Genome edited cancer cell vaccines

Abstract

A cancer vaccine technology is provided which knocks out expression of cell surface immune checkpoint proteins, to facilitate their processing by immune cells, and optionally by knocking-in the expression of cytokines to boost immune response. Non-replicating tumor cells lacking cell surface CD47 are highly effective immunizing agents against subcutaneous mouse melanoma. Whole-cell vaccines inhibited tumor growth, and immunophenotyping showed a dramatic increase in activated effector cell subsets and M1-type macrophages aided by a significant reduction in the tumor-associated macrophage and myeloid derived suppressor cell compartments. A remarkable downregulation of cell surface CD47 was observed in the tumors that did escape after vaccination with genetically modified cells, suggesting the intricate involvement of CD47 in a prophylactic situation. An effective vaccination strategy to increase tumor-specific immune response in solid tumors is provided to improve the outcome of cancer immunotherapy.

Claims

1. A vaccine, comprising: a non-apoptotic senescent or mitosis-arrested irradiated cell which expresses a target antigen, wherein the cell has been genetically edited to: express a serpin protease inhibitor; knockout CD47; knockout PD-L1; and genomically integrate a gene encoding GM-CSF.

2. The vaccine according to claim 1, wherein the target antigen is a tumor-associated neoantigen.

3. The vaccine according to claim 1, in a kit further comprising: a therapy selectively targeting at least one of PD-1 producing cells, CTLA-4 producing cells, LAG-3 producing cells, CD47 producing cells, and PD-L1 producing cells.

4. The vaccine according to claim 1, wherein the non-apoptotic senescent or mitosis-arrested irradiated cell has been CRISPR edited to knockout CD47 expression.

5. The vaccine according to claim 4, wherein the non-apoptotic senescent or mitosis-arrested irradiated cell is a genetically engineered senescent or mitosis-arrested cultured cell having at least one restriction endonuclease recognition site.

6. The vaccine according to claim 5, wherein the non-apoptotic senescent or mitosis-arrested irradiated cell is derived from a neoplastic cell which expresses the target antigen.

7. The vaccine according to claim 6, wherein the neoplastic cell is a melanoma cell.

8. A method for preparing a vaccine, comprising: isolating a live cell which expresses at least one target antigen; genetically editing the live cell to: knockout expression of CD47, knockout expression of PD-L1, genomically integrate a gene encoding GM-CSF, and express a serpin protease inhibitor, while preserving expression of the at least one target antigen; expanding the genetically edited live cell by cell culture; and modifying the cultured live cell by irradiation to ensure that it remains alive and non-apoptotic, but is incapable of proliferating after in vivo administration to a human patient.

9. The method according to claim 8, further comprising obtaining the live cell from a human suffering from a neoplasia, wherein the live cell is a neoplastic cell, and readministering the modified cultured live cell to the human.

10. The method according to claim 9, further comprising administering an immunostimulatory therapy selectively targeting cells producing at least one of CTLA-4, LAG-3, and CD47.

11. The method according to claim 8, wherein the knockout of CD47 comprises knockout of CD47 by CRISPR-Cas9 gene editing.

12. A vaccine, comprising: a non-apoptotic senescent or mitosis-arrested irradiated cell derived from a neoplastic cell which expresses a neoantigen; wherein the cell has been genetically edited to: express a serpin protease inhibitor; genomically integrate a gene encoding GM-CSF; knockout CD-47; and knockout PD-L1; and wherein the cell has been expanded before being made senescent or mitosis-arrested by irradiation; and a formulation adapted for administration to a patient.

13. The vaccine according to claim 1, wherein CD47 is knocked out by a frameshift mutation of a gene producing CD47.

14. The method according to claim 8, wherein the genetic editing to knockout CD47 comprises causing a frameshift mutation of a gene producing CD47 to knock out CD47.

15. The vaccine according to claim 12, wherein CD47 is knocked out by a frameshift mutation of a gene producing CD47.

Description

BRIEF DESCRIPTION OF THE FIGURES

(1) FIGS. 1A and 1B show schematic diagrams of the interaction of CD47.sup.+/ (normal) and CD47.sup./ (genome edited) cancer cells with macrophages.

(2) FIGS. 2A-2E show the knockout of the cd47 gene from the genome of B16F10 cells through CRISPR/Cas9 editing.

(3) FIGS. 3A-3E show profiling of gp75 and CD47; and the effect of CD47 blockade, GM-CSF, and opsonizing antibodies on phagocytosis of CD47.sup./ 3BD9 by BMDMs

(4) FIGS. 4A-4E show in vitro phagocytosis of 3BD9 and B16F10 by BMDMs and in vivo tumor growth after implanting live B16F10 and 3BD9 in mice.

(5) FIGS. 5A-5C show alterations in the tumor microenvironment in mice implanted with CD47.sup./ melanoma cells.

(6) FIGS. 6A-6C show in vivo response to MMC treated melanoma cells.

(7) FIGS. 7A-7C show suppression of tumor growth in vaccinated mice.

(8) FIGS. 8A-8D show tumor-free vaccinated mice show more activated immune phenotypes in Tumor Draining Lymph Nodes (TDLNs).

(9) FIGS. 9A-9E show lymphocyte populations and loss of CD47 expression over time in B16F10 and 3BD9 vaccinated mice.

(10) FIGS. 10A-10F show anti-tumorigenic tumor microenvironment in 3BD9 vaccinated mice with distinct suppressor cell phenotypes.

(11) FIGS. 11A and 11B show a gating strategy for detecting tumor-associated macrophages (TAMs) in the TME.

(12) FIGS. 12A and 12B show immune response in the tumor microenvironment after vaccination with 3BD9 cells.

(13) FIGS. 13A-13C show immune activity in Tumor Draining Lymph Nodes (TDLNs) of vaccinated mice.

(14) FIGS. 14A-14C show homogeneity of suppressive TIL populations in vaccinated mice within a Tumor Micro Environment (TME).

(15) FIGS. 15A-15C show the effect of STING agonist on tumor growth and protection against tumor challenge.

(16) FIGS. 16A-16H show phenotypic changes in the TILs and tumor cells upon STING agonist vaccination.

(17) FIG. 17 shows a bar chart comparing live, mitomycin C-treated, and irradiated cells in a phagocytosis assay.

(18) FIGS. 18A and 18B show depletion of cell surface CD47 by the CRISPR/Cas9 method

(19) FIGS. 19A and 19B show schematic drawings of a vaccination procedure and tumor responses.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

(20) Current cancer treatments are not satisfactory. On the other hand, the body immune system is designed to defend the body against pathogens and dangerous invaders including cancers. However, the body immune system fails to act on cancer cells. The increased expression of CD47 on tumor cells prevents them from being recognized by circulating immune cells such as macrophages. The use of anti-CD47 antibodies to block CD47 on tumor cells has proven effective. Nevertheless, this treatment represents some challenges including difficulty in penetrating antibodies into solid tumors. CD47 is expressed on the surface of many cells and is involved in a range of cellular processes such as apoptosis, proliferation, adhesion, and migration. Thus, the use of anti-CD47 antibodies can have some unpredictable side effects. The present technology utilizes whole-tumor cells as vaccines against cancer. According to a preferred embodiment, CD47-deficient cancer cells serve as both preventive and therapeutic vaccines to facilitate phagocytosis by macrophages, which then present tumor antigens to and activate immune cells such as cytotoxic CD8+ T cells. The activated CD8.sup.+ T cells will attack and kill cancer cells on their own.

(21) The technology harnesses the body immune system to eliminate cancer cells by creating CD47-deficient cancer cell vaccines. Cancer cells can easily escape from the immune system due to the expression of CD47 on their surface. The deletion of CD47 by genome editing enables their phagocytosis by macrophages that activate cytotoxic CD8.sup.+ T cells. The activated CD8.sup.+ T cells attack and kill cancer cells. The adaptive immune response is induced by either a standard cell line vaccine, or by using the patient's own tumor cells that have been genome-edited (to delete CD47) and inactivated by irradiation (to suppress their proliferation). Moreover, it is very powerful for treating metastatic cancer cells where none of current therapies is effective or eliminate recurrent cancer cells after solid tumors are surgically removed.

(22) It is noted that in some cases, it is not the primary tumor which represents the ultimate target of the therapy. That is, in some cases, a tumor is benign or low grade, but is associated with a possible or even likely future change toward increased aggressiveness. Therefore, in such patients, a treatment may be provided that vaccinates the patient against the aggressively transformed cells.

(23) To demonstrate the technology, i.e., the hypothesis of whether CD47 null tumor cells can be formulated into a vaccine eliciting a strong immune response of the host to eliminate cancers from the body, the cd47 gene from B16F10, a mouse melanoma cell line, was knocked out through CRISPR/Cas9 genome editing.

(24) Two guide RNAs were designed that target the exons 1 and 2 of the cd47 gene (SEQ ID: 001, SEQ ID: 002). The guide RNAs were subcloned individually into an all-inclusive plasmid system, PX458 and co-transfected into B16F10 cells. Transfected cells were FACS sorted, and the resultant single cell clones were screened for bi-allelic CD47 knockout by PCR and DNA sequencing (Table 1). The resultant clone was named as 3BD9. The biallelic gene deletion was confirmed through DNA sequencing (FIG. 2A). The absence of CD47 expression on cell surface of 3BD9 was quantified through both flow cytometry (FIG. 2B) and immunofluorescence microscopy (FIGS. 2C-E). The CD47.sup./ 3BD9 cells were used in the subsequent experiments to ascertain their ability to elicit an immune response in the host.

(25) FIGS. 2A-2E show the knockout of the cd47 gene from the genome of B16F10 cells by means of CRISPR/Cas9 editing. (FIG. 2A) DNA sequencing data showing the deletion of a 12 kb fragment from the cd47 gene (SEQ ID: 003). Underlined sequences represent the gRNA target regions, and bold arrows represent cut sites for Cas9. (FIG. 5B) FACS histograms of rat anti-mouse CD47 antibody labeled unedited B16F10 cells (in red), the edited 3BD9 clone (in blue), and a negative control (3BD9 cells without primary antibody treatment) (in orange). Immunofluorescence microscopy was performed to detect the expression of CD47 (green) on the surface of B16F10 (FIG. 2C) and 3BD9 (FIG. 2D). (FIG. 2E), a negative control where the 3BD9 was labeled in the absence of the primary antibody. All cells were counterstained with DAPI (blue). Scale bar: 100 m.

