR ERGs (0.000095 cd-s/m2). Scale bars: 25 μV per 100 ms. The function of rod photoreceptors was evaluated using Scotopic ERGs. Scale bars: 50 μV per 10 ms (9.49 cd-s/m2). Cone photoreceptor function was assessed by Photopic ERGs (9.49 cd-s/m2). Scale bars: 50 μV per 10 ms. The function of cone photoreceptors was assessed by Flicker ERGs (9.49 cd-s/m2). The graphs represent the qualitative assessment of Group 1 (n = 2), Group 2 LT-PDT (n = 1), and Group 2 ST-PDT (n = 2). Scale bars: 50 μV per 100 ms

At pre-PDT, both pSTR and nSTR amplitudes were reduced in the eye with Rb compared with the right eye (contralateral) in the Group 1 and LT-PDT subgroup (Group 2), indicating either tumour-induced damage to RGCs or that the tumour was blocking the passage of light stimuli to the retina, thus preventing proper stimulation of the RGCs. In the ST-PDT subgroup (Group 2), pSTR and nSTR amplitudes were similar between the two eyes, suggesting no impact on RGC function. At post-PDT, pSTR and nSTR amplitudes remained stable in Group 2, while they continued to decline in Group 1 and the LT-PDT subgroup.

The scotopic ERG evaluates rod-mediated vision, with the a-wave reflecting the function of rod photoreceptors and the b-wave reflecting the activity of bipolar cells. Oscillatory potentials also provide insight into the activity of amacrine cells. At pre-inoculation, all scotopic ERG responses were comparable between both eyes in all animals. At pre-PDT, the a- and b-waves amplitudes were reduced in Group 1 (left eye compared to the right contralateral eye) and LT-PDT subgroup (Group 2). The ST-PDT subgroup (Group 2) maintained similar a- and b-wave amplitudes between both eyes. At post-PDT, a further reduction in a- and b-wave amplitudes was observed in Group 1 and the LT-PDT subgroup (Group 2), indicating progressive rod and bipolar cell dysfunction. In contrast, the ST-PDT subgroup demonstrated stable a- and b-wave amplitudes, showing that rod function was preserved.

The photopic ERG assesses cone-mediated vision by measuring the a- and b-waves, with the a-wave reflecting the function of cone photoreceptors and the b-wave reflecting the activity of bipolar cells. At pre-inoculation, photopic ERG responses were similar in both eyes of all experimental groups. At pre-PDT, Group 1 and the LT-PDT subgroup (Group 2) showed reduced response amplitudes in the left eye, while the ST-PDT subgroup maintained similar responses. At post-PDT, the Group 1 and LT-PDT subgroup (Group 2) exhibited further declines in the left eye. However, in the ST-PDT subgroup (Group 2), cone function was preserved, with stable photopic ERG amplitudes.

The Flicker test evaluates cone function using rapid flashes of light. Each flash elicits a trough-to-peak response, which reflects cone activity. At pre-inoculation, minor differences in trough-to-peak amplitude were observed between eyes in Group 1 and the LT-PDT subgroup (Group 2), with the ST-PDT subgroup (Group 2) showing consistent responses. At pre-PDT, a decline in trough-to-peak amplitude was seen in Group 1 and the LT-PDT subgroup (Group 2), while the ST-PDT subgroup (Group 2) maintained similar amplitudes. At post-PDT, Group 1 and the LT-PDT subgroup (Group 2) showed further reductions in trough-to-peak amplitude. In contrast, the ST-PDT subgroup (Group 2) maintained consistent amplitudes.

Comparison of retinal function between both eyes in Group 3 (Supplementary Fig. 5), which included animals without retinoblastoma that subjected to PDT, revealed mostly consistent response amplitudes, suggesting that PDT per se does not affect neuronal activity. The minor differences observed in amplitude are not indicative of retinal damage, as a similar pattern was seen in Group 4, corresponding to the eyes without tumours that were not subjected to PDT.

These qualitative observations suggest that PDT using Px1 is not associated with an additional decline in retinal function in the ST-PDT subgroup (Group 2), as evidenced by relatively stable ERG responses across all tests.

In the LT-PDT subgroup (Group 2), however, PDT failed to prevent retinal dysfunction, as indicated by progressive declines in ERG amplitudes. RGCs, rods, and cones were all negatively affected by the presence of larger tumours, and PDT was unable to inhibit tumour growth significantly in those cases. These findings suggest that PDT with Px1 is more beneficial for treating small Rb tumours, where retinal function can be better preserved.

