WO2025137322A1 - Secreted particle information transfer (spit) system for genetic engineering - Google Patents

Abstract

The invention provides a human cell-based delivery system for performing in vivo genetic modifications. The human cells are engineered to express and secrete one or more components of a gene editing technology which are taken up by target cells in vivo such that the target cells are genetically modified.

Description

SECRETED PARTICLE INFORMATION TRANSFER (SPIT) SYSTEM FOR GENETIC ENGINEERING

CROSS REFERENCE TO RELATED APPLICATIONS

[0001] This application claims the benefit of U.S. provisional application No. 63/612,880 filed December 20, 2023, which is incorporated herein by reference in its entirety.

INCORPORATION BY REFERENCE OF THE SEQUENCE LISTING

[0002] This application contains, as a separate part of disclosure, a Sequence Listing in computer-readable form (Filename: 70015_SeqListing. xml; 7,508 bytes - XML file dated December 19, 2024) which is incorporated by reference herein in its entirety.

FIELD

[0003] The disclosure relates to a human cell-based delivery system for performing in vivo genetic modifications. The human cells are engineered to express and secrete one or more components of a gene editing technology which are taken up by target cells in vivo such that the target cells are genetically modified.

BACKGROUND

[0004] The past decade has seen a revolution in the ability to manipulate the human genome. What was once a herculean challenge is now achieved with remarkable ease to human cell lines in a dish. This transformative progress is largely attributable to the advent of CRISPR-Cas systems, owing to their efficacy, versatility, and broad array of applications.

However, the plethora of tools now available for use in genetic engineering contrasts sharply with constraints in the ability to deliver these tools to the cells of human patients, a necessity for successful clinical translation.

[0005] Two general approaches are currently under development to apply genetic engineering tools to patients. The first are ex vivo genetic engineering approaches, where patient cells are isolated from the body, genetically engineered ex vivo and then transplanted back into the body. Although this approach is clinically efficacious, its application is restricted to cell types that are amenable to this process, primarily hematopoietic cell types. The second approach is to deliver genetic engineering technologies to cells directly in vivo through the use of recombinant viral vectors such as adeno associated virus (AAV, -4.5 kb) or via chemically defined vehicles such as lipid nanoparticles (LNPs, -10-20 kb). However, the successful clinical application of these in vivo technologies has primarily been restricted to the liver and current platforms are limited in the amount of genetic information they can deliver to cells, typically limited to a single gene.

[0006] Virus-like particles (VLPs) and been used as delivery systems ex vivo. However, production of VLPs faces significant manufacturing challenges. These challenges include, for example, (1 ) scalability since adequate quantities of delivery material must be produced ex vivo to achieve the desired therapeutic effect upon administration to a patient which requires large volumes of media and extensive cell cultures to produce sufficient material for therapeutic application, (2) stability since post-production, the biological vectors need purification and storage, which adds significantly to the technical challenge and cost of production, and also many vectors such as retroviral vectors rapidly lose activity at room temperature and 4°C, and degrade further from freeze-thawing processes, providing additional challenges to patient delivery even after purification, and (3) production efficiency since most production methods are based on transient transfection only allowing for collection of biological modalities produced in cells within a short timeframe, but even in cases where stable producer cells lines are generated and can be used, obtaining vectors from stable cell lines requires multiple rounds of purification from the same culture, complicating the scalability and feasibility of producing large quantities of vector from stable cell lines.

[0007] There thus remains a need for genetic engineering products and methods for their production and delivery.

BRIEF SUMMARY

[0008] The present invention provides human producer cells modified to secrete particles comprising one or more components of a genetic editing system, such as a nuclease-based gene editing system or a CRISPR-based gene editing system. The secreted particles, which may be for example, virus-like particles, or “VLPs”, are taken up, or transferred, to a target cell in v/vo that in turn is genetically modified, i.e., the cell's genetic information is manipulated, by the genetic editing system components. This approach to in vivo gene therapy may be referred to herein by the acronym “SPIT” for Secreted Particle Information Transfer. The human producer cells are also referred to herein as “SPIT cells”.

[0009] Also provided is the encapsulation of SPIT cells herein. By transplanting engineered SPIT cells encased in an encapsulation device directly into a patient’s body, the SPIT cells can stably reside and continuously produce and deliver VLPs directly to a patient over an extended period of time, without further processing, handling or storage (FIG. 16A). The capsule's porous membrane supports nutrient and gas exchange, facilitates VLP release, protects SPIT cells from the patient’s immune system, and prevents their distribution throughout the body, thereby enhancing both safety and efficacy.

[0010] In one aspect, provided is a method for in vivo genetic modification of target cells in a subject, the method comprising administering to the subject a composition that includes producer cells and a carrier, where the producer cells comprise one or more heterologous nucleic acids encoding one or more components of a gene editing system as fusion protein(s) with a GAG protein, such that the producer cells secrete VLPs comprising the fusion protein(s) which are taken up by the target cells, thereby genetically modifying the target cells.

[0011] In aspects, the method may also include where the GAG protein is a structural groupspecific antigen (Gag) protein of a retrovirus, such as a murine leukemia virus (MLK), a human immunodeficiency virus, a rotavirus, or a hepatitis B virus, or a derivative of any of the foregoing.

[0012] In aspects, the method may also include where the GAG protein is a non-viral protein, preferably a human protein, selected from an Arc protein, ASPRV1 , a Sushi-Class protein, a SCAN protein such as PGBD1 , and a PNMA protein.

[0013] In aspects, the method may also include where the GAG protein is a non-viral protein selected from PEG10, RTL3, RTL10, and RTL1.

[0014] In aspects, the method may also include where the producer cells encode a fusogen, optionally where the fusogen is VSV-g, RD114a, amphotrophic envelope 4070A, measles F/H, syncytin, or myomaker/myomixer. In aspects, the fusogen is VSV-g or mSynA.

[0015] In aspects, the fusogen is a modified non-retroviral or mammalian endogenous retroviral fusogen which includes a retroviral R-peptide. In aspects, the R-peptide replaces the cytoplasmic tail of the fusogen or is added to the C-terminus of the fusogen cytoplasmic tail, where the protease consensus sequence is located at the N-terminus of the R-peptide fusion. Such fusogens are advantageously protease regulatable, as described in more detail infra. In aspects, the R-peptide is derived from MLV amphotrophic 4070A, human immunodeficiency virus or simian immunodeficiency virus. In aspects, the R-peptide may be a synthetic R- peptide created through the attachment of a peptide or protein, such as GFP, to the cytoplasmic tail of a fusogen, with the consensus sequence for a protease located at the N- terminus of the fused peptide/protein.

[0016] In aspects, the method may also include where the producer cells comprise an expression plasmid comprising the nucleic acid encoding the GAG fusion protein. [0017] In aspects, the method may also include where the producer cells comprise a genetic knock-in encoding one or more of a protease, a pore forming protein, a targeting protein, a transcription factor, and/or an inducible suicide gene such as Caspase 9 or HSV-TK.

[0018] In aspects, the method may also include where the producer cells comprise one or more heterologous nucleic acids encoding a therapeutic protein or a therapeutic RNA.

[0019] In aspects, the method may also include where the producer cells are primary cells. In aspects, the producer cells are pluripotent stem cells, induced pluripotent stem cells (iPSC), hematopoietic stem cells, or hematopoietic progenitor cells. In aspects, the producer cells are neuronal cells, hepatocytes, or fibroblasts. In aspects, the producer cells are a transformed cell line, such as a fibroblast cell line. In aspects of any of the methods and compositions described here, the producer cells are preferably human cells.

[0020] In aspects, the method may also include where the one or more components of a gene editing system includes one or more of a zinc finger nuclease, a transcription activatorlike effector nuclease (TALEN), a Cas nuclease, a single stranded DNA modifying enzyme, a reverse transcriptase, a DNA methylase, a histone acetyltransferase, a deacetylase, and/or a topisomerase.

[0021 ] In aspects, the method may also include where the one or more components of a gene editing system includes one or more components of a CRISPR-Cas gene editing system. In aspects, the one or more components of a CRISPR-Cas gene editing system includes a Cas endonuclease and a guide RNA. In aspects, the VLP contains a Cas/gRNA ribonucleoprotein. In aspects, the method does not comprise introducing exogenous guide RNA to the subject or the target cells.

[0022] In aspects, the method may also include where the target cells are thymus, heart, lung, liver, kidney, intestinal, or spleen cells. The method may also include where the target cells are hematopoietic cells, optionally where the hematopoietic cells are B lymphocytes, T lymphocytes, or myeloid cells.

[0023] In aspects, the method may also include where the producer cells comprise an expression plasmid comprising the one or more heterologous nucleic acids encoding a therapeutic protein.

[0024] Also provided are modified non-retroviral or mammalian endogenous retroviral fusogens which include a retroviral R-peptide. In aspects, the R-peptide replaces the cytoplasmic tail of the fusogen or is added to the C-terminus of the fusogen cytoplasmic tail. In aspects, the R-peptide is derived from MLV amphotrophic 4070A, human immunodeficiency virus or simian immunodeficiency virus. In aspects, the R-peptide is a synthetic R-peptide. In aspects, the synthetic R-peptide may be created through attachment of a peptide or protein, such as GFP, to the cytoplasmic tail of a fusogen, where the consensus sequence for a protease is located at or near the N-terminus of the peptide or protein.

[0025] Also provided is a human producer cell line, where the producer cells comprise one or more heterologous nucleic acids encoding one or more components of a gene editing system as GAG fusion protein(s), and secrete particles comprising the fusion protein(s), preferably VLPs. The human producer cell line may also include where the cells are pluripotent stem cells, induced pluripotent stem cells (iPSC), hematopoietic stem cells, or hematopoietic progenitor cells. In aspects, the human producer cells are a cell line, such as a fibroblast cell line. The human producer cell line may also include where the one or more components of a gene editing system includes a Cas endonuclease and a guide RNA. The human producer cell line may also include where the secreted particles, such as VLPs, contain a Cas-gRNA ribonucleoprotein.

[0026] In aspects, the producer cells in the method are not engineered to encode a fusogen. In aspects, the cells of the human producer cell line are not engineered to encode a fusogen. [0027] In aspects, in the method the GAG fusion protein does not comprise a nucleocapsid domain. In aspects, in the human producer cell line the GAG fusion protein does not comprise a nucleocapsid domain.

[0028] In aspects, in the method the producer cells are engineered to co-express the GAG fusion protein and CD63. In aspects, the cells of the human producer cell line are engineered to co-express the GAG fusion protein and CD63.

[0029] In aspects, the producer cells in the method are encapsulated. In aspects, the cells of the human producer cell line are encapsulated.

[0030] Other technical features may be readily apparent to one skilled in the art from the following figures, descriptions, and claims.

BRIEF DESCRIPTION OF THE DRAWINGS

[0031] FIG. 1A is a schematic showing the general concept of Secreted Particle Information Transfer or “SPIT” for facilitating cell-cell delivery of genetic engineering platforms.

[0032] FIG. 1 B is a schematic of virus-like particle, or “VLP”, components applied to achieve SPIT. Also illustrated is the budding of a VLP from the membrane of a producer cell. POI = Protein of Interest.

[0033] FIG. 1C is a bar graph showing percentage of TdT-i- fibroblasts following the application of different VLPs collected from supernatants of cells transfected with different combinations of Cre with components of retroviral VLPs, Gag-Cre, GAG-Pro-Pol (Gag-Pol), or VSVg (fusogen). (n = 3), Two-Way ANOVA used to compare groups, P<0.0001 = ****.

[0034] FIG. 1 D is a bar graph showing percentage of TdT+ Ai9 293T cells following transfection of plasmids encoding wild type Cre (wtCre, left) or Gag-Cre (right). No statistically significant difference was found between the frequency of TdT+ cells between the two conditions (t-test, n=3).

[0035] FIG. 2A is a bar graph showing percentage of mCherry-i- reporter cells when ABE8 VLPs are produced in the presence (right bar in each pair) or absence (left bar in each pair) of GAG-pro-pol (n = 2-3). Statistical significance determined by ANOVA. ABE8 = SpCas9 Adenine Base Editor version 8. ** = P < 0.005.

[0036] FIG. 2B is a schematic showing differences in design of the Gag-POl vectors used to deliver Cre recombinase versus SpCas9 Adenine Base Editor version 8, “ABE8”. The location where ABE and GAG-Cre are cleaved by the aspartyl protease in GAG-Pro-Pol are indicated by a vertical line and scissors.