(26) TABLE-US-00002 TABLE1 GuideRNAandPCRprimersequencesdesignedforCD47knockoutandscreening ID Sequence Location/Function CRISPRgRNASequences GuideM1 AACCGCCGCCGCGACAACGA Exon1,cd47gene,usedforRNPelectroporation,and SEQID:004 inPX458plasmidforgenedeletion GuideM3 TGCTTTGCGCCTCCACATTA Exon2,cd47gene,usedinPX458plasmidforgene SEQID:005 deletion PCRPrimers MA-FP AGCCAGAGGGAAGGAGTT Forwardprimer,upstreamofGuideM1targetedcd47 SEQID:006 gene MC-RP CCACTTGCCCAAGAAGAG Reverseprimer,downstreamofGuideM3targeted SEQID:007 cd47gene(withMA-FP,ampliconlength:12,986bpin intact,and473bpinadeletedcd47gene)

(27) Opsonizing Antibody Enhances Phagocytosis of CD47.sup./ B16 Cells (3BD9) by Bone Marrow-Derived Macrophages (BMDMs)

(28) While CD47 depletion makes cells more susceptible to phagocytosis by macrophages, efficient phagocytosis requires the presence of an opsonizing antibody that engages the Fc receptors on macrophages. An in vitro phagocytosis assay was performed to determine whether phagocytosis of 3DB9 cells by BMDMs is enhanced in the presence of opsonizing antibody such as TA99, an antibody against glycoprotein 75 (gp75), a common melanoma tumor-associated antigen (Sockolosky et al. 2016). The expression of gp75 on the surface of the B16F10 and 3BD9 was determined by flow cytometry (FIG. 3). Presence of gp75 correlates with the ability of the TA99 antibody to block surface markers on tumor cells and act as an opsonizing agent for phagocytosis.

(29) FIGS. 3A-3E show profiling of gp75 and CD47; and the effect of CD47 blockade, GM-CSF, and opsonizing antibodies on phagocytosis of CD47.sup./ 3BD9 by BMDMs. Flow cytometric analysis of cell surface gp75 expression on live B16F10 and 3BD9 cells (FIG. 3A), cell surface CD47 expression on irradiated B16F10 and 3BD9 cells (FIG. 3B), and cell surface CD47 expression on GVAX (red histogram) and 1FC (CD47.sup./ GVAX; blue histogram) cells with an isotype control (orange histogram) (FIG. 3C). (FIG. 3D) In vitro BMDM phagocytosis of the various types of melanoma cells under different conditions. **p<0.01, ***p<0.001 by one-way ANOVA using GraphPad Prism. (FIG. 3E) The trend shows increase in phagocytosis with the absence of CD47 and the presence of GM-CSF and the TA99 antibody. Data are expressed as meanSEM.

(30) The efficiency of phagocytosis of 3BD9 and B16F10 by BMDMs is shown in FIGS. 4A-B. BMDMs were co-incubated with CFSE-labeled tumor cells in the presence of different opsonizing antibodies, and phagocytosis was analyzed as a percentage of F4/80.sup.+ and CFSE.sup.+ macrophages by flow cytometry. The depleting or blocking of CD47 expression alone on the surface of melanoma cancer cells was found not to lead to a significant increase in their phagocytosis by BMDMs. However, the blocking of gp75 in CD47.sup./ 3BD9 cells with TA99 antibodies resulted in a considerable increase in phagocytosis, suggesting the combinatory effect of CD47 absence and the opsonizing effect of antibody (TA99). In this context, a similar enhancement of phagocytosis of 3BD9 was observed when cells are exposed to a macrophage activating cytokine such as the granulocyte macrophage colony stimulating factor (GM-CSF) (FIG. 3E). The phagocytosis efficiency of CD47 depleted melanoma cancer cells by BMDMs is increased when the cells are exposed to GM-CSF. It also does not appear that the exposure of CD47 depleted cells to both TA99 and GM-CSF can further improve their phagocytosis by BMDMs (FIG. 3C-E).

(31) FIGS. 4A-4E show in vitro phagocytosis of 3BD9 and B16F10 by BMDMs and in vivo tumor growth after implanting live B16F10 and 3BD9 in mice. (FIG. 4A) BMDM phagocytosis of B16F10 cells treated with various opsonizing antibodiespurified anti-CD47 (clone miap301), or anti-gp75 antibodies (TA99). These antibody-dependent conditions were tested alongside un-opsonized conditions with only B16F10 and CD47.sup./ B16 (3BD9) cells. The data shown are the mean (n=3) and the error bars indicate the standard error. *p<0.05, **p<0.01, one-way ANOVA. (FIG. 4B) Flow cytometric analysis of APC-F4/80 and CFSE dual labeled macrophages, depicting the percentage of phagocytosis in each condition. (FIG. 4C) Tumor growth in mice after implanting B16F10 or 3BD9. Points indicate tumor measurements from individual mice (n=8). Data shown are the combination of two independently performed experiments. ***p<0.001, unpaired t test. Error bars indicate standard errors. (FIG. 4D) Survival rate of mice implanted with B16F10 and 3BD9. **p<0.01, Mantel-Cox test. (FIG. 4E) Tumor growth rate after challenge (second tumor implantation with live B16F10 cells) for two mice that were tumor-free for 60 days after initial 3BD9 implantation. p=0.003 by linear regression analysis. All statistical analyses were performed using GraphPad Prism.

(32) These together appear to show that the phagocytosis of melanoma cancer cells is correlated to (i) the blockade of CD47-SIRP- binding, (ii) the engagement of the Fc receptors on macrophages, and (iii) the presence of enhancing cytokines. Translated in vivo, this observation implies that the antigen presenting compartment must be aided by the effector compartment for efficient anti-tumor activity.

(33) CD47 Ablation Delays Melanoma Tumor Growth Significantly In Vivo

(34) To determine whether CD47 ablated melanoma cells can elicit a strong immune response, tumor development was examined by implanting CD47.sup./ 3BD9 cells in syngeneic immunocompetent C57BL/6 mice (Weiskopf et al. 2016). For the mice that developed tumors, growth was delayed by at least 10 days in comparison with the mice implanted with CD47.sup.+/ B16F10. (FIG. 4C). Two of the eight mice (25% of mice) implanted with 3BD9 cells did not develop a tumor up to 60-days post implantation (FIG. 4D). To determine whether the tumor-free mice implanted with CD47.sup./ 3BD9 developed an immune memory against melanoma, a second tumor implantation was performed on Day 61. Interestingly, one mouse (50% of mice challenged) showed significantly delayed tumor growthby about 20 daysin comparison to a control B16F10 implant mouse not previously injected with 3BD9 (FIG. 4E). These experiments were very critical, unveiling the possible elicitation of an immune response by CD47.sup./ tumor cells. Therefore, it appears that CD47.sup./ tumor cells are feasible to use as vaccines against melanoma. A procedure and its results are shown schematically in FIGS. 19A and 19B.

(35) To characterize the immune activity in tumors that do not express CD47, an additional cohort of 15 mice per group that received B16F10 implants, and 3BD9 implants subcutaneously, were used. Examining immune organs of 3BD9 engrafted mice, enlarged spleens and lymph nodes were discovered (data not shown), indicating significant immune cell expansion after implantation. To further confirm the elicitation of an immune response by CD47.sup./ 3BD9, lymphocyte immuno-phenotyping was performed in the tumor microenvironment (TME) and in the tumor-draining lymph nodes (TDLNs) through multicolor flow cytometric analysis. Table 2 lists three panels of antibodies used for immuno-phenotyping: Panel T1 was used for phenotyping the antigen presenting compartment and the tumor cell phenotypes in the tumor microenvironment; Panel SL1 was used for phenotyping the antigen presenting compartment in the TDLNs; and Panel P2 was used for phenotyping the effector cell populations (T cells and NK cells) in both the TME and the TDLNs.

(36) FIGS. 5A-5C show alterations in the tumor microenvironment in mice implanted with CD47.sup./ melanoma cells. Tumors were collected from CD47.sup.+/ B16F10 and CD47.sup./ 3BD9 engrafted mice (n=5 each group) at three stages of tumor developmentsmall (200-300 mm.sup.3), medium (500-600 mm.sup.3), and large (800-900 mm.sup.3). (FIG. 5A) PD-L1 expression on tumor cells, (FIG. 5B) infiltration of regulatory T cells (T-regs), and (FIG. 5C) activated (Ki67.sup.+) effector cells (CD4.sup.+ T cells, CD8.sup.+ T cells, and NK cells) in the tumor microenvironment. *p<0.5, **p<0.01, ***p<0.001 by one-way ANOVA using GraphPad Prism.

