Covalently Surface-Functionalized Porphyrins on Silica Nanoparticles for Efficient Photodynamic Therapy

Abstract

Silica nanoparticles (SiNPs) are widely explored as biocompatible platforms for the delivery of photosensitizers in photodynamic therapy (PDT). In this work, porphyrins bearing amine (PNH₂) or carboxyl (PCOOH) groups were covalently conjugated onto functionalized SiNP surfaces via carbodiimide-mediated amide coupling, yielding the silica–porphyrin nanohybrids H-PNH₂ and H-PCOOH. Successful surface functionalization was confirmed by Fourier-transform infrared spectroscopy, Raman spectroscopy, and X-ray photoelectron spectroscopy. Photophysical studies demonstrated that both nanohybrids retained efficient singlet oxygen (¹O₂) generation. In vitro biological assays revealed a strong dependence of photodynamic activity on the nature of the conjugated porphyrin, with H-PCOOH exhibiting markedly enhanced photocytotoxicity with respect to the free porphyrins, while H-PNH₂ showed an attenuated light-dose response. Notably, H-PCOOH induced pronounced cell death at low light doses (1 J/cm²), with a half-maximal inhibitory concentration (IC₅₀) below 0.3 µM. These findings highlight the potential of silica–porphyrin nanohybrids as efficient photosensitizers for PDT applications.

1. Introduction

Photodynamic therapy (PDT) is a non-invasive therapeutic approach for cancer treatment that combines three key components: a photosensitizer, visible or near-infrared light, and molecular oxygen. Individually, these components are non-toxic; however, when combined, they generate reactive oxygen species (ROS), predominantly singlet oxygen (¹O₂), which induce oxidative stress and promote cancer cell death through direct cellular damage and, in some cases, activation of immune responses [1,2,3]. Compared with conventional treatments such as surgery, chemotherapy, or radiotherapy, PDT offers several advantages, including reduced systemic toxicity, spatial and temporal control of treatment, and a lower risk of resistance [4,5,6].

PDT has been clinically approved for more than two decades, and several photosensitizers are currently in use, including hematoporphyrin derivative (HpD), Photofrin^®, and temoporfin [7]. Most clinically used photosensitizers are based on tetrapyrrolic macrocycles, such as porphyrins, due to their strong absorption in the therapeutic window (600–800 nm), high singlet oxygen quantum yields (Φ_Δ), and tunable chemical structures [8,9,10]. Moreover, their intrinsic fluorescence enables applications in bioimaging and image-guided therapies [11,12]. However, many porphyrin-based photosensitizers suffer from poor aqueous solubility and a strong tendency to aggregate in biological media, which significantly limits their photodynamic efficiency and clinical applicability [13].

Despite its advantages, PDT still faces several limitations, including reduced effectiveness in large or deeply seated tumors, photosensitivity of healthy tissues, and insufficient selectivity of photosensitizers toward cancer cells [14,15]. Consequently, considerable effort has been devoted to developing strategies that enhance photosensitizer delivery, stability, and therapeutic performance. In this context, nanotechnology has emerged as a powerful tool to overcome these limitations by improving solubility, photostability, and tumor accumulation of PDT agents [16,17].

Nanoparticle-based formulations can improve the biocompatibility and circulation time of photosensitizers and may promote their preferential accumulation in tumor tissues through the enhanced permeability and retention (EPR) effect, associated with leaky vasculature and poor lymphatic drainage characteristic of solid tumors [18,19,20]. A wide range of nanomaterials has been explored as photosensitizer carriers, including graphene quantum dots, nanodiamonds, metal–organic frameworks, and metallic nanoparticles such as gold and silver nanoparticles, each offering distinct advantages in terms of optical properties, loading capacity, or multifunctionality [21,22,23,24,25,26,27,28].

Among these platforms, silica nanoparticles (SiNPs) have attracted particular attention due to their excellent biocompatibility, large specific surface area, controllable size, and ease of surface functionalization [29,30,31]. In addition, SiNPs can protect photosensitizers from chemical and biological degradation while allowing for controlled interactions with biological systems [32,33].

Several silica-based porphyrin delivery systems have demonstrated improved photodynamic performance. For instance, encapsulation of S-galactosylated and S-glucosylated porphyrins in amorphous SiNPs enabled efficient internalization in human bladder cancer cell lines and resulted in a 3–5-fold increase in PDT efficacy compared with free porphyrins [34]. Similarly, meso-substituted trans-A₂B₂ porphyrins conjugated to silica-gold core–shell nanoparticles exhibited enhanced anticancer and antibacterial photodynamic activity, attributed to improved cellular uptake and plasmon-assisted photophysical effects [35]. Nevertheless, many of these systems rely on physical encapsulation or complex hybrid architectures.

In contrast, covalent conjugation of porphyrins onto SiNPs represents a more robust strategy to prevent aggregation and preserve photodynamic activity, providing improved stability compared with non-covalent or simple adsorption approaches [35,36]. Functional groups such as amines or carboxylic acids on the porphyrin periphery enable covalent attachment and allow fine-tuning of key parameters, including surface charge, hydrophilicity, and porphyrin loading density. These physicochemical features play a critical role in governing cellular uptake and photodynamic efficiency.

The present study systematically examines how silica nanoparticle surface engineering, combined with controlled covalent immobilization of porphyrins bearing distinct functional groups, influences photodynamic performance. To this end, meso-tetrakis(4-[(2-aminoethyl)amino]-2,3,5,6-tetrafluorophenyl)porphyrin (PNH₂) and meso-tetrakis(4-[(2-carboxyethyl)thio]-2,3,5,6-tetrafluorophenyl)porphyrin (PCOOH) were selected as photosensitizers and covalently conjugated to SiNPs via carbodiimide-mediated coupling. Specifically, PNH₂ was coupled to carboxyl-functionalized SiNPs (SiNP@NHCO@COOH), while PCOOH was conjugated to amine-functionalized SiNPs (SiNP@NH₂), yielding the nanohybrids H-PNH₂ and H-PCOOH, respectively (Figure 1). The resulting systems were thoroughly characterized and evaluated in terms of photophysical properties, ROS generation, cellular uptake, and photocytotoxicity in HeLa cells, highlighting their potential as improved photosensitizers for PDT.