[0037] FIG. 3A illustrates an aspect of the subject matter in accordance with one embodiment.

[0038] FIG. 3B is a bar graph comparing the frequency of recombination found in Ai9 293T cells 3 days, as detected by tdTomato expression after application of VLPs with different retroviral fusogens, when Gag-Pro-Pol is present (+) or absent (-) in VLPs (n=1 ).

[0039] FIG. 3C illustrates an aspect of the subject matter in accordance with one embodiment.

[0040] FIG. 4A is a schematic showing the generation of a recombinant mSynA construct with the R-peptide from amphotrophic 4070A replacing its cytoplasmic tail.

[0041] FIG. 4B is a bar graph comparing the frequency of recombinatiom that occurred in Ai9 293T cells three days after the application of Cre-VLPs produced with different fusogens in the presence (+) or absence (-) of Gag-pro-pol (n=2).

[0042] FIG. 5A is a schematic showing the design of two protease regulatable recombinant VSV-g fusogens. One is made protease regulatable by replacement of its cytoplasmic tail with the cytoplasmic tail/R-peptide of amphotrophic 4070A (VSV-g + ampho tail), with tGFP attached to the C-terminus of the R-peptide to allow detection of its expression in producer cells. The other is made protease regulatable through the creation of a synthetic R-peptide, via the addition of tGFP with the consensus sequence for MLV protease at its N-terminus (VSV-g + tGFP) to the cytoplasmic tail of VSV-g. [0043] FIG. 5B is a bar graph comparing the frequency of recombination (% TdT positive cells) in Ai9 293T cells three days after application of Cre-VLPs produced with the recombinant VSV-g fusogens VSV-g + tGFP, either in the presence (+) or absence (-) of Gag- pro-pol, n=2 for each condition.

[0044] FIG. 5C is a bar graph comparing the frequency of recombination (% TdT positive cells) in Ai9 293T cells three days after application of Cre-VLPs produced with the recombinant VSV-g fusogens VSV-g + ampho tail in the presence (+) of Gag-pro-pol, n=2.

[0045] FIG. 6A is a schematic showing development and packaging of PEG10 VLPs via fusion of a protein of interest (POI) to the C-terminus of Peg10. Peg10 shown as rectangle, POI shown as an oval.

[0046] FIG. 6B is a bar graph comparing the frequency of recombination (% TdT positive cells) in 293T cells three days after application of VLPs produced with the MLV-GAG-Cre or Peg10 VLPs, n=2.

[0047] FIG. 7A is a schematic outlining the experimental procedure for testing VLP-SPIT in vitro using 293T cells transfected with Gag-cre and VSV-g as producer cells and Ai9 293T reporter cells as target cells.

[0048] FIG. 7B is a line graph showing the percentage of TdT-i- Ai9 293T reporter cells over time following co-culture with 293T producer cells (squares) or in the absence of co-culture (circles), n = 3. Statistical significance determined by two-way ANOVA. P<0.01 = **, P<0.001 = ***

[0049] FIG. 8A is a schematic showing the design and orientation of three different “all in one” doxycycline inducible vectors, designated A, B, and C.

[0050] FIG. 8B is a bar graph showing the total frequency of cells positive for GFP and/or TdTomato (% fluorescent) following transfection with Vector A, Vector B, or Vector C, in the presence (right bar in each pair) and absence (left bar in each pair) of doxycycline.

[0051] FIG. 9A is a line graph showing percentage of TdT-i- cells in culture over time when Ai9 reporter 293T cells were co-cultured with 293T cells transfected with the vector in FIG. 9B in the presence (squares) or absence (circles) of doxycycline, n = 3. Statistical Significance determined by two-way ANOVA. *** = P < 0.005, ““ = P < 0.0005. Dox = Doxycycline.

[0052] FIG. 9B is a schematic showing the design of an “all in one" doxycycline inducible VLP-SPIT construct.

[0053] FIG. 10A is a bar graph showing percentage of mCherry-i- TREE reporter 293T cells following application of the indicated amount of supernatant from 293T producer cells secreting ABE VLPs. n = 3. [0054] FIG. 10B is a line graph showing percentage of mCherry+ TREE reporter 293T cells over time following co-culture with 293T producer cells secreting ABE8 RNPs. CRISPR-SPIT constructs were transfected into 293T cells, which were then collected and co-cultured at a 1 :1 ratio with TREE reporter 293T cells (n = 3). Statistical significance determined using two-way ANOVA. * = P < 0.05, * = P < 0.005.

[0055] FIG. 11 A is a schematic illustrating the experimental design for the in vivo experiments. SPIT 293T cells were generated by transfection of Gag-Cre and VSV-g into wild type 293T cells. The SPIT cells were then injected interperitoneally into Ai14 reporter mice. Mice were monitored for 1 .5 weeks post-injection, during which time no mortality or signs of morbidity were noted. After 1 .5 weeks mice were euthanized and Cre recombination determined by TdTomato ("TdT") fluorescence.

[0056] FIG. 11 B is a bar graph showing the total fluorescence (photons/second over background) of different organs (ROI) from experimental (right bar in each pair) and control (left bar in each pair) mice imaged for TdT fluorescence using I VIS, n = 3-5. Experimental mice were injected with cells transfected with SPIT constructs (Gag-Cre and VSV-g); control mice are the same transgenic strain but weren't treated with anything to determine background fluorescence in organs. Statistical significance determined by t-test. P,0.05 = *, P<0.01 = **.

[0057] FIG. 11 C is a bar graph showing percentage of TdT positive hematopoietic cells in the spleen of experimental (right bar in each pair) or control (left bar in each pair) mice, as determined by flow cytometry, n = 3-5. Statistical significance determined using unpaired t- tests.

[0058] FIG. 11 D is a bar graph showing percentage of TdT positive cells in bone marrow from experimental (right bar in each pair) and control (left bar in each pair) mice, as determined by flow cytometry, n = 3-5. Statistical significance determined by unpaired t-tests. BM = Bone Marrow, HSPCs = Hematopoietic Stem and Progenitor Cells, ROI = Region of interest. * = P < 0.05, ** = P < 0.005, *** = P < 0.0005, **** = P < 0.00005.

[0059] FIG. 12 is a bar graph showing frequency of recombination as measured by percentage tdTomato expression in bone marrow (BM) or hematopoietic stem and progenitor cells (HSPCs) of Ai14 mice injected with either 2e8 SPIT cells intra-peritoneally (SPIT) (n=5) or VLPs produced from an equivalent number of 293T cells injected into Ai14 mice retro- orbitally (VLPs) (n=3). VLP mice were analyzed for tdTomato expression 5 months after injection while SPIT treated mice were analyzed 1 .5 weeks after injection. Values of 0 were set to 0.01 so that they could be represented as data points on the graph in log scale. Statistical significance determined by two-way ANOVA. BM = Bone marrow, HSPCs = hematopoietic stem and progenitor cells. ““ = P<0.0001 . [0060] FIG. 13A is a schematic of VSV-g compared to protease regulatable fusogens highlighting the different mechanisms by which the two’s fusogenic activity is regulated.

[0061] FIG. 13B is a schematic showing the co-culture assay performed to detect cell-cell genetic engineering by SPIT.

[0062] FIG. 13C is a schematic showing the different truncated derivatives of GAG-Pol that were tested, with different subdomains indicated. Pro=Protease, RT=Reverse Transcriptase, INT=lntegrase.

[0063] FIG. 13D is a line graph showing the frequency of TdT+ cells detected over time when SPIT cells expressing Amphotrophic 4070A protease regulatable fusogen, GAG-Cre and different GAG-Pol derivatives were co-cultured with Ai14 293T cells at a ratio of 1 :1. The frequency of TdT-i- cells was determined by flow cytometry (n=4, SEM).

[0064] FIG. 13E is a bar graph showing the percent difference in TdT+ cells when using different truncated derivatives of GAG-Pol normalized to wild type GAG-Pol (GAG-Pro-RT-INT) (n=4, SD).

[0065] FIG. 14A is a schematic depicting fusogen-free VLPs being endocytosed by recipient cells and the different outcomes post endocytosis, where either VLPs/cargo are directed to lysosomal degradation or alternatively where cargo from VLPs is able to undergo endosomal escape to enter the cytoplasm of recipient cells.

[0066] FIG. 14B is an electron micrographs of mature and immature retroviral particles treated with detergent, highlighting the greater resistance of immature particles to treatment with detergent, compared to mature particles whose lipid membrane has broken down. Image taken from Wilk et al., J Virol 1999 73: 1931-1940.

[0067] FIG. 14C is a representative image and electron micrographs showing the structural changes that occur in retroviral particles after leaving recipient cells as they mature through protease regulated cleavage events. Taken from Konvalinka et al., Virology 2015 Volumes 479-480: 403-417.

[0068] FIG. 14D is a line graph showing the frequency of TdT+ cells over time as detected by FACs when Ai14 293T cells were co-cultured at a ratio of 1 :1 with SPIT cells expressing either Cre, GAG-Cre or GAG-Cre/GAG-Pol (n=3, SD).

[0069] FIG. 14E is a line graph showing the frequency of TdT+ cells over time as detected by FACS when Ai14 293T cells were co-cultured at a ratio of 1 :1 with control SPIT cells (expressing GAG-Cre/GAG-Pol/membrane GFP) or with SPIT cells co-expressing a tetraspanin (GAG-Cre/GAG-Pol/Tetraspanin) (n=3, SEM). [0070] FIG. 15A is a schematic comparing the formation of VLPs that deliver mRNA through the incorporation of MCP at the C-terminus of GAG, versus VLPs that deliver protein through direct fusion of the protein of interest (POI) to the C-terminus of GAG.

[0071 ] FIG. 15B is a bar graph comparing the frequency of Cre recombination in Ai14 fibroblasts after treatment with VLPs generated using the synthetic GAG derivatives EPN11 and EPN24, when they are engineered to deliver mRNA versus protein (n=3, SD).

[0072] FIG. 15C shows the fold difference in the efficiency of EPN24 in Cre delivery to Ai 14 cells when delivering protein compared to mRNA (n=3, SD).

[0073] FIG. 15D is a schematic of MLV GAG-Cre derivatives tested, highlighting the different subdomains of GAG and different truncated derivatives that were screened for their efficacy in Cre delivery.

[0074] FIG. 15E shows the efficiency of different MLV GAG single domain deletion derivatives in generating VLPs for delivery of Cre to Ai14 fibroblasts. Values are normalized to the efficiency with which wild type MLV-GAG (all domains) was able to induce Cre recombination in cells (n=3, SEM).

[0075] FIG. 15F is a schematic showing different GAG derivatives tested that had multiple domains truncated or deleted. Position of ESCRT recruitment motifs within constructs are indicated (PSAP, PPYL, LYPAL). The positions of the N-terminal domain (NTD) and C-terminal Domain (CTD) within the late domain of GAG are underlined.

[0076] FIG. 15G shows the efficiency of different mult-domain truncated MLV-GAG derivatives in generating VLPs for delivery of Cre to Ai14 fibroblasts. Values are normalized to the efficiency with which wild type MLV-GAG (all domains) was able to induce Cre recombination in cells (n=3, SD).

[0077] FIG. 16A is a schematic outlining the process by which SPIT cells can be generated, packaged into a macroencapsulation device and this device then transplanted subcutaneously into a patient. After transplantation, the patient’s body supplies nutrients and gas exchange to cells, while SPIT cells can deliver genetic engineering enzymes directly to the body via secreted VLPs.

[0078] FIG. 16B is an image of the 20ul device that was used to show the feasibility of macroencapsulation for SPIT delivery through its porous PTFE membrane.

[0079] FIG. 16C is a (Top) brightfield image of Ai14 293T cells cultured together with macorencapsulated SPIT cells and (Bottom) fluorescent image of Ai14 cells expressing tdTomato after co-culture with macroencapsulated SPIT Cells. Scale bar indicates 1000 urn. [0080] FIG. 16D shows the frequency of TdT+ cells detected in culture by flow cytometry after 3 days of co-culture with macroencapsulated SPIT cells (n=1).