(37) TABLE-US-00003 TABLE 2 Antibodies used for immuno-phenotyping lymphocyte populations in TME and TDLNs Concen- tration Antibody in Panel No. Marker Clone Fluorophore Dilution (g/ml) Panel TI - Tumor Cells and Antigen Presenting Compartment - TME 1 CD45 30-F 11 PerCP-Cy5.5 1:300 0.7 2 CDIIb MI/70 Alexa Fluor 488 1:200 1 3 Ly6C HKI.4 Brilliant Violet 785 1:300 0.7 4 Ly6G 1A8 PE 1:200 1 5 CDIIc N4 18 Brilliant Violet 42 1 1:200 1 6 CD47 miap30 1 Alexa Fluor 647 1:100 2 7 PD-LI 10F.9G2 Brilliant Violet 605 1:100 2 8 CDSO 16-IOAI Brilliant Violet 650 1:100 2 9 MHC-11 M5/114.15.2 APC/Fire750 1:200 1 Panel SLI - Antigen Presenting Comartment - TDLN 1 CD45 30-FII PerCP-Cy5.5 1:300 0.7 2 CDIIb MI/70 Alexa Fluor 488 1:200 1 3 Ly6C HKI.4 Brilliant Violet 785 1:300 0.7 4 Ly6G 1A8 PE 1:200 1 5 CDIIc N418 Brilliant Violet 42 1 1:200 1 6 CDSO 16-IOAI Brilliant Violet 650 1:100 2 7 MHC-II M5/114.15.2 APC/Fire750 1:200 1 8 CD206 C068C2 Alexa Fluor 647 1:100 2 Panel P2 - Effector Cell Compartment - TME and TDLN 1 CD45 30-FII PerCP-Cy5.5 1:300 0.7 2 CD8a 53-6.7 Brilliant Violet 510 1:100 2 3 CD69 HI.2 F3 Brilliant Violet 650 1:100 2 4 CD4 GKI.5 Brilliant Violet 785 1:200 1 5 CD25 3C7 APC 1:100 2 6 FoxP3 FJK-16s Alexa Fluor 488 1:100 2 7 CD49b CX5 PE-CF594 1:100 2 8 PD-1 RMPI-14 PE 1:200 1 9 Ki-67 16A8 Brilliant Violet 421 1:200 1

(38) The immuno-phenotyping of the tumor cells revealed a significant increase in cell surface PD-L1 expression as tumor progressed, in B16F10 engrafted mice, suggesting a gradual development of an immunosuppressive environment corresponding to the stage of tumor growth (FIG. 5A). The PDL-1 expressing tumor cells increased from approximately 20% at the early stage (200-300 mm.sup.3) to around 45% when the tumor growth reached to a larger size (approx. 800-900 mm.sup.3) in B16F10 implanted mice. In contrast, PD-L1 expression in CD47.sup./ 3BD9 engrafted mice remained at a low steady level as the tumor grew, suggesting suppression of tumor cell surface PD-L1 expression in these mice. Examining the T cell infiltration revealed an increase in resident CD4.sup.+ CD25.sup.+ FoxP3.sup.+ regulatory T cells (T-regs) in the tumor compartment as tumors grew in both B16F10 and 3BD9 engrafted mice (FIG. 5B). The difference in T-reg infiltration is evident between B16F10 and 3BD9 engrafted mice when the tumors grew to a size of 500-600 mm.sup.3. CD47.sup./ tumors exhibited higher T-regs infiltration in the tumor microenvironment. Interestingly, the T-regs infiltration were statistically at the same level when tumors grew larger in mice implanted with either CD47.sup.+/+ or CD47.sup./ tumor cells. It seems that there is a phase of tumor growth when the host immune system responds to the tumors differently in CD47.sup.+/+ and CD47.sup./ cancer cell engrafted mice.

(39) The immuno-phenotyping of effector immune cells in lymph nodes confirmed this observation. A higher number of the Ki67.sup.+ activated CD4.sup.+ and CD8.sup.+ T cells, as well as natural killer (NK) cells, were discovered in the CD47.sup./ tumor microenvironment (FIG. 5C) in both B16F10 and 3BD9 engrafted mice, the numbers of these immuno-activated cells in lymph nodes seemed to even out when the tumors were relatively small or very large, suggesting their escape from the immune system. A significantly high number of CD4.sup.+ and CD8.sup.+ T cells and NK cells were found in lymph nodes of 3BD9 engrafted mice when the tumors were at a medium size (500-600 mm.sup.3) (p<0.001).

(40) Immunization with Inactivated CD47.sup./ Tumor Cells Protects Mice from a Tumor Challenge.

(41) Upon confirming the immune response of mice to CD47.sup./ cancer cells, it was determined whether inactivated CD47.sup./ cancer cells can be used as a vaccine to protect mice from a tumor challenge. First, 3BD9 cells were chemically inactivated using Mitomycin-C (MMC) which inhibits DNA synthesis, hence rendering the cell non-replicating. Mice were vaccinated with PBS (vehicle control), MMC-treated B16F10 (MMC-B16F10), and MMC-treated 3BD9 (MMC-3BD9), by following an immunization regime depicted in FIG. 6A.

(42) FIGS. 6A-6C show in vivo response to MMC treated melanoma cells. (FIG. 6A) The immunization regime where the red arrows mark the injection of either PBS (vehicle control) or MMC-inactivated melanoma cells (B16F10 or 3BD9). The blue arrow indicates a tumor challenge using live B16F10 cells. Two doses, separated by 7 days, of MMC-treated cells were injected into the left flanks of mice to vaccinate them. The tumor challenge was given 7 days after the 2.sup.nd dose of vaccines. (FIG. 6B) The development of tumors after B16F10 engrafting in MMC-B16F10, MMC-3BD9, and PBS vaccinated mice. **p<0.01 by unpaired t test. (FIG. 6C) Survival rate of vaccinated mice after tumor challenge. n=4 per group. *p<0.5 by Mantel-Cox test.

(43) Immunogenicity in these vaccinated mice was determined after challenging the mice using B16F10 cells on day 7 post the second dose of vaccination. The MMC-3BD9 vaccinated mice exhibited a significantly delayed tumor growth by approximately 5 days (FIG. 6B). In both the MMC-B16F10 and MMC-3BD9 groups, the lifespan of tumor-bearing mice was doubled in comparison to the PBS vaccinated mice. MMC vaccinated mice survived for around 35 days, whereas PBS vaccinated mice only survived for 18 days. However, there was no significant difference in the survival rate of the mice within the B16F10 and 3BD9 groups, suggesting that CD47 does not play a role in this immunogenicity (FIG. 6C). MMC treatment makes cells inherently apoptotic (Terness et al. 2008; Scheffer et al. 2003; Roh et al. 2008; McKenna et al. 2012). CD47 is one of the most essential proteins in the clearance of apoptotic cells. The MMC treatment leads to the rearrangement of CD47 on the cell surface to form islands and helps cellular uptake by macrophages. Thus, the MMC treatment of B16F10 effectively elicited an immune response. These experimental results clearly suggested that MMC-treatment is inappropriate for testing the immunogenic capability of the 3BD9 cells against cancers.

(44) The inactivation of 3BD9 using -irradiation (Roy et al. 2001) was then tested. Irradiation is one of the most commonly used methods for inactivating cells while maintaining their biological integrity Dranoff et al. 1993). Inactivated B16F10 and 3BD9 cells were prepared by -irradiating them using a Cs source at 35 Gy and used them to vaccinate mice before a tumor challenge. As controls, mice vaccinated with PBS and irradiated B16F10 cells were studied alongside. Flow cytometric analysis confirmed the expression of CD47 and the absence of CD47 on irradiated B16F10 and 3BD9 cells, respectively (FIG. 7B). This suggested that the expression of CD47 on cell surface was unaltered post irradiation.

(45) Mice (15 per group) were vaccinated subcutaneously with irradiated 3BD9 or B16F10 cells, on their left flanks and challenged with live B16F10 cells on the same flank 7 days later (FIG. 7A).

(46) FIGS. 7A-7C show suppression of tumor growth in vaccinated mice. (FIG. 7A) The study regime: 15 mice per group were vaccinated with PBS, irradiated B16F10, or irradiated CD47.sup./ 3BD9 cells, and challenged with live B16F10 cells 7 days later. Mice were observed for 90-days post tumor implantation. (FIG. 7B) Tumor growth in mice from the three vaccination regimes. **p<0.01 by unpaired t test. (FIG. 7C) Percentage of tumor free mice after vaccination. ****p<0.0001 by the Mantel-Cox test. Symbol, Irr.: irradiation.

(47) In the irradiated 3BD9 vaccinated group, 40% of the mice (6/15) were tumor-free until day 70 post tumor challenge, and 33% of the mice (5/15) were tumor-free until the end of the regime (90-days post tumor challenge). More significantly, 13% of these mice (2/15) were tumor free at the end of 90-days post tumor challenge (FIG. 7B). For those did develop tumor, a significant delay in tumor growth was observed (FIG. 7C). The average tumor size in irradiated 3BD9 vaccinated mice that developed tumor was around 60 mm.sup.3, whereas it was close to 170 mm.sup.3 in irradiated B16F10 vaccinated mice at day 10. 100% (15 mice) of PBS vaccinated mice developed tumors. These experimental data suggested a strong tumor-specific immune response to CD47 depleted tumor cell vaccines, and it was important to further understand the underlying mechanisms of these responses.

(48) Vaccination with Irradiated CD47.sup./ Tumor Cells Confers Immunity to Mice by Upregulating Antigen Presentation and Increasing Effector Cell Activity

(49) To understand the underlying mechanisms that confer an anti-tumor immune response in the 3BD9 vaccinated mice, multi-parametric immune cell phenotyping was performed to characterize different immune cell subsets in the TME and in the TDLNs of mice using cell-specific protein markers shown in Table 3. A combination of various markers helped us identify specific subsets of cells, their lineages, and correlation with response to vaccines. A comprehensive list of cell types and subsets based on the markers chosen is available in Table 4. The immune mechanisms involved in mice that showed a complete response to vaccination by irradiated 3BD9 cells, i.e., those that did not develop a tumor 90-days post vaccination, were investigated. To that end the immune cell populations in the TDLNs of responders (CD47.sup./ 3BD9 R), and non-responders (CD47.sup./ 3BD9 NR) or those that developed delayed tumors after vaccination were compared.

(50) A significant increase in the M1-type anti-tumorigenic macrophages was observed, and a reduction in the M2-type pro-tumorigenic macrophages. M1 macrophage populations were five times higher and M2 macrophage populations were almost half the amount in the responders when compared to the non-responders. Additionally, the percentage of migratory DCs (Parajuli et al. 2010) and activated DCs that efficiently present antigens to the T cells, were significantly increased (FIG. 8A).

