Radiopaque microsphere-hydrogel composite for extended-release intratumoral immunotherapy in a large animal model

35 min read Original article ↗

Abstract

Intratumoral injection of immune stimulants could potentially reverse the immunosuppressive tumor microenvironment, but the injected drug rapidly washes out into the systemic circulation. This results in systemic toxicity, and reduced efficacy, potentially requiring daily injections to generate an antitumor immune response. To address this problem, we developed an extended-release hydrogel that retains the immunotherapy agent at the injection site, and releases it over several days, even in highly vascular liver tumors. The extended-release hydrogel is a composite material containing drug-loaded microspheres (for sustained, tunable drug release), embedded in a radiopaque cross-linked hydrogel (for local retention). Using these new intratumoral drug carriers, we demonstrate the ability to safely deliver a wide range of different immunotherapy agents (cytokines, oligonucleotides, small molecules) into pig liver tumors, with no acute systemic toxicity. The microsphere gel reduces burst release by 7.8x (compared to microspheres), and increases half-life by > 15x (compared to hydrogel). Thus, the immunotherapy-loaded microsphere gel enables sustained intratumoral delivery of immune stimulants from a single injection, with low systemic toxicity.

Similar content being viewed by others

Introduction

Intratumoral injection of cytotoxic agents [1, 2], immune stimulants [3,4,5], oncolytic viruses [6], and cellular therapies [7,8,9,10] can induce tumor necrosis, as well as antitumor immunity, with regression of uninjected tumors [11,12,13,14,15,16]. However, existing intratumoral drug delivery methods are limited by off target leakage, high burst release, and short drug half-life, resulting in systemic toxicity, and limited intratumoral efficacy. Intratumoral injection into vascular tumors, such as liver metastases, effectively results in an intravenous bolus of drug, because the needle tip can be partially within an intratumoral blood vessel. In addition, given the short half-life of intratumoral injection, daily injections can be required to achieve a good response [14,15,16,17,18]. Repeated intratumoral injections are painful for patients, and they can increase the risk of bleeding, tract seeding, and metastasis [19].

Intratumoral injections frequently leak outside the tumor (along the needle tract, or into intratumoral veins), and this leakage can be reduced by using injectable hydrogels [20, 21]. Hydrogels can also increase the half-life of drug release into tumor. A wide range of injectable hydrogels have been proposed, but there is limited comparison data available for optimal hydrogel selection. Multisidehole needles and catheters can also be used to improve drug delivery into tumors [22, 23].

Drug-loaded microparticles have also been used for intratumoral injection [19, 24], but the microparticles can also leak outside the tumor. Larger microparticles have better intratumoral retention [24], but larger microparticles are more likely to clog the injection needle [25]. Burst release from drug-loaded PLGA microspheres remains an unsolved problem [26].

Here, we present the design, evaluation, and optimization of intratumoral drug carriers that retain the drug at the injection site, and then release the drug over several days, even in highly vascular liver tumors (Table 1). We compare the targeted drug delivery performance of a wide range of different drug carriers. Drug-eluting microspheres embedded in an injectable cross-linked hydrogel (Fig. 1) demonstrated the best results. Retention at the injection site is achieved using a high-molecular-weight cross-linked hydrogel. Extended drug release is achieved using drug-eluting microspheres embedded in the hydrogel. Injectability of the viscous hydrogel is improved by aligning the polymer chains along the long axis of the syringe.

Using these new intratumoral drug carriers, we demonstrate the ability to deliver a wide range of different immunotherapy agents (cytokines, oligonucleotides, small molecules) into liver tumors, with tunable half-life. We evaluate injectability, burst release, biodistribution, half-life, and safety.

Table 1 Criteria for evaluating drug carriers for intratumoral injection

Full size table

Fig. 1

Drug-eluting microsphere gel. Negatively charged drug is loaded onto positively charged quaternary amine microspheres. High molecular weight HA is cross linked with BDDE (1,4-butanediol diglycidyl ether), to improve retention. Cross-linked polymer chains are aligned to improve injectability. The drug-loaded microspheres are then embedded in the cross-linked gel. (Figure created in BioRender.)

Methods

Drug-eluting hydrogels

Hyaluronic acid (HA)

Iohexol 276 mg/ml was dissolved in hot water. Then, 4% (w/v) hyaluronic acid (1.5–2.2 MDa, Thermo Scientific Chemicals) was added, and dissolved overnight at room temperature.

Cross-linked hyaluronic acid (HA)

5% (w/v) hyaluronic acid (1.5–2.2 MDa) was dissolved in 1 N NaOH, final volume 5 ml. 200 µl of BDDE (1,4-butanediol diglycidyl ether, Fisher Scientific) was added, and cross linking was performed in the dark at room temperature for 12–18 h (Supplemental Fig. 1). The cross-linked HA was dialyzed (7 kDa MWCO) into PBS (for cytokines) or deionized water (for other drugs) for 6 h. The cross-linked HA was pumped through a 3-way stopcock 200 times to improve injectability. Drug or drug-loaded microspheres were then mixed into the cross-linked HA, by pumping through the 3-way stopcock. Iohexol was added to make the gel radiopaque.

Poloxamer 407 / hydroxypropylmethylcellulose (HPMC)

Iohexol 276 mg/ml was dissolved in hot water. Then, 22% (w/v) poloxamer 407 (Sigma Aldrich) and 2% (w/v) HPMC (Thermo Scientific Chemicals) was added, and dissolved overnight at 4 °C.

Gelatin / gellan gum

3.5% (w/v) gelatin (type A, 300 Bloom, Sigma Aldrich) + 1% (w/v) gellan gum (Thermo Scientific Chemicals) + iohexol (Thermo Scientific Chemicals) 276 mg/ml were dissolved in water.