DETAILED DESCRIPTION

[0081] A multitude of tools now exist that allow for precise manipulation of the human genome in a myriad of different ways. However, successful delivery of gene editing tools to target cells of human patients remains a significant barrier to clinical implementation. Compared to the limited packaging capacities of contemporary in vivo gene therapy delivery platforms, a human cell's nucleus contains approximately 7.4 billion base pairs of information. The present invention is based in part upon the inventors' insight that human cells harnessed as gene delivery vectors can substantially increase the amount of genetic information available for gene therapies. In addition to harnessing the large packaging capacity of a human cell's nucleus, a delivery system based on human cells further benefits from the potential for longterm engraftment of the cells in human patients and their capacity for more complex genetic programming, such as inducible gene expression, than is possible with other systems.

[0082] LIS 8741340 and US 7186409 describe cell-based therapies that include the creation of stable human transgenic cell lines for subsequent implantation into a subject, either as encapsulated or naked cells, where the cells may produce a therapeutic product. In the gene editing context, US 20220389451 describes a virus-like particle or “VLP” system for delivering gene editing cargo, including CRISPR/Cas9, to a target cell. Similarly, Mangeot et al., J Vis Exp. 2021 Mar 31 ;(169) describes “Nanoblade” technology in which producer cells produce VLPs loaded with Cas9 and sgRNA. US 20210187018 describes a different cell-based delivery vehicle where enucleated cells, referred to as “cytobiologics”, are utilized for delivery of a cargo carried in the lumen or lipid bilayer of the cytobiologic to a target cell in vivo. However, since the cells are enucleated, they do not produce the cargo after administration. WO 2023023528 describes RNA exporter proteins comprising an RNA-binding domain, a membrane-binding domain, and an interaction domain capable of nucleating self-assembly. The RNA exporter proteins are capable of self-assembling into lipid-enveloped nanoparticles (LNs) secreted from a sender cell in which the RNA exporter proteins are expressed, thereby generating a population of LNs comprising a fusogen and exported cargo RNA molecule(s).

[0083] In contrast to the above methods, the SPIT platform described here utilizes human producer cells to deliver gene editing technology cargo proteins, including ribonucleoproteins, directly to target cells in vivo. In aspects, the human producer cells are recombinant primary cells, e.g., induced pluripotent stem cells (iPSC) comprising one or more heterologous genes encoding the protein and/or ribonucleoprotein cargo. The producer cells are delivered to a subject, for example by injection or intravenous infusion, where the cargo proteins are expressed, secreted, and transferred to target cells of the subject, which in turn are genetically modified by the cargo proteins. The SPIT platform offers a number of advantages over existing methods of in vivo gene editing. The human SPIT cells described here are able to deliver larger cargoes and can incorporate genetic logic, for example via the use of inducible gene expression. In addition, human cells have the ability to form stable grafts providing long-term production of gene editing cargo.

SPIT Cells and Cargo

[0084] In aspects, the SPIT producer cells are hematopoietic stem cells or hematopoietic progenitor cells. In aspects, the SPIT producer cells are induced pluripotent stem cells (iPSCs), preferably human iPSCs. In aspects, the SPIT producer cells are neuronal cells, hepatocytes, or fibroblasts. In aspects, the SPIT producer cells are a transformed cell line, such as a fibrobrast cell line, for example 293T cells. Preferably the SPIT producer cells are human cells.

[0085] The present invention provides SPIT producer cells comprising one or more heterologous genes encoding one or more heterologous protein and/or ribonucleoprotein components of a gene editing system, referred to as the “gene editing cargo”.

[0086] In addition to the gene editing cargo, the SPIT producer cells may also comprise one or more additional heterologous genes encoding other protein and/or nucleic acid cargo, such as therapeutic cargo. The additional cargo may include, for example a therapeutic protein. [0087] In addition to cargo intended for delivery to target cells, the SPIT producer cells may also comprise one or more heterologous genes encoding a SPIT system protein or proteins. A SPIT system protein is a heterologous protein that performs a function related to targeting and/or secretion of the SPIT cargo. For example, the SPIT producer cells may also comprise one or more heterologous genes encoding a protease, such as a dimeric and/or aspartyl protease, a pore forming protein such as perforin 2, a targeting protein such as a cell surface receptor protein, a transcription factor, and/or an inducible suicide gene such as Caspase 9 or HSV-TK. In aspects, the heterologous gene encoding the protease encodes the protease as a fusion protein with a GAG protein as defined below. In aspects, a dimeric and/or aspartyl protease or perforin 2 is encoded as a fusion protein with a GAG protein. In aspects, a protease-GAG fusion protein is incorporated into a VLP, as discussed below, where it may perform a variety of functions. For example, the protease may activate a fusogen, and/or regulate the localization, activity, and/or function of one or more other heterologous cargo proteins once VLPs are released from producer cells. In aspects, the SPIT producer cells comprise one or more genetic knock-ins encoding the SPIT system protein(s). A genetic knock-in refers to an insertion of the heterologous nucleic acid into a genetic locus of the producer cells. Alternatively, the SPIT system protein(s) may be encoded on a plasmid.

[0088] In aspects, any one or more of the heterologous genes of the SPIT producer cell may be under the control of an inducible promoter.

[0089] In aspects, the SPIT producer cells comprise one or more heterologous genes encoding nuclease-based gene editing cargo. In aspects, the gene editing cargo includes a nuclease such as a zinc finger nuclease, a transcription activator-like effector nuclease (TALEN), or a Cas nuclease. In aspects, the gene editing cargo includes a single stranded DNA modifying enzyme or “base editor” such as a cytidine base editor (CBE) or an adenine base editor (ABE). In aspects, the gene editing cargo includes a reverse transcriptase for “prime editing” which can accomplish point mutations as well as deletions and insertions. In aspects, the gene editing cargo may also include one or more epigenetic modifying enzymes such as DNA methylase, histone acetyltransferase, deacetylase, and/or topoisomerase.

[0090] In aspects, the SPIT producer cells comprise one or more heterologous genes encoding gene editing cargo selected from one or more of a zinc finger nuclease, a transcription activator-like effector nuclease (TALEN), a Cas nuclease, a single stranded DNA modifying enzyme such as a cytidine base editor or an adenine base editor, a reverse transcriptase, a DNA methylase, a histone acetyltransferase, a deacetylase, a topisomerase, and combinations of any of the foregoing.

[0091] In aspects, the SPIT producer cells comprise one or more heterologous genes encoding CRISPR-based gene editing cargo. The term “CRISPR” refers to a system for genetic modification utilizing a class of enzymes, the clustered regularly interspaced short palindromic repeats (CRISPR)-associated (Cas) endonucleases. Cas proteins contain an endonuclease domain for nucleic acid cleavage and at least one RNA binding domain that interacts with a guide RNA. CRISPR gene editing technology utilizes ribonucleoprotein (“RNP”) complexes of a Cas endonuclease and a synthetic guide RNA (gRNA). The gRNA of the RNP comprises an enzyme-specific region, which binds to the Cas endonuclease, and a region complementary to a target nucleic acid, which may be referred to as the “recognition sequence” or the “guide sequence” of the gRNA. A “target sequence” refers to the sequence of a target nucleic acid that is complementary to the guide sequence of a gRNA. In the context of gene editing, the target sequence may be, for example, a sequence of a genomic DNA. Thus, a “guide RNA” or “gRNA” refers to an RNA molecule that binds to a Cas protein and targets the Cas protein to a target sequence, e.g., within a genomic DNA. Some gRNAs contain two separate RNA molecules, referred to as an “activator-RNA” and a “targeter-RNA”, which may also be referred to as a tracrRNA and a crRNA, respectively. Other gRNAs contain the crRNA and tracrRNA associated as a single RNA molecule and may be referred to as a “single-guide RNA” or an “sgRNA.” In aspects, an sgRNA comprises a crRNA fused to a tracrRNA, optionally via a linker polynucleotide.

[0092] Accordingly, in aspects, the SPIT producer cells comprise one or more heterologous genes encoding CRISPR-Cas gene editing cargo which may include a Gas endonuclease and a gRNA. The gRNA may include one or more of a tracrRNA, a crRNA, or an sgRNA. In aspects, the Gas endonuclease is a Cas9 or a Cas12. In aspects, the Gas endonuclease is a Cas13. In aspects, the endonuclease is a eukaryotic RNA-guided endonuclease, for example a Fanzor protein as described in Saito et al., Nature 2023 620:660-668.

[0093] In aspects, the SPIT producer cells comprise one or more heterologous genes encoding CRISPR-Cas gene editing cargo where the cargo comprises a ribonucleoprotein comprising a Gas endonuclease and a guide RNA. In aspects, the Gas enzyme is Cas9. In aspects, the Cas enzyme is Cas12a, also referred to as Cpf1. In aspects, the gRNA is an sgRNA.

Delivery Platforms

[0094] Delivery of SPIT cargo to target cells may be accomplished in several ways. In one aspect, delivery is achieved by passive release/delivery via fusion with a cell penetrating peptide (“GPP”) or active secretion of GPP fusion protein by utilizing a signal peptide in which the GPP is fused with the component proteins and/or ribonucleoproteins of the gene editing cargo expressed by the SPIT producer cells. In a preferred aspect, delivery is achieved by formation and secretion of a VLP containing the gene editing cargo, and optionally other cargo as discussed above. In some aspects, the VLP may express a fusogen. Accordingly, in aspects the SPIT producer cells comprise one or more heterologous genes encoding one or more gene editing cargo proteins and/or ribonucleoproteins as fusion proteins with a GAG protein. In aspects, the GAG-cargo fusion protein also includes a GPP. In aspects, the GAG- cargo fusion protein also includes a pore forming protein, such as perforin 2. In aspects, the pore forming protein is included as a separate GAG fusion protein. In aspects, where the delivery platform includes fusion with a GAG protein, the SPIT producer cells may also comprise a heterologous gene encoding a fusogen.

[0095] In aspects, the signal peptide is an IL-6 signal peptide, an Igk signal peptide, a preproinsulin signal peptide or an albumin signal peptide.

[0096] In aspects, the GPP is antennapedia, mastoparan, penetratin, SV40 nuclear localization signal, TATp, or transportan. [0097] As used herein, the term “GAG protein” refers to a group of proteins capable of performing the functions of (1 ) translocation and intercalation into a cell membrane and (2) oligomerization with other GAG proteins leading to budding and formation of VLPs. A GAG protein as used herein may refer to a structural group-specific antigen (Gag) protein of a retrovirus, a non-viral protein, or a synthetic protein.

[0098] In aspects, the GAG protein is a structural group-specific antigen (Gag) protein of a retrovirus, such as a murine leukemia virus, a human immunodeficiency virus, a rotavirus, or a hepatitis B virus, or a derivative of any of the foregoing. Gag proteins are the core structural proteins of the retroviral capsid and are capable of independently assembling into virus-like particles (VLPs) both in vivo and in vitro.

[0099] In aspects, the GAG protein is a non-viral protein, preferably a human protein, selected from an Arc protein, ASPRV1 , a Sushi-Class protein, a SCAN protein such as PGBD1 , or a PNMA protein. In aspects, the Arc protein is hARC or dARC1 . In aspects, the PNMA protein is ZCC18, ZCH12, PNM8B, PNM6A, PNMA6EJ2, PMA6F, PMAGE, PNMAI, PNMA2, PNM8A, PNMA3, PNMA4, PNMA5, PNMA6, PNMA7, PNMAI, MOAPI, or CCD8.

[0100] In aspects, the GAG protein is PEG10, RTL3, RTL10, or RTL1.

[0101] In aspects, the GAG protein is a truncated GAG-pro-pol protein including only the Gag and the protease domains.

[0102] In aspects, GAG is a GAG-TEV fusion protein.

[0103] In aspects, GAG is a GAG fusion protein with the aspartyl protease of PEG10.

[0104] In aspects, the GAG protein is a synthetic protein comprising a membrane binding domain, such as the membrane binding domain of Lyn1 , and an oligomerization domain, such as a leucine zipper domain, a coiled-coil domain, a tetratricopeptide repeat domain, an ankyrin repeat domain or a CA domain from a retroviral GAG.

[0105] In aspects, the fusogen is VSV-g, RD114a, amphotrophic envelope 4070A, measles F/H, syncytin, or myomaker/myomixer.

[0106] In aspects, the fusogen may comprise or consist of a fusogenic peptide. In aspects, the fusogenic peptide comprises or consists of the 19 amino acids of human syncytinl at positions 322-340, referred to as “S19” and described by Sudo et al., J Control Release 2017 Jun 10:255:1-11 . In aspects, the fusogenic peptide comprises or consists of Glu-Ala-Leu-Ala, or “GALA”. In aspects the fusogen may be a lipid GALA, a cholesteryl-GALA or a PEG-GALA.

