Development of Adamantane-Conjugated TLR7/8 Agonists for Supramolecular Delivery and Cancer Immunotherapy

Tumor-associated macrophages (TAMs) are often abundant in solid cancers, assuming an immunosuppressive (M2-like) phenotype which supports tumor growth and immune escape. Recent methods have focused on identification of means (e.g., drugs, nanomaterials) that polarize TAMs to a tumor suppressive (M1-like) phenotype; however, reducing the systemic side effects of these therapies and enabling their delivery to TAMs has remained a challenge. Methods: Here, we develop R848-Ad, an adamantane-modified derivative of the toll-like receptor (TLR) 7/8 agonist resiquimod (R848) through iterative drug screening against reporter cell lines. The adamantane undergoes guest-host interaction with cyclodextrin nanoparticles (CDNPs), enabling drug loading under aqueous conditions and TAM-targeted drug delivery. Therapeutic efficacy and systemic side effects were examined in a murine MC38 cancer model. Results: R848-Ad retained macrophage polarizing activity through agonization of TLR7/8, and the adamantane moiety improved drug affinity for the CDNP. In preclinical studies, nanoformulated R848-Ad resulted in a drastic reduction in measurable systemic effects (loss of body weight) relative to similarly formulated R848 alone while arresting tumor growth. Conclusions: The findings demonstrate the ability of strong nanoparticle-drug interactions to limit systemic toxicity of TLR agonists while simultaneously maintaining therapeutic efficacy.

was purified via reversed-phase chromatography. Fraction of similar purity were combined and concentrated to yield the desired intermediate product (45.3 mg, 32.6 % yield). The intermediate (14.7 mg, 0.024 mmol) and 1-adamantylacetylene (4.3 mg, 0.027 mmol, Matrix Scientific) were dissolved in 12 mL of 1:1 mixture of water:t-butanol. To this a solution of sodium ascorbate and copper sulfate was added, prepared by the mixing of 2.4 µL of a freshly prepared solution of sodium ascorbate (10 mg in 50 µL) with 2.4 µL of a freshly prepared solution of copper sulfate (10 mg in 50 µL). The reaction was allowed to react for 16 hours with vigorous mixing. The crude reaction mixture was purified via reversed-phase chromatography. Fraction of similar purity were combined and concentrated to yield the desired product 4 (7.9 mg, 43 % yield). MS (ESI) calculated C 40 H 58 N 8 O 7 , m/z 762.44, found 763 (M + H) + .

Synthesis of 5.
Compound A (144 mg, 0.51 mmol) was dissolved in DMF (5 mL) with N,Ndiisopropylethylamine (2 mL) to this was added Azido-PEG 2 -NHS ester (81.4 mg, 0.40 mmol, Santa Cruz Biotechnology). The reaction was allowed to proceed for 4 h at room temperature. After which, the reaction was concentrated under reduced pressure. The crude reaction mixture was purified via reversed-phase chromatography. Fraction of similar purity were combined and concentrated to yield the desired intermediate product (188 mg, 99% yield). The intermediate (188 mg, 0.40 mmol) and 1-adamantylacetylene (70.4 mg, 0.44 mmol, Matrix Scientific) were dissolved in 50 mL of 1:1 mixture of water:t-butanol. To this a solution of sodium ascorbate and copper sulfate was added, prepared by the mixing of 25 µL of a freshly prepared solution of sodium ascorbate (10 mg in 50 µL) with 25 µL of a freshly prepared solution of copper sulfate (10 mg in 50 µL). The reaction was refluxed for 4 hours with vigorous mixing. The crude reaction mixture was purified via reversed-phase chromatography. Fraction of similar purity were combined and concentrated to yield the desired product 5 (5.6 mg, 2 % yield). MS (ESI) calculated C 34 H 46 N 8 O 4 , m/z 630.36, found 632 (M + H) + .