Drug-eluting microspheres

Loading CpG-STAT3ASOCy3 onto quaternary amine microspheres

CpG-STAT3ASOCy3 was synthesized in the City of Hope DNA/RNA synthesis core facility. CpG-STAT3ASOCy3 was diluted to 0.5 mg/ml in TE buffer, and 1 ml of the solution was added to 120 µl of positively charged quaternary amine microspheres (Q Sepharose, 45–165 μm, Cytiva). Loading was performed for 10 h at room temperature, with periodic vortexing. Loading efficiency was measured by examining the change in fluorescence of the supernatant (see Drug assays section below).

Loading IL-2 onto quaternary amine microspheres

IL-2 (R&D systems, 0.1 µg in 1 ml PBS) was loaded onto 0.1 ml microspheres at room temperature for 2 h, with periodic gentle mixing. Loading efficiency was measured by examining the change in drug concentration in the supernatant (see Drug assays section below).

Loading IFN-γ onto quaternary amine microspheres

IFN-γ (R&D systems, 2 µg in 1 ml PBS) was loaded onto 0.12 ml microspheres at room temperature for 2 h, with periodic gentle mixing. Loading efficiency was measured by examining the change in drug concentration in the supernatant (see Drug assays section below).

Loading 3’,3’-cGAMP onto quaternary amine microspheres

3’,3’-cGAMP (Invivogen, 0.5 mg in 1 ml water) was loaded onto 0.12 ml microspheres at room temperature for 4 h, with periodic mixing. Loading efficiency was measured by examining the change in drug concentration in the supernatant (see Drug assays section below).

Loading erythrosine onto quaternary amine microspheres

Erythrosine (Thermo Scientific Chemicals) was loaded onto microspheres at room temperature. Erythrosine (50 mg in 25 ml water) was loaded onto 2 ml microspheres at room temperature, on a shaker. The supernatant was removed when it turned clear, and more erythrosine (50 mg in 25 ml water) was added. Again, the supernatant was removed when it turned clear, and more erythrosine (25 mg in 25 ml water) was added. Loading efficiency was measured by examining the change in drug concentration in the supernatant (see Drug assays section below).

Loading exosomes onto quaternary amine microspheres

Milk exosomes (CD Bioparticles) in saline were loaded onto microspheres at 4°C. Loading efficiency was measured by examining the change in exosome concentration in the supernatant (see Drug assays section below).

Drug assays

CpG-STAT3ASO (Cy3-labeled) concentration was measured by fluorescence (excite 535 nm, emit 595 nm) or quantitative fluorescence microscopy (excite 550 nm, emit 605 nm). IL-2 and IFN-γ were measured by ELISA. Cross-linked HA was dissolved using hyaluronidase (1 mg/ml), prior to running ELISA. Erythrosine was measured by absorbance at 525 nm. Iohexol was measured by absorbance at 245 nm. 3’,3’-cGAMP was measured by absorbance at 260 nm. Monomethyl auristatin E (MMAE, MedChemExpress) was measured by absorbance at 220 nm. Exosomes were measured by absorbance at 280 nm.

Injectability, rheometry, and gel stability

Hydrogel was placed in a 3 ml syringe, attached to an 18 G × 15 cm Chiba needle, and 39 N of force was applied to the plunger. Injection time (sec/ml) was measured. In addition, rheometry was performed at 37 °C, using a 40 mm cone and plate with peltier plate system (Associated Polymer Labs, Hudson Falls, NY).

Gel stability was measured by injecting 1 ml gel through an 18 G needle into a 15 ml conical tube, adding 7 ml PBS + 34.6% (v/v) glycerol, then shaking at 37 °C at 80 rpm. The supernatant was removed and replaced at multiple time points, and the gel weight was determined by weighing the gel in the tube, and subtracting the previously measured weight of the empty tube. The time required to dissolve 50% of the gel weight was measured.

In vitro drug release half-life

We evaluated in vitro drug release kinetics of various drugs (iohexol, erythrosine, IL-2, IFN-gamma, CpG-STAT3ASO, 3’,3’-cGAMP, MMAE) loaded onto various drug carriers (water, hydrogels, microspheres, or microsphere gels). 0.5 ml of drug-loaded hydrogel or microspheres was eluted into 14 ml of PBS + 34.6% (v/v) glycerol (to match the viscosity of blood), using a dialysis cup, at 37°C and 80 rpm [27]. Eluted drug was measured at multiple time points (see Drug assays section above). Half lives were compared using t-tests, and p < 0.05 was considered statistically significant.

Animals

The Institutional Animal Care and Use Committee at the Lundquist Institute approved all research procedures. All experiments were performed in compliance with relevant guidelines and regulations. All procedures and imaging were performed under general endotracheal anesthesia, using isoflurane. Euthanasia was performed by administering a barbiturate overdose intravenously, as per AVMA guidelines.

Iohexol and erythrosine loaded onto 7 different drug carriers (Table 2) were injected into the livers of Yorkshire pigs (Premier BioSource, Ramona, CA), under ultrasound guidance. Each formulation was injected into 3 different sites in the liver, for a total of 21 injection sites, in a total of 9 pigs (78% male, average age 86 days). 1 ml was injected into each site, over 15 s, as the needle was pulled back 2 cm. 21 G × 15 cm Chiba needle was used for injecting drugs in water, poloxamer 407 / HPMC, and HA (non-cross-linked). 18 G × 15 cm Chiba needles were used for all other drug carriers. Non-contrast CT was performed at multiple time points between 2 min and 2 days post injection, to evaluate burst release, biodistribution, and half-life.

Table 2 Biodistribution, burst release, and half-life of different drug carriers after percutaneous injection of drug into pig liver

Full size table

Liver tumors (undifferentiated carcinomas) were induced in Oncopigs (Sus Clinicals, Chicago), as previously described [28, 29]. Drug-loaded cross-linked hyaluronic acid hydrogels (containing IL-2, IFN-gamma, CpG-STAT3ASO, and MMAE) were injected into Oncopig liver tumors, under ultrasound guidance. Each formulation was injected into 2–3 different liver tumors, for a total of 10 injected liver tumors, in a total of 4 Oncopigs (75% male, average age 94 days). 1 ml of drug was injected into each tumor, over 15 s, as the needle was pulled back through the tumor. CT and necropsy was performed 1 week later, and injection sites were harvested to evaluate the remaining drug concentration at the injection sites.

