Seamlessly Overcoming Biological Barriers with a Small Photosensitizer to Treat Metastatic Tumors with Photodynamic Therapy

Abstract

A breakthrough in the treatment of solid tumors requires primary tumor ablation and remission of metastasis. Photodynamic therapy (PDT) may meet these requirements, but current photosensitizers poorly infiltrate dense tumors and insufficiently stimulate the immune system. We introduce LUZ51, a small bacteriochlorin that features strong near‐infrared absorption, high photostability, moderate lipophilicity, amphiphilicity, and appropriate singlet oxygen quantum yield. LUZ51 has a fast cellular uptake via passive diffusion and EC50s of 5–15 nM range for a light dose of 1 J/cm². It seamlessly crosses biological barriers and accumulates in tumors. Mice with CT26 tumors cured in one treatment (0.15 mg/kg and 20 J/cm²) resisted tumor rechallenge. Treatment of mice with 5 mm orthotopic 4T1 tumors enabled cures, with the associated inhibition of metastasis. The blood half‐life is 82 min and no skin photosensitivity was detected. Small‐size LUZ51 enables tissue‐agnostic PDT, with strong local and systemic effects.

LUZ 51 is a relatively small and amphiphilic photosensitizer that accumulates in poorly‐vascularized tumors, is rapidly internalized by cancer cells, does not elicit skin photosensitivity and triggers robust immunological responses that reduce spontaneous metastasis.

Introduction

Defeating cancer will require combinations of a variety of technologies and pharmaceuticals. Solid tumors represent ∼90% of adult human cancers,^{[ 1 ]} and the pharmaceutical options for their management include chemotherapy, immunotherapy and, less often, photodynamic therapy (PDT). PDT combines the administration of a photosensitizer with the illumination of the tumor after a given drug‐to‐light interval (DLI).^{[ 2} , ^{3 ]} The photosensitizer absorbs light, transfers energy or electrons to molecular oxygen, and generates reactive oxygen species (ROS) in the illuminated volume. PDT is potentially, a tissue‐agnostic therapy because optical fibers enable the delivery of light to most of the body and tumors can be ablated in a variety of organs. Moreover, illumination of photosensitizers in appropriate subcellular locations generates oxidative stress that triggers immunogenic cell death (ICD), characterized by the presentation of tumor antigens in dying cells that activate an adaptive immune response.^{[ 4} , ^{5 ]} PDT has a direct effect in the tumor mass and a systemic effect mediated by the immune system. PDT can be relevant for the treatment of metastatic cancers, which are responsible for 90% of all cancer deaths.^{[ 6 ]}

The directionality of light and ICD make PDT very appealing to treat most cancers with low adverse drug reactions, but this therapy has yet to gain widespread acceptance. Photostable and phototoxic photosensitizers with intense absorption in the red/near‐infrared,^{[ 7 ]} where human tissues are more transparent, were developed and promising PDT clinical results were published.^{[ 5 ]} However, the ablation of large tumors requires photosensitizers that infiltrate all the tumor mass and remain phototoxic at low light intensities and low oxygen concentrations. Seamless crossing of tumor microenvironment (TME) and cell membranes imposes restrictions to the molecular structure of the photosensitizer. Rapid diffusion is favored by small size and cell uptake is enhanced by amphiphilicity.^{[ 8} , ⁹ , ^{10 ]} Photosensitizers need to have favorable absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties, which are difficult to predict from molecular structure. The “sweet spot” of favorable ADMET is populated by drugs with relatively low molecular masses and n‐octanol‐water partition coefficients P _OW from 100 to 10,000 range.^{[ 11 ]} Hence, the eventual adoption of PDT as the first‐line therapy for a wide range of solid cancers requires photosensitizers that have i) intense absorption in the near‐infrared, ii) high ROS quantum yields, iii) photostability, iv) low molecular weight, v) 2 ≤ log P _OW ≤ 4, and vi) amphiphilicity. These molecular properties are expected to enable vii) fast cell uptake and viii) high phototoxicity. Additionally, it is essential that the photosensitizers ix) have low dark phototoxicity, x) accumulate in tumor tissues, and xi) are rapidly cleared from the body to avoid prolonged skin photosensitivity.

The challenge of combining all these properties in one photosensitizer is matched by the challenge of delivering it to all cancer cells in a solid tumor. The enhanced permeation retention effect^{[ 12 ]} which acknowledges that nanostructures with sizes below 300 nm benefit from the large fenestration of tumor blood vessels to extravasate to the TME^{[ 13 ]} has been a main driver for the development of formulations in oncology. Tumor accumulation is expected for nanostructures larger than 20 nm, to ensure a long circulation time, and smaller than 100 nm, to minimize uptake by the macrophage system.^{[ 14 ]} Nevertheless, the accumulation of nanoparticles in tumors is not sufficient for curing most cancers.^{[ 15 ]} Moreover, cell hyperproliferation in solid tumors reduces vascular density, compresses blood and lymphatic vessels, increases interstitial fluid pressure, lowers oxygenation and, consequently, limits drug delivery.^{[ 16 ]} Incomplete drug distribution in solid tumors is often associated with treatment failure.^{[ 17 ]} A photosensitizer with properties (i)–(xi) mentioned above should be delivered in a formulation with nanostructure sizes between 20 and 100 nm, that disassemble in the blood or that easily extravasate from the tumor blood vessels and allow photosensitizer molecules to rapidly diffuse through the TME and be internalized by all cancer cells.

