Disrupting the Interplay between Programmed Cell Death Protein 1 and Programmed Death Ligand 1 with Spherical Nucleic Acids in Treating Cancer

Disrupting the interplay between programmed cell death protein 1 (PD-1) and programmed death ligand 1 (PD-L1) is a powerful immunotherapeutic approach to cancer treatment. Herein, spherical nucleic acid (SNA) liposomal nanoparticle conjugates that incorporate a newly designed antisense DNA sequence specifically against PD-L1 (immune checkpoint inhibitor SNAs, or IC-SNAs) are explored as a strategy for blocking PD-1/PD-L1 signaling within the tumor microenvironment (TME). Concentration-dependent PD-L1 silencing with IC-SNAs is observed in MC38 colon cancer cells, where IC-SNAs decrease both surface PD-L1 (sPD-L1) and total PD-L1 expression. Furthermore, peritumoral administration of IC-SNAs in a syngeneic mouse model of MC38 colon cancer leads to reduced sPD-L1 expression in multiple cell populations within the TME, including tumor cells, dendritic cells, and myeloid derived suppressor cells. The treatment effectively increases CD8+ T cells accumulation and functionality in the TME, which ultimately inhibits tumor growth and extends animal survival. Taken together, these data show that IC-SNA nanoconstructs are capable of disrupting the PD-1/PD-L1 interplay via gene regulation, thereby providing a promising avenue for cancer immunotherapy.


Design of antisense sequences against mouse PD-L1 (mPD-L1)
Antisense DNA sequences targeting mPD-L1 mRNA were not available in the literature. Therefore, sequences were designed for this study that are complementary to mouse PD-L1 mRNA. This sequence information was obtained from the publicly available NCBI GeneBank.
To avoid non-specific, off-target effects, the mouse PD-L1 mRNA was checked for possible homology using a BLAST test. To achieve efficient gene knockdown, mRNA secondary structure was considered because it plays an important role in the accessibility of interacting

DNA synthesis and purification
All sequences were synthesized with phosphorothioate backbones using standard automated solid-phase phosphoramidite DNA synthesis protocols on a MerMade 12 synthesizer (Bioautomation). Universal CPG solid supports and cholesterol-ended supports were used to synthesize linear sequences and cholesteryl-modified sequences, respectively.
The synthesized DNA strands were cleaved from the solid support via incubation in 30 % ammonium hydroxide at 55 °C overnight. Subsequently, excess ammonia was removed by evaporation under nitrogen gas at room temperature. Then, the oligonucleotides were dissolved in distilled water and filtered before purification. An Agilent high-pressure liquid chromatography (HPLC) system with a C4 or C18 column was used to purify cholesterolended and universal-ended sequences, respectively. Gradient triethylammonium acetate and acetonitrile (10 % to 100 % acetonitrile) was used to remove failure strands over 30 min. The purified oligonucleotides were concentrated and lyophilized. The powdered oligonucleotides were incubated with 20 % acetic acid at room temperature for 1 h and then extracted three times using an equal volume of ethyl acetate. The deprotected DNA was lyophilized and reconstituted in 1 mL of deionized water. The sequences were characterized using matrixassisted laser desorption ionization time-of-flight (MALDI-ToF).

Liposome/IC-SNA synthesis
Five mg of 1,2-dioleoyl-sn-glycero-3-phosphocholine (DOPC) powder was resuspended in 5 mL of phosphate-buffered saline (PBS) solution. Liposomes were formed via 5 freeze-thaw cycles, where liquid nitrogen and sonication in a 37 °C water bath were used. The liposomes were then extruded sequentially through polycarbonate filters (T&T Scientific Corp., Knoxville, TN): 200 nm, 100 nm, 80 nm, and finally 50 nm in filter pore size. The liposome concentration was measured using a commercially available phosphatidylcholine assay (Sigma). Cholesterol-ended oligonucleotides were inserted into the double layer of the phospholipid membranes via hydrophobic interactions. Sodium dodecyl sulphatepolyacrylamide gel electrophoresis (SDS PAGE) was run to estimate the average number of strands that could be anchored on a 50-nm liposome (an average of 50 strands per particle S6 was observed). To form SNAs, oligonucleotides were added to liposomes (75:1), and the sample was shaken overnight at 35 °C. The oligonucleotide concentration was determined by UV-vis absorption at 260 nm ( Figure S1). The resulting SNAs were then concentrated to 50 µM by DNA concentration using centrifugation filter units (Millipore), which also removed any unbound DNA. The resulting structures were analyzed to determine their zeta potential and size (Malvern Zetasizer) and DNA loading (gel electrophoresis). and TNF-a was measured to evaluate T cell cytokine release. CD44 and CD62L were stained to detect effector memory T cell subset accordingly.

Safety statement
No unexpected or unusually high safety hazards were encountered.
A) Oligonucleotides can be designed to trigger RNase H-mediated DNA:RNA hybrid degradation. In this case, the sequences were designed to fit into the active cleavage site of the RNase H enzyme. All sequences were 18 bp or less in length. B) The other mechanism of disrupting protein expression with antisense DNA occurs on a translational level, where antisense DNA binds to mRNA and halts the translation machinery. To employ this mechanism, the sequences were designed to avoid binding to preoccupied regions during translation (the gray area). S10 Figure S1. Antisense DNA binding accessibility profile.
The coding sequence (CDS, starts at 84 bp and ends at 956 bp) of mPD-L1 mRNA was simulated to fold using the online Mfold RNA folding tool. The overall length of this segment is 873 bp. The horizontal parameter depicts the single-stranded frequency for a given bp position in 30 simulated folds. Four candidate sequences were synthesized based on the availability of a given 15-bp segment that has a strong binding ability in the middle section.

Figure S2. Synthesis and characterization of IC-SNAs
A) SNAs were synthesized using Sequence B and characterized using Matrix-Assisted Laser Desorption Ionization-Mass Spectrometry (MALDI-MS). B) Subsequently, the SNA was characterized using polyacrylamide gel electrophoresis, C) dynamic light scattering (DLS), and zeta potential. The loading for dye-labeled DNA-based SNAs was measured using electrophoresis. On average, the SNAs were modified with 50 strands per particle. Upon DNA loading, the average increase in diameter was 6.4 nm, and the decrease in zeta potential was 16 mV.