Biodistribution

Biodistribution of radiopaque gel or microparticles (containing iohexol or erythrosine) was evaluated on non-contrast CT obtained at multiple time points between 2 min and 2 days after injection. Retention at the injection site in the liver was evaluated, as well as off-target delivery (to portal vein, hepatic vein, bile duct, peritoneum, or lung).

In vivo burst release

Burst release is the fraction of drug that is released systemically immediately after injection. Intravenous injection has 100% burst release.

For iohexol and erythrosine, burst release was measured on non-contrast CT obtained 2 min after injection. A region of interest (ROI) was drawn around the injection site, and the average Hounsfield units and volume of the ROI was used to calculate the amount of iohexol or erythrosine retained at the injection site, as previously described [23]. Burst release is 100% – % drug retained at the injection site at 2 min.

For IL-2, IFN-gamma, and CpG-STAT3ASO, burst release was calculated from the plasma concentration of drug, before and after injection. Specifically:

$$\begin{array}{l}\text{Plasma drug concentration after injection (mg/ml)} \\\quad= \text{burst release} (\%)\\ \quad \times \text{drug injected (mg)} / \text{plasma volume (ml)}.\end{array}$$

$$\begin{array}{l}\text{Plasma volume (ml)}=\text{pig weight (kg)} \times 65\, \text{ml}/\text{kg} \\ \quad \times (100\% - \text{hematocrit}).\end{array}$$

In vivo drug release half-life

For iohexol and erythrosine, half-life was measured on non-contrast CT obtained at multiple time points between 2 min and 2 days after injection. For IL-2, IFN-gamma, and CpG-STAT3ASO, intratumoral half-life was calculated from the amount of residual drug at the injection site on necropsy, 1 week after injection.

Safety

Safety was evaluated based on labs (CBC, comprehensive metabolic panel), imaging (multiphase-contrast enhanced CT of the abdomen and pelvis), daily clinical evaluation, and vital signs. Imaging and labs were obtained before injection, and 1 week after injection.

Results

Immunotherapy-loaded microspheres

In order to load negatively charged drugs, such as CpG or STING agonists, we used positively charged quaternary amine microspheres (Fig. 2). We demonstrated loading of oligonucleotides (CpG-STAT3ASO), STING agonists (3’,3’-cGAMP), cytokines (IL-2 and IFN-gamma), and exosomes onto these positively charged microspheres (Table 3). Loading was measured based on change in drug concentration in the supernatant, and was confirmed by fluorescence microscopy. These drug-loaded microspheres are designed for both percutaneous and intra-arterial delivery into tumors.

Fig. 2

Immunotherapy-loaded microspheres. Positively charged quaternary amine microspheres can be loaded with negatively charged local immunotherapy agents (such as TLR9 or STING agonists), as well as other therapeutic agents. A. Scanning electron microscopy (SEM) of positively charged quaternary amine microspheres (45–165 μm). B. Photo of packed quaternary amine microspheres (white). C. CpG-STAT3ASOCy3 (fluorescently labeled using Cy3; pink). D. After loading, the supernatant becomes clear, and the settled microspheres become pink, due to CpG-STAT3ASOCy3 loading onto microspheres. E. Fluorescence microscopy shows loading of fluorescently labeled drug on the surface of the microspheres

Table 3 Drug loading onto quaternary amine microspheres

Full size table

Stable and injectable radiopaque hydrogels

157 different gel formulations were evaluated in vitro for injectability, and gel stability after injection. Gel strength is important for retention in vascular tumors, but strong gels can be difficult to inject. The goal is to increase gel stability, while maintaining injectability. Hyaluronic acid (HA), poloxamer 407, gelatin, hydroxypropylmethylcellulose (HPMC), carboxymethylcellulose, methylcellulose, gellan gum, and alginate derived gels were tested. These initial experiments (Supplemental Fig. 2 and Supplemental Table 1) showed that: (A) HPMC improves the stability of poloxamer gels; (B) Gellan gum improves the stability of gelatin gels; and (C) Cross linking (using BDDE) improves the stability of HA gels.

Based on these initial experiments, 4 injectable formulations were selected for further evaluation: (A) 22% poloxamer 407 + 2% HPMC; (B) 3.5% gelatin + 1% gellan gum; (C) 4% hyaluronic acid; and (D) 5% cross-linked hyaluronic acid. Poloxamer 407 / HPMC is thermoresponsive, forming a gel above 20°C, and liquifying below 20°C. The other formulations are gels at both room and body temperature.

Iohexol was added to gels to make them visible on CT. Biodistribution experiments used a high concentration of iohexol (276 mg/ml), measuring 2233 Hounsfield units at 120 kVp, to maximize sensitivity. Drug delivery experiments used a lower concentration of iohexol (50 mg/ml), measuring 405 Hounsfield units, which is adequate for image guidance.

Improving injectability and rheometry

High molecular weight cross-linked hyaluronic acid can be difficult to inject, using standard needles and syringes. Injectability of cross-linked hyaluronic acid can be improved by pumping the gel 200 times through a 3-way stopcock (Fig. 3A). This aligns the polymer chains, resulting in more than 10-fold improvement in injection time through an 18 G needle (Fig. 3B), while preserving gel stability (Supplemental Table 1). Alignment of the polymer chains was demonstrated on scanning electron microscopy (SEM).

Drug-loaded microspheres can be embedded in the cross-linked hyaluronic acid. The drug-loaded microsphere gel flows like a liquid during injection, then solidifies into a well-defined drug depot after percutaneous injection into pig liver (Fig. 3C). Yield stress, which is the minimum stress needed to initiate flow, was measured by rheometry (Fig. 3D).