(51) FIGS. 8A-8D show tumor-free vaccinated mice show more activated immune phenotypes in TDLNs. (FIG. 8A) Macrophage and dendritic cell subsets present in the antigen presenting cell (APC) compartment, (FIG. 8B) ratio of regulatory T cells (T-regs) and activated T cells (Ki67.sup.+), (FIGS. 8C and 8D) CD8.sup.+ and CD4.sup.+ T cell subsets in the T-cell compartment of the tumor-draining lymph nodes (TDLNs). NR: non-responders (mice that grew tumors after vaccination); Rresponders (tumor-free mice after vaccination). n=5 mice per cohort for PBS and R groups. n=4 for NR group. Immuno-phenotyping was performed by flow cytometry. Cell phenotypes are presented as a percentage of their parent cell population. *p<0.5, **p<0.01, ***p<0001 by the unpaired t test performed using GraphPad Prism.

(52) Activated DC population were 6-7 times higher in the responders. These data suggest that the antigen presenting compartment is efficient in the TDLNs of the mice that responded to the 3BD9 vaccine. Moreover, the ratio of regulatory T cells (T-regs) to activated (Ki67.sup.+) T cells was drastically reduced in the responders (FIG. 8B), further confirming the presence of a highly anti-tumorigenic immune response. Analysis of the specific T cell subsets showed that the activated effector cell percentages are much higher in the responders in comparison to the non-responders and the PBS vaccinated mice, leading to the hypothesis that the tumor rejection could be a result of this. It is, however, interesting to note that the T-reg and the exhausted T cell (PD-1+) populations were higher in the 3BD9 vaccinated groups that in the PBS vaccinated groups (FIGS. 8C and 8D), though the ratios were much less in comparison.

(53) TABLE-US-00004 TABLE 3 Markers used for phenotypic characterization of immune and tumor cells. Marker Cellular Location Description Antigen Presenting Compartment/Tumor Compartment CD45 Plasma Membrane Pan leukocyte marker; to differentiate immune cells from other types of cells CD11b Plasma Membrane Macrophage specific marker Ly6C Plasma Membrane Monocyte marker; tumor-associated macrophage marker; part of the Gr1 complex with Ly6G Ly6G Plasma Membrane Monocyte marker; tumor-neutrophilic lineage, part of the Gr1 complex with Ly6C CD11c Plasma Membrane Dendritic cell specific marker CD80 Plasma Membrane Antigen presenting cell (APC) specific activation marker; present in M1-type macrophages and activated DCs CD206 Plasma Membrane M2-type macrophage marker MHC-II Plasma Membrane Antigen presentation capability PD-L1 Plasma Membrane Ligand for PD-1 on exhausted T-cells; present on tumor cells as a defense mechanism, and on APCs CD47 Plasma Membrane Don't eat me signal on tumor cells; interacts with SIRP- on macrophages to contain phagocytosis Effector Cell Compartment CD8a Plasma Membrane Cytotoxic T-cell (CTL) specific marker CD69 Plasma Membrane CTL activation marker CD4 Plasma Membrane Helper T-cell (T.sub.H) specific marker CD25 Plasma Membrane IL2RA protein for identification of CD4.sup.+FoxP3.sup.+ regulatory cells (T-regs) FoxP3 Nucleus Transcription factor expressed by T-regs CD49b Plasma membrane Natural Killer (NK) cell specific marker (only C57BL/6 mice) PD-1 Plasma Membrane Programmed death protein expressed on the surface of exhausted T-cells; deactivation upon interaction with PD-L1 on tumor cells and APCs Ki-67 Nucleus Proliferation marker expressed by activated effector cells (T cells, NK cells) Note: The table lists cell-specific markers used in various combinations as part of two panels that identify specific characteristics of immune cells and tumor cells in the tumor microenvironment and in the tumor-draining lymph nodes.

(54) TABLE-US-00005 TABLE 4 Phenotypic characterization of antigen presenting cells in immune infiltrates. Subset Significance Phenotype Macrophage Subsets Anti-tumorigenic; enhanced IL-2 production CD11b.sup.+ CD80.sup.+ M1-type Macrophages Anti-tumorigenic; efficient antigen CD11.sup.+ CD80.sup.+ MHC-II.sup.hi presentation M2-type Macrophages Pro-tumorigenic; suppressed antigen CD11b.sup.+ CD206.sup.+ MHC-II.sup.low presentation Suppressive Macrophages T cell deactivation CD11b.sup.+ PD-L1.sup.+ Tumor-Associated Macrophage (TAM) Subsets TAM-1 Pro-tumorigenic; inefficient antigen CD11b.sup.+ Ly6C.sup.int MHC-II.sup.hi presentation TAM-2 (MHC-II.sup.hi TAMs) Pro-tumorigenic; normoxic conditions CD11b.sup.+ Ly6C.sup.low MHC-II.sup.hi TAM-3 (MHC-II.sup.low TAMs) Pro-tumorigenic; present in hypoxic CD11b.sup.+ Ly6C.sup.low MHC-II.sup.low conditions; suppressed antigen presentation Dendritic Cell (DC) Subsets\ Activated DCs Anti-tumorigenic; enhanced IL-2 production; CD11c.sup.+ MHC-II.sup.+ CD80.sup.+ efficient antigen presentation Migratory DCs Present mostly in tumor-draining lymph CD11c.sup.hi MHC-II.sup.+ nodes; signifies high DC activity CD11b.sup.+ Regulatory DCs May be pro-tumorigenic CD11c.sup.+ MHC-II.sup.+ CD80.sup.+ Suppressive DCs T cell deactivation CD11c.sup.+ PD-L1.sup.+ Myeloid-Derived Suppressor Cell (MDSC) Subsets Poly-morphonuclear (PMN) Pro-tumorigenic; from Ly6C.sup.hi monocytic CD11b.sup.+ Ly6C.sup.low Ly6G.sup.+ MDSCs (granulocytic) precursors; differ in lineage; Ly6C.sup.hi forms Monocytic (Mo) MDSCs more suppressive than Ly6C.sup.low forms CD11b.sup.+ Ly6C.sup.hi Ly6G.sup.+ Monocyte-derived (Mo- CD11b.sup.hi Ly6G.sup.+ MHC-II.sup.+ derived) MDSCs Cytotoxic T Cell (CTL) Subsets - CD8.sup.+ Activated CTLs Primed tumor-specific CD8.sup.+ T cells CD8.sup.+ CD69.sup.+ Activated/Proliferating CTLs Activated and proliferating CD8.sup.+ T cells; CD8.sup.+ Ki67.sup.+ produce granzymes, perforins, IFN- Exhausted CTLs Deactivated phenotype of CD8.sup.+ T cells; CD8.sup.+ PD-1.sup.+ incapable of cytotoxic killing Helper T Cell (T.sub.H) subsets CD4.sup.+ Regulatory T cells (T-regs) Inhibit proliferation and activation of CTLs CD4.sup.+ CD25.sup.+ FoxP3.sup.+ and T.sub.H cells Activated/Proliferating T.sub.H cells Activated and proliferating CD4.sup.+ T cells; CD4.sup.+ Ki67.sup.+ produce IL-2, IFN- Exhausted T.sub.H cells Deactivated phenotype of CD4.sup.+ T cells; CD4.sup.+ PD-1.sup.+ incapable of cytokine release Natural Killer (NK) Cell Subsets - CD49.sup.+ Activated Natural Killer Cells Primed NK cells; produce granzymes; CD49b.sup.+ Ki67.sup.+ cytotoxic activity Ratios of Cell Types Ratio of T-regs to Ki67.sup.+ T cells High ratio signifies a more suppressive CD4.sup.+ CD25.sup.+ FoxP3.sup.+: environment; no T cell proliferation CD8.sup.+ Ki67.sup.+ Ratio of PD-1.sup.+ to Ki67.sup.+ T cells High ratio signifies a more suppressive CD8.sup.+ PD-1.sup.+: CD8.sup.+ environment; exhausted T cells Ki67.sup.+ CD4.sup.+ PD-1.sup.+: CD4.sup.+ Ki67.sup.+ Ratio of MHC-II.sup.hi TAMs to High ratio signifies a more normoxic CD11b.sup.+ LydC.sup.low: MHC- MHC-II.sup.low TAMs condition and a less suppressive II.sup.hi: CD11b+ Ly6C.sup.low environment MHC-II.sup.low Note: The table lists all the phenotypic subsets used in this project to determine immune cell function associated with tumor rejection and escape. Categorization of these subsets is based on their lineage, function, and specific cell surface markers.

(55) Comparison of tumor and lymphocyte populations in mice vaccinated with irradiated B16F10 and irradiated 3BD9 confirms the vital role of CD47

(56) To elucidate the specific role of CD47 in this immunologic reaction, lymphocyte populations in the TME and in the TDLNs of irradiated B16F10 and 3BD9 vaccinated mice that developed tumors after a tumor challenge were compared. It was important to track the alteration in the immune phenotypes as the tumors grew to understand how the microenvironment and the specific immune responses evolve. To this end, cell types were characterized at three different stages of tumor growthsmall (200-300 mm.sup.3), medium (500-600 mm.sup.3), and large (800-900 mm.sup.3). In the TME, activated and proliferated T and NK cells were examined, and found that the 3BD9 vaccinated mice had significantly higher percentages of Ki67.sup.+ effector cells, especially in the small and medium sized tumors. In contrast, the B16F10 vaccinated mice lost their effector cell populations over time (FIG. 9A).

(57) FIG. 9A09E show lymphocyte populations and loss of CD47 expression over time in B16F10 and 3BD9 vaccinated mice. (FIG. 9A) Regulatory T cells (T-regs), activated (Ki67.sup.+) T cells and activated NK cells in the tumor microenvironment, and (FIG. 9B) macrophage (M1- and M2-type) and dendritic cell subsets in the tumor-draining lymph nodes of mice at different stages of tumor growth in the PBS, 3BD9, and B16F10 vaccinated groups. Comparison of CD47 expression on all tumors (FIG. 9C) at different stages of tumor growth (FIG. 9D) in the PBS, 3BD9, and B16F10 vaccinated mice with representative CD47 expression dot plots (FIG. 9E). In the PBS vaccination cohort, n=5 for all three stages of tumor growth. In the 3BD9 cohort, n=4 (small), n=3 (medium), n=2 (large) and in the B16F10 cohort n=4 (small), n=5 (medium), n=4 (large). Flow cytometric analysis was performed on FlowJo and cell phenotypes are presented as a percentage of the parent cell population. *p<0.5, **p<0.01, ***p<0.001 by unpaired t test performed on GraphPad Prism.