2. Results and Discussion

2.1. Synthesis and Characterization of the Porphyrins

The synthetic access to the functionalized porphyrins PNH₂ and PCOOH involved, in both cases, the nucleophilic aromatic substitution of the para-fluorine atoms of meso-tetrakis(pentafluorophenyl)porphyrin (TPPF₂₀) with ethane-1,2-diamine and 3-mercaptopropanoic acid, respectively, following literature procedures [37,38]. The structures of porphyrins PNH₂ and PCOOH were confirmed by ¹H nuclear magnetic resonance (NMR; Figures S1–S3) and high-resolution mass spectrometry (HRMS; Figure S4).

The ¹H NMR spectra (Figures S2A and S3A) show, in both cases, the expected low-field resonance of the eight β-pyrrolic protons of the porphyrin core. In the aliphatic region, the resonances of the methylene protons from four alkyl chains appear with the expected pattern, and the resonances of the two nitrogen internal protons at δ ~ −3 ppm appear as a broad singlet. In the ¹H NMR spectrum of PNH₂, a broad singlet due to the resonance of NH of the ethylenediamine fragments also appears at δ ~ 6.5 ppm. The analysis of ¹⁹F NMR spectra also confirms the expected loss of four p-fluorine atoms (Figures S1B–S3B).

HRMS data were consistent with calculated molecular formulas (Figure S4). The molecular ion peaks [M + H]⁺ were observed at m/z 1135.315284 (calcd 1135.315966) for PNH₂ and at m/z 1319.074536 (calcd 1319.076376) for PCOOH, providing additional confirmation of the successful functionalization. Altogether, these results validate the successful synthesis of porphyrins bearing amine or carboxyl substituents, which are suitable for subsequent SiNP conjugation.

2.2. Synthesis, Surface Functionalization, and Characterization of SiNPs

SiNPs were synthesized via the Stöber method, in which the hydrolysis and condensation of tetraethyl orthosilicate (TEOS) were catalyzed by ammonium hydroxide in a mixture of ethanol and water [39]. Surface functionalization was subsequently carried out to introduce specific chemical groups: Amine groups were grafted onto the SiNP surface using (3-aminopropyl)triethoxysilane (APTES) to yield SiNP@NH₂ [40], and carboxyl groups were incorporated via reaction with succinic anhydride, producing SiNP@NHCO@COOH [41] (Figure 1).

The formation, size, and monodispersity of the nanoparticles were confirmed by dynamic light scattering (DLS) and electron microscopy (EM), as summarized in

Table 1. Physicochemical properties of SiNPs and functionalized derivatives: hydrodynamic diameter determined by light scattering (D_h) in water and corresponding polydispersity index (PDI), diameter measured by electron microscopy (D), and electrophoretic zeta potential (ζ).

Nanoparticles	Dh (nm) a	PDIDLS	D (nm) b	ζ (mV) a
SiNPs	61 ± 3	0.10 ± 0.04	51 ± 5	−31 ± 3
SiNP@NH2	71 ± 1	0.17 ± 0.03	57 ± 7	23.7 ± 0.1
SiNP@NHCO@COOH	82 ± 3	0.12 ± 0.07	58 ± 6	−30 ± 5

^a D_h, PDI, and ζ are reported as mean ± SD across independent synthesis batches (n = 2–3). ^b D are reported as mean ± SD measured from 50 nanoparticles.

and Figure 2A–F. The particle size and size distribution of all nanoparticles were determined by scanning transmission electron microscopy (STEM; Figure 2A–F), revealing a uniform size distribution and well-defined particle morphology.

As shown in Table 1, DLS measurements revealed a progressive increase in hydrodynamic diameter upon functionalization, from 61 ± 3 nm for SiNPs to 71 ± 1 nm for SiNP@NH₂ and 82 ± 3 nm for SiNP@NHCO@COOH, while EM measurements confirmed that the size of the inorganic core of the particle remained essentially unchanged (Figure 2A–F). The polydispersity index (PDI) remained low for all samples, indicating that the SiNPs maintained a high degree of monodispersity after modification. As previously mentioned, nanoparticles with sizes ranging from 10 to 200 nm benefit from the EPR effect, allowing them to accumulate in tumor cells [42,43]; therefore, the developed nanoparticles have an appropriate size for passive targeting of tumor tissues.

Importantly, surface charge analysis in water at pH 7 further confirmed successful functionalization: SiNPs exhibited a negative zeta potential (ζ; −31 ± 3 mV), due to ionization of surface hydroxyl groups, and amine functionalization reversed the charge to positive (23.7 ± 0.1 mV), due to protonation of the amines, whereas the subsequent carboxylation restored the negative potential (−30 ± 5 mV), due to the presence of charge carboxylates.

Figure 2G shows the Fourier-transform infrared (FTIR) spectra of pristine SiNPs and their surface-modified derivatives [44]. All samples exhibit the characteristic vibrational bands of silica, including the Si–O–Si asymmetric stretching at ~1100 cm⁻¹, the symmetric stretching at ~800 cm⁻¹, and the bending vibrations at ~470 cm⁻¹, together with the Si–OH stretching at ~950 cm⁻¹. In addition, all samples displayed a broad band centered around ~3450 cm⁻¹, attributed to O–H stretching vibrations of surface silanol groups and adsorbed water, as well as a band near ~1640 cm⁻¹ corresponding to the H–O–H bending mode of adsorbed water [45].

Following amine functionalization with APTES, new weak bands emerge in the 2930–2850 cm⁻¹ region, assigned to C–H stretching bands of the methylene groups of the APTES chain. A shoulder around ~1580 cm⁻¹ is consistent with N–H bending vibrations of surface amine groups, although partial overlap with residual water contributions. After this functionalization, the broad band around ~3450 cm⁻¹ may also include contributions from N–H stretching; however, these contributions are difficult to distinguish from the O–H stretching of residual silanols and adsorbed water. Consequently, this spectral region is not considered a reliable diagnostic marker of functionalization [46].