Fig. 3

Injectability and rheometry of cross-linked hyaluronic acid. A. Pumping cross-linked HA through a 3-way stopcock aligns the polymer chains. B. Alignment of polymer chains (shown on SEM images) results in improved hydrogel injectability. (scale bars: 10 μm) C. Drug-loaded microspheres in cross-linked hyaluronic acid gel can be injected through an 18 G needle. Percutaneous injection of the microsphere gel into pig liver forms an adhesive radiopaque drug depot (arrows; photo and CT). D. The microparticle gel is a Bingham plastic, which thins into an injectable liquid when force is applied, and then solidifies below the yield stress. Rheometry shows that the microparticle gel has improved yield stress, compared to microparticles alone.

Optimizing local drug delivery in pig liver, using radiopaque tracers

Compared to humans, mice have 60 times faster drug diffusion, 7 times faster blood perfusion, and typically require 12 times the drug dosage (per kg) [30,31,32]. Pharmacokinetics in rodents is not easily translatable to humans, resulting in failed clinical trials [33]. Therefore, we decided to evaluate drug delivery in pigs, which have similar physiology [30, 31] and drug dosing [32], compared to humans.

Iohexol and erythrosine were used as radiopaque iodinated tracers, to evaluate biodistribution, burst release, and drug release kinetics on CT. Iohexol is a surrogate for neutral, water soluble, small molecule drugs. Erythrosine is a surrogate for negatively charged, water soluble, small molecule drugs.

Injecting iohexol dissolved in water into pig liver results in leakage into portal vein, peritoneum, and bile ducts (Fig. 4). Only 37% of injected iohexol is retained at the injection site, with a half-life of 3.9 min (Table 2).

Drug retention at the injection site in liver can be improved by delivering drug in hydrogel (Table 2). Of the 4 hydrogels tested, only cross-linked HA was retained at the injection site in liver (Fig. 4) for at least 2 days, with no off-target delivery, and low burst release. Half-life of drug release was increased to 5.3 h, and burst release was reduced to 9%. Gelatin / gellan gum and HA (not cross-linked) were retained at the injection site in the liver for at least 2 days, but they also leaked into off-target sites (veins, peritoneum, or bile ducts). Poloxamer 407 / HPMC demonstrated no off-target leakage, but it was not retained at the injection site.

To further increase drug half-life, we tested drug-loaded microspheres, using a radiopaque tracer (erythrosine). Unfortunately, when injected into pig liver, the erythrosine-loaded microspheres immediately leaked into portal vein, peritoneum, and bile ducts, resulting in 93% burst release.

Based on these results, we embedded erythrosine-loaded microspheres into cross-linked hyaluronic acid, resulting in low burst release, as well as an extended half-life of 6.3 days (Table 2, Supplemental Fig. 3). The drug-loaded microsphere gel is a composite material that combines the good local retention properties of the hydrogel, with the favorable drug binding and release properties of the microspheres.

To summarize, drug in water has high burst release, and short half-life, when injected into pig liver. Drug on microspheres have high burst release, but long half-life. Drug in hydrogel has low burst release, and intermediate half-life. Drug in the microsphere gel combines the low burst release of hydrogels, with the long half-life of microspheres.

Fig. 4

High resolution biodistribution on CT. Iohexol in various drug carriers was injected into pig liver (1 ml), and biodistribution was evaluated on CT at multiple time points (2 min to 2 days after injection). Cross-linked HA was retained at the injection site (circles). Water, gelatin / gellan gum, and non-cross-linked HA leaked outside the injection site (arrows), into portal vein, hepatic vein, bile duct, or peritoneum. See Table 2

Intratumoral delivery of cytokines, CpG-STAT3ASO, and MMAE

Cross-linked HA, and drug-loaded microspheres in cross-linked HA were the best drug carriers for delivering radiopaque tracers into pig liver. Next, we evaluated using these drug carriers for targeted delivery of cytokines (IL-2, IFN-gamma), immunostimulatory oligonucleotides (CpG-STAT3ASO), and immunogenic cell death agents (MMAE) into Oncopig liver tumors.

Drugs loaded onto cross-linked HA, or microspheres in cross-linked HA, were injected into pig liver tumors (Supplemental Fig. 4). Drug half-life was measured based on residual drug at the injection site on necropsy, 1 week after injection. Half-life of these immunotherapy agents in vascular liver tumors ranged from 21 h to 5.8 days (Table 4). Burst release was < 1%.

Safety of intratumoral drug delivery was evaluated based on clinical, imaging, and laboratory evaluation. No serious or severe adverse events were seen, per NCI Common Terminology Criteria for Adverse Events (CTCAE) Version 5.0. Specifically, there were no immune-related adverse events, cytokine release syndrome, or infections, and liver function tests remained in the normal range after intratumoral injection. Drug dosage was 6 mcg for IL-2, 100 mcg for IFN-gamma, 1.5 mg for CpG-STAT3ASO, and 0.4 mg for MMAE. Pig weight ranged from 20 to 29 kg.

Multiple drugs can be released at different time points, from a single injection. For example, we loaded CpG-STAT3ASO-microspheres into cross-linked HA gel containing MMAE, resulting in MMAE release with a half-life of 8 h (estimated based on Table 4 and the regression line in Fig. 5), and CpG-STAT3ASO release with a half-life of 140 h (measured).

Table 4 Drug release half-life in vitro versus in pig liver, for a wide range of drugs and drug carriers

Full size table

In vitro assay to predict in vivo drug release and biodistribution

Iohexol (water soluble small molecule) in hydrogel has a half-life of 3–5 h in pig liver. Cytokines (IL-2, IFN-gamma) in cross-linked HA have a half-life of 21–23 h. Negatively charged drugs (CpG-STAT3ASO, erythrosine) loaded onto the microsphere gel have a half-life of around 6 days.

Drug release is always faster in pig liver, compared to in vitro experiments (Table 4; Fig. 5). Drug carriers with minimal drug release in vitro, show substantial drug release in vivo. This is due to active drug release mechanisms only seen in vivo, such as blood flow, and enzymatic and cellular degradation. However, there is a moderate correlation between in vitro and in vivo half-life, with r2 = 0.42. Specifically, in vitro experiments can accurately classify drug carriers into fast (< 30 min), medium (30 min–10 h), and slow (> 10 h) drug release in pig liver.

