Targeting severe acidity for tumor-activatable Interleukin-2 therapy

Summary

Interleukin-2 (IL-2) is a cytokine with curative potential in cancer immunotherapy, but its clinical use is limited by a narrow therapeutic window. Traditional strategies such as polarizing receptor binding or fusing IL-2 with Fc (IL-2-Fc) improve pharmacokinetics and immune selectivity, but systemic toxicity remains a key challenge, while covalent prodrug designs may compromise potency and restrict applicability. Here, we present a non-covalent approach using clinically validated ultra pH-sensitive (UPS) polymers to enable tumor-specific IL-2 activation. The UPS_5.3/IL-2-Fc nanoparticle remains stable at physiological pH, minimizing receptor binding in normal tissues, but dissociates and restores IL-2 activity in severely acidic tumor environments (pH < 5.3). This pH-triggered activation reduces systemic toxicity, resulting in over 100-fold reduction in circulating interferon-γ and prevention of vascular leak syndrome, while preserving antitumor efficacy. Mechanistically, the protective effect relies on both pH-dependent shielding and macrophage clearance. This bioengineering strategy offers a generalizable framework for immune cytokine therapy.

Graphical abstract

Highlights

• •Ultra-pH-sensitive nanoparticle shields IL-2-Fc at neutral pH and releases it at tumor pH
• •Tumor-specific IL-2-Fc release activates cytotoxic lymphocytes against cancer cells
• •pH shielding and macrophage clearance reduce systemic interferon-γ and lung toxicity

Feng et al. develop a pH-sensitive nanoparticle system that releases IL-2-Fc specifically in the severely acidic tumor microenvironment. This strategy attenuates systemic toxicity while preserving antitumor efficacy of cytokine therapy.

Introduction

Interleukin-2 (IL-2) is a potent immunostimulatory cytokine that plays a central role in antitumor immunity, primarily by promoting the proliferation of cytotoxic T cells and natural killer (NK) cells for tumor elimination.^1,2 Proleukin (native IL-2, also known as aldesleukin), the first approved IL-2 therapy, demonstrated curative outcomes in a subset of patients with renal cell carcinoma and has shown synergistic effects when combined with other immunotherapies.^2,3,4 However, low-dose IL-2 preferentially expands immunosuppressive regulatory T cells (Tregs), whereas high-dose administration can induce severe toxicities, including vascular leak syndrome (VLS), multiple organ failure, and systemic cytokine storms, thereby narrowing its therapeutic window and limiting clinical utility.^5,6

Extensive engineering efforts have focused on improving the therapeutic window of IL-2 by enhancing anti-tumor efficacy and/or reducing toxicity. Garcia et al. reported the development of IL-2 superkines with increased affinity for IL-2 receptor beta (IL-2Rβ, or CD122), enabling potent activation and expansion of cytotoxic T cells and NK cells without stimulating Tregs.⁷ As an alternative strategy, fusion of IL-2 to proteins like Fc or albumin has been used to extend IL-2’s half-life and engage additional antitumor mechanisms such as inducing Treg depletion.^8,9 However, systemic administration of IL-2 muteins or fusion proteins still results in dose-limiting toxicity from on-target, off-tumor immune activation. For instance, innate lymphocytes and NK and endothelial cells respond to IL-2 stimulation in healthy tissues, increasing the risk of toxic side effects. Introducing a “shielding” moiety through covalent chemistry or protein engineering has been developed to enhance safety. Nektar Therapeutics established a PEGylated IL-2 prodrug (NKTR-214) with significantly reduced toxicity; however, limited antitumor efficacy due to inefficient IL-2 conversion led to its discontinuation in clinical settings.^10,11,12 Other cleavable shading designs involve complex protein engineering processes, which may disrupt IL-2 structure, introduce immunogenicity, or depend on tumor-specific enzymes such as matrix metalloproteinases, limiting their generality across tumor types.¹³ A critical challenge remains in developing a broadly applicable strategy that effectively preserves IL-2 activity in tumors, while minimizing systemic toxic effects in healthy tissues.

Tumor acidosis is an emerging hallmark of cancer, resulting from the deregulated metabolism of cancer cells (i.e., the Warburg effect) and abnormal vasculature.^{14,15,16,17,18} Previously, we identified severe extracellular acidity (pH_e < 5.3) caused by the polarized secretion of lactic acid from cancer cells, which is markedly lower than the average reported tumor acidity (avg. pH_e = 6.83).¹⁹ This severe acidity has been clinically validated by pegsitacianine, an indocyanine green-conjugated ultra pH-sensitive (UPS) nanoparticle (NP) with a pH transition at 5.3, which showed broad tumor detection in phase 1 and 2 clinical trials involving 160 patients across 11 tumor types.^20,21,22 In response to the acidic tumor microenvironment, pegsitacianine transitions from a micelle to unimer state, activating fluorescence in an “off-to-on” manner for precision tumor imaging, while retaining non-fluorescent status at the neutral pH.^23,24 We hypothesize that the pH-triggered micelle-to-unimer disassembly of UPS NPs can be exploited to switch cytokine activity from “off” to “on” state, thereby masking systemic toxicity in healthy tissues, while enabling selective activation within tumors (Figure 1A). This study aims to establish a nanoengineering strategy, orthogonal to protein engineering and covalent conjugation, for safe and effective IL-2-based cytokine therapy.

Figure 1.

Chemical composition and pH-responsive behavior of UPS/IL-2-Fc NP

(A) Schematic illustrating IL-2-Fc activity control via pH-threshold-dependent encapsulation and release.

(B) Chemical structure of UPS polymers used to formulate pH-activatable IL-2-Fc nanoparticles.

(C–E) pH-dependent characterization of three UPS/IL-2-Fc formulations: nanoparticle size (C), IL-2-Fc release by ELISA (D), and IL-2 activity in HEK-Blue IL-2 reporter cells (E) (n = 3).

(F) Encapsulation stability of IL-2-Fc in different UPS formulations incubated in PBS, cell culture medium, and mouse plasma at 37°C (n = 3).

(G and H) In vivo evaluation of antitumor efficacy and cytokine release in MC-38 tumor-bearing mice treated with various UPS/IL-2-Fc NPs or unencapsulated IL-2-Fc (0.75 mg/kg, n = 5).

Data in (C–G) are presented as mean ± SEM. Heatmap in (H) generated from Z score normalization by column. See also Figure S1.

Results

Formulation of UPS/IL-2-Fc NP

Human IL-2-Fc, composed of IL-2 and the Fc region of human IgG1 linked by a (GGGGS) linker, was selected as the therapeutic payload due to its robust efficacy, reduced stimulation of Treg cells, close similarity to the native human sequence, and the ease of large-scale purification via protein A affinity chromatography.^8,9 Human IL-2 is cross-reactive to mouse IL-2 receptors, allowing for preclinical evaluation in murine models. Three UPS polymers were selected from the UPS library for formulation development: polyethylene oxide-b-poly(ethylpropylaminoethyl methacrylate) (PEO-b-PEPA, or UPS_6.9), polyethylene oxide-b-poly(dipropylaminoethyl methacrylate) (PEO-b-PDPA, or UPS_6.1), and polyethylene oxide-b-poly(dibutylaminoethyl methacrylate) (PEO-b-PDBA, or UPS_5.3), with pH transitions (pH_t) at 6.9, 6.1, and 5.3, respectively (Figure 1B). Dynamic light scattering analysis confirmed micelle-to-unimer phase transitions as a function of polymer protonation (Figure S1A). The assembled micelles exhibited hydrodynamic diameters of approximately 25–30 nm (Figure S1A).