In the design of a photosensitizer with the desired properties, we made a significant departure from conventional wisdom. Rather than pursing the path of adding complexity to fulfil photosensitizer and formulation requirements, we simplified the system. We thus present a low molecular‐weight photosensitizer (MW = 595 Da) with intense absorption in the infrared (ε₇₃₃  = 1.0x10⁵ M⁻¹ cm⁻¹ in ethanol), appropriate singlet oxygen quantum yield (Φ_∆ = 0.39), low photodecomposition quantum yield (Φ_pd = 10⁻⁴), log P _OW = 2.9 and amphiphilicity, that is rapidly internalized by cancer cells and has excellent ADMET properties. It is phototoxic in low nanomolar concentrations at low light doses. In Pluronic 123 (P123) forming micelles with ∼80 nm in diameter, its tumor/peritumor ratio reaches 13 at 24 h post‐intravenous administration. Remarkably, 37.5% of mice with orthotopic 4T1 tumors, characterized by a dense extracellular matrix and a high immunosuppressive environment, were cured in one single treatment when the tumor diameter had reached 5 mm and metastasized. This photosensitizer has another important property, it (xii) triggers immunological responses that control metastases.

Results

Design and Characterization of a Small Bacteriochlorin for PDT

Electron‐withdrawing groups (e.g., F, Cl) in the phenyl rings of meso‐tetraphenyl bacteriochlorins stabilize the macrocycle against oxidation and make bacteriochlorins suitable for use in biological systems.^{[ 18} , ¹⁹ , ²⁰ , ²¹ , ^{22 ]}

We designed diphenyl bacteriochlorins with a carboxamide in one meso position to obtain small bacteriochlorins that are stable and amphiphilic (Figure 1a), in a quest to meet the properties desired for a PDT photosensitizer (Figure 1b). The synthesis of the carboxamide bacteriochlorin, named LUZ51 (MW = 595 Da) was only disclosed in a patent ^{[ 23 ]} and its characterization is presented in the (Figure S1 and S2).

Figure 1.

5‐Methylcarboxamide‐10,20‐bis(2,6‐difluorophenyl) bacteriochlorin, alias LUZ51. a) Molecular structure; b) desired properties; c) absorption (red) and fluorescence (blue) spectra in ethanol, where the fluorescence quantum yield is Φ_F = 0.16; d) singlet oxygen emission at 1270 nm with excitation at 355 nm of LUZ51 in ethanol (grey) or phenalenone in methanol (black), with matched absorbance.

LUZ51 shows the characteristic absorption band of bacteriochlorins at 733 nm with ε ₇₃₃ = 1.0x10⁵ M⁻¹ cm⁻¹ in ethanol, the fluorescence maximum is observed at 737 nm (Figure 1c) and the fluorescence quantum yield is Φ_F = 0.16. This and the photodecomposition quantum yield Φ_pd = 1.5x10⁻⁴, Φ_∆ = 0.39, and log P _OW = 2.9 make LUZ51 photophysical properties (Figure S3–S7 and Table S1) resemble those of redaporfin, a tetraphenylbacteriochlorin issued from our labs that recently completed phase I/II clinical evaluation with promising therapeutic outcomes.^{[ 20 ]} However, LUZ51 has half the molecular weight of redaporfin and its polar group in a meso position makes it more amphiphilic and with a close resemblance with the α₄ atropisomer of redaporfin, which has enhanced cell uptake.^{[ 9 ]} The carboxamide group is also responsible for lowering the P _OW of the bacteriochlorin to ∼800. Overall, LUZ51 meets the design criteria (i)–(vi). The biological properties (vii)–(xii) were evaluated as discussed below.

LUZ51 Passively Diffuses into Cells and Rapidly Accumulates in ER and GA Compartments

Time‐dependent uptake studies with 4T1, B16F10, and CT26 cancer cells, using flow cytometry, assessed the expected enhancement of cell internalization enabled by the small size and amphiphilicity of LUZ51. The highest levels of LUZ51 fluorescence were detected in 4T1 and CT26 cells, while B16F10 cells exhibited comparatively lower levels (Figure 2a–c). Remarkably, 1 h of incubation was enough to observe intense LUZ51 fluorescence from the cells, and detectable cell association required just 5 min of incubation (Figure 2d–f). The uptake of LUZ51 and redaporfin were compared under identical conditions, over from 2 to 24 h incubation periods, and showed that LUZ51 internalization was from 4‐ to 8‐fold higher than that of redaporfin (Figure S8 A). Compared to the α4 atropisomer (the redaporfin atropisomer with the highest cell uptake), LUZ51 was internalized by cancer cells significantly more rapidly, reaching substantially higher levels of cell accumulation between 1 and 8 h (Figure S8B).

Figure 2.

Cellular internalization of LUZ51. a–c) Cellular uptake of LUZ51 (1.85 µM) in 4T1, B16F10, and CT26 cells, evaluated by flow cytometry after varied period of incubations (4, 8, 16, and 24 h); d) cellular uptake of LUZ51 (1.85 µM) in 4T1 cells, evaluated by flow cytometry after short period of incubations (from 5 to 60 min); e) cellular uptake of LUZ51 (0.5 to 4 µM) in U2‐OS cells, evaluated by fluorescence microscopy after varied periods of incubation (1, 2, 4, 8, and 16 h); f) representative images of U2‐OS cells incubated with 0.5 to 4 µM of LUZ51 (red), from 1 to 16 h, followed by nucleus staining with DAPI (blue). Bars indicate the mean ± SEM of 2–3 independent experiments; in all the experiments the fluorescence signal from treated cells was normalized to the untreated cells.

Confocal microscopy of U2‐OS cells expressing green fluorescence protein (GFP) attached to calreticulin, GALT1 or SMAC, which are specific markers for the ER, GA, and mitochondria, respectively confirmed that LUZ51 was internalized, rather than merely associated with the cell membrane. LUZ51 showed a strong colocalization with GFP‐ER and GFP‐GALT1, and to a lesser extent with mitochondria (Figure 3a). No accumulation was observed in the nucleus or cytosol.

Figure 3.