[0107] In aspects, the fusogen comprises or consists of the 122 amino acid G protein of vesicular stomatitis virus, or “VSV-g”, or an operable fragment thereof. [0108] In aspects, the fusogen comprises or consists of as sperm fusogen, such as IZUMO1 , or a muscle fusogen such as myomaker/myomixer.

[0109] In aspects, the fusogen may target a specific type of cell, for example as described in Nakamura et al., Nat Biotechnol. 2004 Mar;22(3):331 -6. In aspects, the fusogen is VSV-g, RD114a, amphotrophic envelope 4070A, measles F/H, syncytin A, myomaker/myomixer.

[0110] In aspects, SPIT producer cells are delivered to target cells by injection.

[0111] In aspects, SPIT producer cells are delivered in a bioencapsulation device which is transplanted into a recipient. The device may be a macroencapsulation device.

VLP-SPIT

[0112] As a proof of concept, the disclosure provides experiments utilizing VLPs as the delivery platform for SPIT cargo. Utilizing human producer cells to manufacture and deliver VLPs in vivo provides numerous advantages over methods where the VLPs are instead produced in vitro and delivered to a subject. For example, VLPs are generally highly unstable and degrade rapidly during the in vitro collection process. In addition, a proportion of the in vitro produced VLPs may also re-transfect producer cells and be lost. This means that, in practice, the amount of VLPs successfully collected from culture supernatant is substantially less than the total amount actually produced by the cells. In contrast, VLPs produced and secreted in vivo provide much higher yield of VLPs for delivery of cargo to target cells, resulting in more efficient gene editing, as discussed in more detail infra. In addition, utilizing SPIT producer cells as described herein affords the further advantage of incorporating genetic logic, such as by incorporating inducible promoters into the heterologous genes of the SPIT producer cells encoding the cargo proteins.

[0113] In aspects, the gene editing cargo comprises a Cas endonuclease and a gRNA, preferably an sgRNA. In accordance with this aspect, the SPIT producer cells comprise heterologous genes encoding a GAG-Cas fusion protein and a gRNA. In aspects, the SPIT producer cells may also comprise a heterologous gene encoding a fusogen. Thus, in accordance with this aspect the SPIT producer cells produce a GAG-Cas fusion protein which complexes with the gRNA to form the ribonucleoprotein complex which is encapsulated in a VLP for delivery to target cells. In aspects, the VLP may include a fusogen.

Cell Therapy

[0114] In aspects, the disclosure provides methods of in vivo genetic engineering that comprise administering a pharmaceutical composition comprising SPIT producer cells, as described herein, to a subject in need of such therapy. In aspects, the SPIT producer cells may be administered by injection or intravenous infusion. In aspects, the subject is a nonhuman vertebrate, for example, a dog, cat, a rodent (e.g., a mouse, a rat, a rabbit), a horse, a cow, a sheep, a goat, a chicken, a duck, or any other non-human vertebrate. In aspects, the subject is a human subject.

[0115] The disclosure also provides a pharmaceutical composition comprising SPIT producer cells as described herein. The pharmaceutical composition may also comprise one or more pharmaceutically acceptable carriers and/or excipients. Pharmaceutically acceptable carriers and excipients are defined, for example, by the U.S. Pharmacopeia-National Formulary (USP- NF) and related regulatory documents. In aspects, the pharmaceutically acceptable carrier is an aqueous solution buffered to physiological pH, saline or other physiologically buffered salt solution, or a cell culture medium, preferably a phenol red free and serum free medium, optionally supplemented with human serum albumin. In aspects, the pharmaceutically acceptable carrier is an aqueous buffered salt solution, such as Ringer's solution formulated for injection, or a phenol-red free cell culture medium such as Dulbecco's Modified Eagle Medium.

[0116] The experiments below describe proof of concept studies of a human cell-based platform for delivery of gene editing technology to target cells for in vivo genetic engineering. The following demonstrates the successful application of the SPIT platform for delivery of Cre and CRISPR/Cas9 gene editing systems to accomplish in vivo editing of a target cells in mice where several organs are genetically modified, including liver, spleen, intestines, peripheral blood, and bone marrow. Also demonstrated is a method for inducibility of the system using a small molecule, doxycycline, illustrating the incorporation of “genetic logic” into the SPIT platform.

[0117] The SPIT approach to in vivo gene therapy is illustrated by the schematic in FIG. 1 A. SPIT includes producer cells (left) are bioengineered to express one or more components of a gene editing system, such as a CRE recombinase or a Cas-gRNA ribonucleoprotein, and secrete it in a form that is taken up by target cells (right) in vivo which are then subjected to gene editing. Both the Cas endonuclease and guide RNA may be expressed by the producer cells such that the secreted particle contains the Cas-gRNA ribonucleoprotein complex, as illustrated in the schematic. Alternatively, the gRNA may be delivered directly to the target cells or tissues for complexation with Cas9 intracellularly or extracellularly.

Example 1 [0118] Preliminary experiments identified virus-like particles (VLPs) as a suitable delivery modality. VLPs were able to be assembled and secreted with a Cre recombinase which was successfully delivered to target cells.

[0119] FIG. 1 B shows a schematic of basic VLP components and illustrates the budding of a VLP from the membrane of a producer cell. In this system the protein of interest, or “POI”, is fused to the C-terminus of the retroviral Group-specific antigen (Gag) protein and a fusogen, VSV-g, is co-expressed as a transmembrane protein. Gag proteins are the core structural proteins of the retroviral capsid and are capable of independently assembling into virus-like particles (VLPs) both in vivo and in vitro. In this context, the expressed fusion protein consisting of POI-GAG, is released from the delivery cells in a virus like particle (VLP) which fuses with the target cell membrane and delivers the POI into the target cell.

[0120] For experiments where VLPs were isolated from the supernatant of producer cells and applied to recipient cells the following procedures were followed. 293T producer cells were obtained by transfection with 2 ug of the appropriate plasmid or plasmids, up to a maximum of 8 ug where multiple plasmids were transfected. Transfections were performed using the PEI MAX® reagent in accordance with the manufacturer's instructions. Producer cells were maintained for 3 days after transfection followed by collection of supernatant which was filtered through a 0.45 urn filter and concentrated using an Amicon 100kDa filter. Reporter cells were prepared on the same day as supernatant collection. Concentrated VLPs were applied to reporter cells just after plating.

[0121] In preliminary experiments, widely available retroviral-based VLP systems were able to package and deliver proteins into cells in vitro. However, such systems are incompatible with in vivo use in humans, at least because they utilize more than two thirds of the viral genome, which increases the risk of some form of recombination occurring in target cells leading to formation of functional, infective, and/or replication competent virus. See NIH Guidelines for Research Involving Recombinant or Synthetic Nucleic Acid Molecules April 2019, Section lll-D-4-a. Accordingly, to develop a suitable VLP platform for in vivo use, a screen was conducted to determine whether any of the viral genes in the VLP system could be eliminated without impairing the ability of the VLP to deliver its cargo.

[0122] A retroviral genome has four key genetic components (1 ) a packaging signal which allows viral RNA to be packaged into viral particles and a nucleocapsid domain which is able to bind to and package it; (2) Gag-pro-pol, a polyprotein comprising Gag which functions in membrane binding and oligomerization, a protease ("pro"), and a reverse transcriptase ("pol"), which are usually separated from each other by proteolytic cleavage during viral formation; (3) a fusogen; and (4) other nonessential accessory proteins that vary substantially between viral species. The first three are essential for the retrovirus life cycle. In considering the construct of a VLP platform for use in vivo, if the construct contains a fusogen, a Gag fusion protein with the protein of interest, and Gag-Pro-pol then it will comprise more than two thirds viral genome and be unsuitable.

[0123] Ai 14 reporter fibroblasts were used as a model system to test the effects of deleting Gag, Gag-Pro, and the fusogen on VLP delivery of a Ore recombinase. In the Ai14 system, Ore recombinase activity is measured by fluorescence of a reporter, TdTomato (Takara Bio USA). Delivery of Cre by VLP produced under the various conditions is measured as percentage TdT positive to reporter cells.

[0124] The following constructs and combinations were tested: (1 ) a construct containing wild-type Cre ("Cre") alone; (2) a Gag-Cre fusion, lacking both pro and pol ("Gag-Cre"); (3) a Gag-Cre fusion and VSV-g; and (4) Gag-Cre, Gag-Pro-Pol and VSV-g. Results for (1)-(4) are shown, left to right, in FIG. 1C. The four bars represent increasing amounts of VLPs as a percentage of producer cell supernatant, 0.05 %, 0.5 %, 5.0 %, and 50 %, n=3 for each condition. As shown in the figure, no difference was found in efficacy of inducing Cre recombination for GAG-Cre VLPs formulated with or without GAG-Pro-Pol. Supernatant from cells transfected with Cre and Gag-Cre (only) were also collected and applied to reporter cells and neither induced recombination.

[0125] Thus, while elimination of the fusogen (VSV-g) completely abolished VLP efficacy, each of the Pro and Pol components of Gag-Pro-Pol could be eliminated without a statistically significant effect on the delivery of Cre (compare Gag-Cre, VSV-g to Gag-Cre, Gag-Pol and VSV-g). Surprisingly, despite the role of Gag-pro-pol in proteolytically releasing a protein of interest from Gag during VLP formation, it was not essential for the packaging, delivery, or activity of Cre by VLPs.

[0126] Since Gag intercalates into the cell membrane it was possible that the Gag-Cre fusion protein would be non-functional in the absence of the protease to cleave Gag from Cre. A preliminary experiment was conducted using Ai9 293T cells transfected with wild type Cre and Gag-Cre plasmids. Recombination in reporter cells was detected three days later by TdT expression. As shown in FIG. 1 D no significant differences were found in the rate of recombination between the two constructs, demonstrating that fusion to Gag does not impair the efficacy of Cre. Thus, Gag-Cre can function for Cre recombination even when it is not cleaved from Gag by Gag-pro-pol.

[0127] Plasmids used in the preceding experiments include VSV-g (12259), GAG-Cre (119971 ) and GAG-Pro-Pol (35614) which were obtained from Addgene (Watertown, MA). Producer cells were prepared and VLPs obtained as described above. Ai14 reporter cells were analyzed by flow cytometry for TdTomato expression three days after application of VLPs. Ai14 fibroblasts and mouse HSCs were generated from Ai14 mice as previously described in Bravo et a!., STAR Protoc. 2021 2(2):100406.

[0128] In order to confirm the results above showing that Gag-pol-pro was not essential for VLP formation and delivery of cargo protein, another experiment was performed in a different model system, this time using the CRISPR-Cas9 adenine base editor (ABE8) as the cargo protein. The model system, TREE, is a transient reporter system described in Standage-Beier et al., Nucleic Acids Res. 2019 47(19):e120. In this system, successful gene editing by ABE8 delivered to the reporter cells via ABE8 VLPs is detected by expression of the fluorescent reporter mCherry. This transient system was introduced into 293T cells to make a permanent reporter cell line by piggybac transposase insertion.

[0129] Two VLP formulations were compared for their ability to deliver ABE8 to recipient cells for gene editing: (1 ) transfection of Gag-ABE8, Gag-Pro-Pol and VSV-g into producer cells; and (2) transfection of Gag-ABE8 and VSV-g into producer cells. The percentages of mCherry-i- cells detected after application of VLPs to reporter cells is shown in FIG. 2A. Plus and minus signs refer to inclusion or exclusion of Gag-Pro-Pol in the VLP formulations, (1 ) and (2), respectively. The four bars represent increasing amounts of VLPs as a percentage of producer cell supernatant, 0.05 %, 0.5 %, 5.0 %, and 50 %, n=3 for each condition. In this experiment, the number of cells successfully edited by ABE8 was substantially reduced when ABE8 VLPs were produced in the absence of Gag-pro-poL Thus, while the viral Gag-pro-pol was not necessary for achieving gene editing in reporter cells with ABE VLPs, inclusion of Gag-pro-pol in VLP production led to a statistically significant increase in the number of cells edited.

[0130] The Gag-ABE8 construct was obtained from Banskota et al., Cell 2022 185(2):250- 256. e16.