(58) The specific effect of the absence of CD47 might be in improving the long-term survival of mice by increasing effector cell populations over time and delaying their exhaustion while keeping the suppressive cells to a minimum. The T-reg populations in the inactivated cell vaccinated mice were consistently observed to be higher than in the PBS vaccinated ones (FIG. 9A). This could be due to the abundance in T cell populations and cytokine release, driving the phenotypic change to a regulatory form. In the TDLNs, an increase in the levels of M1-type macrophages and a reduction in the M2-type macrophages in the small and medium sized tumors developed after irradiated 3BD9 vaccination were detected, which is one of the most expected responses to the lack of CD47 (FIG. 9B). PBS vaccinated mice initially showed a higher population of activated DCs, which reduced drastically over time, and the irradiated 3BD9 vaccinated mice showed much higher levels correlating with tumor progression (FIG. 9B), indicating high immune activity in the lymph nodes. These data correspond to the overall observation that the TME is hot and active in the 3BD9 vaccinated tumors, providing opportunities for further intervention to enhance this effect.

(59) Another very remarkable observation pertaining specifically to the involvement of CD47 in the anti-tumor response, was in the non-TIL compartment of the tumors of vaccinated mice. All tumor implants were B16F10 cells, known to express CD47 on their surface. However, upon analysis of the CD45.sup. (negative for the pan-leukocyte marker) tumor cells in the TME, CD47 was down regulated in 3BD9 vaccinated mice significantly (FIG. 9C). Breaking down this observation and analyzing at three stages of tumor growth, it was interesting to see that when the tumors are small, the B16F10 vaccinated tumors still have a significantly higher expression of CD47 when compared with the 3BD9 vaccinated tumors. However, as the tumor grows, even the B16F10 treated mice begin to lose their CD47 expression (FIG. 9D). Upon closer look, the stark difference in CD47 expression in 3BD9, B16F10 and PBS treated mice is evident in all the replicates analyzed at each stage of tumor growth (FIG. 9E). This observation also supports the corresponding TIL changes in the tumor microenvironment characterized in the above sections. It is worth considering that the tumors seem to evolve against an existing immune response according to the types of primed TILs that have been developing due to vaccination.

(60) Investigating the TME and TDLNs of 3BD9 vaccinated mice reveals actively anti-tumorigenic conditions, with distinct TAM phenotypes.

(61) After determining the role of CD47 in eliciting an immune response in vaccinated mice, the immune activity in the tumor microenvironment and the draining lymph nodes of the mice vaccinated with irradiated 3BD9 cells was investigated. The populations of active antigen presenting cells and proliferating effector cells were investigated. More effector immune cell populations (Ki67.sup.+ T cells and NK cells), but fewer antigen presenting cells in the TME of 3BD9 vaccinated mice as compared to those vaccinated with PBS were detected (FIG. 10A), suggesting an active immune cascade that ends in proliferation of effector cells in the 3BD9 vaccinated mice. To confirm this hypothesis, the antigen presenting cells were interrogated for the expression of the T cell-suppressive marker PD-L1 and found that both DCs and macrophages showed a higher level of PD-L1 expression in the PBS vaccinated mice in comparison with the 3BD9 vaccinated ones (FIG. 10B). There was no statistically significant difference in PD-1 expression on T cells, or in the percentage of T-reg populations between the two groups. However, the ratio of exhausted (PD-11 to activated (Ki67.sup.+) T cells was drastically higher in the PBS vaccinated groups (FIG. 10C), further confirming the finding that the macrophages and dendritic cells in the TME of PBS vaccinated tumors are of a stagnant and suppressive phenotype, not leading to downstream immune cell activation.

(62) FIGS. 10A-10F show anti-tumorigenic tumor microenvironment in 3BD9 vaccinated mice with distinct suppressor cell phenotypes. Multicolor flow cytometric analysis was performed to determine (FIG. 10A) infiltration of effector T, NK cells, macrophages, and DCs, (FIG. 10B) PD-L1 expression in tumor cells, infiltrating macrophages, and DCs, and (FIG. 10C) the distribution of regulatory and activated T cells in tumor microenvironment of 3BD9 vaccinated mice. Additional analysis was performed to characterize MDSC and TAM subsets (FIG. 10D) and the expression of lineage (Ly6G), activation (CD80), and T cell suppression (PD-L1) markers on the surface of MHC-II.sup.hi TAM and MHC-II.sup.low subsets (FIGS. 10E and 10F) in PBS and 3BD9 vaccinated mice at three stages of tumor growthsmall (200-300 mm.sup.3), medium (500-600 mm.sup.3), and large (800-900 mm.sup.3). n=15 mice for PBS, n=9 for irradiated 3BD9. In the PBS vaccination cohort, n=5 for all three stages of tumor growth. In the 3BD9 cohort, n=4 (small), n=3 (medium), and n=2 (large). Immune cell phenotypes are presented as a percentage of the parent cell population. **p<0.01, ***p<0.001 by unpaired t test performed on GraphPad Prism.

(63) The suppressor cellsthe tumor associated macrophages (TAMs) and the myeloid derived suppressor cells (MDSCs), were investigated in the tumor microenvironment. TAMs and MDSCs are responsible for a sizable portion of effector cell suppression (Bronte et al. 2016; Elliott et al. 2017; Italiani et al. 2014). This suppression is active and occurs both in the TME and in the TDLNs (Kumar et al. 2016; Movahedi et al. 2010). Phenotypic characterization of MDSCs and TAMs have been studied in previously. It has been emphasized that common characteristics of tumors that have been irreversible thus far like metastatic potential, downregulation of MHC genes, and overexpression of evasion markers, can be attributed to the specific APC phenotypes in the TME (Ostrand-Rosenberg et al. 2009; Richards et al. 2013). They are often distinguished by their cell surface marker expression (Table 4), and further studied based on mechanisms of action, T cell suppression potential, inhibitory cytokine production, and effect on other cell types. A very specific gating strategy was employed to characterize TAMs based on expression of the granulocytic lineage marker Ly6G, the monocytic marker Ly6C, and the antigen presentation complex MHC-II (FIGS. 11A and 11B). 3BD9 vaccinated tumors were found to have overall significantly less suppressive APCs in their TMEboth the polymorphonuclear and the monocyte-derived MDSCs are lower (FIG. 10D). Event in the TAM compartment, the MHC-II.sup.hi TAMs (TAM-1 and TAM-2) are significantly lower in the 3BD9 vaccinated tumors (FIG. 10D). The MHC-II.sup.low TAMs (TAM-3), however, are significantly higher. These TAMs are usually found in hypoxic conditions, as opposed to their MHC-II.sup.hi counterparts (TAM-2), which are usually present in normoxic conditions (Movahedi et al. 2010). This could also explain the increased neo-vasculature observed near vaccinated tumors and could mean that vaccinated tumors are prone to increased necrosis due to low oxygen, hence their slower growth. The different TAMs also showed a high phenotypic variability, which was unprecedented, and also opens up many avenues for further exploration of this compartment.

(64) FIGS. 11A and 11B show a gating strategy for detecting tumor-associated macrophages (TAMs) in the TME. (FIG. 11A) A pseudo-color scatter plot of MHC-II expression against Ly6C expression by tumor infiltrating lymphocytes. Gates were drawn according to expression levels of the markers based on FMO controls. The populations were categorized based on cells expressing various levels of each marker on their cell surface. (FIG. 11B) A panel of four scatter plots showing the ancestry of back-gating of the cell population to arrive at the graph in (FIG. 11A). Cells were first gated based on their forward scatter (FSC) to filter out doublets and clumps. They were then gated for the expression of the pan-lymphocyte marker CD45 on their cell surface. The next gating was based on cell size, and the last step before categorizing cells as TAMs, was gating for the expression of the macrophage marker CD11 b.

(65) Tumors were harvested from mice at three stages of growthsmall (200-300 mm.sup.3), medium (500-600 mm.sup.3), and large (800-900 mm.sup.3). With increase in tumor size, the MHC-II.sup.hi TAMs in the TME of 3BD9 vaccinated tumors seemed to gradually lose Ly6G expression (FIG. 10E). Ly6G expression signifies a neutrophilic lineage, and hence renders the TME a more inflammation-like environment. This is important to note in the endeavor to change cold tumors into hot tumors. This is opposite to the phenotypes observed in the PBS vaccinated mice. These MHC-II.sup.hi TAMs in the TME also produced minimal T cell stimulatory surface antigens (CD80) in comparison to those on PBS vaccinated mice, which appeared to reduce over time as the tumor grew (FIG. 10E). In the case of the MHC-II.sup.low TAMs, the difference was more distinct. The TAMs in the 3BD9 vaccinated mice maintained a high level of Ly6G expression throughout the tumor growth phase. In contrast, Ly6G expression in the TAMs infiltrating tumors of the PBS vaccinated mice remained almost negligible in the early stages of tumor growth, but became more significant as tumor grows, (FIG. 10F). The latter TAMs also showed a significantly higher level of expression of CD80 in comparison to the MHC-II.sup.low TAMs in the 3BD9 vaccinated tumors (FIG. 10F). Expression of the suppressive PD-L1 on the TAM-2 subtype increased as tumor grew in the 3BD9 vaccinated group and was significantly higher than the PBS vaccinated group. The inverse is true for the TAM-3 subtypes. Using these data, the overall mechanism of immune response in the tumor microenvironment after vaccination with irradiated CD47.sup./ 3BD9 cells can be comprehensively understood (FIGS. 12 and 12B).

(66) After scrutinizing the important immune cell populations in the TME, the draining lymph node was then analyzed to understand the immune activity in vaccinated mice. The TDLNs are T cell repertoires and act as hubs for immune cell activity. APCs of multiple phenotypes from the tumor drain into the nearest lymph nodes, and the environment there facilitates the interactions between APCs and the effector cells (T cells, B cells, NK cells). Profiling the immune cell phenotypes in the TDLNs provided an overview of the type of response that would lead to tumor rejection or escape. In contrast to the TME, TDLNs of the 3BD9 vaccinated mice have significantly high populations of activated APCs, and much lower suppressive populations of APCs. A high population of CD11b.sup. effector and activated DCs and a low population of CD11b.sup.+ regulatory DCs were detected in the 3BD9 vaccinated mice (FIG. 13A), suggesting an efficient T cell stimulating environment in the TDLNs.