Upon carboxyl surface modification with succinic anhydride, the appearance of two bands at ~1565 and ~1550 cm⁻¹ is consistent with overlapping contributions from amide II vibrations and the asymmetric stretching of carboxylate groups, confirming successful conversion of surface amines into succinylated groups. Given the low organic loading relative to the silica core and the proximity of the carbonyl region to strong water-related absorptions, the 1700–1730 cm⁻¹ region was not used as the primary diagnostic marker.

Complementary ¹H NMR analysis, performed after dissolution of the nanoparticles in NaOH/D₂O, provides additional evidence of successful amine functionalization of the silica surface (Figure S5) [47]. The spectrum of SiNP@NH₂ exhibits three distinct resonances assigned to the aminopropyl chain introduced by APTES: a triplet (t, J = 7.0 Hz) at δ 2.37 ppm corresponding to the CH₂ group adjacent to the amine (c), a multiplet (m) between δ 1.41 and 1.24 ppm attributed to the central methylene group (b), and a multiplet (m) between δ 0.35 and 0.19 ppm assigned to the CH₂ group closest to the silica surface (a). Additional signals originate from the solvent and trace amounts of residual ethanol from the nanoparticle washing steps.

For quantitative assessment, trioxane was added as an internal standard, which gave a singlet at δ 5.04 ppm corresponding to six equivalent protons. As each aminopropyl fragment contributes two protons per methylene signal, the integration of these resonances relative to trioxane enabled the estimation of the surface amine density. Based on this analysis, the surface coverage of SiNP@NH₂ was estimated to be approximately 8 APTES molecules per nm², indicating a densely functionalized surface. Previous reports suggest that a compact monolayer corresponds to 2–4 molecules/nm² [48], suggesting that multilayer coverage is present. While this higher surface density could potentially reduce accessibility, complementary colorimetric assays using fluorescamine and ninhydrin (Figure S6) confirmed these groups correspond to accessible primary amines, suggesting multilayer formation does not entirely block surface reactivity for subsequent functionalization.

Regarding carboxylated nanoparticles, this functionalization was also confirmed by ¹H and ¹³C NMR (Figure S7). The ¹H NMR spectrum (Figure S7A) shows a mixture of signals that can be assigned to the initial primary amine and the amide formed by the addition of succinic anhydride. The assignment of the signals was supported by control reactions between APTES and succinic anhydride carried out separately in D₂O and in 0.4 M NaOH [49]. Five key signals were attributed to both amine- and carboxyl-modified sites. A triplet (t, J = 6.8 Hz) at δ 2.88 ppm corresponds to the methylene group adjacent to the amine groups on the amine-modified SiNPs (labeled 3), while another triplet at δ 2.30 ppm indicates the methylene adjacent to the amide groups in the carboxyl-modified SiNPs (labeled c). A singlet at δ 2.19 ppm arises from the methylene groups located between the amide and carboxyl groups (labeled 4 and 5). In agreement with previous reports [49], the two triplets expected for the succinate hydrogens collapsed into a single signal under the basic conditions used to dissolve the silica core. The middle methylene in the propanamine chain appears as a multiplet between δ 1.40 and 1.17 ppm (labeled 2 and b), and a quartet at δ 0.17 ppm (J = 12.8, 8.2 Hz) corresponds to the methylene groups adjacent to the Si atom (labeled 1 and a).

Based on the integration of the signals assigned to the carboxy-ethylamide unit, approximately 50% of the surface amines were converted to carboxyl groups, corresponding to around 4 terminal carboxyl groups per nm² of the nanoparticle surface. These ¹H NMR findings, together with ¹³C NMR signals in the δ 181–171 ppm range for carbonyl groups and methylene resonances between δ 43.85 and 11.56 ppm, provide explicit confirmation of the carboxylation of the silica nanoparticles.

2.3. Synthesis and Characterization of the Hybrids

Silica–porphyrin nanohybrids were successfully prepared through covalent conjugation of tetra-substituted porphyrins to the surface of functionalized silica nanoparticles via amide bond formation, using N-ethyl-N-(3-dimethylaminopropyl)carbodiimide (EDC) and N-hydroxysuccinimide (NHS) coupling chemistry. Specifically, the amino-functionalized porphyrin PNH₂ was coupled to carboxylated silica nanoparticles (SiNP@NHCO@COOH), while the carboxylated porphyrin PCOOH was conjugated to amine-functionalized silica nanoparticles (SiNP@NH₂), yielding the hybrids H-PNH₂ and H-PCOOH (Figure 1), respectively.

These nanohybrids are soluble in DMSO, DMF, or water/DMSO mixtures, which allows their use in subsequent photophysical and biological studies. Neither the porphyrins nor the nanohybrids are soluble or colloidally stable in pure aqueous media, such as water or PBS.

The size distribution and morphology of the hybrids were assessed by EM (Figure 3A–D), confirming that the inorganic silica core was not affected by the surface functionalization process. Both hybrids preserved the spherical morphology and showed no evidence of structural distortion.

Evidence of successful conjugation was obtained from FTIR spectroscopy (Figure 3E,F) [44]. Upon hybrid formation, subtle but consistent spectral changes were observed, including enhanced N–H stretching bands around 3400 cm⁻¹, increased intensity of C–H stretching vibrations in the 2800–3000 cm⁻¹ region, and new features in the 1400–1500 cm⁻¹ range. X-ray photoelectron spectroscopy (XPS) provided more direct confirmation of porphyrin immobilization (Figure 3G). The survey spectra of both hybrids revealed a distinct F1s signal at approximately 688 eV, characteristic of the fluorinated phenyl substituents in the porphyrin structures. As fluorine is absent in the pristine silica nanoparticles, the appearance of this signal unequivocally confirmed the presence of porphyrin moieties at the nanoparticle surface. Taken together, the EM, FTIR, and XPS results demonstrate the successful covalent attachment of porphyrins to silica nanoparticles, yielding well-defined and structurally stable nanohybrids.