LUZ51 is primarily internalized by cells through passive diffusion. a) Representative images of the colocalization of LUZ51 with markers of the ER, GA, and mitochondria in U2‐OS cells. b) Cell uptake of LUZ51 (0.44 µM) in 4T1 cells at 4 °C or 37 °C and at different timepoints. c, d) Cell uptake of LUZ51 (0.44 µM) and Bodipy12 (Ctr +), after a period of incubation of 2 h, upon ATP depletion. e,f) Cell uptake of LUZ51 (0.44 µM) and Bodipy 12 (Ctr +), after a period of incubation of 4 h, in the presence of inhibitors of clathrin‐mediated endocytosis, chloroquine (CQ), and chlorpromazine (CPZ). g,h) Cell uptake of LUZ51 (0.44 µM) and FITC‐albumin (Ctr +), after a period of incubation of 4 h, in the presence of filipin, an inhibitor of caveolin‐mediated endocytosis. Bars indicate the mean ± SEM of 2–3 independent experiments; the fluorescence signal from treated cells was normalized to the untreated cells. Statistical significance was evaluated using two‐way ANOVA * p < 0.05 ** p < 0.01 and **** p < 0.0001.

By design, we expected LUZ51 to have a fast cell uptake. Nevertheless, measurable cell internalization in just 5 min is intriguing. We investigated cell uptake mechanisms to understand the fast transport across the cell membrane. Temperature influences transport mechanisms that rely on the fluidity of the cellular membrane, such as endocytosis and passive diffusion.^{[ 24 ]} At 4 °C, the cellular uptake of LUZ51 was approximately 10 times lower than at 37 °C, but some internalization was detected at the lowest temperature (Figure 3b). LUZ51 fluorescence in cells increases rapidly over the first 3 h of incubation at 37 °C, and then more gradually. This temperature‐dependent influx suggests that internalization is influenced by the fluidity of the cell membrane (e.g., passive diffusion or endocytosis).

We distinguished between LUZ51 active and passive transport using 20 mM of 2‐deoxy‐D‐glucose (2‐DG), a glucose analogue that inhibits glycolysis, to deplete ATP (Figure S9). LUZ51 uptake did not change with partial ATP depletion (Figure 3c), although the uptake of Bodipy 12 (Figure S10), which is known to be internalized by energy‐dependent mechanisms,^{[ 25 ]} was significantly inhibited (Figure 3d). This suggests that LUZ51 internalization occurs through an energy‐independent mechanism.

To further exclude active transport mechanisms, namely endocytosis, LUZ51 uptake was evaluated in the presence of inhibitors of clathrin‐mediated endocytosis (chloroquine and chlorpromazine) and caveolin‐mediated endocytosis (filipin). The presence of these inhibitors markedly decreased cell uptake of endocytosis positive controls, namely Bodipy 12 and FITC‐albumin, but had no significant impact on LUZ51 internalization (Figure 3e–h). Hence, LUZ51 uptake does not occur through clathrin‐ or caveolin‐mediated endocytosis. Taken together, these findings indicate that passive diffusion is the primary mechanism driving LUZ51 internalization. Property (vii, fast cell uptake) was verified to an extent that exceeded our best expectations.

LUZ51 is Not Toxic in the Dark but is Very Phototoxic at Low Light Doses

Dark cytotoxicity studies revealed that LUZ51 has negligible toxicity in the dark, with EC₅₀ > 80 µM after 24 h of incubation in 4T1 cells (Figure S11E). Its phototoxicity was initially evaluated in 4T1, B16F10, and CT26 cancer cells following a 24 h incubation. At a low light dose, LD = 1 J/cm², we obtained EC₅₀ = 5.2, 15, and 15 nM for 4T1, B16F10, and CT26 cells, respectively (Figure 4a–c and Table 1). Increasing LD to 5 J/cm², decreased EC₅₀ for CT26 by approximately 5‐fold, to EC₅₀ = 3.3 nM. A more modest EC50 reduction was observed for 4T1 cells (4.0 nM). This may be related to differences in uptake and to more extensive bleaching at the lower concentrations (Figure S11) and Table S2). EC₅₀ values of a few nanomolar at low light doses signal that LUZ51 is a very phototoxic photosensitizer. When this is combined with its low toxicity in the dark, we obtain an unprecedented “phototherapeutic index” >25,000.^{[ 20} , ^{26 ]}

Figure 4.

LUZ51 exhibits strong phototoxicity even after short incubation periods. a–c) Phototoxicity of LUZ51 in 4T1, B16F10, and CT26 cells following a 24 h incubation and activation with LD of 1 J/cm²; d‐f) phototoxicity of LUZ51 in 4T1, B16F10, and CT26 following a 4 h incubation and activation with LD of 1 J/cm², g–i) phototoxicity of LUZ51 in 4T1, B16F10, and CT26 following a 0.5 h incubation and activation with LD of 1 J/cm²; j–l) phototoxicity of LUZ51 in 4T1, B16F10, and CT26 when illumination (1 J/cm²) was performed immediately after the addition of LUZ51. Bars indicate the mean ± SEM of 2–3 independent experiments. Statistical significance was evaluated using two‐way ANOVA in comparison to untreated cells (Ctrl), ** p < 0.01 and **** p < 0.0001.

Table 1.

EC₅₀ of LUZ51 activated at 1 J/cm².

Dose light	Time of incubation	4T1	B16F10	CT26
1 J/cm2	24 h	5.2 nM	15 nM	15 nM
		(95% CI:	(95% CI:	(95% CI:
		4.8 – 5.6 nM)	13.8 – 16.9 nM)	13.9 – 15.3 nM)
1 J/cm2	4 h	9.5 nM	16 nM	10 nM
		(95% CI:	(95% CI:	(95% CI:
		8.3 – 10.8 nM)	15.2 – 16.8 nM)	9.6 – 10.2 nM)
1 J/cm2	0.5 h	28 nM	26 nM	18 nM
		(95% CI:	(95% CI:	(95% CI:
		26.6 – 29.2 nM)	24.2 – 28.8 nM)	17.1 – 18.9 nM)
1 J/cm2	0 h	1.8 µM	2 µM	1.2 µM
		(95% CI:	(95% CI:	(95% CI:
		1.52 – 2.19 µM)	1.8 – 2.5 µM)	0.9 – 1.6 µM)

LUZ51 phototoxicity was also assessed after shorter incubation periods, namely between 4 and 0.5 h. After 4 h of incubation, LUZ51 maintained its low EC50: for example, in 4T1 cells, an EC50 of 9.5 nM was observed at 1 J/cm², and an EC50 of 7.4 nM was observed at 5 J/cm² (Figures 4d and S11, Tables 1 and S12). Remarkably, we obtained EC₅₀ = 18 to 28 nM with just 30 min of incubation, which is consistent with the rapid cellular internalization of LUZ51 (Figure 4e and Table 1). When LUZ51 is activated extracellularly (immediately after its addition to cells), its EC50 increases from 1.2 to 2 µM (Figure 4f). Overall, shorter incubation times led to a modest increase in EC₅₀, consistent with a slight reduction on cellular uptake. EC₅₀ values remained within the 5 – 28 nM range for all cell lines incubated for 30 min or more, at 1 J/cm², underscoring the potent phototoxicity of LUZ51 at short incubation times.