A bifurcated microfluidic device (NanoAssemblr NxGen) was employed with a 1:1 flow rate ratio and a total flow rate of 10 mL/min to enable efficient mixing and reproducible production of UPS/IL-2-Fc NPs without precipitation. For UPS_5.3/IL-2-Fc and UPS_6.1/IL-2-Fc NP, UPS_5.3 and UPS_6.1 micelles (pH 7.4) were individually mixed with IL-2-Fc to form stable NPs. Complete encapsulation and uniform particle size were achieved when the UPS_5.3 to IL-2-Fc weight ratio exceeded 5:1 (Figures S1B–S1D). In contrast, micelle-protein mixing procedures failed to produce stable UPS_6.9/IL-2-Fc NP. To address this, an acidic UPS_6.9/IL-2-Fc solution (pH 4.5) was mixed with pH 8 phosphate-buffered saline (PBS), adjusting the final pH to 7.4 within the microfluidic device. Encapsulation efficiencies for all three formulations were verified to exceed 98% using fast protein liquid chromatography (FPLC).

pH-controlled IL-2-Fc release from UPS/IL-2-Fc NP

The pH-dependent disassembly of UPS/IL-2-Fc NP was characterized by dynamic light scattering (Figure 1C). All three formulations remained in the micelle form (∼30 nm) above their respective pH_t and dissociated into unimer form (<10 nm) below those thresholds. pH-triggered IL-2-Fc release was assessed using enzyme-linked immunosorbent assay (ELISA), with PBS with 1% bovine serum albumin (BSA) as the diluent. IL-2-Fc remained stably encapsulated above the transition pH for each formulation, but was released below it, with 50% release observed at pH 5.5 for UPS_5.3, pH 6.5 for UPS_6.1, and pH 7.1 for UPS_6.9 (Figure 1D). The slight elevation (0.1–0.4 pH units) in the 50% release point relative to micelle disassembly (Figure S1A) suggests that IL-2-Fc loading affects the pH transition behavior of UPS micelles.

The bioactivity of each UPS/IL-2-Fc NP was evaluated using an IL-2 receptor-expressing reporter cell line (HEK-Blue IL-2 cells). UPS_5.3 encapsulation suppressed IL-2 activity above pH 5.8 and restored it below pH 5.5, demonstrating pH-threshold-modulated cytokine activation (Figure 1E). While UPS_6.1 and UPS_6.9 formulations also showed acidity-triggered activation profiles, both exhibited residual IL-2 activity at higher pH values. To further assess formulation integrity, the stability of each formulation was tested in three biological environments: PBS, cell culture medium, and mouse plasma. ELISA was used to quantify IL-2-Fc leakage (Figure 1F). The UPS_5.3/IL-2-Fc NP remained stable across all tested conditions, showing no significant release of IL-2-Fc in 6 h and less than 20% leakage even after incubation for 48 h (Figure S1E). In contrast, UPS_6.1/IL-2-Fc NP maintained stability in PBS but exhibited over 20% leakage in mouse plasma after 6 h. UPS_6.9/IL-2-Fc NP showed poor stability in all environments, with greater leakage in protein-containing media compared to PBS. These results indicate that UPS_5.3/IL-2-Fc NP offers superior serum stability and represents the most promising candidate for in vivo application.

Antitumor efficacy and systemic toxicity of the three UPS/IL-2-Fc NP formulations were evaluated in the MC-38 murine colorectal cancer model, using tumor growth inhibition and systemic cytokine release following intravenous administration as primary endpoints. All formulations demonstrated anti-tumor efficacy comparable to that of unencapsulated IL-2-Fc, consistent with restored IL-2 activity in the acidic tumor microenvironment (Figure 1G). However, only UPS_5.3/IL-2-Fc NP effectively suppressed systemic cytokine levels after repeated dosing (Figure 1H). Based on these findings, targeting severe tumor acidity using the UPS_5.3 formulation represents a viable strategy for cytokine therapy with a broad therapeutic window.

UPS_5.3 polymer interacts with IL-2 and blocks receptor binding

To investigate the interaction between IL-2-Fc and UPS_5.3, we individually mixed IL-2, Fc, and IL-2-Fc with UPS_5.3 micelles at neutral pH and analyzed the mixtures by FPLC (Figure 2A). UPS_5.3 selectively interacted with IL-2 but not with Fc, as evidenced by the disappearance of IL-2 and IL-2-Fc peaks, while the Fc peak remained unchanged (Figure 2B). Additional FPLC and Förster resonance energy transfer (FRET) analyses confirmed that the encapsulation of IL-2-Fc relies on the interaction between IL-2 and micelle core (Figure S2A). When IL-2 was incubated with 5kDa poly(ethylene glycol) (PEG5K), it eluted as a distinct peak, similar to free IL-2, indicating minimal interaction. In contrast, IL-2 incubated with UPS_5.3 micelles was entirely retained in the void volume, suggesting strong association with the hydrophobic poly(methyl methacrylate) (PMMA) core of the polymer (Figure S2B). To further probe the spatial proximity between IL-2-Fc and the PMMA core, we performed FRET-based proximity assays using cyanine 3 (Cy3)-labeled IL-2-Fc and cyanine 5 (Cy5)-labeled UPS_5.3 (labeled on the PMMA block). At pH values above the polymer transition pH, where micelles form, we observed both quenching of Cy3 emission and a strong hetero-FRET signal (Cy3 excitation, Cy5 emission), indicating <5 nm proximity between IL-2-Fc and the micelle core (Figures S2C and S2D). Notably, this encapsulation occurred only at neutral pH and not under acidic conditions, as evidenced by the loss of FRET signal and FPLC analysis performed at pH 5, which showed distinct peaks of IL-2-Fc and UPS_5.3 (Figure S2E). Together, these results indicate that encapsulation is driven by interactions between IL-2 and the PMMA segment at neutral pH.

Figure 2.

pH-responsive blocking and restoration of IL-2 receptor binding by UPS_5.3

(A) Schematic of the experimental design to assess the intrinsic interaction between UPS_5.3 micelles and proteins.

(B) FPLC chromatograms showing individual proteins or UPS_5.3/protein mixtures.

(C) Molecular docking model of the DBA dimer (blue) with IL-2, highlighting predicted interaction sites.

(D–F) Biolayer interferometry analysis of IL-2-Fc binding to IL-2Rα, IL-2Rβ, and IL-2Rγ, with and without UPS_5.3 NP encapsulation.

(G) Schematic illustrating ex vivo assessment of IL-2-Fc binding to immune cells from various organs.

(H) Quantification of IL-2-Fc binding to lymphocytes isolated from liver, lung, spleen, and tumor (n = 5).

Data are presented as mean ± SEM. See also Figures S2 and S3.

To further characterize this interaction, we conducted molecular docking simulations to screen the entire protein surface and quantify binding affinity scores. As shown in Figure S2F, dimers of the UPS_5.3 repeat unit (DBA2) exhibited significantly stronger binding to IL-2 than dimers from other UPS variants (DPA2 and EPA2), correlating with the superior in vitro stability and reduced in vivo toxicity observed for the UPS_5.3 formulation (Figures 1E–1H). Notably, the calculated binding affinities ranged from approximately −3 to −4 kcal/mol, which is weaker than typical drug-target interactions (−5 to −6 kcal/mol). These results suggest that the encapsulation mechanism is distinct from high-affinity, receptor-specific interactions. Instead, it relies on weak, multivalent contacts that collectively drive stable loading into the micelle core (Figure 2C).