2.4. Photophysical Characterization

The photophysical characterization of the porphyrins and the hybrids was performed in dimethylformamide (DMF) (Figure 4 and

Table 2. Photophysical data of PNH₂, H-PNH₂, PCOOH, and H-PCOOH in DMF: absorption and emission maxima (${\mathsf{\lambda }}_{\mathrm{abs}}^{\max }$ and ${\mathsf{\lambda }}_{\mathrm{em}}^{\max }$), fluorescence quantum yield (Φ_F), and singlet oxygen quantum yield (Φ_Δ).

Compound	${\mathsf{\lambda }}_{\mathbf{abs}}^{\mathbf{max}}$ (nm)	${\mathsf{\lambda }}_{\mathbf{em}}^{\mathbf{max}}$ (nm)	ΦF (%)	ΦΔ (%)
PNH2	422, 512, 546, 592, 647	657, 714	5.2	25
H-PNH2	425, 516, 545, 592, 652	656, 719	5.4	17
PCOOH	414, 506, 533, 581, 646	654, 706	6.3	72
H-PCOOH	415, 508, 533, 582, 648	656, 705	7.4	44

). All four compounds exhibited characteristic Soret and Q bands typical of porphyrins [50]. Compared to the parent compound TPPF₂₀ (Figure S8), both PNH₂ and PCOOH show red shifts in their absorption and emission spectra, indicating electronic perturbations due to the introduction of either an amino group or a thioether group directly onto the meso-fluorophenyl groups. These effects were more pronounced for PNH₂, consistent with its stronger electron-donating nature.

Both nanohybrids retained the characteristic Soret and Q bands of the porphyrins, indicating that the macrocyclic electronic structure was largely preserved upon conjugation. A noticeable broadening of the Soret band, together with enhanced scattering, was observed, consistent with porphyrin immobilization onto the nanoparticle surface and the resulting heterogeneous local environments. In free porphyrins, the main fluorescence emission appears at approximately 710 nm, with a less intense secondary maximum near 655 nm; however, this pattern becomes inverted after conjugation to SiNPs. This change in the relative emission intensities suggests modifications in the local environment and electronic interactions of the porphyrin chromophore upon hybrid formation, possibly associated with reduced structural flexibility following surface immobilization.

The fluorescence quantum yield (Φ_F) showed a modest increase upon hybridization: PNH₂ increased from 5.2% to 5.4%, and PCOOH from 6.3% to 7.4% (Figure S9A), indicating that singlet-state radiative and non-radiative decay rates remained essentially unchanged. This result is consistent with the limited effects observed in the absorption and emission spectra upon conjugation to the SiNPs. Conversely, Φ_Δ values decreased upon conjugation, with PNH₂ reducing from 25% to 17%, and PCOOH from 72% to 44% (Figure S9B). This effect could arise from a reduction in the probability of collisional interaction between the porphyrin and the dissolved molecular oxygen associated with a decrease in the accessible area of interaction. Nonetheless, both hybrids retained sufficient ¹O₂ generation to induce photocytotoxicity.

A rough estimate of the amount of porphyrin per nanoparticle can be obtained from the absorption spectrum. Assuming the molar absorption coefficient for PNH₂ and H-PNH₂ remains the same in the Soret band, it is estimated that the hybrid contains approximately 0.28 porphyrins/nm², corresponding to 3.4% (w/w) of porphyrin in H-PNH₂. Using the same assumption, H-PCOOH is estimated to contain approximately 0.10 porphyrins/nm², corresponding to 1.4% (w/w) of porphyrin in this hybrid.

2.5. Biological Studies

The photodynamic efficiency of the hybrids H-PNH₂ and H-PCOOH, together with their free porphyrin counterparts (PNH₂ and PCOOH), was evaluated in HeLa cells over a range of concentrations and light fluences (Figure 5). For the hybrids, the indicated concentrations correspond to the porphyrin concentration estimated from the porphyrin weight percentage in the hybrid material. Control experiments confirmed that neither the compounds in the dark nor light irradiation alone affected cell viability (Figure S10), demonstrating that the observed cytotoxicity arises only from light-activated photosensitizers. Quantitative half-maximal inhibitory concentration (IC₅₀) values, photocytotoxicity indices (PIs), and curve fitting parameters are summarized in Table S1.

For the free porphyrins PNH₂ and PCOOH, irradiation induced a pronounced reduction in cell viability, with near-complete cell death observed at submicromolar concentrations under 5–10 J/cm². Two-way ANOVA revealed that, for both compounds, concentration and light fluence each had highly significant effects on cell viability, as well as a significant interaction effect between concentration and fluence (all p < 0.001). PNH₂ consistently exhibited slightly higher photocytotoxicity than PCOOH, despite its lower Φ_Δ value, underscoring the dominant role of cellular uptake over intrinsic photophysical efficiency. This behavior is further reflected in the high PI values, which increased sharply with irradiation dose and exceeded 300 for PNH₂ at 10 J/cm², confirming strong light selectivity and negligible dark toxicity.

The nanohybrids displayed distinct and markedly different response profiles compared to their free counterparts. For H-PNH₂, two-way ANOVA showed that both concentration and fluence individually had significant effects on cell viability (p = 0.003 and p < 0.001, respectively). In contrast, no statistically significant interaction effect was observed (p > 0.05), consistent with a flatter dose–response behavior upon conjugation. H-PNH₂ exhibited reduced photodynamic potency relative to free PNH₂, with IC₅₀ values remaining above 5 μM at 1 J/cm² and decreasing to the submicromolar range only at higher irradiation doses.

In contrast, H-PCOOH exhibited a strong dependence on light fluence (p < 0.001), while neither concentration nor the interaction between concentration and fluence had statistically significant effects (p > 0.05). Cell viability dropped below 20% even at the lowest concentration tested (0.3 μM) at all irradiation doses. As a result, dose–response fitting was not feasible for H-PCOOH. This behavior is accompanied by consistently high PI values (>64 at all fluences), highlighting both its high efficacy and excellent light selectivity.