[0131] FIG. 2B shows a schematic of vectors used to deliver Cre and ABE8 in the preceding experiments. As shown in the figure, GAG-ABE8 is engineered slightly differently than GAG- Cre. The Gag-ABE8 construct has several nuclear export signals (3x NES) upstream of the site cleaved by the protease within Gag-Pro-Pol. These NES cause Gag-ABE8 to be localized in the cytoplasm of producer cells. Subsequently, during VLP formation, Gag-Pro-Pol cleaves ABE8 from Gag at a site after the NES signals, permitting ABE8 to localize to the nucleus in the target cells.

[0132] In contrast, the GAG-Cre construct has a structure of the Gag cleavage site, SV40NLS-Cre. Genetic recombination is not improved by adding Gag-Pro-Pol into VLPs since this construct lacks a 3x NES signal. [0133] Together, the data from these experiments show that the protease within Gag-Pro-Pol is not essential either for VLP to deliver their cargo, e.g., Cre or CRISPR gene editing proteins. This finding is significant because Gag-pro-pol is a polyprotein expressed in retroviruses that contains Gag, protease, an RNA binding domain, a reverse transcriptase and an integrase. Its use in vivo could lead to the development of a functional virus via an unintended recombination event. Therefore, the ability to eliminate this polyprotein from the VLP is important for its translation to clinical use.

[0134] Another key insight from these data is that, although the protease is not essential, it can improve efficacy depending on the construct utilized. Thus, inclusion of a protease can be used to regulate the activity of a protein of interest and/or the VLPs.

[0135] FIG. 3A is a schematic illustrating formation of VLPs when GAG-Pro-Pol is included (Left) or not (Right). In VLPs where GAG-Pro-Pol is not included the protein of interest (POI) remains fused to GAG, which is capable of intercalating into and leaving the membrane.

[0136] As discussed above, a key insight of the previous experiments was that inclusion of a protease can be used to regulate activity of a cargo protein and/or the VLP. For example, a protease may be used to cleave the cytoplasmic tails of fusogens when VLPs are released from the cellular membrane of producer cells. This would allow the fusogens to be active only after the VLP has budded from the cell surface, thereby preventing cell-cell fusion of producer cells, which is a significant cause of toxicity.

[0137] Unlike VSV-g, which is often loaded into the cellular membrane in an active state, the activity of retroviral fusogens is regulated by a cytoplasmic tail, referred to as an R-peptide. While attached to the fusogen the R-peptide inhibits its fusogenic activity. The R-peptide contains a consensus sequence for the retroviral protease of GAG-pro-pol and is cleaved from the fusogen following release of viral particles from the cell membrane. Once the peptide is cleaved fusogenic activity is no longer inhibited, allowing the fusogen to facilitate fusion of viral particles with recipient cells.

[0138] Notably, the protease cleavage site of the R-peptide varies between different retroviral species and retroviral fusogens from one species cannot typically be cleaved by the retroviral protease of another species. To make fusogens from different retroviral species compatible, the R-peptide from one species can be swapped onto the fusogen of another species. For example, the fusogen RD114A (Simian immunodeficiency virus (SIV) fusogen) was generated by replacing the R-peptide of RD114 with the R-peptide from MLV Amphotrophic 4070A.

[0139] To test the protease regulatable nature of alternative viral fusogens, Cre-VLPs were produced by transfection of GAG-Cre and different retroviral fusogens into 293T cells either with or without Gag-pro-pol. VLPs were applied to Ai9 293T reporter cells and analyzed for Cre recombination by TdT expression 3 days later. As shown in FIG. 3B, each of the retroviral fusogens tested, RD114A and Ampho4070A (left to right) were only able to deliver Cre to reporter cells when Gag-Pro-Pol was included (+) in VLP production. Thus, both the native MLV fusogen (amphotrophic 4070A) and the SIV fusogen RD114A were effective for delivery of Cre by VLPs with their fusogenic activity being regulated by MLV Gag-pro-poL

[0140] Regulation of VSV-g and retroviral fusogens by different mechanisms is illustrated schematically in FIG. 3C. The schematic highlights how retroviral fusogens only become fusogenic after protease cleavage by the retroviral protease found in Gag-pro-pol, once VLPs form and leave the cell membrane.

[0141] In general, the use of retroviral elements in the SPIT system is undesirable due to the potential immunogenicity of viral components. Accordingly, further experiments were conducted to identify a non-viral fusogen for the SPIT system.

[0142] mSynA is a fusogen acquired from retroviruses during mammalian evolution. VLPs were prepared via transfection of GAG-Cre and mSynA into 293T cells. However, by three days post-transfection there was almost 100% cytotoxicity of the producer cells which had fused into large syncytia. Without being bound by any particular theory, it may be that the protease regulatable tail of mSynA had diverged from its retroviral origins to render mSynA constitutively active. To test this, the cytoplasmic tail of mSynA was replaced with the R- peptide from the MLV amphotrophic 4070A fusogen to produce an mSynA-ampho 4070A construct. FIG. 4A shows a schematic representation of the generation of the mSynA-ampho 4070A construct.

[0143] The mSynA-ampho 4070A and wild-type mSynA constructs were transfected into 293T cells. Unlike the wild-type construct, the modified construct was not cytotoxic. In addition, the mSynA-ampho 4070A construct produced a protease regulatable fusogen, as shown in FIG. 4B. VLPs produced by transfection of 293T cells with GAG-Cre and either wild-type mSynA or mSynA-ampho 4070A constructs in the presence (+) or absence (-) of Gag-pro-pol were applied to Ai9 reporter cells and the frequency of recombination after three days was measured as percent TdT positive cells. No recombination occurred in the absence of Gag- pro-pol for VLP produced with the mSynA-ampho 4070A constructs (right pair of bars), compared to 0.745% TdT-i- cells in the presence of Gag-pro-pol. The wild-type construct produced 0.13% TdT-i- cells in the absence of Gag-pro-pol.

[0144] Having demonstrated that addition of the R-peptide to mSynA made it protease regulatable, a series of experiments was performed to test whether this strategy could be used to produce protease regulatable fusogens from non-retroviral species and to test whether addition of other protease cleavable sequences could be added to the fusogens' cytoplasmic tails to the same end, creating synthetic R-peptides.

[0145] VSV-g was selected as a model fusogen due to its broad tropism and widespread use in recombinant viral platforms. VSV-g is derived from vesicular stomatitis virus (VSV), which is from a different family of viruses, Rhabdoviridae, than retroviruses, Retroviridae.

[0146] The R-peptide from MLV amphotrophic 4070A and turbo GFP (tGFP) with the consensus sequence for MLV protease found at the N-terminus of tGFP, were added to the cytoplasmic tail of VSV-g. tGFP was also placed at the C-terminus of the R-peptide from MLV in order to detect expression of this construct in producer cells.

[0147] FIG. 5A is a schematic showing the design of these two protease regulatable recombinant VSV-g fusogens, referred to as VSV-g + ampho tail and VSV-g + tGFP.

[0148] Murine leukemia virus VLPs that packaged Cre were produced by transfection of 293T cells with each fusogen, VLPs were harvested and applied reporter cells. Recombination was evaluated as % TdT positive cells. FIG. 5B compares the frequency of recombination (% TdT positive cells) in Ai9 293T cells three days after application of Cre-VLPs produced with the recombinant VSV-g fusogen, VSV-g + tGFP, either in the presence (+) or absence (-) of Gag- pro-pol, n=2 for each condition. VLPs produced in the absence of Gag-pro-pol showed very low (0.04%) recombination. In contrast, VLPs produced in the presence of Gag-pro-pol had significantly higher rates of recombination.

[0149] FIG. 5C compares the frequency of recombination (% TdT positive cells) in Ai9 293T cells three days after application of Cre-VLPs produced with the recombinant VSV-g fusogens VSV-g + ampho tail in the presence (+) of Gag-pro-pol, n=2. VLPs produced in the presence of Gag-pro-pol showed high rates of recombination. Addition of tGFP to the cytoplasmic tail of VSV-g + ampho was used to detect expression.

[0150] These results demonstrate that the addition of the amphotrophic 4070A tail (with tGFP at its C-terminus) or the addition of tGFP directly (with the MLV protease consensus sequence at its N-terminus) were able to regulate the activity of VSV-g and that removal of these cytoplasmic tail sequences facilitated their fusogenic activity. Taken together the results shown in foregoing experiments demonstrate the feasibility of making protease regulatable fusogens both utilizing the retroviral R-peptide as well as other amino acid sequences/proteins added to the cytoplasmic tail of a non-retroviral fusogen, such as VSV-g to create a synthetic R-peptide.

[0151] In order to determine whether the MLV Gag could be replaced with a non-viral Gag homolog for delivery of a protein cargo, constructs were made to fuse Cre to the the C- terminus of PEG10. The Cre-PEG10 constructs were transfected into 293T producer cells along with VSV-g. FIG. 6A shows a schematic of the experiment.

[0152] VLPs produced from Cre-PEG10 and VSV-g transfected producer cells were harvested and applied to reporter cells. VLPs produced from MLV Gag-Ore and VSV-g transfected cells served as a positive control. Recombination was evaluated as % TdT positive cells. Results are shown in FIG. 6B. The results show that PEG10 could successfully be modified to deliver Ore to reporter cells by this approach. Although Cre-PEG10 was not as efficient as MLV Gag-Ore, these results nevertheless demonstrate the feasibility of utilizing endogenous retroviral Gag homologs in place of the retroviral Gag in the SPIT system.

[0153] Next, the Gag-Cre VLP constructs were used to test the SPIT system in a proof-of- concept in vitro experiment. Gag-Cre and the VSV-g fusogen were transfected into wild type 293T cells (wt293T) followed one day post transfection by mixing these producer cells with reporter Ai9 293T cells at a ratio of 1 :1. Cre recombination in the reporter cells over time was assessed by Td tomato expression. A schematic of the experimental design is shown in FIG. 7A. Ai9 293T cells were generated through knockin of the Ai9 reporter construct into the AAVS1 locus by homology dependent repair and puromycin selection (2 ug/ml).

[0154] The VSV-g (12259) plasmid was obtained from Addgene (Watertown, MA). For this and similar co-culture experiments, 293T cells were plated at a concentration of 9e4 cells/cm^A2 one day before transfection on a 12 well plate. Cells were transfected with 500 ng Gag-Cre plasmid or 100 ng of VSV-g using PEI MAX® in accordance with the manufacturer's instructions. Cells were collected 1 day after transfection and then co-cultured with reporter 293T cells at a ratio of 1 :1 in 12 well plates at a concentration of 90,000 cells/cm^A2. Cells were split and analyzed by FACs once every two days post co-culture and split 1 :1 .

[0155] The results, shown in FIG. 7B, indicated that an average of 14 % of cells became TdTomato positive after 6 days of co-culture with SPIT cells, compared to 0.095% when reporter cells were cultured alone (P < 0.0001 at day 6, ANOVA). The same experiment was performed using primary mouse fibroblasts as the target cell. With primary cells, an average of 0.79% of cells became positive for Td Tomato expression after 4 days of co-culture, compared to 0.02% when reporter cells were cultured alone (ANOVA, P = 0.0013).

[0156] Next, the inventors tested whether it is feasible to add genetic logic to the SPIT system by incorporating an inducible promoter. A doxycycline system was used as a proof of concept. FIG. 8A illustrates the construction of several all-in-one inducible vectors with different promoter/gene orientations constructed for this experiment. Efficacy for doxycycline regulatable gene expression was determined using TdTomato and GFP. For experiments where a doxycycline inducible plasmid was used, doxycycline was added to cells at a concentration of 2ug/mL Doxycycline inducible elements were synthesized by Gene Universal Inc. (Newark DE) from previously published sequences. The different doxycycline inducible plasmids utilized in this study were then constructed from this synthesized vector using Gibson Assembly®.

[0157] As shown in FIG. 8B, each of the three vectors performed similarly, indicating that the orientation of genes and promoters in the cassette did not have a significant effect on gene expression. Vector A was selected to produce the inducible SPIT cassette.

[0158] Next, Gag-Cre and VSV-g were placed under the control of doxycycline using an inducible promoter construct. The construct was transfected into wild-type 293T cells followed by co-culturing of the transfected cells with reporter 293T cells in the presence or absence of doxycycline. As shown in FIG. 9A, an average of 4% of cells were positive for Td Tomato expression when cells were co-cultured in the presence of doxycycline, compared to only 0.06% when cells were co-cultured in its absence (ANOVA, P = < 0.0001), demonstrating proof-of-principal for the incorporation of genetic logic into a SPIT platform.