Molecular docking also revealed that DBA dimers preferentially bind regions overlapping with the IL-2 receptor alpha (IL-2Rα) domain, followed by beta (IL-2Rβ), with minimal interaction at the gamma (IL-2Rγ) site (Figure 2C). To validate these results, biolayer interferometry was used to assess IL-2-Fc binding to IL-2 receptor subunits with and without UPS_5.3 encapsulation (Figures 2D–2F). These data demonstrate that UPS_5.3 encapsulation abolishes IL-2 binding to all three receptors at neutral pH by interacting with the IL-2 molecule at receptor binding sites.

To demonstrate the pH-dependent binding of UPS_5.3/IL-2-Fc NP to IL-2 receptor-expressing cells, the formulation was incubated with immune cells isolated from the liver, lung, spleen, and tumors of MC-38 tumor-bearing mice under different pH conditions at 4°C (Figure 2G). The gating strategy for identifying the relevant immune cell populations is provided in Figure S3. In CD8⁺ T cells, NK cells, CD4⁺ Th cells, and CD4⁺ Treg cells, UPS_5.3 encapsulation completely abrogated IL-2-Fc binding at neutral pH. In contrast, binding was fully restored following acidic pH-triggered release (Figure 2H). Although absolute binding levels varied among cell types and tissues due to differences in receptor expression, the pH-dependent modulation of IL-2 activity was consistently observed across all groups.

Pharmacokinetics, biodistribution, and tumor-specific release of IL-2-Fc in vivo

Pharmacokinetic and biodistribution studies were conducted to evaluate UPS_5.3/IL-2-Fc NP and its selectivity to release IL-2-Fc in tumors (Figure 3A). A customized ELISA method was developed to distinguish between free and encapsulated IL-2-Fc (Figures S4A and S4B). Briefly, plasma samples diluted with PBS containing 1% BSA did not trigger additional IL-2-Fc release from NPs, allowing measurement of free IL-2-Fc. In contrast, dilution with IP lysis buffer (containing NP-40) and 1% BSA disrupted the NPs, enabling quantification of total IL-2-Fc. The pharmacokinetic profiles revealed a marked reduction in free IL-2-Fc levels following UPS_5.3/IL-2-Fc NP treatment compared to unencapsulated IL-2-Fc, while total IL-2-Fc levels remained comparable between the two groups (Figure 3B). The elimination half-lives (second phase) for total IL-2-Fc were 4.6 ± 1.4 h for IL-2-Fc and 4.1 ± 2.1 h for UPS_5.3/IL-2-Fc NP. Area-under-the-curve (AUC) analysis indicated that free IL-2-Fc exposure in the UPS_5.3/IL-2-Fc NP group was less than 10% of that in the unencapsulated group, confirming stable encapsulation and reduced premature release during circulation (Figure 3C). Importantly, UPS_5.3 encapsulation hinders FcRn-mediated recycling of IL-2-Fc, as evidenced by the absence of free IL-2-Fc in the blood following UPS_5.3/IL-2-Fc NP administration (Figures 3B and 3C), which would otherwise be triggered during FcRn engagement within acidified endosomes.

Figure 3.

Tumor-specific release of IL-2-Fc by UPS_5.3/IL-2-Fc NP

(A) Schematic illustrating measurement of free and total IL-2-Fc levels and the concept of tumor-specific activation.

(B) Pharmacokinetic analysis of free and total IL-2-Fc following intravenous administration of UPS_5.3/IL-2-Fc NP or unencapsulated IL-2-Fc alone (n = 5).

(C) AUC comparison of free and total IL-2-Fc concentrations (n = 5).

(D) PET/computed tomography imaging and quantification in MC-38 tumor-bearing mice 24 h after intravenous injection of ⁶⁴Cu-labeled IL-2-Fc or UPS_5.3/IL-2-Fc NP (n = 3).

(E) Schematic depicting the FRET-based design for assessing IL-2-Fc encapsulation status in vivo.

(F) FRET-based analysis of IL-2-Fc encapsulation and release in the spleen, lung, liver, and tumor following systemic administration. Scale bar, 50 μm.

Data in (B–D) are shown as mean ± SEM. See also Figures S4 and S5.

Biodistribution of the IL-2-Fc payload was quantified using ⁶⁴Cu-labeled IL-2-Fc. Dodecane tetraacetic acid (DOTA) was conjugated to IL-2-Fc to chelate ⁶⁴Cu²⁺ (radiation half-life = 12.7 h). FPLC analysis confirmed that DOTA conjugation and ⁶⁴Cu²⁺ chelation did not affect UPS_5.3 encapsulation (Figures S4C and S4D). Positron emission tomography (PET) imaging showed that both free IL-2-Fc and UPS_5.3/IL-2-Fc NP accumulated in MC38 tumors (Figure 3D). Quantitative analysis of resected organs 24 h after injection indicated that UPS_5.3 encapsulation did not significantly alter the overall biodistribution of IL-2-Fc, with only minor differences observed in the spleen (Figure 3D).

To assess the in vivo status of UPS_5.3/IL-2-Fc NP, IL-2-Fc was labeled with Cy3 and the UPS_5.3 polymer with Cy5, allowing simultaneous monitoring of Cy3, Cy5, and FRET signals. Co-localization of Cy3, Cy5, and FRET indicates that IL-2-Fc remains encapsulated within the UPS_5.3 NP, whereas spatial separation of Cy3 and Cy5 with loss of FRET reflects IL-2-Fc release (Figure 3E). In normal tissues such as the spleen, lung, and liver, strong overlap of Cy3, Cy5, and FRET signals confirmed that IL-2-Fc remained encapsulated (Figure 3F). By contrast, in tumor tissues, Cy3 and Cy5 signals were spatially separated, and no FRET signal was detected, indicating successful release of IL-2-Fc from UPS_5.3 NPs in the tumor microenvironment.

To establish the mechanistic role of pH in triggering release, MC38 tumor-bearing mice were pretreated with sodium bicarbonate to raise intratumoral pH, followed by intravenous injection of dual-labeled UPS_5.3/IL-2-Fc NPs (Figure S5A). Tumor sections collected 24 h later showed strong co-localization of IL-2-Fc and polymer signals, indicating suppressed release under buffered conditions (Figure S5B). By contrast, untreated tumors exhibited clear Cy3/Cy5 separation, consistent with pH-triggered micelle disassembly. To rule out alternative mechanisms such as protease-mediated degradation, we repeated the experiment in mice pretreated with broad-spectrum inhibitors of matrix metalloproteinases and cathepsins (batimastat and GB111). In all protease-inhibited groups, IL-2-Fc and UPS_5.3 signals remained spatially separated in the tumor, mirroring the untreated condition and distinct from the buffered group (Figure S5B). These findings support that micelle disassembly and IL-2-Fc release are governed primarily by acidic pH rather than enzymatic activity.

Taken together, UPS_5.3 encapsulation does not alter the overall pharmacokinetics and biodistribution of IL-2-Fc, but enables tumor-specific release of the payload.