Cellular internalization of the porphyrin derivatives was assessed by widefield (Figure S11) and confocal (Figure S12) fluorescence microscopies. Free PNH₂ displayed a strong intracellular fluorescence signal predominantly localized in the cytoplasm, whereas PCOOH showed no significant intracellular emission under the same conditions. This difference is consistent with the higher photocytotoxicity of PNH₂ and can be attributed to the positive charge of its terminal amine groups, which favor interactions with the negatively charged cell membrane [51]. In contrast, both nanohybrids exhibited combined intracellular and membrane-associated fluorescence, suggesting uptake via endocytic pathways and partial surface association. Given the relatively low porphyrin surface coverage (0.1–0.3 porphyrins/nm²) and the reduced surface charge of the nanohybrids compared with the precursor nanoparticles, differences in cellular uptake between the hybrids are less pronounced. Consequently, the enhanced photocytotoxicity of H-PCOOH relative to H-PNH₂ is primarily attributed to its higher ¹O₂ generation efficiency (Φ_Δ, Table 1).

In vitro ROS generation induced by H-PCOOH upon light activation was assessed using confocal fluorescence microscopy (Figure 6). HeLa cells were incubated with the fluorescent probe MitoSOX Red in the absence or presence of H-PCOOH, and images were acquired before and after blue light irradiation (450–490 nm, 60 s). In Figure 6A, the H-PCOOH nanohybrid is visualized in red, MitoSOX-associated fluorescence in green, and cell nuclei in blue.

MitoSOX emission primarily reports intracellular superoxide radical anion ($O_2^{•-})$ formation. Superoxide can be directly generated by type I photosensitization mechanism or arise from a secondary wave of ROS generation triggered by ¹O₂ [52]. Under non-irradiated conditions, confocal images show low basal and comparable MitoSOX fluorescence in both control and H-PCOOH-treated HeLa cells. Conversely, blue light irradiation results in a marked increase in ROS-associated MitoSOX fluorescence in H-PCOOH-treated cells, whereas untreated control cells show a decrease in MitoSOX emission. In Figure 6B, quantification of the mean fluorescence intensity per cell further supports these observations.

Statistical analysis by two-way ANOVA reveals that the presence of H-PCOOH, irradiation, and their interaction are all statistically significant (p < 0.0001), demonstrating that ROS generation is specifically associated with the photoactivation of the nanohybrid. Together, these results support the ability of H-PCOOH to induce intracellular oxidative stress in vitro upon irradiation, while remaining largely inactive under dark conditions.

Overall, these results demonstrate that covalent conjugation to silica nanoparticles profoundly alters photodynamic response profiles in a porphyrin-dependent manner. While immobilization attenuates the activity of PNH₂, it leads to a highly potent and fluence-dominated photodynamic response in the case of H-PCOOH, illustrating how surface chemistry governs the balance between uptake, photophysics, and biological efficacy.

3. Materials and Methods

3.1. Chemicals

All reagents and solvents were purchased from commercial suppliers and used as received, unless otherwise stated. DMF was dried over 4 Å molecular sieves for 24 h, and tetrahydrofuran (THF) was dried over sodium wire and benzophenone.

3.2. Equipment

¹H, ¹³C, and ¹⁹F NMR spectra were acquired using a Bruker Avance III spectrometer (Bruker, Bremen, Germany) (¹H: 300 or 400MHz, ¹³C: 75 MHz, ¹⁹F: 282 MHz). HRMS data were obtained using a Bruker SolariX XR (Bruker) with electrospray ionization (ESI). FTIR spectra were recorded using a Bruker VERTEX 70 spectrometer (Bruker). Ultraviolet–visible (UV-Vis) spectra were collected using Jasco V-660 (Jasco, Tokyo, Japan) and Shimadzu UV-2600i instruments (Shimadzu, Kyoto, Japan). Fluorescence emission spectra were recorded using a Horiba Fluorolog-3 spectrofluorometer (Horiba, Tokyo, Japan) equipped with a Xe lamp. ¹O₂ phosphorescence was measured using a DeltaFlex Time-Correlated Single Photon Counting system (Horiba, SN: 17043) coupled with a pulsed 405 nm laser diode (Becker & Hickl, BDS-SM-405 FBE) and a near-infrared detector (Hamamatsu, H10330C-45, Hamamatsu City, Japan). DLS and ζ analyses were performed using a Malvern Zetasizer Nano ZS ZEN3600 (Malvern Panalytical Ltd., Worcestershire, UK). EM was conducted on JEOL JEM-2200FS (JEOL, Tokyo, Japan) or JEM-2100 microscopes (JEOL, Tokyo, Japan) at 200 kV. XPS spectra were recorded at a pass energy of 20 eV and a normal emission angle. Widefield fluorescence microscopy images were acquired using an EVOS M5000 Imaging System (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). Confocal fluorescence microscopy images were acquired using a laser-scanning confocal microscope (Leica TCS-SP5, Leica, Wetzlar, Germany) equipped with a continuous Ar-ion laser (456, 476, 488, 496, and 514 nm), a helium–neon laser (633 nm), and a Ti:sapphire laser (Spectra-Physics Mai Tai BB, Spectra-Physics, Santa Clara, CA, USA, 710–990 nm, 100 fs, 82 MHz). Imaging was performed using a 63× water-immersion objective (HCX PL APO CS 63.0×/1.20 NA, Leica, Wetzlar, Germany). Photocytotoxicity assays were performed using an LED-based irradiation system operating at 420 nm (UniLight 420, max optical power 26 mW/cm², Sarspec, Vila Nova de Gaia, Portugal), and cell viability was determined using a POLARstar OPTIMA microplate reader (BMG Labtech, Ortenberg, Germany).

3.3. Synthesis

3.3.1. Synthesis of Meso-Tetrakis(pentafluorophenyl)porphyrin (TPPF₂₀)

TPPF₂₀ was synthesized according to the procedure described by Gonsalves et al. [53]. ¹H NMR and ¹⁹F NMR confirmed the purity (Figure S1).

¹H NMR (400 MHz, CDCl₃) δ: 8.93 (d, J = 2.6 Hz, 8H, β-H) ppm (Figure S1A). ¹⁹F NMR (376 MHz, CDCl₃) δ: −136.49 (dd, J = 23.6, 8.2 Hz, 8F, o-F), −151.13 (t, J = 20.6 Hz, 4F, p-F), −161.26 (m, 8F, m-F) ppm (Figure S1B).