Gel stability > 30 days in vitro predicts good biodistribution in pig liver (retention at the injection site, and no leakage to off-target sites), although results should be interpreted cautiously, given the small number of gels tested. See Table 2 and Supplemental Table 1.

Fig. 5

Drug release in vitro and in vivo. A. In vitro experiments show that cross-linked HA hydrogel slows drug release. Very little drug is released from drug-loaded microspheres and microsphere gel, in vitro. Here, erythrosine, which is negatively charged, water soluble, and radiopaque, is used as the model drug. B. Correlation between in vitro vs. in vivo drug release kinetics, for a wide range of drugs and drug carriers (Table 4). Dashed lines show the cutoffs for fast, medium, and slow drug release

Discussion

The liver is immunosuppressed, compared to other organs, limiting the effectiveness of checkpoint inhibitors in patients with liver metastases [34, 35]. Intratumoral injection of immune stimulants could potentially reverse the immunosuppressive tumor microenvironment, but the injected drug rapidly washes out into the systemic circulation. This results in systemic toxicity, and reduced efficacy, requiring daily injections to generate potent antitumor immune responses.

To address this problem, we systematically evaluated a wide range of different drug carriers, including hydrogels and microparticles, for image-guided percutaneous delivery of immunotherapy agents into liver tumors. Iohexol dissolved in water rapidly washes out from vascular tumors, with 63% burst release, and 3.9 min half-life. Delivering drug in hydrogel reduces the burst release to < 15%, and increases half-life to 3–23 h, depending on the drug and the gel. Half-life can be further increased to 6 days, using drug-loaded microspheres in hydrogel.

The microsphere gel is a composite material that combines the favorable drug binding and release properties of drug-eluting microspheres, with the good local retention properties of hydrogels. Electrostatic binding between negatively charged drugs, and positively charged microspheres, translates into sustained drug release. The cross-linked hydrogel forms a cohesive drug depot at the injection site, preventing immediate washout into intratumoral veins, resulting in low burst release. Injectability of the high molecular weight cross-linked hyaluronic acid was improved by pumping the gel through a 3-way stopcock, which aligns the polymer chains, resulting in more than 10-fold improvement in injection time. The drug-loaded microsphere gel flows like a liquid during injection, then solidifies into a well-defined drug depot after percutaneous injection into vascular tumors.

A wide range of therapeutic payloads can be delivered into tumors using the microsphere gel, including DNA and RNA therapeutics [36] (CpG-STAT3ASO), STING agonists (3’-3’-cGAMP), cytokines (IL-2 and IFN-gamma), immunogenic cell death agents (MMAE), and exosomes. Sustained intratumoral drug release can be achieved from a single injection. Retention at the injection site, with low burst release, translates into low systemic toxicity. The radiopaque gel allows easy evaluation of drug distribution on CT.

Toxic drugs can be safely delivered into vascular tumors, using the microsphere gel, given the low burst release and long half-life. No adverse events were seen after injection of cytokines (IL-2, IFN-gamma), oligonucleotides (CpG-STAT3ASO), or immunogenic cell death agents (MMAE) into pig liver tumors. Drug that are too toxic to be given systemically, could potentially be safe and effective when delivered locally. Drug targeting is achieved using image guidance. The gel and drug stay at the injection site, allowing the interventional radiologist to “paint” tumors with drug.

Multiple drugs can be released at different time points, from a single injection. Here, we demonstrate release of an immunogenic cell death agent (MMAE) with an estimated half-life of 8 h, followed by release of an immune stimulant (CpG-STAT3ASO) with a half-life of 140 h. The modular design of the microsphere gel, with separation of local retention (gel) and drug release (microspheres), enables each drug to have separately tuned release kinetics. This programmable time- and location-dependent drug delivery system could be used to provide the right stimulus to the immune system at the right time. Here, we release the cell death agent first, followed by STAT3 inhibition to augment antigen presentation.

Many drug-eluting microspheres, such as Sandostatin LAR depot, are FDA-approved for subcutaneous or intramuscular injection, to provide sustained systemic drug release [37]. However, these microspheres are not suitable for percutaneous delivery into vascular tumors. In our experiments, injection of drug-loaded microspheres into pig liver resulted in only 7% retention at the injection site.

Approved drug-loadable microspheres (LC beads, Boston Scientific), which are supplied without drug, are negatively charged, and can only be loaded with positively charged drugs, such as doxorubicin. There are no FDA-approved positively charged drug-loadable microspheres. Thus, we used positively charged quaternary amine microspheres, which can be loaded with a wide range of negatively charged drugs (oligonucleotides, STING agonists), as well as exosomes, and protein drugs. This expands the range of drugs that can be loaded onto microspheres.

Intratumoral drug delivery has been performed using various injectable hydrogels, including thermoresponsive gels [11, 38,39,40,41], multidomain peptides [22], nanosilicate gels [42], alginate [43], chitosan [44, 45], and hyaluronic acid [46,47,48]. Here, we used an iterative design approach, starting with in vitro evaluation of 157 different gel formulations, and ultimately testing in vivo drug delivery using 4 different hydrogels. Only cross-linked HA was retained at the injection site in liver, with no off-target delivery, and low burst release. Hyaluronic acid is currently FDA-approved for intra-articular injection for osteoarthritis, and as an injectable filler for cosmetic use.

This study has several limitations. First, intratumoral half-life of IL-2, IFN-gamma, and CpG-STAT3ASO was estimated from a single time point: the amount of residual drug at the injection site on necropsy, 1 week after injection. Additional time points will be needed to confirm these estimated half-lives. The half-life of erythrosine in the microsphere gel was longer than expected, and it was extrapolated from the 2-day post injection time point. Second, therapeutic efficacy and longer term safety outcomes were not evaluated in this study. Third, manufacturability, stability and storage, GLP toxicology, and regulatory issues will need to be addressed prior to translation to human trials.

