Establishing a Robust and Reliable Response from a Potent Osmium-Based Photosensitizer Via Lipid Nanoformulation†

Abstract

Osmium (Os) based photosensitizers (PSs) are a unique class of nontetrapyrrolic metal-containing PSs that absorb red light. We recently reported a highly potent Os(II) PS, rac-[Os(phen)₂ (IP-4T)](Cl)₂ , referred to as ML18J03 herein, with light EC₅₀ values as low as 20 pm. ML18J03 also exhibits low dark toxicity and submicromolar light EC₅₀ values in hypoxia in some cell lines. However, owing to its longer oligothiophene chain, ML18J03 is not completely water soluble and forms 1-2 μm sized aggregates in PBS containing 1% DMSO. This aggregation causes variability in PDT efficacy between assays and thus unreliable and irreproducible reports of in vitro activity. To that end, we utilized PEG-modified DPPC liposomes (138 nm diameter) and DSPE-mPEG₂₀₀₀ micelles (10.2 nm diameter) as lipid nanoformulation vehicles to mitigate aggregation of ML18J03 and found that the spectroscopic properties important to biological activity were maintained or improved. Importantly, the lipid formulations decreased the interassay variance between the EC₅₀ values by almost 20-fold, with respect to the unformulated ML18J03 when using broadband visible light excitation (P = 0.0276). Herein, lipid formulations are presented as reliable platforms for more accurate in vitro photocytotoxicity quantification for PSs prone to aggregation (such as ML18J03) and will be useful for assessing their in vivo PDT effects.

Figure 1.

Chemical structure of the osmium-based PS rac-[Os(phen)₂(IP-4T)](Cl)₂ (ML18J03) and graphical representations of its respective lipid formulations, DSPE-mPEG2000 micelles (mic-ML18J03) and PEG-modified DPPC liposomes (lipo-ML18J03). Lipid formulations are not drawn to scale. Figure prepared using an academic license of Biorender

Figure 2.

Longitudinal study of ML18J03 activity toward SK-MEL-28 human melanoma cells compared over eleven different assays, each performed in triplicate. Normoxic data points are represented by solid symbols, and hypoxic (1% O₂) data is represented by open symbols (connecting lines used to guide the eye). Treatments were dark or light (fluence=100 J cm⁻²; irradiance 20 mW cm⁻²) with cool white visible, 523 nm green, or 633 nm red LEDs. Note that the visible treatment is the most potent light condition due to contributions from the blue wavelengths of the spectrum.

Figure 3.

Comparison of activity toward SK-MEL-28 human melanoma cells using two different serial dilution methods for free ML18J03. Tips in dilution method #1 (solid symbols and lines) were exchanged only at 0.1 nM, whereas tips in dilution method #2 (open symbols and dotted lines) were exchanged at 30 μM and between each dilution below 600 nM. Differences in photocytotoxicities as a result of the two dilution methods only manifest for sub-μM EC₅₀ values.

Figure 4.

Dose-response (±SD) curves for SK-MEL-28 human melanoma cells treated with mic-ML18J03 at 11 mol% PS loading (a) or lipo-ML18J03 at 0.6 mol% PS loading (b) in normoxia. Treatments were dark (black filled circles) or light (fluence=100 J cm⁻²; irradiance 20 mW cm⁻²) with cool white visible (blue squares), 523 nm (green inverted triangles), 633 nm (red triangles), or 733 nm (purple crosses). The irradiance for 733 nm was 5 mW cm⁻².

Figure 5.

Min-max plots depicting the interassay distribution of EC₅₀ values of free and lipid formulated ML18J03 in SK-MEL-28 human melanoma cells using visible light (a), green light (b) and red light (c) excitation. d) The relationship between excitation light and statistical significance between variances demonstrates that lipid formulation of ML18J03 reduces interassay variability most for visible light excitation. Statistical significance between the free and lipid formulated groups was performed using a two-tailed t-test with Welch’s correction and statistical significance between the variances of groups was calculated using an F test.

Figure 6.

Hydrodynamic diameters (a) and polydispersity indices (b) of lipo-ML18J03 during storage at 4°C and incubation at 37°C for up to 11 days. Values are mean ± S.D.

Figure 7.

(a) Hydrodynamic diameters of DSPE-mPEG₂₀₀₀ micellar formulations of ML18J03 at varying mol% of the PS. Replicates were identical on the 55 mol% data point. (b) Relative fluorescence emission intensities (λ_em=740 nm, λ_ex=448 nm) of ML18J03 formulated in micelles at varying mol% of the PS. Hydrodynamic diameters (c) and polydispersity indices (d) of mic-ML18J03 (11 mol%) in storage at 4°C and incubation at 37°C for up to 11 days. Values are mean ± S.D.

Figure 8.

Comparison of the steady-state UV-vis absorption (a) and emission (b) spectra of unformulated and formulated ML18J03. Samples were prepared at 20 μM ML18J03 in PBS and argon-sparged prior to measurement. Emission spectra were collected at their excitation maxima (λ_ex=472 nm). Lipo-ML18J03 contained 0.6 mol% PS, and mic-ML18J03 contained 11 mol% PS.

Figure 9.

Comparison of the normalized excited state TA spectra (λ_ex=355 nm) of unformulated and formulated ML18J03 obtained over the first 5 μs following the excitation pulse (a). Normalized TA decay signals measured at 660 nm (b). Samples contained 20 μM ML18J03 in argon-sparged PBS except for free ML18J03 in (a), which was collected in argon-sparged MeCN due to its weak TA signal in PBS solution.