3.3.2. Synthesis of Meso-Tetrakis(4-[(2-aminoethyl)amino]-2,3,5,6-tetrafluorophenyl)porphyrin (PNH₂)

PNH₂ was prepared according to the procedure reported by Rodrigues et al. [37], with minor modifications as described below: TPPF₂₀ (25.0 mg, 25.7 µmol) was dissolved in N-methyl-2-pyrrolidone (NMP; 0.5 mL) in a small vial, followed by the addition of an excess of ethane-1,2-diamine (17.2 μL, 10 equiv.). The reaction mixture was irradiated in a microwave oven (800 W, 1 bar, 200 °C, Mileston, MicroSYNTH, Milestone Srl, Sorisole (BG), Italy) for 9 min. The crude mixture was redissolved in CH₂Cl₂/MeOH (95:5) and washed with aqueous KHCO₃ solution. The organic layer was dried (Na₂SO₄) and the solvent removed under reduced pressure. The product was directly crystallized from a CHCl₃/MeOH mixture.

¹H NMR (300 MHz, DMSO-d₆) δ: 9.23 (s, 8H, β-H), 6.44 (s, 4H, NH-PhF₄), 3.59 (t, J = 6.5 Hz, 8H, H_a), 2.96 (t, J = 6.3 Hz, 8H, H_b), −3.13 (s, 2H, NH) ppm (Figure S2A). ¹⁹F NMR (282 MHz, DMSO-d₆) δ: −166.63 (d, J = 18.3 Hz, 8F, o-F), −184.35 (d, J = 19.1 Hz, 8F, m-F) ppm (Figure S2B). HRMS (ESI) m/z [C₅₂H₃₉F₁₆N₁₂ + H]⁺: observed 1135.315284, calculated 1135.315966 (Figure S4). FTIR (KBr): 3385, 2955, 1656, 1101 cm⁻¹ (Figure 3E). Yield: 70%.

3.3.3. Synthesis of Meso-Tetrakis(4-[(2-carboxyethyl)thio]-2,3,5,6-tetrafluorophenyl)porphyrin (PCOOH)

The synthesis of PCOOH was performed according to the procedure described by Hewage et al. [38]. Briefly, a suspension of TPPF₂₀ (50 mg), diethylamine (DEA, 150 μL), and ethyl acetate (EtAc, 3.0 mL) in DMF (0.5 mL) was reacted with excess 3-mercaptopropionic acid (50 mg) for 4 h. The crude mixture was separated by preparative TLC using 5% CH₃OH/CH₂Cl₂.

¹H NMR (300 MHz, DMSO-d₆) δ: 9.41 (s, 8H, β-H), 3.42–3.37 (m, 8H, H_a), 2.48–2.44 (m, 8H, H_b), −3.19 (s, 2H, NH) ppm (Figure S3A). ¹⁹F NMR (282 MHz, DMSO-d₆) δ: −131.02 (dd, J = 26.8, 11.0 Hz, o-F), −135.90 (dd, J = 26.8, 11.1 Hz, m-F) ppm (Figure S3B). HRMS (ESI) m/z [C₅₆H₃₁F₁₆N₄O₈S₄ + H]⁺: observed 1319.074536, calculated 1319.076376 (Figure S4). FTIR (KBr): 3450, 2934, 1593, 1467 cm⁻¹ (Figure 3F). Yield: 40%.

3.3.4. Synthesis of Silica Nanoparticles (SiNPs)

The synthesis of SiNPs was based on the Stöber method [39,54]. Briefly, in a polypropylene flask, ethanol (86 g), water (9 g), and ammonium hydroxide solution (1.5 mL) were added, and the mixture was stirred at 650 rpm and 50 °C. TEOS (4.5 mL) was added dropwise and allowed to stand overnight. After cooling, the dispersion was centrifuged (11,515× g, 10 min), washed three times with ethanol, and dried at 60 °C overnight. This procedure yielded 0.8–0.9 g of SiNPs with a D_h of 61 ± 3 nm and a ζ of −31 ± 3 mV.

FTIR (KBr): 3450 (ν O–H), 1100 (ν_a Si–O–Si), 950 (ν Si–OH), 800 (ν_s Si–O–Si), 470 (δ Si–O–Si) cm⁻¹ (Figure 2G).

3.3.5. Surface Amination of Silica Nanoparticles (SiNP@NH₂)

For the amination of nanoparticles (SiNP@NH₂), the procedure described by Perro et al. [40] was followed. Initially, a dispersion of SiNPs (0.25 g) in ammonium hydroxide solution (14 mL) was sonicated for 30 min. Subsequently, under an inert atmosphere, APTES (64 μL) diluted in ethanol (1 mL) was added and mixed at room temperature for 24 h. The resultant dispersion was centrifuged (11,515× g, 10 min), washed three times with ethanol, and dried at 60 °C overnight. This procedure resulted in 0.18–0.21 g of SiNP@NH₂ with a D_h of 71 ± 1 nm and a ζ of 23.7 ± 0.1 mV.

¹H NMR (400 MHz, NaOH/D₂O) δ: 2.37 (t, 2H, H_c), 1.41–1.24 (m, 2H, H_b), 0.35–0.19 (m, 2H, H_a) ppm (Figure S5). FTIR (KBr): 3450 (ν O–H, ν N–H), 2930–2850 (ν C–H), 1580 (δ N–H), 1100 (ν_a Si–O–Si), 950 (ν Si–OH), 800 (ν_s Si–O–Si), 470 (δ Si–O–Si) cm⁻¹ (Figure 2G and Figure 3F).

The presence of amine groups on the surface of the silica nanoparticles was confirmed using two complementary assays. Specifically, ninhydrin and fluorescamine assays were employed to assess the surface amine content.

Fluorescamine Assay. The assay for quantifying amine groups was adapted from Chen and Zhang [55]. A calibration curve was prepared by mixing 570 μL of 0.1 M borate buffer with 60 μL of an aqueous APTES solution (0–100 μM) and then adding 270 μL of fluorescamine in acetonitrile (10 μM) for detection. Fluorescamine reacts with primary amines to form a fluorescent product that is detectable at 478 nm upon excitation at 392 nm. For quantifying amines on SiNP@NH₂, a nanoparticle suspension (0.05 mg/mL) was used instead of APTES.