The remarkable phototoxicity of LUZ51 reflects the success of the design criteria (i) – (vi) in combination with (vii).

Low‐dose LUZ51‐PDT Cures Balb/c Mice with Subcutaneous CT26 Tumors

LUZ51 was designed to be amphiphilic and liposoluble. Consequently, it requires a formulation for intravenous (iv.) administration. Our preclinical studies with redaporfin initially employed a formulation consisting of a 1:5 mixture of Cremophor EL with ethanol,^{[ 27} , ^{28 ]} but more recently a P123‐based formulation improved its bioavailability, and tumor‐to‐muscle (T/M) and tumor‐to‐skin (T/S) ratios.^{[ 29 ]} This formulation led to a mean micelle size of 22 ± 1 nm, while sulfonyl‐substituted phthalocyanines P123‐micelles exhibited sizes around 26 nm.^{[ 29} , ^{30 ]}

Micelle sizes of the P123 formulation of LUZ51 were determined by dynamic light scattering. Immediately after preparation the micelle sizes were 25 ± 5 nm. After filtration through a 0.2 µm cellulose acetate filter and 1 h stabilization, the micelles increased to 32 ± 4 nm. Within 24 h, the micelle size further increased to 79 ± 1 nm and remained stable for at least 7 days (Figure S12A). The LUZ51‐P123 formulation was prepared 1 day in advance of the in vivo studies and used within 2 days after preparation.

It is important to explain that the cell uptake studies presented above employed LUZ51 in cell culture medium containing <1% DMSO, but the in vivo studies employed the LUZ51‐P123 formulation. Figure S12B shows that the LUZ51‐P123 formulation lowers LUZ51 cell uptake, but it remains quite impressive. At this stage, it remains unclear whether intact P123 micelles containing LUZ51 are internalized by cells or whether LUZ51 is transferred from the micelles to the cells through the cell culture medium. Two hypotheses may explain the reduced uptake of LUZ51 in P123 micelles: (i) Pluronic P123 is a triblock copolymer of polyethylene oxide (PEO) and polypropylene oxide (PPO) that self‐assembles into amphiphilic micelles in water with the PEO (polar) blocks forming an outer corona, and this hydrophilic shell may hinder close interactions with the cell membrane; ii) LUZ51 may be located in the hydrophobic PPO core of the P123 micelles, which lowers drug release, slows down membrane‐transfer dynamics and limits partitioning into the cell membrane.^{[ 31 ]} The reduced cell uptake of LUZ51‐P123 in vitro is probably inconsequential in vivo because it is expected that a large fraction of P123 micelles disassemble, or at least destabilize, within minutes after iv) administration, due to dilution below the critical micellar concentration, protein interactions, and shear‐induced destabilization.

PDT at DLI = 15 min gave the best results for redaporfin.^{[ 27 ]} We conducted comparable studies with LUZ51 in male Balb/c mice bearing subcutaneous CT26 tumors. Mice were treated with escalating doses of LUZ51 (0.075, 0.10, and 0.15 mg/kg) and tumor exposure to LD = 20 J/cm² at DLI = 15 min. A dose‐related response was observed. At the lowest dose (0.075 mg/kg), PDT induced mild oedema and necrosis, the median survival of mice increased from 7 days to 12 days (p = 0.0067) but no cures were observed. Mice treated with 0.10 or 0.15 mg/kg showed pronounced oedema and extensive necrosis 1 day after PDT. The dose of 0.10 mg/kg extended the median survival to 18 days (p = 0.0009) and 14% of the mice had complete tumor remission. PDT with 0.15 mg/kg extended the median survival to 45 days (p = 0.0003) and 50% of mice were cured (i.e., tumor‐free for at least for 45 days) (Figure 5a–c).

Figure 5.

Vascular‐PDT (DLI 15 min) using LUZ51 for the treatment of subcutaneous CT26 tumors. a–c) Mice were submitted to vascular‐PDT with escalating doses of LUZ51, 0.075, 0.1, or 0.15 mg/kg, and a LD of 20 J/cm²: a) survival curves; b) tumor volume for each individual mouse; c) representative images after PDT treatment. d–f) Mice were submitted to vascular‐PDT with 0.10 or 0.15 mg/kg of LUZ51, and a LD of 40 J/cm²: d) survival curves; E) tumor volume for each individual mouse; f) representative images after PDT treatment. The significant level between the PDT groups vs. untreated mice was evaluated using Long‐rank (Mantel‐Cox) test versus Ctrl, ** p < 0.01; *** p < 0.001.

The absence of treatment‐related lethality encouraged increasing LD to 40 J/cm². The cure rate of the 0.10 mg/kg dose increased to 57%, but that of the 0.15 mg/kg dose did not offer an additional increase (Figure 5d–f). Light doses can still be increased but, for small animals, this comes with a risk of lethality. Photoacoustic tomography performed 6 h post‐iv. administration of LUZ51 in mice bearing CT26 tumors (Figure S13) revealed detectable photosensitizer in the tumor and encouraged the increase of DLI to 6 h to look for higher cell uptake and selectivity. This PDT protocol increased the median survival of mice from 11 to 22 days, with one mouse remaining tumor‐free for at least 60 days (Figure S14).

