Bispecific antibody targeting of lipid nanoparticles

Abstract

Lipid nanoparticles (LNP) are the most clinically advanced non-viral gene delivery system. While progress has been made for enhancing delivery, cell specific targeting remains a challenge. Targeting moieties such as antibodies can be chemically-conjugated to LNPs however, this approach is complex and has challenges for scaling up. Here, we developed an approach to generate antibody-conjugated LNPs that utilizes a bispecific antibody (bsAb) as the targeting bridge. As a docking site for the bsAb, we generated LNPs with a short epitope, derived from hemagglutinin antigen (HA), embedded in the PEG component of the particle (LNP^HA). We generated bsAb in which one domain binds HA and the other binds different cell surface proteins, including PD-L1, CD4, CD5, and SunTag. Non-chemical conjugation of the bsAb and LNP resulted in a major increase in the efficiency and specificity of transfecting cells expressing the cognate target. LNP/bsAb mediated a 4-fold increase in in vivo transfection of PD-L1 expressing cancer cells, and a 26-fold increase in ex vivo transfection of quiescent primary human T cells. Additionally, we created a universal bsAb recognizing HA and anti-rat IgG2, enabling LNP tethering to off-the-shelf antibodies such as CD4, CD8, CD20, CD45, and CD3. By utilizing a molecular dock and bsAb technology, these studies demonstrate a simple and effective strategy to generate antibody-conjugated LNPs, enabling precise and efficient mRNA delivery.

Figure 1.. Generation of a linear epitope binding bispecific antibody.

a. Schematic of bsAb^HA-SunTag and depiction of activity in bringing epitope expressing cell types together. SP:signal peptide, scFV: single chain variable fragment. b. Schematic outlining generation of the cell populations expressing the targets of the bsAb^HA-SunTag used in the experimental conditions. MEL leukemia cells were transduced with a lentiviral vector (LV) expressing mCherry or GFP. MEL^mCherry and MEL^GFP were subsequently transduced with a LV expressing delta-NGFR^HA and delta-NGFR^Suntag respectively. MEL^{mCherry/NGFR-HA} and MELGFP/NGFR-Suntag were mixed in 1:1 ratio. c. The relative frequency of MEL^{mCherry/NGFR-HA} and MEL^{GFP/NGFR-Suntag} cells was assessed by florescent microscopy and flow cytometry. Shown is a representative image (20x) of the cell mix (top) and dot plot from the flow cyometry (bottom). d. Microscopy image of MEL^{GFP/NGFR-Suntag} and MEL^{mCherry/NGFR-HA} cell cultures incubated with 1 mg/ul of purified bsAb^HA-SunTag or unrelated bsAb. Image is representative from 3 independent cell-cell aggregation experiments. e. Quantification (left histogram) and visualization (representative images on the right) of clumps of MEL^{eGFP/NGFR-Suntag} and MEL^{mCherry/NGFR-HA} cells measured by Amnis^® Image-Stream analysis. Experiment was repeated 3 independent times.

Figure 2.. bsAb-coupled LNP mediates enhanced ligand-specific cell transfection.

a. Schematic of the LNP^HA with bound bsAb (left) and depiction of bsAb-mediated uptake through ligand binding. b. Schematic showing how LNP^HA can be redirected to different cell types with different bsAb. c. Schematic showing the bidirectional lentiviral vector (Bid.LV) used to generate PD-L1^Suntag/BFP reporter cell line (top). Representative dot-plot of human erythroleukemia cells (K562) transduced with Bid.LV PD-L1^Suntag/BFP, K562^PD-L1/SunTag (Bottom). d. Size analysis by DLS of SM-102 LNP (green) and SM-102 LNP^HA (purple) particles (left). Histograms showing mRNA encapsulation efficiency in LNP and LNP^HA as measured by RiboGreen assay (right). Graphs are representative of all LNP-mRNA particles generated in the study. e. Flow cytometry analysis of GFP and and PD-L1 expression in K562^PD-L1/SunTag cells transfected with 10 ng of GFP mRNA in LNP or LNP^HA with or without bsAb^HA-Suntag. Shown are representative dotplots. Data representative of 5 independent experiments, with n=3 biological replicates for each experiment. f. Graphs show the mean±s.d. percentage of GFP-positive K562^PD-L1/SunTag cells from (e). Analysis of GFP expression was performed 24 hours post-transfection. Two-way anova and Tukey’s multiple comparison post-test, **p<0.001; ****p<0.00001 (n=3). Data representative of 5 independent experiments. g. Flow cytometry analysis of GFP and and PD-L1 expression in 293T^PD-L1/SunTag cells transfected with 10 ng of GFP mRNA in LNP or LNP^HA with or without bsAb^HA-Suntag. Representative dotplots shown (n=3–8, 3 independent experiments) (left). Graph (right) shows fold change of GFP mean florescence intensity (MFI) of PD-L1^Suntag positive cells treated with bsAb^HA-Suntag compared to untreated cells. h. Flow cytometry analysis of GFP and and PD-L1 expression in K562^PD-L1/SunTag cells transfected with 10 ng of GFP mRNA in Dlin-MC3 LNP^HA or ALC-0315 LNP^HA in absence (left column) or presence (right column) of bsAb^HA-Suntag. Analysis performed 24 hours post transfection. i. Graphs showing percentage (left) and MFI (right) of GFP positive cells from (h). Statistical analysis: two-way anova and Tukey’s multiple comparison post-test,; ***p<0.0001 (n=3) Data representative of one independent experiment.