[0159] The ability of the SPIT system to deliver CRISPR/Cas ribonucleoproteins for gene editing was evaluated next using a TREE reporter cell line, discussed above. Briefly, in this system, successful gene editing is detected via expression of a fluorescent reporter, in this case mCherry. The sgRNA for the TREE reporter system (164413) and the plasmids GAG- ABE (181751 ) and ABE (164415) were obtained from Addgene (Watertown, MA). To generate TREE 293T reporter cells, the ABE TREE reporter construct (164411 ) was cloned from its original vector into a Piggybac® plasmid with a puromycin selection cassette (pb-TREE). In addition, the fluorescent reporter activated by ABE gene editing was changed from GFP to mCherry during cloning. TREE 293T reporter cells were generated by Piggybac® insertion of the pb-TREE reporter construct into cells by chemical transfection of the pb-TREE plasmid together with a hyPBase plasmid. The hyPBase plasmid is described in Proc Natl Acad Sci USA 108:1531 (2011 ). Reporter cells were then selected for by puromycin selection (2 ug/ml). Reporter cells were split 3 days after the application of VLPs and cultured for an additional 3 days, after which cells were analyzed for mCherry expression by flow cytometry.

[0160] FIG. 10A shows data from applying the supernatant (VLPs) from producer cells transfected with Gag-ABE8, Gag-Pro-Pol and VSV-g onto TREE reporter cells, demonstrating that these VLPs work for gene editing.

[0161] Gag-ABE8, Gag-pro-pol and VSV-g plasmids were transfected into wild-type 293T cells and the transfected cells were co-cultured with TREE 293T reporter cells at a ratio of 1 :1. FIG. 10B shows that an average of 4.7 % of cells were positive for mCherry expression six days post co-culture, compared to 0.073% for reporter cells were cultured alone (ANOVA, P < 0.0001 ). This evidences that the SPIT system successfully produced VLPs that could deliver an ABE RNP for gene editing via cell-to-cell transfer.

[0162] Having established that the SPIT system was effective for cell-to-cell genetic engineering in vitro, a series of experiments was designed to test its application in vivo.

[0163] Two cell types that are easily amenable to genetic modification by plasmid transfection, C57BL6/j mouse embryonic fibroblasts (MEFs) and human 293T cells, were selected as potential vectors for SPIT. The efficacy of each cell type as a SPIT vector was compared through transfection of a luciferase expression plasmid into the cells followed by intra-peritoneal (IP) injection of transfected cells into mice, and tracking luciferase expression over time, the inventors found that both cell types could transiently engraft and express a transfected gene for at least two days in vivo. There were no statistically significant differences in the persistence of luciferase expression found between the cell types transplanted or the immunological setting of the host they were transplanted into i.e. , xenogeneic, allogeneic, or syngeneic hosts.

[0164] The human 293T cells were selected for further study. FIG. 11 A illustrates the experimental design. SPIT 293T cells were generated by transfection of Gag-Ore and VSV-g into wild type 293T cells and the transfected cells were IP injected into Ai14 reporter mice. Mice were monitored for 1 .5 weeks post-transplantation, during which time they exhibited no mortality or signs of morbidity. After 1 .5 weeks mice were euthanized and Ore recombination determined by TdTomato ("TdT") expression.

[0165] Ai 14 and C57BL6/j mice were either purchased from Jackson Laboratories or bred inhouse, while CD1 mice were purchased from Charles River Laboratories. For experiments tracking luciferase expression in vivo 1.5e7 293T cells or 2.9e6 C57BL6/j MEFs (ATCC) were plated onto 15 cm plates one day before transfection. Cells were transfected one day after playing with 10 ug of a luciferase expression vector. Twelve hours post-transfection cells were collected from plates using TrypLE (Thermo Fisher Scientific). 2e7 cells were re-suspended in 200 ul of PBS and injected intraperitoneally into mice using a 22-gauge needle. For in vivo SPIT experiments 6.5e6 293T cells were plated onto a 15 cm plate two days before transfection. Two days pos- plating each plate was transfected with 9 ug of GAG-Cre and 1 ug of VSV-g, using 60 ug of PEI max per transfection. Twelve hours post transfection cells were collected and 2e8 cells were re-suspended in 400 ul of PBS and injected into Ai14 reporter mice intraperitoneally using a 22-gauge needle. Following dissection of SPIT-treated mice images of the peritoneum were taken using Xcite-GR fluorescence flashlight and appropriate filter. For quantitative imaging of organs, an I VIS® imager (PerkinElmer) was used with an excitation of 535 nm and emission of 590 nm. For luciferase imaging experiments mice were injected with 0.15 mmol of Akaluciferase and spectral luminescence was measured after 10 minutes using an IVIS imager. Akaluciferase is described in Bozec et al., Neurooncol Adv. 2020 Oct 10;2( 1 ). The amount of radiance from each organ or mouse was determined using Aura imaging software (Spectral Instruments).

[0166] Dissection of mice revealed clear TdTomato expression in multiple organs and tissues of SPIT-treated mice including: the diaphragm, liver, spleen, perigonadal fat, and in some cases the intestines. For a quantitative analysis of Cre-mediated recombination, solid organs were extracted from mice and the intensity of TdTomato fluorescence in each organ measured with an IVIS imager. Results are shown in FIG. 11 B. Statistical analysis of the fluorescence intensity of each organ indicated a significantly higher amount of TdTomato fluorescence in the livers (mean ROI/Background control = 1.3, SPIT = 4, p = 0.025, n = 3-5) and spleens (mean ROI/Background control = 0.88, SPIT = 2, p = 0.003) of SPIT-treated mice compared to untreated controls. Notably, the intestines of some SPIT-treated mice also showed significantly increased TdTomato expression compared to controls by IVIS imaging, albeit with considerable sample to sample variability.

[0167] To ascertain whether SPIT delivered Cre recombinase systemically and to validate Ore recombination at a single cell level, cells from the spleen, peripheral blood (PB), and bone marrow (BM) of mice were collected for flow cytometric analysis. In the spleen, as shown in FIG. 11 C, the 1.2-fold increase in fluorescence intensity measured by IVIS corresponded to 6.42% of splenocytes (CD45+) detected as positive for Tdtomato expression by flow cytometry, compared to only 0.021% of cells from the control. Further examination of specific hematopoietic lineages within the spleen found that 1.8% of B-cells (B220+), 1 % of T-cells (CD4+/8+), and 3.7% of Myeloid Cells (Ly6G+/GR1 +) in the spleen were positive for TdTomato expression in SPIT-treated mice, compared to less than 0.1% of cells in controls.

[0168] PB analyses of SPIT-treated mice revealed significantly lower levels of Cre recombination than found in the spleen, nonetheless some TdTomato expression could clearly be detected. An average of 0.04% of CD45+ PB cells were positive for TdTomato expression in SPIT-treated mice, compared to 0% in control mice. In contrast significantly higher levels of recombination could be detected in cells from the BM, as shown in FIG. 1 1 D, with an average of 2.89% of BM cells positive for TdTomato expression in SPIT-treated mice, compared with 0.029% of cells in the control group.

[0169] Notably, a significantly higher proportion of hematopoietic stem and progenitor cells (HSPCs/ Lineage-, cKit+, Seal +) were positive for TdTomato expression compared to the general population of cells in the BM, with an average of 37% of HSPCs in SPIT-treated mice positive for TdTomato expression, compared to 0% in the control. Collectively, these results convincingly demonstrate that SPIT is capable of delivering genetic engineering enzymes in vivo, in an immunocompetent setting. Achieving not only local but also a degree of systemic delivery of Cre recombinase to cells throughout the body.

[0170] For flow cytometry, splenocytes and peripheral blood cells were stained with the following antibodies for 30 minutes at 4 C: FITC CD11 b (M1/70; eBioscience), FITC-GR1/Ly- 6G (1 A8; BioLegend), APC-CD4 (RM4-5; Invitrogen), APC-CD8 (53-6.7; Invitrogen), APO Efluor 780-B220 (RA3-6B2; Invitrogen), BV421 -CD45.2 (104; Invitrogen). Cells for the bone marrow were first stained with the following combination of biotinylated antibodies for 30 minutes at 4 C: Gr-1 (RB6-8C5; Biolegend), Ter-119 (TER-119; Invitrogen), CD4 (RM4-5; Biolegend), CD8 (54-6.7; BioLegend), B220 (RA3-6B2; Biolegend), IL-7R (A7R34; Biolegend). After which cells were washed with PBS and then stained in the following cocktail: BV421 -ckit (2B8; Biolegend), FITC-Sca1 (D7; Biolegend), APC/Efluor780-streptavidin (Biolegend). After staining with antibodies cells were re-suspended in PBS that with propidium iodide at a concentration of 1 ug/ml, after which cells were analyzed using a BD FACs Aria BD flow cytometer.

[0171] VLPs secreted in vivo by SPIT producer cells result in significantly higher transfection efficiencies compared to in vitro prepared and injected VLPs, as evidenced by the following proof of concept experiment. In this experiment, mice were injected with either SPIT producer cells or VLPs produced ex vivo using the same constructs (GAG-Cre, VSV-g), producer cell type (293T cells) and equivalent number of cells (~2e8 cells) as the SPIT producer cells. SPIT cells were administered by intraperitoneal injection (~2e8 cells) and VLPs were administered by direct retro-orbital (RO) injection.

[0172] As shown in FIG. 12, significantly lower levels of Cre recombination were observed for ex vivo produced VLPs compared to SPIT producer cells, indicating that SPIT was significantly more efficacious for in vivo delivery of Cre to Ai14 mice than the direct retro-orbital (RO) injection of VLPs produced ex vivo. An average of 0.04% of BM cells and 0.12% of HSPCs were found to be positive for tdTomato expression following direct injection of VLPs into Ai 14 mice, compared to 2.9% of BM cells and 37% of HSPCs, for mice administered SPIT cells. Statistical analysis by two-way ANOVA found the difference in TdT-i- cells between the two conditions to be statistically significant (P=<0.0001 ). Notably however, VLP treated mice were analyzed for recombination in the BM 5 months after injection, compared to after only 1.5 weeks in the case of SPIT treated mice. Frequencies were compared in separate experiments; SPIT was compared to prior attempts by the inventors to target HSCs in vivo using VLPs produced in vitro. The low efficacy of in vitro produced VLPs for targeting long term HSCs was one motivation for investigating alternative approaches for in vivo delivery. [0173] In addition, when VLPs were directly injected into mice RO only a small amount of tdTomato expression could be detected visually in the liver, with no tdTomato expression detected in other organs (data not shown). Without being bound by any theory, the reduced efficacy of VLPs compared to SPIT cells may be a result of VLP loss during ex vivo production. The production of recombinant viral vectors for clinical translation in general has proven to be a major challenge in the gene therapy field, due to the difficulties of scaling the production and purification of recombinant viral vectors. These data show that a SPIT based approach can avoid similar challenges via direct injection of producer cells.

[0174] The foregoing experiments show the successful application of SPIT technology for utilizing human cells as vectors for in vivo genetic engineering. Incorporating a VLP system, SPIT technology is demonstrated to be effective for delivery of both Cre and a CRISPR-Cas9 adenine base editor for cell-cell genetic engineering. In addition, the experiments above show that genetic logic can be incorporated into a SPIT platform. Further, these experiments provide the first demonstration of human cells as vectors for in vivo genetic engineering in immunocompetent mice. The results described here underscore the immense promise that human cells have as vectors for the delivery of a diverse array of genetic engineering tools in vivo, including transcription factors (such as OKSM), telomerases, zinc fingers, TALENs and CRISPR-Cas systems, among others.

[0175] While a number of different VLP systems have been developed that can deliver mRNA to cells, these systems have struggled to achieve gene editing with CRISPR-Cas systems unless an sgRNA is supplied to cells exogenously. In contrast, the ABE experiments above demonstrate that SPIT technology can be utilized to deliver the RNP directly to target cells, without the need for exogenous guide RNA.