Ninhydrin Assay. The assay was adapted from Miller and Shantz [56]. A calibration curve was prepared by reacting 2 mL of APTES in methanol (0–20 μM) with 1 mL of 0.2 M ninhydrin in methanol. The solutions were stirred at 65 °C for 1 h, and then centrifuged (11,515× g, 10 min). After cooling to room temperature, 0.2 mL of the supernatant was diluted with 0.6 mL of methanol and analyzed by UV-Vis. Absorbance was measured at 580 nm to quantify all samples. To quantify the primary amines on SiNP@NH₂, a nanoparticle suspension (12.5 mg/mL) was used instead of the APTES solution.

3.3.6. Surface Carboxylation of Silica Nanoparticles (SiNP@NHCO@COOH)

Nanoparticles were carboxylated with succinic anhydride using a protocol adapted from Williams et al. [41]. SiNP@NH₂ (100 mg) and succinic anhydride (0.5 g) were dispersed in 5 mL of DMF and stirred overnight at room temperature. The mixture was centrifuged (11,515× g, 10 min), washed three times with ethanol, and dried at 60 °C overnight. This procedure yielded approximately 70 mg of SiNP@NHCO@COOH with a D_h of 82 ± 3 nm and a ζ of −30 ± 5 mV.

¹H NMR (400 MHz, NaOH/D₂O) δ: 2.88 (t, J = 6.8 Hz, H₃), 2.30 (t, J = 6.8 Hz, H_c), 2.19 (s, H₄, H₅), 1.40–1.17 (m, H₂, H_b), 0.17 (q, J = 12.8, 8.2 Hz, H₁, H_a) ppm (Figure S7A). ¹³C NMR (101 MHz, D₂O) δ: 181.13 (C=O), 175.52 (C=O), 171.08 (C=O), 43.85 (CH₂), 42.61 (CH₂), 33.99 (CH₂), 33.10 (CH₂), 32.36 (CH₂), 26.71 (CH₂), 23.38 (CH₂), 11.81 (CH₂), 11.56 (CH₂) ppm (Figure S7B). FTIR (KBr): 3450 (ν O–H, ν N–H), 2930–2850 (ν C–H), 1580(δ N–H), 1565–1550 (δ N–H, ν C–N, ν_s COO⁻), 1100 (ν_a Si–O–Si), 950 (ν Si–OH), 800 (ν_s Si–O–Si), 470 (δ Si–O–Si) cm⁻¹ (Figure 2G and Figure 3E).

3.3.7. Synthesis of H-PNH₂

The H-PNH₂ nanohybrid was obtained by covalent conjugation of SiNP@NHCO@COOH with PNH₂ via amide bond formation, mediated by EDC coupling chemistry. SiNP@NHCO@COOH (40 mg), EDC (10 mg), and NHS (10 mg) were dissolved in THF (15 mL) and stirred for 2 h at room temperature. PNH₂ (8 mg) was added, and the mixture was stirred for 48 h. The mixture was centrifuged (11,515× g, 10 min), washed three times with CH₂Cl₂, twice with water, and dried at 60 °C overnight. This procedure yielded approximately 50 mg of H-PNH₂.

FTIR (KBr): 3405, 2955, 1657, 1376, 1099, 961, 801, 465 cm⁻¹ (Figure 3E).

3.3.8. Synthesis of H-PCOOH

The H-PCOOH nanohybrid was synthesized via EDC/NHS-mediated amide coupling between PCOOH and SiNP@NH₂. Briefly, PCOOH (8–10 mg), EDC (10 mg), and NHS (10 mg) were dissolved in THF (15 mL) and stirred at room temperature for 2 h. Subsequently, SiNP@NH₂ (40 mg) were added, and the reaction mixture was stirred for 48 h. The resulting suspension was centrifuged (11,515× g, 10 min), washed three times with CH₂Cl₂, twice with water, and dried at 60 °C overnight. This procedure yielded approximately 40 mg of H-PCOOH.

FTIR (KBr): 3426, 2955, 1641, 1465, 1100, 947, 799, 467 cm⁻¹ (Figure 3F).

3.4. Optical Characterization

3.4.1. Fluorescence Quantum Yield

Φ_F values were determined in DMF using a comparative method with meso-tetraphenylporphyrin (TPP) as the reference compound (Φ_F = 11%) [57]. Emission spectra were recorded in DMF upon excitation at around 420 nm, with absorbance at the excitation wavelength kept below 0.2 to minimize inner-filter effects. The fluorescence quantum yield was determined using the following formula:

$${\mathsf{\Phi }}_{F,sample}={\mathsf{\Phi }}_{F,TPP} \times {\displaystyle \frac{I_{sample}}{I_{TPP}}} \times {\displaystyle \frac{1-{10}^{-A_{TPP}}}{ 1-{10}^{-A_{sample}}}}$$

(1)

where $A$ is the absorbance at the excitation wavelength and $I$ is the integrated fluorescence intensity over the 560–800 nm range.

3.4.2. Singlet Oxygen Quantum Yield

Φ_Δ values were determined in DMF using a relative method, with TPP as the reference photosensitizer (Φ_Δ = 64%) [58]. The phosphorescence emission spectrum of ¹O₂ was recorded upon excitation of the photosensitizer at 405 nm, with absorbance at the excitation wavelength maintained below 0.2. The Φ_Δ values were calculated using the following equation:

$${\mathsf{\Phi }}_{∆,sample}={\mathsf{\Phi }}_{∆,TPP} \times {\displaystyle \frac{I_{sample}}{I_{TPP}}} \times {\displaystyle \frac{1-{10}^{-A_{TPP}}}{1-{10}^{-A_{sample}}}}$$

(2)

where $A$ is absorbance at 405 nm, and $I$ is integrated phosphorescence intensity between 1220 and 1320 nm.

3.5. In Vitro Biological Assays

All solutions of porphyrins or nanohybrids used in the following experiments were prepared in culture medium (Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin) containing ≤1% (v/v) DMSO.