Lymphocyte activation and antitumor efficacy of UPS_5.3/IL-2-Fc NP in tumors

The binding of IL-2-Fc to CD8⁺ T cells and NK cells and their activation within the tumor microenvironment were assessed (Figure 4A). Following a two-dose treatment regimen, IL-2-Fc binding to CD8⁺ T and NK cells was analyzed 1 day after each dose, and lymphocyte activation was evaluated 2 days post-treatment (Figure 4B). Fluorescein isothiocyanate-labeled IL-2-Fc was used to quantify cell binding, and unlabeled IL-2-Fc was used to assess immune activation. The gating strategy for cell population identification is provided in Figure S6. UPS_5.3/IL-2-Fc NP treatment resulted in effective binding of IL-2-Fc to both CD8⁺ T cells and NK cells in tumors, with no significant difference compared to free IL-2-Fc, particularly after the second injection (Figure 4C, left panel). The abundance of CD8⁺ T cells and NK cells, as well as the population of granzyme B⁺ Ki67⁺ cytotoxic cells, remained unchanged after the first dose but increased significantly following the second dose in both IL-2-Fc and UPS_5.3/IL-2-Fc NP treatment groups (Figure 4C, right panel). While wild-type IL-2 at low doses can expand Tregs due to high IL-2Rα expression, high-dose IL-2-Fc, as used here, favors effector cell activation and may also deplete Tregs via Fcγ receptor-mediated depletion.⁸ Flow cytometry confirmed that both IL-2-Fc and UPS_5.3/IL-2-Fc NP significantly increased the CD8⁺/Treg ratio in tumors after the second dose (Figure S7A), indicating a favorable immune shift. Together, these results confirm efficient tumor-localized release of IL-2-Fc from UPS_5.3 NPs and its ability to activate cytotoxic lymphocytes to the same extent as free IL-2-Fc.

Figure 4.

UPS_5.3/IL-2-Fc NP binds cytotoxic immune cells in tumors and elicits strong antitumor responses

(A) Schematic illustrating tumor-specific activation of UPS_5.3/IL-2-Fc NP in response to severe acidity, promoting cytotoxic immune cell stimulation.

(B) Treatment schedule for in vivo evaluation of NK and CD8⁺ T cell binding and activation.

(C) Binding and activation of CD8⁺ T cells and NK cells by UPS_5.3/IL-2-Fc NP (red) in MC-38 tumors, showing comparable effects to unencapsulated IL-2-Fc (blue, n = 5).

(D) Dose-dependent antitumor efficacy of UPS_5.3/IL-2-Fc NP in the MC-38 tumor model (n = 7–8).

(E) Anti-tumor effect in 4T1 orthotopic tumors with starting size >400 mm³. UPS_5.3/IL-2-Fc NP was administered at 2.25 mg/kg for three doses (n = 6).

(F) Combination therapy in the B16F10 model using UPS_5.3/IL-2-Fc with either anti-PD-1 or PolySTING. UPS_5.3/IL-2-Fc NP was administered at 2.25 mg/kg, twice (n = 5–7).

Data are shown as mean ± SEM. ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. See also Figures S6 and S7.

We further evaluated the anti-tumor efficacy of UPS_5.3/IL-2-Fc NP in both immune-hot and immune-cold tumor models. In immune-hot MC-38 tumors, a dose-dependent response was observed (Figure 4D). At 0.25 mg/kg, UPS_5.3/IL-2-Fc NP inhibited tumor growth, with 1 of 7 animals achieving complete remission. Increasing the dose to 0.75 mg/kg improved the cure rate to 3 of 8 animals. Further dose escalation to 2.25 mg/kg resulted in 5 of 7 animals achieving complete tumor remission. Comparable efficacy was seen in the immune-hot CT-26 model with well-established tumors (average size >150 mm³), where 3 of 9 animals achieved complete remission following treatment at 2.25 mg/kg (Figure S7B). In lung metastatic LL2 model, treatment with both UPS_5.3/IL-2-Fc NP and free IL-2-Fc significantly reduced pulmonary metastatic burden compared to PBS (Figure S7C). This functional evidence supports the presence of sufficient acidity in metastatic lesions to activate the UPS_5.3 platform and enable IL-2-Fc release. In the immune-cold 4T1 orthotopic model with well-established tumor (average size >400 mm³), UPS_5.3/IL-2-Fc NP shows significant tumor growth inhibition; however, no cures were achieved (Figure 4E). In the immune-cold B16F10 melanoma model, we evaluated UPS_5.3/IL-2-Fc NP as a monotherapy and in combination with either anti-PD-1 checkpoint blockade or an NP-formulated stimulator of interferon genes (STING) agonist (PolySTING) that activates the STING pathway.²⁵ Anti-PD-1 monotherapy had minimal impact on tumor progression, whereas UPS_5.3/IL-2-Fc NP alone significantly suppressed tumor growth. Notably, combining UPS_5.3/IL-2-Fc NP with anti-PD-1 produced a synergistic effect, further enhancing tumor inhibition (Figure 4F). PolySTING, which promotes type I interferon (IFN-I) release, dendritic cell cross-priming, and T cell activation, also amplified therapeutic outcomes. Its combination with UPS_5.3/IL-2-Fc NP resulted in a synergistic response, demonstrating robust tumor growth inhibition in the immune-cold B16F10 model.

UPS_5.3 blocks IL-2 interaction with NK cells in normal tissues with reduced toxicity

Clinical IL-2 therapy is associated with systemic toxicities, including cytokine storm, vascular leak syndrome (VLS), and multiple organ failure, with VLS being the major dose-limiting toxicity.^5,6,26 Although endothelial damage is strongly linked to VLS pathogenesis,²⁷ the specific immune cell population responsible for disrupting vascular integrity following IL-2-Fc administration has not been fully elucidated. In the C57BL/6 mouse model, a triple dose of IL-2-Fc (0.75 mg/kg IL-2 equivalent) induced VLS, as evidenced by a significant increase in lung weight (Figure 5A). We investigated the immune cell dependency of VLS using depletion/blocking antibodies and genetically modified mice (Figure 5A). IL-2-Fc failed to induce VLS in mice with NK cell depletion (via NK1.1 and CD90.2 antibodies) or with CD122 (IL-2Rβ) blockade. In contrast, VLS still occurred in mice lacking functional T and B cells (Rag1^−/− and Rag2^−/−), lacking other T cell types (e.g., γδ-T cells or NKT cells), or with CD25 (IL-2Rα) blockade, suggesting that NK cell activation via CD122 is the key immunological driver of VLS. Based on this, we hypothesize that UPS_5.3 NP encapsulation blocks IL-2-Fc interaction with NK cells in normal tissues, thereby mitigating systemic toxicity (Figure 5B).

Figure 5.

UPS_5.3 encapsulation blocks IL-2-Fc binding to NK cells in normal tissues, reducing systemic toxicity

(A) Treatment regimen and assessment of cell and receptor dependency in IL-2-Fc-induced vascular leak syndrome (VLS).

(B) Schematic illustrating UPS_5.3-mediated inhibition of NK cell activation by IL-2-Fc in normal tissues.

(C) Treatment schedule for in vivo evaluation of NK cell binding and activation in major organs.

(D) Comparison of IL-2-Fc binding to NK cells and NK cell counts in the spleen, liver, and lung from UPS_5.3-IL-2-Fc NP (red) and unencapsulated IL-2-Fc (blue). (n = 5–7).

(E–G) Improved safety profile of UPS_5.3/IL-2-Fc NP in C57BL/6 mice, demonstrated by stable body weight (E, n = 5), >100-fold reduction in systemic IFN-γ (F, n = 5), and absence of lung edema (G, n = 3).

n = 3 for (A); n = 5 for (D–G). Data are presented as mean ± SEM. ∗∗∗∗p < 0.0001. See also Figures S8–S11.