(67) FIGS. 12A-12C show the immune response in the tumor microenvironment after vaccination with 3BD9 cells. The chart (FIG. 12B) depicts the immune activity as observed by analysis of the immune phenotypes infiltrating tumors post vaccination with CD47.sup./ 3BD9 cells. Green arrows and red arrows near cell subsets depict anti-tumorigenic and pro-tumorigenic cell types, respectively. Black arrows in the figure depict lineage, blue arrows depict activation, and red lines depict inhibition. Cell types are divided into sections based on their phenotypic characteristics. The antigen presenting cells subset shows activated M1-type macrophages and activated dendritic cells; the activated effector compartment includes Ki67.sup.+ T cells and NK cells; the exhausted T cell compartment consists of PD-1.sup.+ T cells; the tumor associated macrophage compartment includes the MHC.sup.hi TAMs (TAM-2) and MHC.sup.low TAMs (TAM-3); and the myeloid derived suppressor cells are shown in a separate section. (FIG. 12A).

(68) FIGS. 13A-13C show the immune activity in TDLNs of vaccinated mice. Multicolor flow cytometric phenotyping of (FIG. 13A) the dendritic cell and macrophage populations, (FIG. 13B) the ratio of MHC.sup.hi/MHC.sup.low TAMs, and (FIG. 13C) the effector cell populations (CD4.sup.+ T cells, CD8.sup.+ T cells, and NK cells) in the TDLNs of PBS and 3BD9 vaccinated mice. In the PBS vaccination cohort, n=15. In the 3BD9 cohort, n=9. Immune cell phenotypes are presented as a percentage of the parent cell population. *p<0.5, **p<0.01,***p<0.001 by unpaired t test performed on GraphPad Prism.

(69) M1-polarized macrophages showed a slight increase, but it was interesting to note that the 3BD9 vaccinated mice had higher populations of M2 macrophages. Conversely, one of the most potent subsets of MDSCs, the monocyte-derived MDSCs (CD11b.sup.hi Ly6G.sup. MHC-II.sup.+), were significantly lower in the TDLNs of the CD47.sup./ 3BD9 vaccinated mice (FIG. 13A). Higher levels of MHC.sup.low TAMs were detected in the TME than MHC.sup.hi TAMs in the TDLNs of the 3BD9 vaccinated mice, and the ratio of MHC.sup.hi/MHC.sup.low TAMs in these TDLNs was significantly lower (FIG. 13B). This may suggest, as mentioned above, that the migrating and antigen presenting macrophages are altering phenotypes to MHC.sup.low and expressing a defective phenotype due to tumors becoming more hypoxic. The next analysis was the effector cell populations in the TDLNs. The presence of significantly high populations of primed and proliferating T cell subsets was observed in the 3BD9 vaccinated lymph nodes (FIG. 13C). PD-1 expression on the CD4.sup.+ T cells also seemed to be higher in these mice, which could be a correlation to the increase in the M2-type macrophages and inefficient antigen presentation and T cell suppression. There was also an increased number of natural killer cells in the TDLNs (FIG. 13C).

(70) Vaccinated Mice Maintain Homogenous TIL Populations Throughout Tumor Growth Period.

(71) Examining the macroscopic changes occurring the TME after vaccination was important to provide a baseline for developing combination therapies. The alterations to immune cells' infiltration into tumors after vaccination are not limited to the difference in phenotypes of TILs. The levels or amounts of the several types of TILs in the TME is also a crucial factor in the immune modulation potential of vaccines. Many anti-tumor therapies rely on the percentages of TILs in tumors, and tumors often show extreme heterogeneity in infiltrating lymphocytes at various stages of development. An important effect of vaccination before a tumor implant is the maintenance of stable TIL amounts throughout the development of a tumor. In particular, the percentage of suppressive phenotypes of all TILs remains constant at all stages of tumor growth in vaccinated mice, when compared to vehicle controls. Analyzing levels of TAMs and MDSCs (FIGS. 14A-B), as well as T-regs and PD-1.sup.+ T cells (FIG. 14C) at three different stages of tumor growth, revealed that there is stability among these phenotypes in both the B16F10 and 3BD9 vaccinated tumors in comparison with the PBS vaccinated tumors. Since most immune checkpoint blockade therapies target these suppressive phenotypes, this is an important observation that could potentially alter dose escalation issues and minimize modulations in therapy.

(72) FIGS. 14A-14C show homogeneity of suppressive TIL populations in vaccinated mice. Tumors at different stages of growth were analyzed for suppressive TIL populations in PBS vaccinated (black lines), irradiated B61F10 vaccinated (blue lines), and irradiated 3BD9 vaccinated (pink lines) mice. (FIG. 14A) Tumor-associated macrophages (TAMs), (FIG. 14B) myeloid derived suppressor cells (MDSCs), and (FIG. 14C) T-regs (CD4.sup.+CD25.sup.+FoxP3.sup.+) and PD-1.sup.+ T-cells are shown. The quantification was done at three stages of tumor growthsmall (200-300 mm.sup.3), medium (500-600 mm.sup.3), and large (800-900 mm.sup.3). In the PBS vaccination cohort, n=5 for all three stages of tumor growth. In the CD47.sup./ 3BD9 cohort, n=4 (small), n=3 (medium), and n=2 (large). Immune cell phenotypes are presented as a percentage of the parent cell population. *p<0.5, **p<0.01, ***p<0.001 by unpaired t test performed on GraphPad Prism.

(73) Preliminary Characterization of a STING Agonist as a Vaccine Adjuvant

(74) The addition of adjuvants to vaccination regimes helps increase tumor specific immune cell infiltration to protect against an upcoming tumor challenge. To this end, addition of a STING pathway agonista cGAMP analog, 23-c-di-AM(PS)2 (Rp,Rp)was attempted as an adjuvant with irradiated whole-cell tumor vaccines following the immunization regime depicted in FIG. 15A. The adjuvant caused a significant delay in tumor growth (FIG. 15B) but did not enhance the overall effects of the CD47 depleted whole cell tumor vaccine. At the end of the 90-day observation regime, only 7% (1/15) of the mice were tumor free in the group that received the STING agonist, in comparison with the group that received only the irradiated 3BD9 vaccines without the adjuvant that had 33% (5/15) tumor-free mice (FIG. 15C).

(75) FIGS. 15A-15C show the effect of STING agonist on tumor growth and protection against tumor challenge. (FIG. 15A) The study regime: 15 mice per group were vaccinated with either vehicle (PBS), irradiated 3BD9, or irradiated 3BD9 with a STING agonist adjuvant. 7 days later, there were given a tumor challenge using live B16F10 cells. (FIG. 15B) Tumor growth in mice from the three vaccination regimes. **p<0.01, ***p<0.001 by unpaired t test. (FIG. 15C) Percentage of tumor free mice after vaccination. ****p<0.0001 by the Mantel-Cox test.

(76) Modulations to the Anti-Tumor Immune Response by STING Agonist

(77) The STING agonist treatment in this study was a single subcutaneous dose of 10 g per mouse along with the 0.5 million irradiated cells as part of the vaccination regime. It was sought to be determined whether the TIL phenotypes are altered by the addition of the adjuvant on the vaccinated groups. To study the immediate infiltration of the antigen presenting cells (DCs being the direct targets of the STING pathway agonist), this analysis was performed when the tumors were starting to grow and had reached a size of 200-300 mm.sup.3. As expected, there was a remarkable increase in the DCs infiltrating tumors and priming T cells in the TDLNs (FIGS. 16A and 16B). This enhancing effect of the STING agonist was observed on macrophages, which showed a significant increase in their activation and tumor infiltration. Otherwise, all through the various stages of tumor growth, the TILs of the 3BD9 groups vaccinated with a STING agonist behaved very similar to their non-adjuvant counterparts in that they showed an increase in the MHC.sup.low TAMs, and reduction in MHC.sup.hi TAMs (FIG. 16C). Furthermore, they also showed a significant increase in activated (Ki67.sup.+) T cells as well as regulatory T cells in the effector cell compartment (FIG. 16D).

(78) FIGS. 16A-1H show phenotypic changes in the TILs and tumor cells upon STING agonist vaccination. Flow cytometric immuno-phenotyping analysis of the macrophage and dendritic cell subsets in the (FIG. 16A) tumor microenvironment and the (FIG. 16B) draining lymph nodes of small tumors (200-300 mm.sup.3; n=5 for all three cohorts), and (FIGS. 16C and 16D) the TAM and T cell compartments in the tumor microenvironment of the all mice vaccinated with PBS, irradiated B16F10 with STING agonist, and irradiated 3BD9 with STING agonist (PBS vaccination cohort, n=15. Both STING agonist cohorts n=14). Additional analysis of (FIG. 16E) PD-L1 and (FIG. 16F) CD47 expression on tumors, (FIG. 16G) PD-L1 expression on dendritic cells, and (FIG. 16H) the presence of activated (Ki67.sup.+) NK cells in the tumor microenvironment was performed using multicolor flow cytometry for groups vaccinated with PBS, irradiated B16F10, irradiated B16F10 with STING agonist, irradiated 3BD9, and irradiated 3BD9 with STING agonist (PBS vaccination cohort, n=15. irradiated B16F10, n=13; irradiated 3BD9, n=9. Both STING agonist cohorts n=14). Immune cell phenotypes are presented as a percentage of the parent cell population. *p<0.5, **p<0.01, ***p<0.001 by unpaired t test performed on GraphPad Prism.