LUZ51‐PDT Activates Antitumor Immunity

We assessed immunological memory after LUZ51‐PDT by rechallenging Balb/c mice cured with 0.15 mg/kg and 20 or 40 J/cm², with the subcutaneous administration of 350,000 CT26 cells. Tumor growth in rechallenged mice was delayed and tumor rejection was observed in 3 out of 7 mice (43% rejection rate, p = 0.0016). The tumor growth delay increases the median survival from 18.5 days to 31 days (Figure 6a, b). LUZ51‐PDT triggers a strong immunological memory, which is a response regulated by the adaptive arm of the immune system and typically mediated by T‐cells.

Figure 6.

Vascular‐PDT (DLI = 15 min) using LUZ51 induces immunological memory and is dependent on T cells. a, b) Tumor protection following rechallenge with 350000 CT26 cells in Balb/c mice that remained tumor‐free for 42 or 60 days after PDT: a) survival curves and b) tumor volume for each individual mouse. (c‐e) PDT efficacy (0.15 mg/kg LUZ51 and LD 20 J/cm²) in WT versus nude Balb/c bearing CT26 tumors: c) survival curves; d) tumor volume for each individual mouse and; e) representative images after PDT treatment. The significance level between the different PDT groups and untreated mice was evaluated using Long‐rank (Mantel‐Cox) test, PDT vs. Ctrl * p < 0.05, ** p < 0.01, *** p < 0.001; WT versus nude (# p < 0.05).

We confirmed the involvement of T cells in LUZ51‐PDT comparing Balb/c wild‐type (WT) and nude mice. Nude mice lack a thymus and are unable to produce T cells, rendering them immunodeficient. Necrosis was more pronounced in WT mice and 67% of these mice remained tumor‐free for at least 60 days. In contrast, no cures were observed in nude mice, although their median survival increased from 7 to 21 days (p = 0.0123) (Figure 6c–e). This survival benefit is attributed to direct tumor cell killing by oxidative stress, with potential contributions from components of the innate immune system, such as neutrophils. The mechanism of T‐cell activation requires further investigation, but our observations are consistent with the antitumor immunity described for redaporfin.^{[ 27} , ²⁸ , ^{32 ]}

Biodistribution and Pharmacokinetics of LUZ51 in Highly Aggressive Tumors

The orthotopic 4T1 breast cancer model in Balb/c mice closely mimics human stage IV triple‐negative cancer, and is characterized by poor vascularization, immunosuppressive TME, dense and fibrotic extracellular matrix (desmoplasia) and high metastatic potential to various organs.^{[ 33} , ^{34 ]} We used this challenging model to evaluate the biodistribution and pharmacokinetics of LUZ51 and, subsequently, to test the limits of LUZ51‐PDT.

Fluorescence imaging after iv. administration of 1.2 mg/kg of LUZ51 in female Balb/c mice bearing orthotopic 4T1 tumors, offers a broad perspective of the distribution and elimination of this photosensitizer (Figure S15). LUZ51 fluorescence was observed as early as 15 min post‐administration in the upper right quadrant of the abdominal cavity, indicating rapid liver accumulation. After 1 h, the fluorescence signal around this area becomes more diffuse, likely reflecting accumulation in the spleen. Two days after administration, LUZ51 fluorescence is much less intense and comes mostly from the lower part of the abdomen, suggesting a role of the urinary system in the elimination.

Biodistribution in Balb/c mice bearing orthotopic 4T1 tumors was followed for 3 days. Mice were sacrificed at varied time points after the iv. administration of LUZ51 (1.2 mg/kg), and the photosensitizer was quantified in blood, tumor, kidneys, liver, spleen, lungs, muscle, and skin (Figure 7a). LUZ51 pharmacokinetics fits the one‐compartment model with a half‐life t_1/2 = 82 min (Figure 7d) and a volume of distribution V _d = 0.14 L/kg. These values indicate a relatively fast elimination from the body and a moderate distribution to peripheral tissues. 15 minutes after administration, the liver showed the highest concentration of LUZ51 (16.37 µg/g tissue), followed by the spleen (6.58 µg/g tissue), and lungs (6.53 µg/g tissue). LUZ51 was also detected in 4T1 tumors (0.93 µg/g tissue) and kidneys (2.61 µg/g tissue), but not in muscle or skin. At later time points, LUZ51 levels gradually decreased in the liver, spleen, and lungs (Figure 7a). The pronounced accumulation of LUZ51 in the liver, spleen, and lungs aligns with their central roles in the reticuloendothelial system. The size of the nanoparticles in the LUZ51‐P123 formulation, 79 nm is close to the size required for recognition and removal by the macrophage system (∼200 nm), and some accumulation in the reticuloendothelial system was expected. LUZ51 levels in 4T1 tumors increase from 0.25 to 3 h post‐administration. Interestingly, LUZ51 levels are still 2 µg/g in the tumor 24 h post‐administration, but decreased to 0.04 µg/mL in the blood. This reflects accumulation and retention of LUZ51 in tumors.

Figure 7.

PK and BD of LUZ51 after iv administration (1.2 mg/kg) in female Balb/c mice bearing orthotopic 4T1 tumors. a) Quantification of LUZ51 in lysates obtained from the indicated organs, and b) in blood, inferred from a calibration curve based on LUZ51 fluorescence. The amount of LUZ51 was normalized to each tissue weight (g). c) LUZ51 in single cell suspensions obtained from 4T1 tumors at the indicated time points after iv. administration, measured by flow cytometry; d) Blood pharmacokinetics of LUZ51. Bars indicate the mean ± SEM of 4 − 5 mice; the fluorescence signal from treated cells was normalized to the untreated cells. Significance level of the difference between the Ctr and LUZ51 was evaluated via One‐way ANOVA **p < 0.01, ***p < 0.001, and ***p < 0.0001.