IL-2-Fc binding and activation of NK cells in major normal tissues were evaluated following a two-dose treatment regimen. NK cell binding was assessed 1 day after each treatment, and NK cell activation was measured 2 days after each treatment in the liver, lung, and spleen (Figure 5C; gating strategy shown in Figure S8). Unencapsulated IL-2-Fc exhibited strong binding to NK cells in the spleen and lung but showed weaker NK binding in the liver (Figure 5D, left panels). In contrast, UPS_5.3-encapsulated IL-2-Fc displayed minimal binding to NK cells in the spleen. Similarly, IL-2-Fc binding to NK cells in the lung was substantially reduced with UPS_5.3 encapsulation. In the liver, however, the difference between UPS_5.3/IL-2-Fc NP and unencapsulated IL-2-Fc was less pronounced. By day 5, two days after the second injection, IL-2-Fc treatment induced marked NK cell proliferation in the spleen, lung, and liver, consistent with systemic immune activation (Figure 5D, right panels). In contrast, UPS_5.3/IL-2-Fc NP treatment resulted in significantly lower NK cell proliferation in the spleen, lung, liver, and lymph nodes (Figures 5D and S9A). Moreover, IL-2-Fc treatment caused a sustained increase in the CD8⁺/Treg ratio in peripheral blood, similar to the tumor, whereas UPS_5.3/IL-2-Fc maintained a stable CD8⁺/Treg ratio over a 2-week period (Figure S9B). These results demonstrate that UPS_5.3 encapsulation limits IL-2-Fc exposure and activation in normal tissues, particularly in NK cells, thereby reducing systemic immune stimulation and associated toxicity.

The maximum tolerated dose of IL-2-Fc in C57BL/6 mice was established as three doses of 0.75 mg/kg administered at 1-day intervals, which resulted in significant body weight loss (∼20%) and 100% mortality (Figure 5E). At the same IL-2-Fc equivalent dose, mice treated with UPS_5.3/IL-2-Fc NP exhibited no detectable weight loss. This reduced toxicity was further supported by decreased cytokine release following UPS_5.3 NP encapsulation. IL-2-Fc treatment led to high circulating levels of IFN-γ 24 h after the second injection, along with elevated levels of other cytokines at multiple time points. In contrast, UPS_5.3/IL-2-Fc NP treatment resulted in a >100-fold reduction in systemic IFN-γ and significantly lower serum concentrations of IL-6, MCP-1, and tumor necrosis factor-α (Figures 5F and S10A–S10D). VLS, which represents the primary dose-limiting toxicity, was not observed following UPS_5.3/IL-2-Fc NP treatment at 0.75 mg/kg, as lung weight remained unchanged (Figure 5G). In comparison, IL-2-Fc treatment caused more than a 2-fold increase in lung weight, indicating severe vascular leakage and edema. Since UPS_5.3 encapsulation induced a modest increase in spleen accumulation, we also examined spleen weight as an additional indicator of systemic toxicity. Notably, UPS_5.3/IL-2-Fc NP did not cause an increase in spleen weight compared to IL-2-Fc treatment (Figures S10E and S10F).

Histological analysis was performed on mice treated with different formulations 1 day after the third IL-2-Fc dose (Figure S11). In the lungs, IL-2-Fc treatment induced peribronchial inflammation, alveolar filling, and patchy focal non-specific pneumonitis. In the liver, it caused portal inflammation. In immune organs such as the thymus, IL-2-Fc treatment led to architectural disorganization, evidenced by the loss of cortical structures. In the bone marrow, IL-2-Fc induced myeloid hyperplasia, indicated by an increased myeloid-to-erythroid ratio. In contrast, UPS_5.3/IL-2-Fc NP treatment did not trigger such inflammatory changes; the histology of normal tissues closely resembled that of untreated controls, demonstrating reduced systemic toxicity.

To summarize the therapeutic window, both IL-2-Fc and UPS_5.3/IL-2-Fc NP showed robust antitumor activity at a dose of ≥0.25 mg/kg in the MC38 model. However, IL-2-Fc reached its maximum tolerated dose at 0.75 mg/kg due to systemic toxicity, whereas UPS_5.3/IL-2-Fc NP could be safely administered at doses as high as 6.75 mg/kg without observable toxicity. This highest dose was constrained by formulation limitations rather than adverse effects. These results indicate that UPS_5.3 encapsulation expands the therapeutic window by at least 9-fold, enabling more effective treatment while maintaining a favorable safety profile.

The protective effect of UPS_5.3 encapsulation depends on host macrophages

Macrophages play a key role in NP uptake and clearance.^28,29 After intravenous injection, UPS_5.3/IL-2-Fc NPs predominantly accumulate in the red pulp region of the spleen, which is enriched with macrophages (Figure 6A). In contrast, unencapsulated IL-2-Fc accumulates in both the red pulp and white pulp regions, the latter being rich also in lymphocytes. Flow cytometry assay reveals that UPS_5.3/IL-2-Fc NP formulation leads to increased IL-2-Fc uptake by macrophages in the spleen and liver but not tumor, particularly following the second dose (Figure S12). We hypothesize that macrophage uptake and clearance of intact UPS_5.3/IL-2-Fc NPs contribute to the attenuated IL-2 toxicity in normal tissues. To test this causal relationship, C57BL/6 mice were treated with an anti-CSF1R antibody to deplete monocyte-derived macrophages. In IL-2-Fc-treated mice, body weight loss was unaffected by macrophage depletion. In contrast, UPS_5.3/IL-2-Fc NP-treated mice exhibited significant body weight loss by day 6 following macrophage depletion, demonstrating that the protective effect depends on the presence of host macrophages (Figure 6B). Additional pharmacokinetic analysis of UPS_5.3/IL-2-Fc NP revealed no substantial change in circulating levels of free or total IL-2-Fc following macrophage depletion, suggesting that the protective role of macrophages primarily arises from tissue-resident macrophage-dependent degradation clearance rather than altered systemic circulation (Figures S12C and S12D).

Figure 6.

UPS_5.3 NP encapsulation of IL-2-Fc leverages macrophage clearance to reduce systemic toxicity

(A) Treatment regimen and histological analysis showing differential IL-2-Fc distribution in the spleen following nanoparticle or free IL-2-Fc administration. W, white pulp; R, red pulp. Scale bars, 100 μm.

(B) Body weight changes in C57BL/6 mice treated with IL-2-Fc or UPS_5.3/IL-2-Fc NP, with or without macrophage depletion using anti-CSF1R (n = 3). IL-2-Fc dose: 0.75 mg/kg per injection.

(C) PBMC-engrafted humanized mice treated with different IL-2-Fc formulations (n = 5).

(D) CD34⁺ cell-engrafted humanized mice treated with different IL-2-Fc formulations (n = 5). IL-2-Fc dose for (C and D): 2.25 mg/kg per injection.

(E) Circulating cytokine profiles measured in both humanized mouse models post-treatment (n = 3).

Data are presented as mean ± SEM. ∗p < 0.05, ∗∗p < 0.01. See also Figure S12.

To better assess the translational potential and safety of UPS_5.3/IL-2-Fc NP in a human immune context, IL-2-Fc and UPS_5.3/IL-2-Fc NP were evaluated in two humanized mouse models. Mice were irradiated with 100 cGy (X-ray irradiation using an X-RAD 320 irradiator) and intravenously injected with either human peripheral blood mononuclear cells (PBMCs) or CD34⁺ hematopoietic stem cells. The PBMC model generated a human immune system primarily composed of lymphocytes and neutrophils, with few macrophages, whereas the CD34⁺ model established both lymphoid and myeloid lineages. In this PBMC model, UPS_5.3/IL-2-Fc NP treatment led to significant body weight loss compared to untreated controls (Figure 6C). Notably, the UPS_5.3 NP provided partial protection, with no mortality (0 of 7 deaths) compared to the high mortality rate (6 of 7 deaths) in the IL-2-Fc-treated group. In the CD34⁺ model, UPS_5.3/IL-2-Fc NP treatment offered strong protection, with no significant body weight loss even after six injections at 2.25 mg/kg (Figure 6D). Blood cytokine profiles were analyzed in both humanized models to assess systemic inflammatory responses. Consistent with the body weight data, UPS_5.3/IL-2-Fc NP showed reduced toxicity in the CD34⁺ model compared to the PBMC model, underscoring the role of myeloid cells in mediating toxicity response (Figure 6E). These findings not only reveal a mechanistic contribution of macrophage clearance to safety in the context of human immune system but also demonstrate the feasibility of this strategy for translation into human immunotherapy settings.

