Resomer^® RG 752 H, Poly(D,L-lactide-co-glycolide) copolymer, monomer ratio 75:25, Mw=4000–15000 Da, Poly(vinyl alcohol) (PVA), Mw=9000-10000, 80% hydrolyzed, dichloromethane (DCM), acetonitrile HPLC grade (>99.93%), trifluoroacetic acid (TFA) HPLC grade (≥99%), Phosphate-buffered saline (PBS), hydrogen peroxide solution (30% w/w), and Tris(2,2′-bipyridyl)dichlororuthenium(II) hexahydrate were purchased from Merck KGaA (Darmstadt, Germany). Cyclosporine A was obtained from ADOOQ Bioscience (CA, USA). DiO’ (3,3’-Dioctadecyloxacarbocyanine Perchlorate) was purchased from ThermoFisher Scientific (Dreieich, Germany).

Multifunctional theranostic nanoparticles (tPLGA NPs) were synthesized using a reproducible double emulsion (W/O/W) method (32), designed to co-encapsulate both hydrophilic and hydrophobic components. TEM revealed the successful formation of spherical nanostructures with an average diameter of 133 ± 40 nm (Figure 2a). Within the PLGA matrix, the successful encapsulation of high-density inorganic nanoparticles was visible as darker, hypointense regions. To confirm the co-encapsulation and distribution of the magnetic components, STEM-EDX mapping was performed. Elemental maps clearly verified the presence of both iron (Fe) and manganese (Mn) within individual PLGA nanoparticles (Figure 2b). Interestingly, the mapping suggested a degree of spatial heterogeneity, with Fe-based NPs tending to cluster towards the core and Mn-based NPs localizing more peripherally.

Dynamic light scattering (DLS) analysis indicated a hydrodynamic diameter (D_h) of 210 ± 3 nm with a low polydispersity index (PDI) of 0.12 ± 0.03, confirming a monodisperse population in aqueous solution (Figure 2c). The larger size observed by DLS compared to TEM is expected, as DLS measures the hydrodynamic diameter, which includes the polymer corona and associated hydration shell, while TEM visualizes the dehydrated state (37). The nanoparticles exhibited a significant negative surface charge (ζ-potential) of -44 ± 1 mV, which is indicative of high colloidal stability and is crucial for preventing aggregation upon systemic administration. The successful incorporation of all components was further validated by FTIR (Supplementary Figure S1). The spectrum of the final tPLGA NPs was dominated by the characteristic peaks of the PLGA polymer, including C-H stretching vibrations below 3000 cm^-1, a strong carbonyl (C=O) stretch at approximately 1700 cm^-1 and C-O stretching bands between 1000–1300 cm^-1. The presence of CycA was confirmed by the persistence of its characteristic C-H bending vibrations around 1500 cm^-1. Furthermore, a subtle peak around 550 cm^-1, attributable to the metal-oxygen bonds of the iron and manganese oxides, confirmed the successful loading of the inorganic nanocrystals within the final composite structure.

A primary objective of this work was to engineer a nanoparticle system capable of acting as a dual-mode T₁-T₂ contrast agent for non-invasive MRI monitoring. This required careful optimization of the magnetic payload. The magnetic properties of the tPLGA NPs, base of their MR performance, were investigated using magnetometry, which revealed a mixed superparamagnetic-paramagnetic profile, resulting from the combination of superparamagnetic Fe₃O₄ and paramagnetic MnO NPs, characterized by the absence of hysteresis and S-shaped magnetization curves without saturation (Supplementary Figure S2).

To achieve an optimal balance for dual-mode MR imaging, a systematic study was conducted by varying the Fe: Mn ratio (Table 1). While a higher Mn content (e.g., 1:2 Fe: Mn ratio) was found to be ideal for maximizing T₁ contrast, this led to significant synthetic challenges and a lower overall metal loading. Consequently, a 1:1 Fe: Mn ratio was selected as a pragmatic compromise, ensuring sufficient concentrations of both metals for effective T₁-T₂ contrast while maintaining manufacturability. The relaxometric properties of the optimized tPLGA NPs were evaluated at clinical field strengths of 1.5 T and 3.0 T (Figures 2d, h). The nanoparticles demonstrated excellent performance as a T₂ contrast agent (13, 38) with a transverse relaxivity of 250 mM^-1s^-1 at 1.5 T and 537 mM^-1s^-1 at 3.0 T. These values are substantially higher than those of the commercial iron-based agent Feridex^® (r₂ = 41 and 93 mM^-1s^-1 at 1.5 T and 3.0 T, respectively). Concurrently, the NPs exhibited strong T₁ contrast, with a longitudinal relaxivity of 3.0 mM^-1s^-1 at 1.5 T and 4.5 mM^-1s^-1 at 3.0 T (Figure 2d), surpassing the manganese-based agent Teslascan^® (r₁ = 1.6 and 1.5 mM^-1s^-1, Figure 2h) (36). The unexpected increase in r₁ at the higher field strength may arise from the complex interplay and coupling effects between the T₁ and T₂ moieties within the nanocomposite, a phenomenon previously observed in intricate dual-mode systems (39).