3.4 Px1-based PDT reduces tumour mass in the early stages of Rb

The IVIS Lumina system was employed to confirm the presence of tumour masses of varying sizes in the animals and to monitor their progression. Figure 5A illustrates tumour progression across the different study groups at pre- and post-PDT. Animals from Group 1, with Rb in the left eye, did not receive PDT treatment, leading to substantial tumour growth in all animals. Figure 5B shows the ratio of the bioluminescence signal intensity in the left eye of each group. The untreated left eyes of Group 1 exhibited the largest increase of bioluminescence signal (2.0 ± 0.8, n = 5), followed by the LT-PDT subgroup (Group 2) (1.5 ± 0.5, n = 4). The ST-PDT subgroup (Group 2) showed tumour shrinkage (0.4 ± 0.4, n = 6), indicating a more favourable response to PDT in early-stage tumours. For example, one animal in Group 1, which was not submitted to PDT, showed a 4.8-fold increase in tumour size between the pre-PDT and the first post-PDT observations, highlighting the aggressive progression of untreated Rb. Despite the intervention, animals in the LT-PDT subgroup (Group 2) experienced tumour growth, specifically, the tumour expanded 1.5-fold and 2.0-fold in the first and second post-PDT examinations, respectively. These findings indicate that PDT was ineffective in halting tumour progression in cases with more advanced tumours, suggesting a limitation of the treatment in late-stage disease. In contrast, the ST-PDT subgroup (Group 2), comprising animals with smaller, early-stage tumours, demonstrated better outcomes. In post-PDT, most animals exhibited a reduction in tumour size. Notably, two animals showed reductions in tumour size superior to 90% (specifically, 96% and 91%). After ensuring optimal care and adherence to animal welfare guidelines, a second imaging examination was performed on some animals a week later. The treated left eyes in Group 1 showed an average bioluminescence signal of 1.7 ± 0.5 (n = 2), indicating a slight decrease compared to the first post-PDT evaluation. However, this difference was not statistically significant, and the reduction in the number of animals evaluated may have contributed to this result.

Fig. 5

Bioluminescence signals showing tumour progression in animals of the control, LT-PDT and ST-PDT subgroups. A Each column of images is representative of each condition, and each line of images represents timepoints. The colour scale shows radiance (photons.s−1.cm−2.sr−1). B Mean variation of bioluminescence 2 days after PDT and 8 days after PDT in ST-PDT and LT-PDT. Bioluminescence is presented as the ratio of counts versus the counts of the previous time point

The LT-PDT subgroup (Group 2) showed an average bioluminescence signal of 2.0 ± 0.7 (n = 2), confirming the nil effect of PDT in advanced disease. Unfortunately, the ST-PDT subgroup (Group 2) showed an average bioluminescence signal of 1.2 ± 0.6 (n = 3), suggesting tumour control over the short term. This observation indicates that, although PDT can effectively reduce tumour size, there is a potential for tumour regrowth in the days following treatment. Such findings imply that a single PDT session may not be sufficient to eradicate all tumour cells, as regrowth was observed within one week of treatment.

In summary, PDT treatment was more effective in reducing tumour mass in animals with smaller, early-stage Rb compared to those with advanced tumours. The reduction in small tumour sizes was significant, with minimal tumour regrowth in most cases. However, a modest regrowth was observed one-week post-PDT, indicating that multiple PDT sessions may be required for long-term tumour control. These findings highlight the potential of PDT for treating early-stage Rb, while also emphasising the need for further research into optimising treatment protocols for advanced-stage tumours.

3.5 Px1-based PDT does not induce death by apoptosis in healthy retinal cells and arrests the proliferation of the tumour cells

Considering that our results indicated the potential efficacy of PDT treatment in reducing tumour mass in animals with smaller and early-stage Rb, compared to those with advanced tumours, immunohistochemistry was performed on retinal slices from the left eyes to evaluate both the therapeutic effect of PDT (group 1 vs group 2) and its safety (group 3 vs group 4), allowing the evaluation of the effects of Px1-based PDT on both Rb (putative therapeutic effects) and healthy retinal tissue (safety profile assessment).

Apoptotic cell death was assessed using Caspase-3 immunolabelling (Figs. 6A and B). No Caspase-3-positive cells were detected in the retinas of the four experimental groups, indicating that PDT does not induce apoptosis in healthy retinal cells (Fig. 6A). However, Caspase-3-positive cells were found in the tumour regions of both eyes with Rb (Group 1) and eyes with Rb treated with PDT (Group 2). No significant changes were detected regarding Caspase-3 immunoreactivity between these experimental groups, suggesting that PDT-induced tumour cell death (Fig. 6B) was not significant. Nevertheless, Ki67, a marker of cell proliferation, was used to assess tumour growth dynamics (Figs. 6C and D). No Ki67-positive cells were detected in healthy retinas (Group 4) or PDT-treated retinas (Group 3) (Fig. 6C). However, eyes with Rb tumours (Group 1) exhibited a high abundance of Ki67-positive cells, mirrored by increased Ki67 immunoreactivity, reflecting the aggressive proliferation. Notably, the Ki67-immunoreactivity was significantly reduced (3.2 × 10⁶ ± 7.1 × 104 AU) in the eyes with Rb treated with PDT (Group 2) compared to the eyes with Rb (Group 1; 5.9 × 10⁶ ± 8.4 × 104 AU; p = 0.0015) (Fig. 6D). This suggests that PDT treatment effectively arrested tumour cell proliferation.