Discussion

Natural killer cells and innate lymphocytes present a central dilemma in cytokine therapy: while they are essential for antitumor immunity, they are also broadly distributed across normal tissues and serve as key mediators of cytokine-induced toxicity. Extensive protein engineering efforts have been dedicated to improving the efficacy of IL-2-based therapies. One major approach involves receptor biasing, as exemplified by IL-2 superkines and Neo-2/15, which selectively activate CD122⁺CD132⁺ cytotoxic lymphocytes while limiting CD25-mediated Treg expansion.^30,31 Another strategy aims to enhance IL-2 function in acidic tumor environments by introducing acid-resistant mutations that preserve cytokine activity at low pH.^32,33 These engineered variants have demonstrated improved antitumor activity and reduced systemic toxicity, partly due to the use of lower doses. However, they still show on-target, off-tumor effects such as pulmonary edema and weight loss. Protease-cleavable IL-2 prodrugs have also been developed to restrict cytokine activation to tumors by masking receptor-binding domains with tumor-specific cleavage motifs.^13,33 While these approaches improve therapeutic precision, they face limitations related to protease heterogeneity, immunogenicity, and tumor-type specificity. Achieving precise control over IL-2 activity within tumors while minimizing systemic exposure remains a major clinical challenge.

Over the past decade, our laboratory has developed a UPS NP system that functions like a pH-responsive “transistor,” undergoing rapid phase transitions within a narrow pH window of less than 0.25 units. By targeting broadly existing tumor acidosis, UPS NPs exhibit a sharp on/off response: they dissociate into unimers in acidic tumor environments while remaining as stable micelles at neutral pH. The UPS_5.3 nanoplatform targets pH 5.3, a more distinct and specific trigger compared to pH 7.4, and has been clinically validated in phase 1/2 trials using indocyanine green-labeled UPS_5.3 nanoprobes (pegsitacianine or ONM-100), which achieved broad tumor detection in human patients. Building on these findings, we employed UPS_5.3 NPs for tumor-activatable delivery of IL-2-Fc, engineered to stimulate cytotoxic lymphocyte proliferation specifically within tumors. At physiological pH, UPS_5.3/IL-2-Fc NPs remain stable and do not bind IL-2 receptors. In contrast, under acidic tumor conditions, UPS micelles dissociate effectively, restoring IL-2-Fc binding to its receptors and enabling selective activation within the tumor microenvironment. This study demonstrates a nanoengineering strategy to spatially restrict cytokine activation through a pH-triggered mechanism, which provides a generalizable framework for safe cytokine delivery with potential applicability to other immunotherapies and combinatorial treatment strategies.

NP encapsulation was originally developed to improve the solubility and circulation of therapeutic agents, as exemplified by formulations like Doxil. In our study, UPS_5.3/IL-2-Fc NP achieved a circulation profile comparable to that of free IL-2-Fc, but through a distinct mechanism. Whereas free IL-2-Fc relies on FcRn-mediated recycling for extended half-life, UPS_5.3/IL-2-Fc NP uses PEGylation and increased hydrodynamic size to evade renal clearance and reduce nonspecific uptake. Although both formulations show prolonged circulation, only UPS_5.3 maintains the cytokine in an inactive state until release is triggered by tumor acidity. These findings highlight that our platform redefines the role of NP encapsulation beyond simply extending half-life.

Biodistribution is another critical factor in NP-based delivery. While NPs have long been envisioned as vehicles for targeted payload delivery to tumors, recent studies have challenged this paradigm. Many NPs are poorly accumulated in tumors and are instead preferentially taken up by macrophages throughout the body, including within tumor tissue.^28,34,35 Our platform shifts the focus from physical targeting to spatially regulated activation. Although positron emission tomography (PET) imaging revealed similar organ-level distribution for free IL-2-Fc and UPS_5.3/IL-2-Fc NP, their activation profiles diverge significantly. UPS_5.3/IL-2-Fc NP remains inactive and intact within normal tissues until macrophage uptake, yet releases its cytokine payload specifically in acidic tumor environments. These findings support the concept of UPS_5.3 as a pH-activatable ON/OFF switch for tumor-selective cytokine delivery with minimized systemic toxicity. While this platform is highly compatible with IL-2-Fc, other cytokines, including IL-2 variants, may also benefit from similar encapsulation strategies. However, successful implementation will require case-by-case design and optimization.

Limitations of the study

Several limitations remain in this study. First, while molecular simulations and functional assays support the encapsulation of IL-2-Fc within the micelle core, high-resolution structural analysis such as cryo-electron microscopy imaging is not currently feasible. Additional biophysical studies are needed to determine whether micelle encapsulation induces conformational changes that may affect protein integrity. Second, although no abnormal immune activation was observed in mice, future studies in human-relevant systems will be valuable to study immunogenicity, premature release, and systemic safety. Third, while free IL-2-Fc serves as a clinically relevant benchmark, the advantages of this pH-responsive platform compared to protease-cleavable IL-2 prodrugs remain to be defined, as no protease-activated IL-2 construct is currently available for direct comparison. Fourth, this study focuses on short-term treatment outcomes; future work should investigate long-term therapeutic durability, immune memory development, and potential resistance mechanisms before successful clinical translation.

In summary, the pH-activatable UPS_5.3/IL-2-Fc NP system offers a method to mask systemic toxicity while maintaining antitumor efficacy. This approach provides a promising platform for the continued development of safe and effective cancer immunotherapies.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Jinming Gao (jinming.gao@utsouthwestern.edu).

Materials availability

All reagents generated in this study are available from the lead contact with a completed materials transfer agreement.

Data and code availability

• •Data: All data supporting the findings of this study are available within the main text and the supplementary materials. No standardized datasets (e.g., sequencing, proteomics, and microarray) were generated.
• •Code: No custom code or computational pipelines were generated or used in this study.
• •Other: Additional materials or resources related to this work can be obtained from the lead contact upon request.

Acknowledgments

We thank Tian Zhao and Zirong Chen from Onconano Medicine for the technical support in nanoparticle formulation and evaluation. We thank the following funding agencies for the support of this work: National Cancer Institute grants U54CA244719 (J.G.), R01 CA289258 (J.G.), Rally Foundation 24CDN03 (T.Y.), and Mendelson-Young endowment in cancer therapeutics (J.G.).

Author contributions

Conceptualization, J.G. and Q.F.; methodology, Q.F., R.P., J.G.L., W.L., Z.S., Q.W., S.Y., and V.S.B.; investigation, Q.F., R.P., J.G.L., W.L., Z.S., Q.W., S.Y., V.S.B., G.H., K.T., M.C., and T.Y.; writing – original draft, Q.F.; writing – review & editing, J.G.; funding acquisition, J.G. and T.Y.; supervision, J.G. and B.D.S.

Declaration of interests

B.D.S. and J.G. are scientific co-founders and advisors of OncoNano Medicine, Inc. Q.F., W.L., Z.S., G.H., and J.G. are named inventors on a patent related to this technology (WO/2023/087029).