3.5.1. Cellular Uptake

HeLa cells (European Collection of Authenticated Cell Cultures) were seeded in Ibidi^® μ-Slide 8-well glass-bottom plates and incubated for 24 h at 37 °C and 5% CO₂. Cells were then incubated for an additional 12 h with solutions containing 10 μM of porphyrin equivalents, either as free porphyrins (PNH₂ and PCOOH) or incorporated into nanohybrids (H-PNH₂ or H-PCOOH).

The porphyrin concentration in dispersed hybrid nanoparticles was determined by UV–Vis spectroscopy using the solution-phase molar absorptivity coefficient, assuming that incorporation into the nanoparticles does not significantly alter the extinction coefficient. After incubation, cells were washed with Dulbecco’s phosphate-buffered saline (DPBS), and nuclei were stained with Hoechst 33342 (1 μg/mL) and membranes with Alexa Fluor 633 (5 μg/mL).

Widefield fluorescence images were collected using two sets of parameters (two channels): one for nuclear staining (λ_ex = 340–380 nm, λ_em > 425 nm) and one for porphyrin fluorescence (λ_ex = 515–560 nm, λ_em > 590 nm).

Confocal fluorescence images were collected using three channels: one for nuclear staining (λ_ex = 800 nm, λ_em = 390–500 nm), one for porphyrin fluorescence (λ_ex = 514 nm, λ_em = 600–720 nm), and one for the porphyrin and membrane staining (λ_ex = 633 nm, λ_em = 650–700 nm).

3.5.2. ROS Generation

HeLa cells were seeded in Ibidi^® μ-Slide 8-well glass-bottom plates and incubated for 24 h at 37 °C and 5% CO₂. Cells were then incubated for an additional 24 h with H-PCOOH (5 μM). After incubation, the cells were washed with DPBS, incubated with MitoSOX Red (300 nM) and Hoechst 33342 (1 μg/mL) for 30 min at 37 °C and 5% CO₂, and washed again with DPBS.

Confocal fluorescence images were collected using three channels: one for nuclear staining (λ_ex = 800 nm, λ_em = 400–560 nm), one for porphyrin fluorescence (λ_ex = 514 nm, λ_em = 620–720 nm), and one for the MitoSOX fluorescence (λ_ex = 514 nm, λ_em = 570–620 nm).

3.5.3. Dark Toxicity

HeLa cells were seeded in 96-well plates at a density of 1.5 × 10⁴ cells per well in 200 μL of culture medium and incubated for 36 h. Cells were then treated with the test compounds at concentrations ranging from 0 to 20 μM and incubated for an additional 24 h under the same conditions, protected from light.

Cell viability was assessed using the PrestoBlue assay by adding the reagent to a final concentration of 10% (v/v) and incubating for 1–2 h at 37 °C. Fluorescence was measured using a microplate reader (λ_ex = 530 nm, λ_em = 590 nm). Cell viability was expressed as a percentage relative to untreated control cells.

3.5.4. Photocytotoxicity

HeLa cells were seeded and cultured as described for the dark toxicity assays. The culture medium was replaced with fresh medium containing the test compounds (free porphyrins up to 1 μM and nanohybrids up to 5 μM of porphyrin equivalents), and cells were incubated for 12–16 h. Before irradiation, all wells were carefully washed with Dulbecco’s phosphate-buffered saline (DPBS).

Irradiation was performed using a 420 nm LED light source at an irradiance of 10 mW/cm², delivering light doses of 1, 5, and 10 J/cm² across the entire plate. Following irradiation, the cells were incubated in fresh culture medium for 24 h at 37 °C.

Cell viability was evaluated using the PrestoBlue assay as described above. To include untreated dark controls on the same plate, the last three columns containing untreated cells were covered with opaque black material to prevent light exposure during irradiation. Cell viability was expressed as a percentage relative to untreated cells kept in the dark.

3.5.5. Statistical Analysis

All viability experiments were performed with n = 2 independent biological replicates, each measured in at least three technical replicates. Data are presented as mean ± standard deviation (SD). Statistical analysis was carried out using two-way analysis of variance (ANOVA) to evaluate the effects of concentration and light fluence on cell viability.

Dose–response curves were generated by plotting normalized cell viability as a function of photosensitizer concentration on a logarithmic scale and fitted using nonlinear regression with a four-parameter logistic dose–response model (variable slope). Half-maximal inhibitory concentration (IC₅₀) values were calculated from the fitted curves and are reported with 95% confidence intervals.

In vitro ROS generation was analyzed using two-way ANOVA followed by Dunnett’s multiple comparisons test, with the non-irradiated control used as the reference for all comparisons.

Differences were considered statistically significant for p < 0.05. All analyses were performed using GraphPad Prism version 10.2 (GraphPad Software, San Diego, CA, USA).

4. Conclusions

This study reports the successful synthesis of two silica–porphyrin nanohybrids, H-PNH₂ and H-PCOOH, obtained through the covalent attachment of functionalized porphyrins to surface-modified silica nanoparticles via carbodiimide-mediated amide coupling. The use of complementary amine- and carboxyl-functionalized building blocks enabled the efficient immobilization of porphyrins onto the surface of silica nanoparticles. Structural and surface characterization by DLS, electron microscopy, FTIR, XPS, and NMR confirmed the formation of nanohybrids.

Photophysical characterization demonstrated that the characteristic absorption and emission features of the porphyrins were largely preserved after conjugation. Importantly, ¹O₂ generation confirmed that both nanohybrids retained significant photoactivity, indicating that covalent surface immobilization of the porphyrins on the SiNPs does not compromise their optical properties.

The photodynamic performance of H-PNH₂ and H-PCOOH was evaluated in HeLa cells and compared with that of the corresponding free porphyrins. All compounds exhibited clear light-dose-dependent cytotoxicity, with negligible dark toxicity, confirming that cell death arises specifically from photodynamic activation. H-PCOOH exhibited very high photodynamic activity, with IC₅₀ values below the lowest tested concentration at all irradiation fluences. In contrast, H-PNH₂ showed reduced activity compared with the free porphyrin, indicating that the nature of the porphyrin substituents and their interaction with the SiNP strongly influence photodynamic efficiency.

These results provide design principles for next-generation PDT nanomaterials and may guide the development of future theragnostic platforms.