In view of the recognized difficulty of drugs to infiltrate the TME, we evaluated the association/internalization of LUZ51 by cancer cells in 4T1 tumors. Tumors collected at various time points were enzymatically digested, processed into single‐cell suspensions, and the cells were analyzed by flow cytometry. LUZ51‐associated fluorescence was high 15 min after administration (Figure 7c) and continued to rise, peaking at 6 h. This fluorescence likely reflects both the internalization of LUZ51 into cells (cell uptake) and its association with cell membranes (cell surface binding).

T/S ratios predict skin photosensitivity and T/M ratios reflect selective retention of photosensitizer in the tumor. Table 2 shows that these ratios increased gradually, peaking at 24 h post‐administration with values of T/S = 5.1 ± 1.8 and T/M = 13.2 ± 2.3. Skin photosensitivity was evaluated exposing small areas of depilated skin on the back of mice to a solar simulator for 15 or 30 min, either 24 h or 72 h after iv. administration of 0.15 mg/kg LUZ51. The exposure of control mice (no LUZ51) to the solar simulator for 15 min did not have a visible effect, but the 30 min exposure led to the delayed observation of mild erythema that resolved spontaneously. Similar exposure of mice after administration of LUZ51 did not lead to observable differences (Figure S16). There is no skin photosensitivity associated with the administration of 0.15 mg/kg LUZ51. Taken together, these data show that LUZ51 meets the desired properties (x, accumulation in tumors) and (xi, rapid clearance from the body).

Table 2.

Tumor‐to‐skin (T/S) and tumor‐to‐muscle (T/M) ratios of LUZ51 (± SEM) in 4T1 tumors implanted in the mammary fat pad of Balb/c mice.^a)

Time (h)	T/S	T/M
0.25	2.9 ± 0.7	4.8 ± 1.3
3	3.5 ± 0.8	6.5 ± 0.8
6	3.8 ± 0.5	7.8 ± 1.9
24	5.1 ± 1.8	13.2 ± 2.3
72	0.97 ± 0.24	3.5 ± 1.3

^a) n = 5, except for 0.25 h where one outlier was censured

LUZ51‐PDT Significantly Increases the Survival of Balb/c Mice Bearing Orthotopic 4T1 Tumors

Orthotopic 4T1 tumors present the challenges already mentioned and, additionally, PDT is impaired by the proximity of vital organs (liver, spleen, lungs, kidneys) where the photosensitizer is also present. This limits the drug and light doses that can be delivered without lethality. We started with iv. administration of 0.15 mg/kg LUZ51 and a light dose of 20 J/cm² at DLI = 15 min. Astonishingly, all mice responded to treatment, the median survival increased from 21 days in control to 38 days in PDT‐treated mice (p = 0.016), and one treated mouse remained tumor‐free for at least 65 days (Figure 8a,b). The cured mouse was rechallenged with live 4T1 cells and exhibited a significant delay in the development of a new tumor, indicating the presence of immunological memory (Figure S17). When we increased the dose to 0.19 mg/kg, the median overall survival increased to 41 days and 3 out of 8 mice remaining tumor‐free for at least 65 days (Figure 8c–e). The necrotic area was large and 2 mice died of treatment‐related effects with this drug dose.

Figure 8.

Vascular‐PDT (DLI = 15 min) using LUZ51 for the treatment of orthotopic 4T1 breast tumors. A,B) 0.15 mg/ kg of LUZ51 and a LD of 20 J/cm²: a) survival curves, b) tumor volume for each individual mouse; C–E) 0.19 mg/ kg of LUZ51 and a LD of 20 J/cm²: c) survival curves, d) tumor volume for each individual mouse, e) representative images after PDT treatment. The significant level between the PDT groups vs. untreated mice was evaluated using Long‐rank (Mantel‐Cox) test vs. Ctrl, * p < 0.05.

MRI of the lungs revealed large metastases in most untreated mice (e.g., 3 in 6 mice with metastases at day 21). The first MRI‐detectable metastases of the PDT‐treated group were observed on day 42 (Figure 9a, c). All untreated mice exhibited splenomegaly, which was significantly reduced in PDT‐treated mice and absent in cured mice (Figure 9b, d). These findings suggest that LUZ51‐PDT may trigger immunological responses that help control metastases.

Figure 9.

LUZ51‐based PDT significantly reduces lung metastases. Mice were treated with PDT using LUZ51 (0.15 mg/kg, 0.2 J/cm²), and MRI scans were acquired weekly until the end of the study. Representative MRI images of the a) lungs and of the b) spleen from an untreated mouse (Ctr), a PDT‐treated mouse with tumor regrowth (PDT), and a completely tumor‐free mouse (PDT cured). c) Number of mice in each group with large metastases. d) Quantification of spleen size. Statistical significance between Ctr and LUZ51‐PDT groups was assessed by one‐way ANOVA (**p < 0.01; **** p < 0.0001).

Discussion

The concern with the “molecular obesity” of current photosensitizers motivated the design of a lean bacteriochlorin inspired in halogenated bacteriochlorins,^{[ 18} , ²⁰ , ^{35 ]} characterized by intense absorption in the near‐infrared, photostability, and appropriate singlet oxygen quantum yields. Redaporfin (F₂BMet), FBMet and LUZ51 have four difluorophenyl, four fluorophenyl, and two difluorophenyl substituents in the macrocycle, respectively, and have Φ_pd = 1.0 × 10⁻⁵,^{[ 20 ]} 8 × 10⁻⁵,^{[ 35 ]} and 1.5 × 10⁻⁴,^{[ 23 ]} respectively. Photodecomposition increases along this series, but remains below that of temoporfin in methanol: water, Φ_pd = 63 × 10⁻⁵.^{[ 36 ]} Abdicating of two fluorophenyl groups lowered the photostability of the photosensitizer but did not compromise its use. The electron‐withdrawing effect of the carboxamide group contributes for photostability and imparts amphiphilicity. LUZ51 meets the 6 chemical properties ambitioned: i) ε₇₃₃ = 1.0x10⁵ M⁻¹ cm⁻¹, ii) Φ_∆ = 0.39, iii) Φ_pd = 1.5 × 10⁻⁴, iv) MW = 595 Da, v) log P _OW = 2.9, vi) amphiphilicity.