(79) STING Agonist Increases Immune Evasion Marker Expression on Tumor Cells and Reduces NK Cell Activation

(80) Next, to understand the mechanisms due to which the STING agonist seemed to dampen the anti-tumor effect and sustained tumor rejection due to the 3BD9 vaccines, the cell types in the tumor microenvironment in mice vaccinated with B16F10 and 3BD9 were studied both with and without the STING agonist as an adjuvant. A significant increase in PD-L1 expression on tumor cells as well as dendritic cells in 3BD9 vaccinated mice was discovered, along with a STING agonist (Scarlett et al. 2012; Tran Janco et al. 2015) (FIGS. 16E, 16G and 16H). This effect almost seemed to be a reversal from the less suppressive phenotype in the 3BD9 vaccinated mice to the more suppressive phenotype of the vehicle vaccinated mice. In the groups vaccinated with irradiated 3BD9, the stark reduction in tumor cell surface CD47 expression also seemed to be reversed upon the inclusion of a STING agonist in the vaccine (FIG. 16F). These observations suggest that the overall increase in the activated antigen presenting cells, also simultaneously leads to the tumors building more defenses in response to the specific cytokines and tumor-specific immune attacks. An increase in PD-L1 expression on DCs is suggestive of gradual loss of antigen presenting and T cell activation capability (Benencia et al. 2012; Keirsse et al. 2017; Dudek et al. 2013). Similarly, with increase in Type I IFNs and activated macrophages, the tumors started producing more PD-L1 and CD47 to combat the response. Another interesting observation was the significant reduction in the activated NK cells in STING agonist groups (FIG. 16H). NK cells are one of the most potent anti-tumor effector cells, and this observation is essential in understanding the downregulation of tumor specific immune response upon addition of a STING adjuvant to the whole-cell tumor vaccines.

(81) FIG. 17 shows in vitro phagocytosis of CFSE-stained live, MMC-treated, and gamma irradiated CD47.sup.+/+ (WT) and CD47.sup./ cells by APC-F4/80-stained mouse bone marrow-derived macrophages. Phagocytosis was measured as a percentage of CFSE+F4/80+ macrophages and quantified by flow cytometry. The graph shows consolidated data from three independent experiments. Statistical analyses were performed on GraphPad Prism. **p<0.01, ***p<0.001.

(82) FIGS. 18A and 18B show depletion of cell surface CD47 by the CRISPR/Cas9 method. CRISPR components were transfected into GVAX and MC-38 cells by electroporation, and single cell clones were sorted out post CRISPR knockout. Data shown on the histograms are for single cell clones GVAX-IFC (FIG. 18A), and MC38 IKC (FIG. 18B). Red peaks represent the CD47 expression on unedited WT cells (CD47.sup.+/+)positive control. Orange peaks represent the isotype control, and blue peaks represent the CD47-depleted samples, in each case. CD47 staining was performed using an Alexa Fluor 647-tagged rat anti-mouse CD47 antibody (clone miap301). Samples were run on the BD FACS Aria II flow cytometer and analysis was performed using the FlowJo software.

(83) All these outcomes proved important for charting out the specifics of an anti-tumor response after vaccination with irradiated CD47.sup./ tumor cells, which is pictorially represented in FIGS. 12A, 12B, 19A, and 19B. Therapies targeting PD-1 expressing T cells, and the regulatory compartment of the T cells would be very effective in enhancing tumor rejection. Also, the use of cytokines like GM-CSF to selectively increase the M1-type macrophages (Italiani et al. 2014) and reduce the MHC.sup.low TAM populations (Movahedi et al. 2010) can be employed as adjunct therapies to CD47-based vaccination. The expression of CTLA-4, TIM-3 and LAG-3 usually corresponds to the expression of PD-1 on T cells (Anderson et al. 2016; Wherry et al. 2015; Yi et al. 2010) and these can be potential combinations with the CD47 target. (Kooreman et al. 2018).

(84) CD47, an extensively explored immune evasion target, has always been studied in the context of active blockade using a therapeutic antibody treatment. This cell surface antigen has the potential to alter immune response dramatically, as it functions at one of the earliest and most crucial stages of an immune response cascade (Nilsson et al. 2009). The correct cocktail of factors in the body that harnesses cytokine release, macrophage activation, and the presence of tumor-specific antibodies can vividly enhance the already positive effects of CD47 as an immunotherapeutic target. In parallel, vaccinations, as a method of eliciting immune responses, are imperative in developing anti-tumor therapies. The use of inactivated non-replicating tumor cells as vaccines utilizes the immunogenic potential of whole tumor cellsit circumvents the extensive processing of neoantigens and intra-tumoral components to develop an effective vaccine concoction. Tumor cells lacking CD47 expression were hypothesized to elicit a strong immune response against tumors. These cells would aid the accumulation of anti-tumor effector cells that can protect the body from tumor development (preventive vaccines) or even eliminate tumors (therapeutic vaccines) from the body. Tumor cells lacking CD47 confer a lasting immune response, even in live formats. The prophylactic feasibilities of using inactivated CD47 depleted cancer cells was also demonstrated.

(85) CD47.sup./ inactivated tumor cells can serve as a vaccine to confer a lasting immunity in mice, where 33% (5/15) vaccinated mice remained tumor-free at the end of 90-days post tumor challenge. The mechanisms underlying the action of these vaccines have been elucidated. Extensive profiling the tumor infiltrating lymphocytes and the cells in the draining lymph nodes, revealed remarkable differences in the immune response orchestrated by vaccination. In the mice that exhibited a complete response to vaccination and did not develop tumors for 90 days post a tumor challenge, a significant increase in M1-type and corresponding decrease in the M2-type macrophages were found. The activated effector T cells (CD4.sup.+ and CD8.sup.+) were significantly higher, as expected. However, there was also a remarkable increase in the number of regulatory T cells as well as PD-1 expression on T cells. Anti-tumor immunity is conferred by the large numbers of highly activated effector cells as well as the anti-tumorigenic macrophages stimulated by the vaccine. (Kooreman et al. 2018).

(86) Enhancement in the populations of anti-tumorigenic macrophages and activated effector cell types in mice vaccinated with CD47.sup./ whole-cell vaccines in comparison with their CD47.sup.+/+ counterparts confirmed the vital role played by CD47 in the efficacy of these vaccines. There was also unexpected extreme downregulation of cell surface CD47 by the tumor cells after vaccination with irradiated CD47.sup./ cells. The vaccinated tumors express almost no CD47, suggesting that the tumors might be reacting to the specific types of immune cell priming and response, which originated from cells that did not have cell surface CD47. This also suggests that the CD47 depletion from tumor cells had a very specialized effect on the anti-tumor immune response.

(87) In the mice that showed delayed outgrowth of tumors after 3BD9 vaccination, there were significantly more effector cells infiltrating the tumor, which accounted for a delayed outgrowth. At the same time the activated antigen presenting cells (macrophages and DCs) were reduced in number, suggesting that multiple doses may prove more effective. It seemed that the regulatory T cell populations and the PD-1 expression on CD4.sup.+ T cells were elevated, signifying the anti-tumor and suppressive function of these cells leading to eventual tumor escape in these mice. An increase in tumor infiltrating NK cells in the CD47.sup./ vaccinated mice was constantly observed. Other studies have shown the dependence of NK cell-associated cytotoxicity on CD47 expression (Kim et al. 2008; Soto-Pantoja et al. 2014), and the present experiments confirmed that the absence of CD47 leads to higher amounts of active NK cell populations, leading to tumor containment and sustained rejection. Presence of prominent levels of functionally activated NK cells keeps the immune response consistently anti-tumorigenic Guillerey et al. 2016; Lowry et al. 2017). Other principal factors in a suppressive environment are TAMs and MDSCs. The MHC-II.sup.low TAMs were significantly elevated, suggesting that the tumors became more hypoxic in the CD47.sup./ 3BD9 vaccinated mice. These TAMs were also found to express high levels of Ly6G, the neutrophil lineage marker. Aside from the MHC-II.sup.low TAMs, all other suppressor cells, including the PMN- and Mo-MDSCs, were highly downregulated in the TME of CD47.sup./ vaccinated tumors, proving that they have an overall anti-tumorigenic environment.

(88) An important consideration in the design of therapeutic regimes for actively growing tumors is the intensity of immune activity in the tumors (Lehmann et al. 2017; Lee et al. 2012). Cold tumors, hence, are often hard to treat just by therapeutics, and this also one of the reasons solid tumors in the first place are harder to treat with immunotherapeutic agents than liquid or systemic tumors. The immune infiltration was found not only to increase, but also a consistent composition was maintained throughout the phases of tumor growth after vaccination. The extreme heterogeneity that is characteristic of melanoma tumors, is reversed upon vaccination, and tumors harvested at the three stages of growth show more homogeneity in TIL populations, especially in the suppressive subsets. These tumors also show a gradual increase in the effector cell compartment as tumor progresses, an effect opposite to the one seen in the tumors of mice vaccinated with inactivated B16F10 cells, which distinguishes the two types of vaccines and emphasizes the involvement of CD47 as a target in this regime.

(89) The stimulator of interferon genes (STING) is a protein present on the endoplasmic reticulum of cells and is activated by a stream of preceding signaling mechanisms that involve the recognition of pathogen- or tumor-associated DNA by dendritic cells (DCs). One of the most crucial players in this pathway is the cyclic GMP-AMP (cGAMP) synthase (cGAS) that produces cyclic dinucleotides, commonly annotated as 23-cGAMP that activated STING directly (Sokolowska et al. 2018; Li et al. 2016). The activation of STING leads to the production of Type-I interferons (IFNs): IFN- and IFN-. These IFNs act both as a self-stimulatory signal that helps in the maturation of DCs and a T cell proliferation signal that leads to the priming, maturation, and proliferation of CD8.sup.+ cytotoxic T cells (CTLs). STING agonists have been used in the recent years to mediate anti-tumor immunity in solid tumors (Curran et al. 2016; Bullock et al. 2017). Also, the use of adjuvants to vaccines has been names as one of the most crucial elements in the design of vaccines (Overwijk et al. 2017; Bullock et al. 2017). STING has also been shown previously to be involved in the CD47 blockade-based mechanism of anti-tumor activity (Liu et al. 2015). For the purpose of this study, 23-c-di-AM(PS)2 (Rp,Rp) VacciGrade, a cyclic dinucleotide cGAMP analog, that directly activates the STING protein, was used in the vaccine solution as an adjuvant to elicit a tumor-specific immune response.