Example 2

[0176] Truncated GAG-Pro-Pol Vectors to Facilitate SPIT

[0177] Initial SPIT approaches utilized VSV-g as a fusogen to facilitate the uptake of viruslike particles (VLPs) into recipient cells and their fusion with the cell membrane to release cargo into the cytoplasm. Although effective, VSV-g can be highly cytotoxic and immunogenic. This challenge has previously prevented the development of stable lentiviral producer cell lines using VSV-g as a fusogen [Ferreira et al., Biotechnol. J. 2021 16, e2000017]. One alternative to address these challenges is to use alternate protease-regulated fusogens as provided herein, which are less cytotoxic and have proven effective for producing lentivirus from stable cell lines (FIG. 13A). [0178] The use of these protease regulatable fusogens necessitates the expression of the GAG-Pol polyprotein (comprising MLV retroviral GAG), Protease (PRO), Reverse Transcriptase (RT), and Integrase (INT). To reduce the potential for recombination or mutation that could lead to live, replication-competent viruses arising in a patient’s body, truncated GAG-Pol was developeD, removing genes not essential for SPIT functionality (FIG. 130). Various GAG-Pol derivatives were transfected, along with GAG-Cre and the protease- regulatable fusogen Amphotropic 4070A, into 293T cells to generate SPIT cells. The SPIT cells were then co-cultured with Ai14 293T cells at a 1 :1 ratio, and the frequency of TdT+ cells was measured on days 2 and 5 post co-culture by FACS (FIG. 13B).

[0179] Comparisons found no significant difference in efficacy between GAG-Pro-RT-INT and GAG-Pro-RT derivatives. However, removing RT led to a 93% reduction in TdT-i- cell frequency (FIG. 13D-E). RT is crucial for activation of the dimeric protease found within GAG-Pol in other retroviral species such as HIV, with the activation of the dimeric protease dependent on dimerization of RT within GAG-Pol [Chagas et al., Protein Sci. 2024 33, e5080]. It is contemplated herein that MLV protease activation is regulated by a similar mechanism and that removal of RT significantly reduces the efficiency of MLV protease activation during VLP formation due to loss of RT dimerization.

[0180] To enhance protease activity without incorporation of functional MLV reverse transcriptase, it is contemplated herein that partially truncated RT retains some dimerization capability to activate MLV protease. RT was truncated to 2/3 of its length (RT 1 -223aa) and the co-culture experiment was repeated. This RT derivative retained 38% of the wild-type GAG-Pol activity, significantly improving the capability for cell-cell engineering (FIG. 13D-E).

Example 3

[0181] Fusogen Free SPIT

[0182] To reduce cytotoxicity and unintended cell-cell fusion possibilities potentially associated with using fusogens for SPIT, alternative approaches were developed to facilitate SPIT-mediated delivery in the absence of any fusogen. Notably, expression of GAG/GAG-Cre alone was sufficient to cause budding of VLP particles from a cell membrane. Screens were performed showing the efficiency of Ore delivery to recipient cells when either Ore, GAG-Cre or GAG-Cre/Gag-Pol were transfected to generate SPIT cells.

[0183] Although GAG-Cre expression alone was sufficient to cause VLP budding and release of particles from producer cells, expression of GAG-Pol, in particular the protease found within GAG-Pol, can significantly alter VLP structure [Konvalinka, J. et al., supra]. During VLP maturation GAG-Cre is a single protein. However, after VLP release from the cell membrane, the action of the protease found withing GAG-Pol digests GAG into its subunits, matrix (MA), Late (LA), Capsid (CA) and Nucleocapsid (NC), in addition to releasing the cargo protein (Cre) from GAG. Once cleaved into its mature form, the physical shape of the Capsid core is altered which can be visually detected by electron microscopy (FIG. 14C). One approach that can be applied to differentiate between immature and mature viral particles is the use of a mild treatment of detergent on viral particles, where immature capsids are detergent resistant due to oligimerization of GAG at the cell membrane protecting the integrity of the viral particle (FIG. 14B). In contrast, once digested into its subdomains by a protease, the mature capsid is no longer detergent resistant (FIG. 14B) (Wilk et al., supra).

[0184] Once released from SPIT producer cells, VLPs can be taken up by recipient cells via endocytosis. Typically, this endocytosis process leads to degradation of endosomal components via the lysosomal degradation pathway. However, some endosomal components can undergo a process of endosomal escape into the cytosol (FIG. 14A). Due to the detergent resistant nature of immature VLPs, it is contemplated herein that immature VLPs are unable to efficiently undergo endosomal escape, due to the increased rigidity of their membranes compared to mature VLP particles. Thus by incorporating the retroviral protease into VLPs, significantly higher rates of cell-cell delivery of cargo proteins into recipient cells can be achieved because VLPs/GAG are able to be cleaved into mature form which is not detergent resistant. In addition, cargo proteins are freed from GAG, which can also potentially increase their ability to undergo endosomal escape. Thus, incorporation of the protease into VLPs is contemplated to lead to higher levels of endosomal escape of cargo proteins delivered via SPIT even in the absence of any co-expression of fusogens.

[0185] To demonstrate this, either Cre, Gag-Cre or GAG-Cre/GAG-Pro-Pol were transfected into wild type 293T cells to generate SPIT 293T cells. One day after transfection, these SPIT 293T cells were then collected and co-cultured with Ai14 293T cells at a ratio of 1 :1 and rates of Cre delivery to recipient cells/recombination were measured based on TdT expression by flow cytometry. After 5 days of co-culture, an average of only 0.002% and 0.001% TdT+ cells could be detected when Cre or Gag-Cre were transfected to generate SPIT cells respectively, while an average 0.13% TdT-i- cells was detected when Gag-Cre/Gag-Pol were co-transfected to generate SPIT cells (FIG. 14D). These results clearly demonstrated that co-transfection of GAG-Pol/MLV protease can enhance delivery of cargo proteins into recipient cells in the absence of any fusogens.

[0186] Notably, when initially screening VLPs for SPIT as described above, GAG-Cre/GAG- Pol expression was found to be insufficient to facilitate Cre delivery to recipient cells when VLPs were isolated from the supernatant, filtered and then applied to Ai 14 reporter fibroblasts. However, when co-culturing cells delivery of Cre to reporter cells via SPIT was detected. The difference in results between SPIT mediated delivery by direct co-culture, versus isolation and application of purified VLPs from producer cells, highlights the significant difference that direct cell-cell mediated delivery of genetic engineering enzymes can have on delivery compared to traditional isolation, purification and application of VLPs to cells or an organism.

Example 4

[0187] CD63 Enhances Fusogen-Free SPIT Delivery

[0188] After successfully achieving fusogen-free SPIT delivery through the co-expression of GAG-Cre and GAG-Pol, whether additional genes could further enhance endosomal escape of cargo proteins and improve fusogen-free SPIT delivery was investigated. Tetraspanins, a family of four-pass transmembrane proteins, have been implicated in exosome function, cellcell signaling, and endosomal sorting. Prior studies have suggested that tetraspanins may facilitate alternative trafficking of endosomes or may increase membrane fluidity, leading to enhanced endosomal escape of cargos delivered via lentiviruses, exosomes, or VLPs [Boker, et al., Mol. Ther. 2018 26, 634-647; Joshi et al., ACS Nano 2020 14, 4444-4455], Three tetraspanins (CD9, CD63, and CD81 ) were screened for their potential to enhance SPIT- mediated delivery when co-expressed with GAG-Cre and GAG-Pol.

[0189] To evaluate the impact of tetraspanin co-expression, 293T cells were transfected with either GAG-Cre/GAG-Pol/Membrane GFP (control, with Membrane GFP used as an additional membrane protein to normalize for transfection efficiency/expression of an additional membrane protein) or GAG-Cre/GAG-Pol/tetraspanin. These SPIT cells were subsequently cocultured with Ai14 293T reporter cells at a 1 :1 ratio for five days. Delivery to recipient cells was assessed using FACS to detect TdT expression.

[0190] The results revealed that co-expression of CD9 or CD81 reduced the frequency of fusogen-free SPIT delivery, while co-expression of CD63 enhanced it. Specifically, cultures with control SPIT cells (GAG-Cre/GAG-Pol/Membrane GFP) produced an average of 0.11% TdT+ cells. In contrast, co-expression of CD63 led to an average of 0.24% TdT+ cells, representing more than a 100% increase in delivery efficiency (FIG. 14E). These findings demonstrate that CD63 co-expression enhances SPIT-mediated delivery of cargo proteins to recipient cells in the absence of any fusogen expression.

Example 5

[0191] SPIT mediated Delivery of Protein Outperforms mRNA Delivery [0192] SPIT relies on the fusion of a desired cargo protein to GAG to facilitate cell-cell delivery of the cargo. An alternative approach involves packaging mRNA into VLPs, which can then be translated into the desired protein upon delivery to recipient cells. While this mRNA- based approach provides opportunities to regulate gene expression through mRNA stability and regulatory mechanisms, mRNA requires capture by GAG molecules for incorporation into VLPs. However, GAG molecules can induce VLP budding even in the absence of mRNA, resulting in many empty VLPs devoid of the desired cargo. In contrast, direct fusion of cargo proteins to GAG ensures incorporation into VLPs at a 1 :1 ratio with GAG (FIG. 15A). Based on these differences, it is contemplated herein that direct fusion of proteins to GAG is significantly more efficient for SPIT-mediated delivery compared to mRNA-based methods.

[0193] To demonstrate this, two previously published synthetic GAG derivatives, EPN11 and EPN24, [Horns, F. et al. Cell 2023 186, 3642-3658 e3632] were compared for their ability to deliver Cre mRNA (via mRNA packaging) versus Cre protein (via direct fusion). For mRNA delivery, an MS2-tagged Cre mRNA with MCP (MS2 Coat Protein) fused to the synthetic GAG molecules was used to facilitate mRNA incorporation into VLPs. For protein delivery, Cre was directly fused to the C-terminus of EPN11 and EPN24. VLPs were generated by transfecting 293T cells with either synthetic GAG molecules, Cre mRNA (for mRNA delivery), or GAG-Cre fusion protein (for direct fusion), alongside VSV-g. Cells were cultured in 24-well plates, and after three days, the supernatant was collected, filtered using a 0.45 pm filter, and applied to Ai 14 fibroblasts. Three days after treatment, fibroblasts were harvested, and the frequency of TdT+ cells was assessed by flow cytometry.

[0194] The results demonstrated that protein delivery via direct fusion to GAG significantly outperformed mRNA delivery for both synthetic GAG derivatives. For EPN1 1 , mRNA delivery resulted in an average of 0% TdT-i- cells, while protein delivery via direct fusion resulted in an average of 29% TdT-i- cells. Similarly, for EPN24, mRNA delivery resulted in an average of 0.7% TdT+ cells, whereas protein delivery via direct fusion resulted in an average of 45.6% TdT+ cells (FIG. 15B). Protein delivery through direct fusion to the C-terminus of EPN24 was approximately 80-fold more efficient than mRNA delivery in achieving Cre delivery to recipient cells (FIG. 15C).

[0195] These results highlight the superior efficiency of SPIT-mediated delivery using direct protein fusion to GAG compared to mRNA-based delivery. Direct fusion ensures that each VLP contains the desired cargo at a 1 :1 ratio with GAG, avoiding the inefficiencies of empty VLPs associated with mRNA packaging as well as constraints in the number of mRNA molecules that can be packaged due to the larger size of nucleic acids compared to amino acids. Example 6

[0196] Minimal GAG Derivatives to Achieve SPIT

[0197] Retroviral GAG consists of four different recognized subdomains that play roles in budding of viral particles from a cell and the viral life cycle (FIG. 15D). 1 ) The matrix domain (MA), the N-terminus of this domain begins with a myristylation signal sequence MGQAVT (SEQ ID NO: 1 ) (MGXXXS/T is the signal for post-translational myristylation, SEQ ID NO: 2), it also contains a polybasic amino acid motif of RKRR (SEQ ID NO: 3) which targets binding to the cell membrane. In addition, the MA domain also contains a ESCRT recruitment sequence PSAP (SEQ ID NO: 4). 2) The late domain (LA) of GAG is primarily required for ESCRT recruitment and contains two motifs that facilitate ESCRT recruitment, LYPAL (SEQ ID NO: 5) and PPPY (SEQ ID NO: 6). ESCRT factors are involved in excision of particles from the cell membrane. 3) A capsid domain (CA) that is primarily involved in oligomerization. 4) A nucleocapsid domain (NC) which is involved in recruitment of viral RNA into particles [Welker, L. et al., Viruses 202 13, 1559; Freed, E.O. Virology 1998 251 , 1-15; Freed, E.O. Nat. Rev. Microbiol. 2015 13, 484-496]. In addition to these roles in viral formation, these domains also play other roles in the viral life cycle, including replication and integration into the genome of cells.