Fig. 6

Px1-based PDT induces a decrease in Ki67-positive tumour cells. Representative images from immunochemistry using the cell death marker Caspase-3 in red (A, B) and the marker of proliferation Ki67 in red (C, D), with DAPI nuclei staining in blue (Y79-GFP-luc cells are observed in green). Retinal cryosections were obtained from Groups 1, 2, 3 and 4. Graphic representation of the immunoreactivity of Caspase 3-positive (B) and Ki67-positive (D) tumour cells found in the tumours of Groups 1 and 2. Results are presented as the mean and standard error of n = 3 to 4 animals. Scale bar = 100 μm. Statistical significance is represented with ** for p < 0.1

3.6 Px1-based PDT has no significant impact on glial cells and neurons

Microglial activation and distribution were assessed using the Iba-1 antibody (Figs. 7A and B). Px1-based PDT did not induce significant changes in microglia number in healthy retinas (Groups 3 and 4) (Fig. 7C). No microglia were observed in the outer plexiform layer (OPL) across any experimental group. PDT-treated eyes (Group 3) showed a slight increase in the total number of microglial cells (all retinal layers) (20.0 ± 4.2 cells per field) compared to control (Group 4) eyes (14.9 ± 2.6 cells per field), although not statistically significant. This tendency to increase the microglia numbers was more pronounced in the ganglion cell layer and inner plexiform layer (GCL + IPL) of PDT-treated eyes (Group 3), while the inner nuclear layer (INL) showed no changes. In contrast, Rb-affected eyes (Groups 1 and 2) exhibited a marked increase in microglial cell number and notable morphological changes reflecting an inflammatory response triggered by the presence of the tumour (Fig. 7A). Microglial cells transformed from a ramified to an amoeboid shape in response to the tumour. Notably, infiltration of blood-derived macrophages may also contribute to the increase of Iba-1-positive cells observed in the retina.

Fig. 7

Px1-based PDT induces a slight increase in the total number of microglial and immune cells. Representative images from immunohistochemistry using the microglial/immune cell marker Iba-1 in red (A and B), the astrocyte/reactive Müller cell marker GFAP in red (C) and the retinal ganglion cells marker Brn3a in red (D and E), with DAPI nuclei staining in blue. Retinal cryosections were obtained from Groups 1, 2, 3 and 4. Graphic representation of the number of Iba-1 (B) or Brn3a cells (E) in Groups 3 and 4. Results are presented as the mean and standard error of n = 3 to 4 animals. Scale bar = 100 μm

Glial fibrillary acidic protein (GFAP) labelling was used to detect astrocytes and/or Müller cells, which are key components of the retinal glial network (Fig. 7C). Müller cells exhibit minimal GFAP expression in healthy retinas, but their reactivity increases in response to injury. In most animals, PDT did not induce changes in GFAP reactivity. Nevertheless, in some animals, a tendency for increased GFAP immunoreactivity was noted in PDT-treated eyes (Group 3) compared to control retinas (Group 4) (Fig. 7C). In eyes with Rb (Group 1), increased macroglial reactivity was noted, characterised by elevated GFAP expression, which may reflect astrogliosis and increased Müller cell reactivity indicative of reactive gliosis in response to the tumour. In eyes with Rb treated with PDT (Group 2), a decrease in GFAP immunoreactivity (Fig. 7C) was observed in some of the studied animals.

Brn3a immunolabelling (a well-established marker for RGCs) was used to assess the effect of PDT per se on the number of RGCs, as well as its putative therapeutic impact in retinas with tumours (Fig. 7D and E). No significant changes were observed in control retinas (Group 4) compared to PDT-treated retinas (Group 3), suggesting that PDT per se is safe and does not affect the number of RGCs (Fig. 7E). In Rb-affected eyes, both untreated (Group 1) and PDT-treated (Group 2), no Brn3a-positive cells were detected when the tumour was in contact with RGCs (data not shown). In the eyes with tumours located in the choroid/sclera, Brn3a-positive cells could be observed, although in decreased number compared to the Group 3 or Group 4 experimental groups (Fig. 7D).