Based on studies with fluorophenyl bacteriochlorins, we looked for a “sweet spot” in hydrophilicity–lipophilicity and amphiphilicity to optimize cell uptake and intracellular tropism. The hydrophilic sulfonic bacteriochlorin F₂BOH (alias LUZ10, log P _OW = –1.4), shows poor cellular internalization and predominantly localizes in endocytic vesicles.^{[ 37} , ³⁸ , ^{39 ]} The cell uptake of redaporfin (alias LUZ11 or F₂BMet, log P _OW = 1.9), is higher than that of F₂BOH and that of the very lipophilic heptylsulfamoylphenyl bacteriochlorin Cl₂BHep (log P _OW = 4.5).^{[ 38 ]} Lipophilic photosensitizers, including temoporfin (log P _OW = 5.5),^{[ 40 ]} verteporfin (log P _OW = 1.6),^{[ 41 ]} and hypericin (log P _OW = 3.4)^{[ 42 ]} are predominantly internalized via passive diffusion^{[ 9} , ^{43 ]} and preferably locate in the ER, Golgi apparatus, and mitochondria. In contrast, larger and/or more hydrophilic photosensitizers, such as AlPcS4 and LUZ10 are primarily internalized by endocytosis, a process that is slower than passive diffusion and involves accumulation in endocytic vesicles.^{[ 37} , ^{43 ]} The role of amphiphilicity in cell uptake was recently highlighted by redaporfin atropisomers and strapped porphyrins, where the conformers with higher amphiphilicity showed uptake levels orders of magnitude higher than their less amphiphilic counterparts.^{[ 9 ]} LUZ51 was designed to have a modest lipophilicity and marked amphiphilicity, which foster rapid and efficient cellular uptake, and convey tropism to membrane‐enriched organelles. This was fully verified and LUZ51 excels in (vii) fast cell uptake.

In vitro phototoxicities of LUZ51 after 24 h of incubation across a panel of cell lines, EC₅₀ ≈ 10 nM @ 1 J/cm², compare favorably with those of potent clinically‐tested photosensitizers. Redaporfin α₄ atropisomer has EC₅₀ ≈ 200 nM @ 0.2 J/cm² in 4T1 cells, temoporfin has EC₅₀ = 26 nM @ 1 J/cm² in MDA‐MB‐231 cells,^{[ 44 ]} TLD143 (a ruthenium complex) has EC₅₀ = 21 nM @ 45 J/cm² in CT26.WT cells,^{[ 45 ]} and verteporfin has EC₅₀ ≈ 250 nM @ 1 J/cm² for 4T1 cells.^{[ 46 ]} Redaporfin and TLD143 have low dark toxicities, but temoporfin has dark toxicity: EC50 = 3.3 µM. Once the difference in phototoxicity between LUZ51 and redaporfin α₄ atropisomer is corrected for the light dose (1 versus 0.2 J/cm²), the remaining difference can be assigned to their cell uptakes. Seamless passive diffusion through biological barriers, nanomolar phototoxicity under 1 J/cm² after just 30 min of incubation and EC₅₀ > 80 µM in the dark after 24 h of incubation, fulfil two additional properties required for a good photosensitizer: viii) low dark cytotoxicity and iv) high phototoxicity.

Tumor‐to‐skin and tumor‐to‐muscle ratios are important to gauge the potential for skin photosensitivity and for damage to peritumoral regions during PDT. LUZ51 reaches T/S = 5.0 ± 1.8 and T/M = 13.1 ± 2.2 in orthotopic 4T1 tumors 24 h after iv. administration in the Pluronic P123 formulation. These high ratios are consistent with the expectation that P123 micelles are disassembled at this timepoint. Redaporfin in the more persistent Cremophor formulation reached T/S = 3.3 ± 1.1 and T/M = 4.6 ± 1.0 in a subcutaneous CT26 tumor model 48 h after iv. administration.^{[ 28 ]} Photofrin in a lung carcinoma model gave, 24 h after iv. administration, T/S = 1.6 and T/M = 10.0.^{[ 47 ]} Temoporfin, in the Foscan formulation and 4 h post‐administration, reached T/S = 4.1 and T/M = 9.4.^{[ 48 ]} Verteporfin in a liposomal formulation reached T/S = 3.9 ± 1.3 and T/M = 19.2 ± 4.6 at 24 h post‐administration.^{[ 49 ]} In addition to offering some of the best T/S and T/M ratios, LUZ51 has a relatively fast elimination from the body, t_1/2 = 82 min. Given the very favorable biodistribution and pharmacokinetics, it is not surprising that 24 h post‐iv. administration of 0.15 mg/kg of LUZ51 we did not observe skin phototoxicity in mice exposed for 30 min (180 J/cm²) to a solar simulator. LUZ51 does extremely well in x) accumulation in tumors and xi) rapid clearance from the body.

The advantage of the intense absorption of bacteriochlorins in the near infrared is manifested in PDT of tumors with diameters ≥ 5 mm, because they recapitulate the dependence of the optical penetration depth in tissues with wavelength.^{[ 50 ]} We recall that, whereas we have ε₇₃₃ = 1.0x10⁵ M⁻¹ cm⁻¹ for LUZ51, the corresponding values for verteporfin, temoporfin, Photofrin, and TLD143 are ε₆₈₆ = 3.4x10⁴,^{[ 51 ]} ε₆₅₀ = 3.0x10⁴, ε₆₃₀ = 1.2x10³,^{[ 20 ]} and ε₄₁₀ ≈ 3.5x10⁴ M⁻¹ cm⁻¹,^{[ 52 ]} respectively.