Figure 3.. LNP^HA/bsAb increases in vivo targeting of PD-L1+ tumor and CD4 T cells.

a. Schematic of the bsAb^HA-Suntag and its binding domain on K562^PD-L1-SunTag (bottom left) and bsAb^HA/PD-L1 (top right) and its binding domain on K562^PD-L1-SunTag (bottom right). Flow cytometry analysis of K562^PD-L1-SunTag treated with 10 ng of GFP mRNA LNP^HA incubated with either bsAb^HA-Suntag or bsAb^HA-PD-L1. GFP and PD-L1 levels were measured after 24 hours. b. Graphs show the mean±s.d. percentage of GFP+ K562^mPDL1 transfected with 20 ng GFP mRNA LNP or LNP^HA +/− bsAb^HA/PD-L1 (n=3 biological replicates, 2 independent experiments). Two-way anova and Tukey’s multiple comparison post-test, ***p<0.0001; ****p<0.00001. c. Graphs show the percentage of transfected PD-L1+ and PD-L1− B16F10^{dsRed-LSL-GFP} melanoma cells following intratumoral injection of melanomas with matching concentrations of Cre mRNA encapsulated in LNP^HA or LNP^HA+bsAb^HA-PD-L1. Analysis was performed by flow cytometry. Each dot is a separate mouse. Two-way anova and Tukey’s multiple comparison post-test, **p<0.001. Schematic of the experiment shown (top). d. Graphs show the percentage of transfected PD-L1+ and PD-L1− ID8^{dsRed-LSL-GFP} ovarian cancer cells following i.p. injection of ovarian tumor bearing mice with matching concentrations of Cre mRNA encapsulated in LNP^HA or LNP^HA+bsAb^HA-PD-L1. Analysis was performed by flow cytometry. Each dot is a separate mouse. Two-way anova and Tukey’s multiple comparison post-test, **p<0.001. Schematic of the experiment shown (top). e. Graphs show the percentage of transfected PD-L1+ and PD-L1− KC^{dsRed-LSL-GFP} pancreatic cancer cells following i.p. injection of pancreatic tumor bearing mice with matching concentrations of Cre mRNA encapsulated in LNP^HA or LNP^HA+bsAb^HA-PD-L1. Each dot is a separate mouse. Two-way anova and Tukey’s multiple comparison post-test, **p<0.001. Schematic of the bsAb^HA-CD4 (top left) bound to its target, CD4, on murine T cell. f. Schematic of the bispecific antibody recognizing HA and CD4. g. Flow cytometry analysis of tdTomato and CD4 expression on total splenocytes isolated from Ai14 mice and transfected with 50 ng Cre mRNA encapsulated in LNP, LNP^HA, LNP^HA+bsAb^HA-CD4. Cultures included IL-2 and anti-CD3/CD28 beads to activate/expand T cells. Shown are representative dotplots. h. Flow cytometry analysis of splenocytes isolated from Ai14 mice i.v. injected with Cre mRNA encapsulated in LNP or LNP^HA and coupled with or without bsAb^HA-CD4_. Spleens collected 3 days post injection. Dotplots are representative of n=3 mice group, 2 independent experiments. i. Graphs show percent transfection of CD4+ and CD4− splenocytes from (h) (n=3 mice, 2 independent experiments). j. Graphs show in vivo targeting efficiency of the different LNP formulations calculated by the ratio of tdTomato+ CD4+ to CD4− splenocytes from (h). k. Graphs show the percentage of GFP+ resting human T cells treated with 200 ng GFP mRNA encapsulated in LNP or LNP^HA with or without bsAb^HA-CD5. Expression was measured 24 hours post transfection. Two-way anova and Tukey’s multiple comparison post-test, **p<0.001; ***p<0.0001(n=4). Data representative of 3 independent experiments. l. Graphs showing the GFP MFI of resting human T cells treated as in (k). Two-way anova and Tukey’s multiple comparison post-test, ***p<0.0001(n=4). Data representative of 3 independent experiments.