In conclusion, drug delivery into vascular liver tumors can be improved by using an extended-release microsphere gel, which flows like a liquid during injection, then solidifies into an intratumoral drug depot (Fig. 6). The microsphere gel is a composite material containing drug-loaded microspheres (for sustained, tunable drug release), embedded in a radiopaque cross-linked hydrogel (for local retention). A wide range of therapeutic payloads can be safely delivered into tumors, including oligonucleotides, STING agonists, cytokines, immunogenic cell death agents, and exosomes. The microsphere gel reduces burst release by 7.8x (compared to microspheres), and increases half-life by > 15x (compared to hydrogel), thus enabling sustained intratumoral drug release from a single injection, with low systemic toxicity.

Fig. 6

Improved intratumoral drug retention using a microsphere gel. When drug is injected percutaneously into a liver tumor, the drug rapidly washes out via the hepatic and portal veins (top row). When drug is loaded onto a microsphere gel, it is retained at the injection site for several days (bottom row). The cross-linked gel sticks together and prevents burst release into a vein, and the embedded microspheres bind the drug and provide sustained release.

Data availability

Data will be made available on request.

References

  1. Kuang M, Lu MD, Xie XY, Xu HX, Xu ZF, Liu GJ, et al., Lencioni R. Ethanol ablation of hepatocellular carcinoma Up to 5.0 cm by using a multipronged injection needle with high-dose strategy. Radiology. 2009;253:552–61.

    Article  PubMed  Google Scholar 

  2. Nief C, Morhard R, Chelales E, Adrianzen Alvarez D, Bourla Bs I, et al. Polymer-assisted intratumoral delivery of ethanol: Preclinical investigation of safety and efficacy in a murine breast cancer model. PLoS ONE. 2021;16:e0234535.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  3. Shyr CR, Liu LC, Chien HS, Huang CP. Immunotherapeutic agents for intratumoral immunotherapy, Vaccines (Basel), 11 (2023).

  4. Murthy R, Sheth R, Tam A, Gupta S, Subbiah V, Janku F, et al. 397 Intra-tumor immunotherapy injections utilizing image guidance in interventional radiology: clinical trial experience at a tertiary care cancer center. J Immunother Cancer. 2020;8:A241–2.

    Google Scholar 

  5. Gong N, Alameh MG, El-Mayta R, Xue L, Weissman D, Mitchell MJ. Enhancing in situ cancer vaccines using delivery technologies. Nat Rev Drug Discov. 2024;23:607–25.

    Article  CAS  PubMed  Google Scholar 

  6. Shalhout SZ, Miller DM, Emerick KS, Kaufman HL. Therapy with oncolytic viruses: progress and challenges. Nat Rev Clin Oncol. 2023;20:160–77.

    Article  PubMed  Google Scholar 

  7. Gu X, Zhang Y, Zhou W, Wang F, Yan F, Gao H, et al. Infusion and delivery strategies to maximize the efficacy of CAR-T cell immunotherapy for cancers. Exp Hematol Oncol. 2024;13:70.

    Article  PubMed  PubMed Central  Google Scholar 

  8. Papa S, Adami A, Metoudi M, Beatson R, George MS, Achkova D, et al. Intratumoral pan-ErbB targeted CAR-T for head and neck squamous cell carcinoma: interim analysis of the T4 immunotherapy study. J Immunother Cancer. 2023;11:e007162.

    Article  PubMed  PubMed Central  Google Scholar 

  9. Burger MC, Forster MT, Romanski A, Strassheimer F, Macas J, Zeiner PS, et al. Intracranial injection of natural killer cells engineered with a HER2-targeted chimeric antigen receptor in patients with recurrent glioblastoma. Neuro Oncol. 2023;25:2058–71.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  10. Kitahara M, Mizukoshi E, Terashima T, Nakagawa H, Horii R, Iida N, et al. safety and long-term outcome of intratumoral injection of OK432-stimulated dendritic cells for hepatocellular carcinomas after radiofrequency ablation. Transl Oncol. 2020;13:100777.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Som A, Rosenboom JG, Wehrenberg-Klee E, Chandler A, Ndakwah G, Chen E, et al., Percutaneous intratumoral immunoadjuvant gel increases the abscopal effect of cryoablation for checkpoint inhibitor resistant cancer. Adv Healthc Mater. 2024;13:e2301848.

    Article  PubMed  Google Scholar 

  12. Ghani MA, Bangar A, Yang Y, Jung E, Sauceda C, Mandt T, et al. Treatment of hepatocellular carcinoma by multimodal in situ vaccination using cryoablation and a plant virus immunostimulant. J Vasc Interv Radiol. 2023;34:1247–e12571248.

    Article  PubMed  PubMed Central  Google Scholar 

  13. Zhang Y, Rahman MM, Clark PA, Sriramaneni RN, Havighurst T, Kerr CP, et al. In situ vaccination following intratumoral injection of IL2 and Poly-l-lysine/Iron Oxide/CpG nanoparticles to a radiated tumor site. ACS Nano. 2023;17:10236–51.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  14. Hong WX, Sagiv-Barfi I, Czerwinski DK, Sallets A, Levy R. Neoadjuvant Intratumoral Immunotherapy with TLR9 Activation and Anti-OX40 Antibody Eradicates Metastatic Cancer. Cancer Res. 2022;82:1396–408.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  15. Anfray C, Mainini F, Digifico E, Maeda A, Sironi M, Erreni M, et al. Intratumoral combination therapy with poly(I:C) and resiquimod synergistically triggers tumor-associated macrophages for effective systemic antitumoral immunity. J Immunother Cancer. 2021;9:e002408.

    Article  PubMed  PubMed Central  Google Scholar 

  16. Chu Y, Li R, Qian L, Liu F, Xu R, Meng F, et al. Tumor eradicated by combination of imiquimod and OX40 agonist for in situ vaccination. Cancer Sci. 2021;112:4490–500.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  17. Moreira D, Sampath S, Won H, White SV, Su YL, Alcantara M, et al. Myeloid cell-targeted STAT3 inhibition sensitizes head and neck cancers to radiotherapy and T cell-mediated immunity. J Clin Invest. 2021;131(2):e002408.