Immunogenic CT26 subcutaneous tumor are frequently used in PDT and were used for a first assessment of LUZ51 efficacy. LUZ51‐PDT cured 57% of Balb/c mice with implanted CT26 tumors with 0.10 mg/kg and 40 J/cm². Verteporfin‐PDT at 1 mg/kg and 120 J/cm² failed to cure mice with this tumor model.^{[ 53 ]} Tookad‐PDT achieved cure rates of up to 75% at 9 mg/kg and 30 J/cm².^{[ 54 ]} TLD1433‐PDT offered a 67% cure rate but this required intratumoral injection of 53 mg/kg and 190 J/cm.^{2 [ 45 ]} A 100% cure rate was achieved with the α₄ atropisomer of redaporfin using 0.45 mg/kg and 40 J/cm².^{[ 9 ]} Other photosensitizers under preclinical evaluation, such as bisamino silicon IV phthalocyanine (BAM‐siPc, 1.3 mg/kg, 60 J/cm²),^{[ 55 ]} and zinc(II) or platinum(II) sulfonyl‐substituted phthalocyanines (1.5 mg/kg, 90 J/cm²), also mediated tumor cures in this model but required high doses.^{[ 30 ]} The high potency of LUZ51 observed in vitro is amplified in vivo by the use of deeply‐penetrating near‐infrared light. Furthermore, the mice cured with LUZ51‐PDT resisted rechallenge with CT26 cells, and nude mice had a poorer response to the therapy. This is cogent evidence that LUZ51‐PDT (xii) triggers immunological responses.

A breakthrough in PDT requires more than incremental improvements over other photosensitizers. LUZ51 should be able to elicit cures in tumor models where this is not normally possible. Balb/c mice with orthotopic 4T1 tumors proved very difficult to cure with iv. administration of photosensitizers followed by one light delivery session, once tumor diameters reached 5 mm and lung metastasis occur.^{[ 56 ]} PDT with Photofrin,^{[ 57 ]} verteporfin,^{[ 46} , ^{58 ]} or redaporfin^{[ 16 ]} did not succeed in this tumor model. A recent study reported tumor control for 20 days with cyanine‐carboranes using 6DT1 breast cancer cells injected in the mammary fat pad of FVB mice, but this required 6 PDT sessions performed every 48 h.^{[ 59 ]} Cures of subcutaneous 4T1 tumors with PDT have been reported, but this tumor model does not translate the challenges of drug infiltration in the TME, its metastatic potential or the presence of vital organs beneath the tumor. In fact, when orthotopic 4T1 tumors attain 5 mm, the complete resection of the tumor does not cure the animals because lung metastases do not regress.^{[ 16 ]}

Orthotopic 4T1 tumors are characterized by immunosuppressive TME, high interstitial fluid pressure, poor vascularization, desmoplasia, and a challenging location for PDT. Current photosensitizers failed to cure mice with such tumors because of poor infiltration of the photosensitizer in the TME, limited phototherapeutic window and insufficient immune stimulation.^{[ 56 ]} LUZ51‐PDT of orthotopic 4T1 tumors cured 37.5% of Balb/c mice with 0.19 mg/kg and 20 J/cm², when the tumor size was 5 mm. The excellent response of 4T1 tumors to LUZ51‐PDT is due to tumor accumulation enabled by the P123 formulation, small size, and enhanced amphiphilicity of the photosensitizer that facilitate infiltration in the TME, and exceptional phototoxicity when illuminated in the near‐infrared. Additionally, the remission of lung metastasis and the normalization of the spleen size, indicate a very strong immune stimulation.

Conclusion

LUZ51 was rationally designed to exhibit optimal physicochemical properties, namely: i) intense near‐infrared absorption, ii) appropriate ROS quantum yield, iii) photostability, iv) low molecular weight, v) moderate lipophilicity, and vi) amphiphilicity. This design led to a photosensitizer with vii) fast cell uptake, viii) high cytotoxicity, ix) low dark cytotoxicity, x) high accumulation in tumors, xi) rapid clearance from the body, and capable of xii) triggering strong immunological responses. This makes LUZ51 an exceptional candidate for PDT, especially for tumors characterized by poor vascularization, immunosuppressive microenvironments, and high desmoplasia.

LUZ51‐PDT at low drug (0.10 mg/kg) and light (40 J/cm²) doses significantly increases the overall survival of mice bearing tumors of moderate aggressiveness, such as CT26 tumors (57% cures in one treatment). Moreover, 43% of the cured mice resisted rechallenge with the same tumor cells, which demonstrates a strong and lasting immunological response. LUZ51‐PDT of highly aggressive 4T1 orthotopic tumors elicited cures with one treatment of metastasized 5‐mm primary tumors. This is a challenging tumor model because illumination of the mammary gland affects surrounding organs and reduces the phototherapeutic window, and ablation of the primary tumor must be accompanied by clearance of the metastasis.

The clinical acceptance of PDT is limited by the perception that it is a local therapy associated with skin photosensitivity and of limited efficacy for large tumors. LUZ51 can change this perception. It accumulates in poorly‐vascularized tumors, is rapidly internalized by cancer cells, does not elicit skin photosensitivity and triggers robust immunological responses that reduce spontaneous metastasis. Small molecules have been overlooked in the quest for multifunctional and multitask photosensitizers, but they are a valuable pharmaceutical asset in the management of tumors where delivery of the pharmaceuticals to the cancer cells is limited by infiltration in tumor mass.

Author Contributions

Luis G. Arnaut, Mariette M. Pereira, and Lígia C. Gomes‐da‐Silva conceived the research. Sara M. A. Pinto and João P. A. Campos purified and characterized samples. Lígia C. Gomes‐da‐Silva, Mafalda Penetra, and Bárbara Lima performed the biological experimental work. Fábio A. Schaberle and João P. A. Campos made the photophysical studies. Luis G. Arnaut and Lígia C. Gomes‐da‐Silva provided supervision. Luis G. Arnaut and Lígia C. Gomes‐da‐Silva wrote the manuscript.

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Penetra M., Lima B., Campos J. P. A., Pinto S. M. A., Schaberle F. A., Pereira M. M., Arnaut L. G., Gomes‐da‐Silva L. C., Angew. Chem. Int. Ed.. 2026, 65, e09121. 10.1002/anie.202509121

Data Availability Statement

The data that support the findings of this study are available in the Suppporting Information of this article.