Figure 4.. Comparison of cell transfection by LNP^HA/bsAb and thiol-Mal-conjugated LNP.

a. Flow cytometry analysis of GFP and PD-L1 expression in K562^mPDL1 cells transfected with the indicated LNP. Graph (left) shows the mean±s.d. of GFP+ cells. Dotplots (right) are representative from the experiment. All groups used matched concentrations of LNP-RNA (10 ng). GFP expression was measured 24 hours post transfection. (n=3 per group, 3 independent experiments performed) b. Flow cytometry analysis of GFP and CD5 expression in K562^hCD5 transfected with 10 ng GFP mRNA encapsulated in the indicated LNP. Graph shows the mean±s.d. of GFP mean florescence intensity (MFI) of GFP+ cells. (n=3 per group, 3 independent experiments performed) c. Flow cytometry analysis of GFP and CD5 expression in resting human T cells treated with 200 ng GFP RNA encapsulated in SM-102 LNP (left), SM-102 LNP^HA (center) with or without bsAb^CD5-HA or with LNP^MAL-hCD5. Representative dotplots are shown (n=3 per group, 3 independent experiments performed). d. Graph showing the mean±s.d. of the percentage of GFP+ resting human T cells treated as in (c). One-way anova and Tukey’s multiple comparison post-test, ***p<0.0001; Data representative of 3 independent experiments with n=3 per group. e. Graph showing the mean±s.d. GFP MFI of resting human T cells treated as in (c). Statistical analyis: One-way anova and Tukey’s multiple comparison post-test, **p<0.001 ***p<0.0001 (n=3) ; Data representative of 3 independent experiments. f. Flow cytometry analysis of GFP and CD5 expression in activated human T cells treated with 200 ng GFP mRNA encapsulated in SM-102 LNP or SM-102 LNP^HA with or without bsAb^CD5-HA, or LNP^MAL-hCD5 or LNP^MAL-IgG. Expression was measured 24 hours post-transfection. Data representative of 3 independent experiments (n=3 per group per experiment). g. Graph showing the mean±s.d. percentage of GFP+ activated human T cells treated as in (f). One-way anova and Tukey’s multiple comparison post-test, ***p<0.0001 (n=3); Data representative of 3 independent experiments. (n=3 per group, 3 independent experiments performed) h. Graph showing the mean±s.d. GFP MFI of activated human T cells treated as in (g). (n=3 per group, 3 independent experiments performed) i. Graph showing the mean±s.d. percent of GFP+ activated human T cells treated as in (g) at 3 days post transfection. (n=3 per group, 3 independent experiments performed) j. Graph showing the mean±s.d. GFP MFI in activated human T cells treated as in (g) at 3 days post transfection. (n=3 per group, 3 independent experiments performed)

Figure 5.. Tether bsAb platform redirect LNP^HA towards target cell type in presence of monoclonal antibody.

a. Schematic depicting the bsAb^HA-ratIgG2b tethering antibody binding to LNP^HA particles and a monoclonal antibody with the recognized constant region (Rat IgG2b). b. Flow cytometry analysis of GFP expression in K562^mPDL1 cells transfected with 10 ng of GFP mRNA encapsulated in SM-102 LNP or SM-102 LNP^HA with or without bsAb^HA/PD-L1 or with bsAb^HA/ratIgG2b with or without PD-L1 monoclonal antibody. GFP expression was measured 24 hours post transfection. Graphs show the mean±s.d. of the percentage of GFP+ cells. c. Representative dotplots and graphs of total splenocytes treated with Cre mRNA encapsulated in LNP or LNP^HA with or without bsAb^HA-IgG2b and the indicated monoclonal antibody. Analysis was performed 48 hours after LNP treatment. Data representative of 3 independent experiments.