    Article  Google Scholar 

  18. Hammerich L, Marron TU, Upadhyay R, Svensson-Arvelund J, Dhainaut M, Hussein S, et al. Systemic clinical tumor regressions and potentiation of PD1 blockade with in situ vaccination. Nat Med. 2019;25(5):814–24.

    Article  CAS  PubMed  Google Scholar 

  19. Lu X, Miao L, Gao W, Chen Z, McHugh KJ, Sun Y, et al. Engineered PLGA microparticles for long-term, pulsatile release of STING agonist for cancer immunotherapy. Sci Transl Med. 2020;12:eaaz6606.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  20. Sheth R.A., Wehrenberg-Klee E, Patel SP, Brock KK, Fotiadis N, de Baere T. Intratumoral Injection of Immunotherapeutics: State of the Art and Future Directions. Radiology. 2024;312:e232654.

    Article  PubMed  PubMed Central  Google Scholar 

  21. Mikhail AS, Morhard R, Mauda-Havakuk M, Kassin M, Arrichiello A, Wood BJ. Hydrogel drug delivery systems for minimally invasive local immunotherapy of cancer. Adv Drug Deliv Rev. 2023;202:115083.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  22. Munoz NM, Williams M, Dixon K, Dupuis C, McWatters A, Avritscher R, et al. Influence of injection technique, drug formulation and tumor microenvironment on intratumoral immunotherapy delivery and efficacy. J Immunother Cancer. 2021;9:e001800.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Pedersoli F, Mohammad IS, Patel AK, Kessler J, Chao C, Liu B, et al. Bioinspired intratumoral infusion port catheter improves local drug delivery in the liver. Sci Rep. 2024;14:27782.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  24. Bakker RC, Lam M, van Nimwegen SA, Rosenberg A, van Es RJJ, Nijsen JFW. Intratumoral treatment with radioactive beta-emitting microparticles: a systematic review. J Radiat Oncol. 2017;6:323–41.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  25. Sarmadi M, Behrens AM, McHugh KJ, Contreras HTM, Tochka ZL, Lu X, et al. Modeling, design, and machine learning-based framework for optimal injectability of microparticle-based drug formulations. Sci Adv. 2020;6:eabb6594.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Yoo J, Won YY. Phenomenology of the initial burst release of drugs from PLGA microparticles. ACS Biomater Sci Eng. 2020;6:6053–62.

    Article  CAS  PubMed  Google Scholar 

  27. Boas FE, Kemeny NE, Sofocleous CT, Yeh R, Thompson VR, Hsu M, et al. Bronchial or Pulmonary Artery Chemoembolization for Unresectable and Unablatable Lung Metastases: A Phase I Clinical Trial. Radiology. 2021;301:474–84.

    Article  PubMed  PubMed Central  Google Scholar 

  28. Nurili F, Monette S, Michel AO, Bendet A, Basturk O, Askan G, et al. Transarterial embolization of liver cancer in a transgenic Pig model. J Vasc Interv Radiol. 2021;32:510–e517513.

    Article  PubMed  PubMed Central  Google Scholar 

  29. Boas FE, Basturk O, Yin Z, Yarmohammadi H, Elkhadragy L, Chiang J, et al. Undifferentiated carcinomas in a large animal model of liver cancer. Anim Model Exp Med, (2026).

  30. West GB, Brown JH, Enquist BJ. A general model for the origin of allometric scaling laws in biology. Science. 1997;276:122–6.

    Article  CAS  PubMed  Google Scholar 

  31. Schmidt-Nielsen K. Scaling: Why is animal size so important? New York: Cambridge University Press; 1984.

    Book  Google Scholar 

  32. Nair AB, Jacob S. A simple practice guide for dose conversion between animals and human. J Basic Clin Pharm. 2016;7:27–31.

    Article  PubMed  PubMed Central  Google Scholar 

  33. Mak IW, Evaniew N, Ghert M. Lost in translation: animal models and clinical trials in cancer treatment. Am J Transl Res. 2014;6:114–8.

    PubMed  PubMed Central  Google Scholar 

  34. Yu J, Green MD, Li S, Sun Y, Journey SN, Choi JE, et al. Liver metastasis restrains immunotherapy efficacy via macrophage-mediated T cell elimination. Nat Med. 2021;27:152–64.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. Varanasi SK, Chen D, Liu Y, Johnson MA, Miller CM, Ganguly S, et al. Bile acid synthesis impedes tumor-specific T cell responses during liver cancer. Science. 2025;387:192–201.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  36. DeFranciscis V, Amabile G, Kortylewski M. Clinical applications of oligonucleotides for cancer therapy. Mol Ther. 2025;33:2705–18.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  37. Wan B, Bao Q, Burgess D. Long-acting PLGA microspheres: Advances in excipient and product analysis toward improved product understanding. Adv Drug Deliv Rev. 2023;198:114857.

    Article  CAS  PubMed  Google Scholar 

  38. Kim J, Archer PA, Manspeaker MP, Avecilla ARC, Pollack BP, Thomas SN. Sustained release hydrogel for durable locoregional chemoimmunotherapy for BRAF-mutated melanoma. J Control Release. 2023;357:655–68.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  39. Delgado JF, Negussie AH, Varble NA, Mikhail AS, Arrichiello A, Borde T, et al. In vivo imaging and pharmacokinetics of percutaneously injected ultrasound and X-ray imageable thermosensitive hydrogel loaded with doxorubicin versus free drug in swine. PLoS ONE. 2024;19:e0310345.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  40. Seo HW, Kim DY, Kwon DY, Kwon JS, Jin LM, Lee B, et al. Injectable intratumoral hydrogel as 5-fluorouracil drug depot. Biomaterials. 2013;34:2748–57.

    Article  CAS  PubMed  Google Scholar 

  41. Bhatta R, Han J, Liu Y, Bo Y, Wang Y, Nguyen D, et al. Injectable extracellular vesicle hydrogels with tunable viscoelasticity for depot vaccine. Nat Commun. 2025;16:3781.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  42. Albadawi H, Zhang Z, Keum H, Cevik E, Nagalo BM, Gunduz S, et al. Percutaneous delivery of oncogel for targeted liver tumor ablation and controlled release of therapeutics. Adv Mater. 2024;36:e2406080.