A key therapeutic strategy of this platform is to mitigate IRI by generating oxygen in situ. The ability of the MnO component to produce O₂ in a redox-active environment was first assessed using a Ru(bpy)₃-based fluorescence quenching assay (Figure 2f). In the presence of H₂O₂, a significant, time-dependent decrease in fluorescence intensity was observed exclusively for the sample containing tPLGA NPs, confirming catalytic O₂ generation (30, 40, 41). This was further corroborated by relaxometry, where the r₁ value of MnO-containing NPs increased from 2.8 to 4.4 mM^-1s^-1 in the presence of H₂O₂, consistent with the reduction of MnO and release of paramagnetic Mn²⁺ ions that accompanies oxygen production (30) (Figure 2e).

The second therapeutic function is the controlled delivery of the immunosuppressant CycA. The encapsulation of the hydrophobic drug was highly efficient (>85%), even with the co-loading of inorganic MNPs. The choice of PLGA with a 75:25 lactic-to-glycolic acid ratio was intended to slow polymer degradation and prolong drug release (10, 32). Release kinetics studies over 15 days showed a sustained release profile, with the tPLGA NPs releasing approximately 20% of their payload, compared to 30% for control PLGA NPs lacking MNPs (Figure 2g). This suggests the MNP-loaded matrix creates a more tortuous path for drug diffusion and polymer degradation, enabling extended release (18, 32, 42). Kinetic modeling of the release data showed the best fit with the Korsmeyer-Peppas model (K_m= 0.882 ± 0.215; n=0.575 ± 0.052; AIC = 46.83; R²=0.902), yielding a release exponent (n) of 0.575. This value, falling between 0.5 and 1, indicates an anomalous, non-Fickian release mechanism driven by a combination of drug diffusion and polymer matrix erosion (43), which is ideal for long-term, controlled immunosuppression.

The clinical translation of any nanomedicine hinges on its safety. The cytocompatibility of the tPLGA NPs was evaluated in both rat pancreatic β-cells (RIN-m) and human peripheral blood mononuclear cells (PBMCs). Control formulations without CycA showed no significant cytotoxicity at metal concentrations up to 15 µg/mL (Figures 3a, b), a dose well above that required for imaging or therapeutic effects. This confirms the safety of the nanoparticle vehicle itself. Subsequently, the cytotoxicity of free CycA was compared to that of the encapsulated drug (Figures 3c, d). In both cell lines, free CycA induced significant dose-dependent toxicity. In contrast, the tPLGA NPs demonstrated a markedly improved safety profile, with significant toxicity only appearing at the highest concentration tested (50 µg/mL). This crucial result highlights a major advantage of nanoencapsulation: the ability to shield cells from the off-target toxicity of the free drug, thereby widening the therapeutic window.

Effective therapeutic action requires nanoparticle internalization. Cellular uptake was confirmed using confocal and fluorescence microscopy after incubating RIN-m cells with DiO-labeled tPLGA NPs. The images revealed efficient and homogenous uptake, with nanoparticles predominantly localizing in the cytoplasm (Figures 4a, b). This confirms that the NPs can effectively enter target cells to deliver their therapeutic payload and to serve as intracellular labels for MRI monitoring.

Having established the individual functionalities, the integrated theranostic performance of tPLGA NPs was evaluated in relevant cell models. RIN-m cells incubated with tPLGA NPs (containing 2.2 µg/mL Mn and 2.9 µg/mL Fe) generated significant MRI contrast. A notable T₂ shortening effect (T₂ = 184 ms vs 936 ms for water) and a detectable T₁ effect (T₁ = 2,291 ms vs 2,387 ms for control) were observed, confirming the potential for in vitro cell tracking (Figure 4c). The capacity of the tPLGA NPs to remediate hypoxic conditions was then assessed in RIN-m cells. Using a hypoxia-sensitive fluorescent kit, a significant dose-dependent reduction in fluorescence in cells treated with tPLGA NPs was observed, indicating successful O₂ generation and alleviation of hypoxia (Figures 4d, e).

Finally, the immunosuppressive efficacy of the formulation was tested. Activated PBMCs treated with tPLGA NPs showed a profound, dose-dependent reduction in the secretion of both IL-2, a key cytokine for T-cell proliferation, and IFN-γ, a major pro-inflammatory cytokine (Figures 5a, b) (44). Critically, at equivalent CycA concentrations, the tPLGA NPs demonstrated a significantly greater suppressive effect than the free drug. For instance, at a concentration of 0.4 µg/mL, the tPLGA NPs markedly reduced cytokine levels, whereas free CycA had a negligible effect. This enhanced efficacy, achieved at non-toxic concentrations (Supplementary Figures S3, S4), highlights the benefit of the nanoparticle delivery system, which can potentiate the drug effect and suppress key pathways of allograft rejection (45, 46).

The proposed t-PLGA NP formulation is designed for initial ex vivo application, ensuring that the primary drug release occurs locally at the pancreatic graft site, thereby establishing a robust protective microenvironment prior to transplantation. Subsequent administrations will be delivered systemically, necessitating careful consideration of targeting specificity. In the context of transplantation, an optimal balance must be achieved between local and systemic immunosuppression; while systemic approaches more comprehensively address the multifactorial mechanisms underlying graft rejection, localized immunosuppression minimizes systemic exposure to potent immunosuppressive agents and their associated toxicities. Targeting specificity may be enhanced through exploitation of the magnetic properties of the nanoparticles via magnetic guidance systems, or through incorporation of targeting ligands in future iterations of the platform. Regardless of the approach employed, comprehensive in vivo preclinical studies will be required to establish the optimal dosing regimen that maximizes therapeutic efficacy while minimizing toxicity in this clinical setting.