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT 4o for grammar and spell-check assistance. After using this tool or service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
PerCP Rat Anti-Mouse CD45	BioLegend.	Cat# 103129 RRID:AB_893343
Pacific Blue™ anti-mouse CD4 Antibody	BioLegend	Cat# 103126 RRID:AB_493535
BV786 Hamster Anti-Mouse CD3e	BioLegend	Cat# 564379 RRID:AB_2738780
APC-R700 Rat Anti-Mouse CD8a	BioLegend	Cat# 564983 RRID:AB_2739032
PE anti-mouse CD25 Antibody	BioLegend	Cat# 102007 RRID:AB_312856
FITC anti-human/mouse Granzyme B Antibody	BioLegend	Cat# 515403 RRID:AB_211457
Brilliant Violet 605™ anti-mouse Ki-67 Antibody	BioLegend	Cat# 652413 RRID:AB_2562664
Biotin (Goat) anti-human IL-2	BioLegend	Cat# 517605 RRID:AB_256427
Human IL-2 antibody	BioLegend	Cat# 500302 RRID:AB_315089
APC-Cy™7 Mouse Anti-Mouse NK-1.1	BD Biosciences	Cat# 560618 RRID:AB_1727569
InVivoMAb anti-mouse PD-1 (CD279)	BioXcell	Cat# BE0146 RRID:AB_10949053
InVivoMAb anti-mouse Thy1.2 (CD90.2)	BioXcell	Cat# BE0066 RRID:AB_1107682
InVivoMAb anti-mouse CD122 (IL-2Rβ)	BioXcell	Cat# BE0298 RRID:AB_2687820
InVivoMAb anti-mouse CD25 (IL-2Rα)	BioXcell	Cat# BE0012 RRID:AB_1107619
InVivoMAb anti-mouse TCRβ	BioXcell	Cat# BE0102 RRID:AB_10950158
InVivoMAb anti-mouse TCR γ/δ	BioXcell	Cat# BE0070 RRID:AB_1107751
InVivoMAb anti-mouse CD4	BioXcell	Cat# BE0003-1, RRID:AB_1107636
InVivoMAb anti-mouse CD8α	BioXcell	Cat# BE0061, RRID:AB_1125541
Chemicals, peptides, and recombinant proteins
Streptavidin-HRP	Genscript	M00091
InVivoMAb recombinant human IgG1 Fc	BioXcell	BE0096
Human IL-2 R alpha/CD25 Protein, Fc Tag	Acro Biosystems	ILA-H5251-100ug
Human IL-2 R beta/CD122 Protein, Fc Tag	Acro Biosystems	ILB-H5253-50ug
Human IL-2 R gamma/CD132 Protein, Fc Tag	Acro Biosystems	ILG-H5256-100ug
Cyanine3 NHS ester	Lumiprobe Corporation	21020
Cyanine5 NHS ester	Lumiprobe Corporation	23020
DOTA-NHS-ester	TargetMol	T17842
NHS-Fluorescein (5/6-carboxyfluorescein succinimidyl ester) mixed isomer	Thermofisher Scientific	46409
ProLong™ Diamond Antifade Mountant with DAPI	Thermofisher Scientific	P36962
eBioscience™ Fixable Viability Dye eFluor™ 506	Thermofisher Scientific	65-0866-14
QUANTI-Blue™ Solution	Invivogen	rep-qbs
HEK-Blue™ CLR Selection	Invivogen	hb-csm
Normocin™	Invivogen	ant-nr-1
Puromycin	Invivogen	ant-pr-1
Critical commercial assays
BD™ Cytometric Bead Array (CBA) Mouse Inflammation Kit	BD Biosciences	552364
LEGENDplex™ HU Proinflam. Chemokine Panel 1 (13-plex)	BioLegend	740984
LEGENDplex™ Human CD8/NK Panel (13-plex)	BioLegend	741187
Experimental models: Cell lines
B16F10	ATCC	CRL-6475
MC38	ATCC	discontinued
CT-26	ATCC	CRL-2638
4T1	ATCC	CRL-2539
HEK-Blue™ IL-2 Cells	Invivogen	hkb-il2
Experimental models: Organisms/strains
C57BL/6J	The Jackson Laboratory	000664
BALB/cJ	The Jackson Laboratory	000651
NOD.Cg-Prkdcscid/J	The Jackson Laboratory	001303
B6.129S7-Rag1tm1Mom/J	The Jackson Laboratory	002216
B6.Cg-Rag2tm1.1Cgn/J	The Jackson Laboratory	008449
B6.129S6-Del(3Cd1d2-Cd1d1)1Sbp/J	The Jackson Laboratory	008881
Human CD34+ cells	Stemcell Technology	70008.5
Human PBMC	Carter Blood Center	N/A

Experimental model and study participant details

Mice

All animal experiments were performed with ethical compliance and approval from the Institutional Animal Care and Use Committee of the University of Texas Southwestern Medical Center. Female (6–10 weeks) mice were purchased from Jackson Laboratory. All mice were maintained in specific pathogen free animal facility. Temperature (68–79°F), humidity (30–70%), and light/dark cycle (lights between 6 a.m. and 6 p.m.) were controlled. Mice were euthanized at the end of the experiment with CO₂ asphyxiation with cervical dislocation as a physical secondary assurance method.

Cells

cell lines were purchased from American Type Culture Collection (ATCC) or Invivogen. Cells were cultured in high glucose Dulbecco’s modified Eagle’s medium supplemented with 10% fetal bovine serum, 100 U/mL penicillin, 100 U/mL streptomycin and 1×GlutaMax under 5% CO₂ at 37°C. HEK-Blue IL-2 cells were cultured and used according to manufacturer’s protocol (Invivogen).

Method details

IL-2-FC and modified IL-2-Fc

IL-2-Fc was obtained from Genscript with customized protein sequences. Briefly, the IL-2-Fc sequence consists of human wild-type IL-2 linked to the human IgG1 Fc region via a GSSSS linker. This sequence was cloned into a pcDNA3.4 vector and expressed in Freestyle 293-F cells. The supernatant was collected, and IL-2-Fc was purified using a Protein A affinity column according to standard protocols. Protein concentration was determined by UV-vis absorption and HPLC, and quality was confirmed by SDS-PAGE and size-exclusion chromatography.

For FITC, Cy3, or Dodecane tetraacetic acid (DOTA) conjugation, IL-2-Fc was incubated with DOTA-NHS-ester in PBS (pH 8) at a 4:1 molar ratio at 4°C overnight. Unconjugated dye or DOTA was removed by filtration using Millipore ultrafiltration membranes with a 10 kDa molecular weight cutoff.

Chelation of ⁶⁴Cu²⁺ to DOTA-IL-2-Fc was performed by adjusting the pH to 6.0–6.5 with 4 M ammonium acetate buffer and incubating for 15 min at 37°C. Unbound ⁶⁴CuCl₂ was removed by three rounds of centrifugal membrane filtration with a 10 kDa molecular weight cutoff.

Syntheses OF UPS polymers and dye conjugated UPS polymers

UPS polymers were synthesized as previously described using an atom transfer radical polymerization (ATRP) method. UPS_6.9, UPS_6.1 and UPS_5.3 were synthesized from ethylpropylaminoethyl methacrylate (EPA), dipropylaminoethyl methacrylate (DPA), and dibutylaminoethyl methacrylate (DBA) monomers, respectively through ATRP reaction using PEG-Br as the initiator. Three primary amino groups (aminoethyl methacrylate or AMA-MA) were introduced into each polymer chain during the polymerization.