Synthesis of 6.
Compound A (148 mg, 0.520 mmol) was dissolved in acetic anhydride (86.5 mL) with N,N-diisopropylethylamine (0.174 mL). The reaction was allowed to proceed for 30 minutes at room temperature. After which, the reaction was concentrated under reduced pressure to yield the desired product 6 (169.9 mg, 99 % yield). MS (ESI) calculated C 17 H 21 N 5 O 2 , m/z 327.17, found 328 (M + H) + .

Synthesis of 7.
Compound B (26.4 mg, 0.100 mmol) was dissolved in acetic anhydride (0.100 mL) with N,N-diisopropylethylamine (0.850 mL). The reaction was allowed to proceed for 30 minutes at room temperature. After which, the reaction was concentrated under reduced pressure to yield the desired product 7 (29.9 mg, 84% yield). MS (ESI) calculated C 19 H 25 N 5 O 2 , m/z 355.20, found 356 (M + H) + .

Synthesis of 9.
Compound C (5.6 mg, 0.02 mmol) was dissolved in acetic anhydride (0.1 mL) with N,N-diisopropylethylamine (0.1 mL). The reaction was allowed to proceed for 30 minutes at room temperature. After which, the reaction was concentrated under reduced pressure to yield the desired product 9 (6.2 mg, 99 % yield). MS (ESI) calculated for C 24 H 27 N 5 O, m/z 401.22, found 402 (M + H) + .
Synthesis of 10. 1-Adamantaneacetic acid (28 mg, 0.14 mmol), HATU (53 mg, 0.14 mmol), and NHS (17 mg, 0.14 mmol) was dissolved in DMF (5 mL) with N,N-diisopropylethylamine (0.1 mL) under Ar. After 30 mins, Compound C (40 mg, 0.11 mmol) dissolved in minimal DMF and were added and the reaction was allowed to proceed for 30 minutes at room temperature. After which, the reaction was concentrated under reduced pressure. The crude reaction mixture was purified via reversed-phase chromatography. Fraction of similar purity were combined and concentrated to yield the desired product 10 (59 mg, 78 % yield).   Figure S8).
Visualization of drug binding. Visualization of R848 derivative interactions with TLR7 (protein databank number: 5gmh) was performed with the aid of PyMOL software package [4]. The crystal structure of monkey TLR7-R848 was chosen as a surrogate for the human receptor due to the high quality data available and conservation of protein structure across species. TLR7 exists in vivo as a dimeric species; however, the PDB data file used for drug optimization is a monomer projection and requires workup to display the dimeric component prior to analysis. Figure S1. Crystal structure of R848 derivative binding. (A) For R848, the tertbutyl tail is cleared from contact with the native protein structure. (B) Following direct modification by adamantane (Generation 1, compound (2)), drug accessibility of the binding site is statically hindered (indicated: yellow arrow). (C) The incorporation of suitable linkers, such as PEO 5 shown (Generation 2, compound (5)), enable access of the R848 core structure to the binding site (indicated: yellow arrow, 1) while displacing adamantane to a less sterically hindering location (indicated: yellow arrow, 2).     The activity of R848-Ad@CDNP was examined in HEK-Blue mTLR7 (A) and HEK-Blue hTLR8 (B) cells at a fixed concentration of 10 nM R848-Ad and an increasing CDNP concentration (0-250 μg/mL). CDNP controls (black, left) did not activate TLR7 or TLR8 activity. CDNP increased R848-Ad activity, relative to the free drug. Increased activity was attenuated at high CDNP concentrations (molar ratio CD/Ad >> 1). Results represent the mean ± s.d. of the absorbance at = 662 nm; N=4; ****p<0.0001, ***p<0.001, **p<0.01, *p<0.01 vs free drug (0 μg/mL CDNP). Figure S7. Change in body weight of mice following treatment at day 0, 3, and 6 with CDNP vehicle controls (solid black, left), R848-Ad (dashed orange, middle), or R848-Ad@CDNP (solid orange, right). Mean ± s.e.m; N = 10; **P < 0.01 (Friedman, Dunn's multiple comparison) relative to vehicle control. Figure S8. 1H NMR spectra of R848-Ad. The final product, R848-Ad, was determined via 1 H NMR and HPLC (at both 250 and 315 nm) to be spectroscopically pure (>95%).