[0198] Experiments were performed to determine whether any of these viral sequences could be eliminated, with the goal of reducing the number of viral sequences that are required to accomplish SPIT, eliminate unnecessary domains of GAG that play other roles in the viral lifecycle, such as replication and integration, and to develop a minimal GAG derivative that could facilitate cell-cell engineering.

[0199] First, it was tested whether any single domains of GAG could be eliminated and still achieve VLP production and delivery of Cre to Ai14 fibroblasts. Different truncations of GAG were performed with either the MA, LA, CA or NC, domains of GAG removed (FIG. 15D). VLPs were then produced through co-transfection of truncated GAG derivatives into 293T cells along with VSV-g. After three days the supernatant was collected from transfected cells, filtered through a 0.45um filter and was then applied to Ai14 fibroblasts. Three days after addition of supernatant Ai 14 fibroblasts were analyzed for TdT expression by flow cytometry.

[0200] Comparing the different truncations, truncation of MA was found to completely abrogate any delivery of Cre by VLPs to reporter cells, in contrast truncation of LA and NC lead to a 10 fold and 2-fold reduction in the delivery of Cre to reporter cells respectively (FIG. 15E). In contrast, truncation of CA from GAG led to an average of a 3-fold increase in the frequency of Cre recombination in reporter cells. These results demonstrated that, although MA is essential for VLP formation and delivery, LA, CA and NC can all be eliminated with elimination of CA improving delivery efficiency.

[0201] Having found single domains could be eliminated from GAG and the NC domain is not required to achieve SPIT, delivery of cargo proteins to recipient cells with GAG derivatives consisting of MA-Cre, MA-LA-Cre and MA-synLA-Cre where the LA domain is substituted with a synthetic LA domain (synLA) consisting of only the ESCRT recruitment motifs PSAP (SEQ ID NO: 4), LYPAL (SEQ ID NO: 5) and PPPY (SEQ ID NO: 6), that is PSAPLYPALPPPY (SEQ ID NO: 7) was tested. Performing the same assay as described for single domain derivatives, MA-Cre alone was insufficient for achieving any Cre recombination with VLPs, while derivatives consisting of MA and LA were found to facilitate Cre delivery to reporter cells. MA- LA-Cre induced cre recombination at 20% the efficacy of wild type GAG-Cre, while MA-synLA- Cre was able to deliver Cre at 10% of the efficiency of wild type GAG-Cre (FIG. 15G). These results demonstrated that these truncated derivatives can facilitate VLP production and cellcell engineering via SPIT.

Example 7

[0202] Macroencapsulation of SPIT cells

[0203] To demonstrate the feasibility of macroencapsulation for SPIT cell delivery into a patient’s body, a commercially available Theracyte device was used which features a porous 0.45pm PTFE membranecx. SPIT cells were generated by transfecting 293T cells with GAG- Cre and VSV-g and then encapsulating the SPIT cells within a 20pl Theracyte device (Figure 16B). The device was placed in a well containing Ai14-293T cells to test if SPIT delivery could be facilitated via a macroencapsulation device.

[0204] Three days after co-culturing with the macroencapsulated SPIT 293T cells, Cre recombination was observed in Ai14 293T cells using fluorescent microscopy and flow cytometry (FIG. 16C-D). A total of 2.54% of cells were detected as TdT+ by flow cytometry, confirming proof-of-principle for macroencapsulation-mediated SPIT delivery of genetic engineering enzymes.

[0205] It will be appreciated that the present invention is set forth in various levels of detail in this application. In certain instances, details that are not necessary for one of ordinary skill in the art to understand the invention, or that render other details difficult to perceive may have been omitted. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting beyond the scope of the appended claims. Unless defined otherwise, technical terms used herein are to be understood as commonly understood by one of ordinary skill in the art to which the disclosure belongs.

[0206] The abbreviations used herein have their conventional meaning within the chemical and biological arts.

[0207] Unless defined otherwise, technical and scientific terms used herein have the same meaning as commonly understood by a person of ordinary skill in the art. See, e.g., Singleton et al. dictionary of Microbiology and Molecular Biology, 2nd ed., J. Wiley & Sons (New York, NY 1994); Green and Sambrook, Molecular Cloning, A Laboratory Manual, 4th ed., Cold Springs Harbor Press (Cold Springs Harbor, NY 2012). Any methods, devices and materials similar or equivalent to those described herein can be used in the practice of this disclosure. The following definitions are provided to facilitate understanding of certain terms used frequently herein and are not meant to limit the scope of the present disclosure.

[0208] The terms “nucleic acid”, “polynucleotide”, and “oligonucleotide” are used interchangeably herein to refer to polymers of deoxyribonucleotides or ribonucleotides in either single-, double- or multiple-stranded form, or complements thereof. The term “nucleotide” refers, in the usual and customary sense, to a single unit of a polynucleotide, i.e., a monomer. Nucleotides can be ribonucleotides, deoxyribonucleotides, or modified versions thereof. Examples of polynucleotides include single and double stranded DNA, single and double stranded RNA, and hybrid molecules having mixtures of single and double stranded DNA and RNA. RNA may include messenger RNA (mRNA), small interference RNA (siRNA), short hairpin RNA (shRNA), micro RNA (miRNA), guide RNA (gRNA), CRISPR RNA (crRNA), and transactivating RNA (tracrRNA). DNA may include plasmid DNA (pDNA), minicircle DNA, genomic DNA (gNDA), and fragments thereof. The term “duplex” in the context of polynucleotides refers, in the usual and customary sense, to double strandedness.

[0209] The term “about”, unless otherwise specified herein, means a range of values including the specified value, which a person of ordinary skill in the art would consider reasonably similar to the specified value. In aspects, about means within a standard deviation using measurements generally acceptable in the art. In aspects, about means a range extending to +/- 10% of the specified value, and including the specified value.

[0210] The phrases “at least one”, “one or more”, and “and/or”, as used herein, are open- ended expressions that are both conjunctive and disjunctive in operation. The terms “a”, “an”, “the”, “first”, “second”, etc., do not preclude a plurality. For example, the term “a” or “an” entity, as used herein, refers to one or more of that entity. As such, the terms “a” (or “an”), “one or more” and “at least one” can be used interchangeably herein. [0211] Also as used herein, “and/or” refers to and encompasses any and all possible combinations of one or more of the associated listed items, as well as the lack of combinations when interpreted in the alternative (“or”).

[0212] As used herein, the term “comprises/comprising” does not exclude the presence of other elements, components, features, regions, integers, steps, operations, etc. Additionally, although individual features may be included in different claims, these may possibly advantageously be combined, and the inclusion in different claims does not imply that a combination of features is not feasible and/or advantageous. The transitional phrase “consisting essentially of” (and grammatical variants) is to be interpreted as encompassing the recited materials or steps “and those that do not materially affect the basic and novel characteristic(s)” of the recited embodiment. Thus, the term “consisting essentially of” as used herein should not be interpreted as equivalent to “comprising.”

Claims

CLAIMS What is claimed is:

1 . A method for in vivo genetic modification of target cells in a subject, the method comprising administering to the subject a composition comprising producer cells and a carrier, wherein the producer cells comprise one or more heterologous nucleic acids encoding one or more components of a gene editing system as fusion protein(s) with a GAG protein, such that the producer cells secrete a virus-like particle ("VLP") comprising the fusion protein(s) which are taken up by the target cells, thereby genetically modifying the target cells.

2. The method of claim 1 , wherein the GAG protein is a structural group-specific antigen (Gag) protein of a retrovirus, such as a murine leukemia virus, a human immunodeficiency virus, a rotavirus, or a hepatitis B virus, or a derivative of any of the foregoing.

3. The method of claim 1 , wherein the GAG protein is a non-viral protein, preferably a human protein, selected from an Arc protein, ASPRV1 , a Sushi-Class protein, a SCAN protein such as PGBD1 , and a PNMA protein.

4. The method of claim 1 , wherein the GAG protein is a non-viral protein selected from PEG10, RTL3, RTL10, and RTL1.

5. The method of any one of claims 1 to 4, wherein the producer cells encode a fusogen, optionally wherein the fusogen is VSV-g, mSynA, RD114a, amphotrophic envelope 4070A, measles F/H, syncytin, or myomaker/myomixer.

6. The method of claim 5, wherein the fusogen is a protease regulatable non-retroviral or mammalian endogenous retroviral fusogen comprising a retroviral R-peptide.

7. The method of claim 6, wherein the R-peptide replaces the cytoplasmic tail of the fusogen or is added to the C-terminus of the fusogen cytoplasmic tail.

8. The method of claim 6 or 7, wherein the R-peptide is derived from MLV amphotrophic 4070A, human immunodeficiency virus or simian immunodeficiency virus.

9. The method of claim 6 or 7, wherein the R-peptide is a synthetic R-peptide comprising a consensus sequence for a protease located at or near its N-terminus.

10. The method of any one of claims 5 to 8, wherein the fusogen is VSV-g or mSynA.

11 . The method of any one of claims 1 -5 or 10, wherein the producer cells comprise an expression plasmid comprising the nucleic acid encoding the GAG fusion protein.

12. The method of any one of claims 1 -5 or 10-1 1 , wherein the producer cells comprise a genetic knock-in encoding one or more of a protease, a pore forming protein, a targeting protein, a transcription factor, and/or an inducible suicide gene such as Caspase 9 or HSV-TK.

13. The method of any one of claims 1 -5 or 10-12, wherein the producer cells comprise one or more heterologous nucleic acids encoding a therapeutic protein.

14. The method of claim 13, wherein the producer cells comprise an expression plasmid comprising the one or more heterologous nucleic acids encoding the therapeutic protein.

15. The method of any one of claims 1 -5 or 10-14, wherein the producer cells are primary cells.

16. The method of claim 15, wherein the producer cells are pluripotent stem cells, induced pluripotent stem cells (IPSC), hematopoietic stem cells, or hematopoietic progenitor cells.

17. The method of any one of claims 1 -5 or 10-16, wherein the producer cells are human cells.

18. The method of any one of claims 1 -5 or 10-17, wherein the one or more components of a gene editing system comprises one or more of a zinc finger nuclease, a transcription activatorlike effector nuclease (TALEN), a Gas nuclease, a single stranded DNA modifying enzyme, a reverse transcriptase, a DNA methylase, a histone acetyltransferase, a deacetylase, and/or a topisomerase.

19. The method of any one of claims 1 -5 or 10-18, wherein the one or more components of a gene editing system comprises one or more components of CRISPR-Cas gene editing system.

20. The method of claim 19, wherein the one or more components of CRISPR-Cas gene editing system comprises a Cas endonuclease and a guide RNA.

21. The method of claim 20, wherein the VLP contains a Cas/gRNA ribonucleoprotein.

22. The method of claim 21 , wherein the method does not comprise introducing exogenous guide RNA to the subject or the target cells.

23. The method of any one of claims 1-5 or 10-22, wherein the target cells are thymus, heart, lung, liver, kidney, intestinal, or spleen cells.

24. The method of any one of claims 1-5 or 10-22, wherein the target cells are hematopoietic cells, optionally wherein the hematopoietic cells are B lymphocytes, T lymphocytes, or myeloid cells.

25. A human producer cell line, wherein the producer cells comprise one or more heterologous nucleic acids encoding one or more components of a gene editing system as GAG fusion protein(s), and secrete particles comprising the fusion protein(s).

26. The human producer cell line of claim 25, wherein the cells are pluripotent stem cells, induced pluripotent stem cells (iPSC), hematopoietic stem cells, or hematopoietic progenitor cells.

27. The human producer cell line of claim 25 or 26, wherein the one or more components of a gene editing system comprises a Cas endonuclease and a guide RNA.

28. The human producer cell line of any one of claims 25 to 27, wherein the secreted particles contain a Cas-gRNA ribonucleoprotein.

29. The method of claim 1 or the human producer cell line of claim 25, wherein the producer cells are not engineered to encode a fusogen.

30. The method or the human producer cell line of any preceding claim, wherein the GAG fusion protein does not comprise a nucleocapsid domain.

31. The method or the human producer cell line of any preceding claim, wherein the producer cells are engineered to co-express the GAG fusion protein and CD63.

32. The method or human producer cell line of any preceding claim, wherein the producer cells are encapsulated.