For dye conjugation, UPS_5.3 polymers and dye-N-hydroxysuccinimide ester (1.5 molar equivalents to the primary amino group) were dissolved in dimethylformamide and stirred at room temperature for 12 h. Dye-conjugated copolymer was purified by ultracentrifugation (10 kDa cutoff) with methanol (10 cycles) and water (3 cycles) to remove unreacted dye molecules. The purified product was lyophilized and stored at −80°C.

Formulation OF UPS/IL-2-Fc nanoparticles

UPS/IL-2-Fc nanoparticles were formulated using microfluidic mixing with a bifurcated microfluidic device (Nanoassemblr Nxgen, flow rate ratio 1:1, total flow rate 10 mL/min) for efficient mixing. Ten to one weight ratio of UPS polymer to IL-2-Fc was used in all experiments. For UPS_5.3/IL-2-Fc NP and UPS_6.1/IL-2-Fc formulations, UPS_5.3 and UPS_6.1 micelles were individually mixed with IL-2-Fc to create UPS_5.3/IL-2-Fc NP and UPS_6.1/IL-2-Fc NP, respectively. For UPS_6.9/IL-2-Fc NP, an acidic UPS_6.9/IL-2-Fc solution (pH 5) was mixed with pH 8 phosphate-buffered saline to achieve a final pH of 7.4.

FPLC analysis of protein and protein loaded nanoparticles

Fast protein liquid chromatography (FPLC) analysis was conducted using an AKTA Pure chromatography system equipped with a Superdex 200 Increase 10/300 GL column. Phosphate-buffered saline (PBS) was used as the mobile phase at a flow rate of 0.75 mL/min.

ELISA assay for detection of free and total IL-2-Fc

Flat-bottomed 96-well plates (Corning) were precoated with 100 μL/well of rat anti-human IL-2 antibody (2 μg/mL in PBS) and incubated overnight at 4°C. After washing with PBS-T (PBS +0.05% Tween), plates were blocked with 200 μL/well of 2% BSA in PBS for free-protein analysis or IP Lysis Buffer for total-protein analysis at room temperature (RT) for 2 h. Serum samples were diluted to 5–20 ng/mL in 1% BSA in the appropriate buffer (BSA+ PBS for free and BSA+IP for total). Biotin-conjugated goat anti-human IL-2 detection antibody (0.1 μg/mL in 1% BSA) was added at 100 μL/well, followed by Streptavidin-HRP (0.4 μg/mL in 1% BSA, 100 μL/well). The reaction was developed with TMB-ELISA substrate (100 μL/well) for 10–15 min in the dark and stopped with 100 μL/well of 4N sulfuric acid. Optical densities were measured at 450 nm (signal) and 570 nm (background).

Characterization OF PH-ACTIVATABLE UPS/IL-2-Fc nanoparticles

To evaluate the acidity threshold response, we performed titrations of UPS/IL-2-Fc and measured size distribution, IL-2-Fc release, and activity in IL-2 reporter cell lines. Briefly, a UPS/IL-2-Fc NP solution (pH 7.4, 150 mM NaCl) was incrementally acidified with HCl. After each HCl addition, an aliquot was transferred to a separate tube.

For pH-dependent nanoparticle dissociation, the size of each aliquot was measured directly using dynamic light scattering. To assess pH-dependent IL-2-Fc release, each aliquot was mixed with 1% BSA in PBS at the corresponding pH to stabilize dissociated unimers, then adjusted to neutral pH for ELISA analysis. For pH-dependent IL-2 activity, neutralized UPS/IL-2-Fc aliquots were prepared as described above, and activity was tested using HEK-Blue reporter cells according to the manufacturer’s protocol.

In vitro stability of IL-2-Fc encapsulation IN UPS nanoparticles

UPS/IL-2-Fc was added to each environment (PBS, DMEM +10% FBS, and mouse plasma), aliquoted, and incubated at 37°C. At predetermined time points, the free and total concentrations of IL-2-Fc were measured using the ELISA method described above.

Flow cytometry analysis

Single cell suspensions were obtained from cell culture. Fc receptor was blocked with anti-FcγIII/II (clone 2.4G2) for 15 min, followed by staining with selective antibodies of cell surface markers and live/dead dyes. Intracellular markers was stained after cell permeabilization with True-Nuclear transcription factor buffer set (BioLegend). Data were collected on Beckman CytoFLEX flow cytometer and analyzed by CytoExpert (Beckman coulter).

Tumor growth

C57BL/6J mice were inoculated with 1.5 × 10⁶ MC38 or 1.5 × 10⁵ B16F10 tumor cells on the right flank. Balb/c mice were inoculated with 1 ×10⁶ 4T1 tumor cells in the fourth, right mammary fat pad. Tumor volumes were measured with a caliper by the length (L) and width (W) and calculated as tumor volume = L×W×W/2. Tumors with volume 50–100 mm³ were used for in vivo experiments unless otherwise specified.

Molecular docking of IL-2 with DBA dimer

Molecular docking of the DBA dimer to IL-2 was performed using AutoDock Vina. The IL-2 structure (PDB ID: 1m47) was obtained from the Protein DataBank, and the DBA ligand was prepared using ChemOffice 2013. The docking space was defined as a cubic region centered at (6.75, 25.891, 14.189) with dimensions (47.5, 47.5, 47.5). Within this space, 216 small cubes of size (15, 15, 15) were generated with a 50% overlap, which served as potential binding pockets. All binding pockets were screened, and the top 10 docking results across all pockets were selected for the final docking figure. The IL-2 receptor-binding surface was labeled based on previously reported cryo-EM structures.

EX VIVO binding of IL-2-Fc to different lymphocytes

MC-38 tumor-bearing mice were euthanized, and single-cell suspensions were prepared from the liver, spleen, lung, and tumor tissues using either physical dissociation with a 70 μm cell strainer or enzymatic digestion with collagenase IV and DNase I. Cells were then incubated with different IL-2-Fc formulations at their original concentration or treated with low pH, followed by BSA binding, and returned to neutral pH at 4°C for 30 min. After incubation, cells were stained with lymphocyte markers and analyzed by flow cytometry.

Cytokine release

Systemic cytokine level was measured using either mouse CBA kit for mouse cytokine or Human multiplex cytokine assay according the manufacturer’s manual.

PET-CT imaging

MC-38 tumor bearing mice was created with subcutaneous injection of MC38 cells. After tumor size reaches 50–100 mm³, mice were injected with ⁶⁴Cu labeled IL-2-Fc or UPS_5.3/IL-2-Fc NPs. The in vivo images were taken by a Siemens Inveon PET/CT Multi-Modality System. The ex vivo quantification of IL-2-Fc in each organ was performed on a PerkinElmer 2480 wizard2 3″ automatic gamma counter.

Humanized mouse models

Two humanized mouse models were used in this experiment: the HuPBMC engraftment model and the HuCD34⁺ cell engraftment model. Briefly, NOD-SCID mice were irradiated 100 cGy (X-ray irradiation using an X-RAD 320 irradiator) prior to engraftment. From huPBMC mouse model, huPBMC was injected intravenously to irradiated mouse. For the HuCD34⁺ model, 1×10⁵ CD34⁺ hematopoietic stem cells were intravenously injected into recipient mice. Twelve weeks after engraftment, humanized mice with over 30% reconstitution of human CD45⁺ cells were used for tumor studies.

Quantification and statistical analysis

Data were analyzed using GraphPad Prism 8.3. One-way Analysis of variance (ANOVA) or unpaired two-tailed t tests were used to analyze data unless specified. p < 0.05 was considered statistically significant. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. n.s. not significant.

Published: January 20, 2026