    Article  PubMed  PubMed Central  Google Scholar 

  43. Yang X, Huang C, Wang H, Yang K, Huang M, Zhang W, et al. Multifunctional nanoparticle-loaded injectable alginate hydrogels with deep tumor penetration for enhanced chemo-immunotherapy of cancer. ACS Nano. 2024;18:18604–21.

    Article  CAS  PubMed  Google Scholar 

  44. Gierlich P, Donohoe C, Behan K, Kelly D.J., Senge M.O., Gomes-da-Silva L.C. Antitumor Immunity Mediated by Photodynamic Therapy Using Injectable Chitosan Hydrogels for Intratumoral and Sustained Drug Delivery. Biomacromolecules. 2024;25:24–42.

    Article  CAS  PubMed  Google Scholar 

  45. Mei E, Chen C, Li C, Ding X, Chen J, Xi Q, et al. Injectable and Biodegradable Chitosan Hydrogel-Based Drug Depot Contributes to Synergistic Treatment of Tumors. Biomacromolecules. 2021;22:5339–48.

    Article  CAS  PubMed  Google Scholar 

  46. Gao Q, Jiang Y, Li X, Chen H, Tang S, Chen H, et al. Intratumoral injection of anlotinib hydrogel combined with radiotherapy reduces hypoxia in lewis lung carcinoma xenografts: assessment by micro fluorine-18-fluoromisonidazole positron emission tomography/computed tomography hypoxia imaging. Front Oncol. 2021;11:628895.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  47. Catania G, Rodella G, Vanvarenberg K, Preat V, Malfanti A. Combination of hyaluronic acid conjugates with immunogenic cell death inducer and CpG for glioblastoma local chemo-immunotherapy elicits an immune response and induces long-term survival. Biomaterials. 2023;294:122006.

    Article  CAS  PubMed  Google Scholar 

  48. Ju HJ, Kim MJ, Kim S, Son KE, Lee MY, Kim HS, et al. Enhanced anticancer effects of intratumorally injected electrostatic doxorubicin-loaded click-type crosslinked hyaluronic acid hydrogel. Adv Ther-Germany. 2024;7:2400246.

    Article  CAS  Google Scholar 

Download references

Funding

Open access funding provided by SCELC, Statewide California Electronic Library Consortium. This research was funded by City of Hope, NIH/NCI (P30 CA033572 and UM1 CA186717), the Stephenson fund, and the COH-Caltech Biomedical Research Initiative (Merkin Institute).

Author information

Authors and Affiliations

  1. Department of Radiology, City of Hope Medical Center, Duarte, CA, USA

    Anup Kumar Patel, Imran Shair Mohammad, Jonathan Kessler, Julie Ressler, Cherng H Chao, Taedo Jake Choi, Chandana Lall & F. Edward Boas

  2. Department of Immuno-Oncology, City of Hope Medical Center, Duarte, CA, USA

    Marcin Kortylewski

  3. Department of Hematology & Hematopoietic Cell Transplantation, City of Hope Medical Center, Duarte, CA, USA

    Steven T Rosen

  4. Lundquist Institute, Torrance, CA, USA

    Catalina Guerra

Authors

  1. Anup Kumar Patel
  2. Imran Shair Mohammad
  3. Marcin Kortylewski
  4. Steven T Rosen
  5. Jonathan Kessler
  6. Julie Ressler
  7. Cherng H Chao
  8. Taedo Jake Choi
  9. Chandana Lall
  10. Catalina Guerra
  11. F. Edward Boas

Contributions

AKP, ISM, MK, and FEB designed experiments. AKP, ISM, CG, and FEB performed experiments. AKP, ISM, and FEB analyzed the data. AKP and FEB wrote the manuscript. All authors reviewed the manuscript.

Corresponding author

Correspondence to F. Edward Boas.

Ethics declarations

Ethical standards

Clinical trial number: not applicable.

Conflict of interests

MK is a co-founder of Twin Peaks, Forta and AptaDiR Therapeutics with stock options. He is a co-inventor of CpG-STAT3ASO strategy with pending patent application. STR is founder and chair of the scientific advisory board at SLAM bio and Noval Biotech. He is on the scientific advisory board of Cizzle. He is on the board of directors of Zylem. JK received a grant from Metavivor, and consulting fees from Johnson and Johnson, Boston Scientific, and Galvanize Therapeutics. CHC received support for attending meetings from Boston Scientific and Stryker. He is a co-investigator on studies funded by Instylla and CrannMed. FEB received research grants from City of Hope, Guerbet, Stephenson fund, COH-Caltech Biomedical Research Initiative (Merkin Institute), Society of Interventional Radiology, Society of Interventional Oncology, NIH/NCI, Department of Defense, Thompson family foundation, and Brockman Medical Research Foundation. He received research supplies (investigator-initiated) from Boston Scientific, and attended research consensus panels with the Society of Interventional Oncology. He is a co-founder of Claripacs, LLC, and an investor in Labdoor, Qventus, CloudMedx, Xgenomes, and Solugen. He has patent applications covering intratumoral drug delivery devices, immunoembolization, Y90 dosimetry, endoluminal ablation, CT metal artifact reduction, and other inventions.

Additional information

Publisher’s note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Patel, A.K., Mohammad, I.S., Kortylewski, M. et al. Radiopaque microsphere-hydrogel composite for extended-release intratumoral immunotherapy in a large animal model. Drug Deliv. and Transl. Res. (2026). https://doi.org/10.1007/s13346-026-02176-9

Download citation

  • Received:

  • Accepted:

  • Published:

  • Version of record:

  • DOI: https://doi.org/10.1007/s13346-026-02176-9

Keywords

Profiles

  1. Anup Kumar Patel View author profile
  2. Marcin Kortylewski View author profile
  3. Steven T Rosen View author profile