(90) The addition of a STING pathway agonist to therapeutic regimes is shown to increase the infiltration of dendritic cells into the TME and help the development of effective anti-tumor T cell responses. In this study, however, the addition of a STING pathway agonist as a vaccine adjuvant did delay the tumor development significantly, but also seemed to dampen the effect of the vaccine and cause an inferior prognosis. The addition of the adjuvant significantly increased not only dendritic cell infiltration into tumors but also macrophage infiltration. The most glaring reasons for the seemingly inverse effect of the STING agonist on vaccine efficacy are a drastic increase in PD-L1 expression on tumor cells and DCs, and a remarkable reduction of activated NK cells in these mice. Studies have found previously the very alarming role DCs seem to play in the TME and non-specific activation of the DC compartment could be leading to an unprecedented effect in combination with the CD47-based regime (Benencia et al. 2012; Keirsse et al. 2017; Dudek et al. 2013; Merad et al. 2013). Additionally, the STING agonist seemed to reverse the earlier observed CD47 reduction on tumor cells and showed a significantly higher expression in comparison with the CD47.sup./ cell vaccine regime without the adjuvant. This adjuvant study was a preliminary, single-dose study, and the effects observed would help design a better method of employing the anti-tumor capabilities of STING pathway stimulation.

(91) The present findings point to the fact that as a certain type of immune cell population increases in percentage, the regulatory or suppressive phenotypes of that population also increase correspondingly. An increase in M2-type and TAM populations was observed as compared to the M1-type and the activated subsets when macrophage populations are enlarged. Similarly, elevated amounts of activated T cell populations were found after CD47.sup./ vaccination, but also correspondingly found elevated levels of regulatory subsets of these T cells. In the STING adjuvant-treated groups, there was extremely high infiltration of DCs, but the PD-L1 expression on these DCs also increase dramatically, leading to a suboptimal scenario.

(92) Cell Culture. The B16F10 (mouse melanoma) cell line was obtained from ATCC. The cells were cultured using high glucose DMEM (Gibco, MD) supplemented with 10% v/v Fetal Bovine Serum (FBS) (Sigma-Aldrich, MO). Mouse bone marrow derived macrophages (BMDMs) were obtained by flushing out the bone marrows of untreated mice, and culturing for 7 days in RPMI 1640 (Gibco, MD) supplemented with 10% v/v FBS.

(93) CD47 knockout using CRISPR. The spCas9(BB)-2A-GFP (PX458) plasmid (Addgene, MA) containing the gRNA sequence targeting CD47, and the Cas9 and an eGFP separated by a T2A sequence was transfected into the B16F10 cells using the Viafect transfection kit (Promega, WI). The successfully transfected single cells were sorted into 96-well plates using a BD FACS Aria II sorter. Single cells were expanded and tested for biallelic knockout of CD47 using PCR, T7E1 mismatch assay, and Sanger's sequencing. Absence of the protein expression in genome edited cells was reconfirmed by immunofluorescence microscopy and flow cytometric analysis. Rat anti-mouse CD47 antibody clone miap301 (BD Biosciences, NJ) was used as a primary antibody to determine the absence of CD47 expression in edited cells. This antibody was used at a final concentration of 10 g/mL (1:50 dilution). Goat anti-rat IgG tagged with AlexaFluor-488 (2 mg/mL) (Sigma-Aldrich, MO) was used as a secondary antibody. This antibody was used at a final concentration of 10 g/mL (1:200 dilution).

(94) The required gRNA sequence (in the form of a ligated double stranded DNA, Invitrogen, CA) was cloned into the cloning site of the plasmid under a U6 promoter, using restriction digestion. The cloning site was flanked by restriction sites for Bbsl enzyme (New England Biolabs, MA). Two gRNAs were used, targeting Exons 1 and 2 of the cd47 gene. The distance between the cut sites for these gRNAs was approx. 12.9 kb. Confirmation of knockout was done using a PCR primer set (MA-FP/MC-RP) (Eurofins Genomics, KY) flanking the cut sites. The deletion of the 12 kb fragment was visually analyzed by PCR (BioRad, CA) and gel electrophoresis.

(95) In vitro phagocytosis assay. Macrophages were extracted from bone marrows flushed out from the femurs of nave mice, plated on 10 cm dishes. The growth medium was supplemented with 10 g/ml GM-CSF. The cells were cultured for 7 daysthe growth medium was replaced every 3 days. 510.sup.4 macrophages were co-cultured with 110.sup.5 Carboxyfluorescein (CFSE)-labeled B16F10 tumor cells in RPMI 1640 (Gibco, MD) for 2 hours at 37 C. and 5% CO.sub.2 in the presence of 0.5 g opsonizing antibodies, anti-CD47 miap301 (BD Biosciences, NJ) and anti-gp75 TA99 (BioXCell, NH). Macrophages were then stained with APC-tagged F4/80 (Biolegend, CA). Phagocytosis analysis was done on a BD FACS Aria II flow cytometer.

(96) Vaccination and animal study. 7-week-old female C57BL/6 mice (Jackson Laboratory, ME) were housed in a pathogen-free facility in the vivarium of Binghamton University. All animal study procedures were approved by the Institutional Animal Care and Use Committee (IACUC) at Binghamton University.

(97) 510.sup.5 B16F10 cells were implanted on the left flank of 7-week-old female C57BL/6 mice to induce the tumor development. The tumors were measured using calipers every alternate day after tumor growth was observed.

(98) To prepare whole cell vaccines, CD47.sup./ B16F10 cells (Referred to as CD47.sup./ 3BD9) and CD47.sup.+/+ B16F10 cells (Referred to as B16 WT) were irradiated with 35Gy gamma irradiation using a Cs source (University of Rochester Medical Center).

(99) Mice (15 animals per group) were vaccinated with 510.sup.5 irradiated 3BD9 or B16 WT cells subcutaneously on the left flank and were challenged with 510.sup.5 live B16F10 cells 7 days later. Two more cohorts of 15 mice each were vaccinated as mentioned above with 3BD9 and B16 WT along with 10 g/mouse of a STING pathway agonist, 23-c-di-AM(PS)2 (Rp,Rp) VacciGrade (InVivoGen).

(100) Tumors and TDLNs at three different stages of tumor growthsmall (200-300 mm.sup.3), medium (500-600 mm.sup.3), and large (800-900 mm.sup.3)were collected from 5 mice per group after euthanasia by CO.sub.2 inhalation. For the mice that did not develop tumors, TDLNs alone were collected 90-days post tumor challenge. Organs were enzymatically digested and made into single cell suspensions for immunostaining.

(101) Immunophenotyping. To determine the types of tumor infiltrating lymphocytes (TILs) and TDLN lymphocytes, single cell suspensions of the tumors and TDLNs were stained using two multicolor panels covering the APC compartmentmacrophages (M.sub.), dendritic cells (DCs), myeloid derived suppressor cells (MDSCs), and monocytes; and the effector cell compartmentcytotoxic T cells (CTLs), helper T cells (T.sub.H cells), memory cells, regulatory T cells (T-regs), natural killer (NK) cells, and activated effector cells. All pre-conjugated antibodies were purchased from Biolegend unless otherwise specified. Samples were run on LSR Fortessa flow cytometers (University of Rochester Medical Center Flow Core Facility) and analyzed using the FlowJo software v10 (TreeStar).

(102) Statistical Analyses. All statistical analyses were performed on GraphPad Prism. The non-parametric Mantel-Cox test was used for survival and tumor-free mice data. For other correlative analyses either a one-way ANOVA or an unpaired t test was performed based on the number of groups being compared. The Shapiro-Wilk test was used to determine population distribution when necessary. 95% confidence interval was used in all analyses to accept or reject the null hypothesis.

(103) Nucleofection of RNPs. For knocking out CD47 from GVAX cells (GM-CSF producing B16F10 cells), the ribonucleoprotein (RNP) method was used. gRNA (in the form of mRNA, Integrative DNA Technologies, IA)) was complexed with purified Cas9-3NLS protein in vitro (obtained from Integrative DNA Technologies, IA), and the mixture was added to GVAX cells and electroporated using the Neon Electroporation System (Invitrogen, CA). The cells were analyzed by flow cytometry (BD FACS Aria II) 6-9 days after electroporation, and CD47.sup./ cells were sorted for single cell colonies into 96-well plates using the BD FACS Aria II Cell Sorter. GVAX 1FC was chosen as the CD47.sup./ clone for further studies.

(104) T7E1 mismatch assay. PCR products from CRISPR-edited cells were subjected thermal denaturation at 95 C. and cooled down at the rate of 0.1 C./min to 85 C., and then at the rate of 1 C./min to 25 C. Once cooled, 1U of the T7E1 enzyme (New England Biolabs, MA) was added to the mix and incubated at 37 C for 15 minsthe final volume of the mix was 20 l. The reaction was stopped using 1.5 l of 0.25M EDTA, and the digested were run on an agarose gel to check for indels in the cut site.

(105) Immunofluorescence microscopy. Cells were plated on individual wells of a 6-well plate and maintained until they reached 70-80% confluency. The cells were washed with 1 D-PBS (Gibco, MD) once. Then they were fixed with 4% paraformaldehyde (PFA) at room temperature for 15 minutes, and washed twice with 1 D-PBS, followed by blocking with 0.5% BSA for 30 mins at room temperature. After blocking, the cells were incubated with primary antibody for 1 hour at 4 C. and washed once with 1-DPBS. Secondary antibody was then added to the cells and incubated for 1 hour at 4 C. After another wash with 1-DPBS, the cells were counterstained with a 1:1000 dilution of DAPI (Invitrogen, CA) at room temperature for 15 minutes. The cells were again washed twice with 1 D-PBS and viewed using a fluorescence microscope (Nikon, NY).

(106) In vitro phagocytosis of GM-CSF producing B16F10 cells (GVAX). GVAX is a type of B16F10 mouse melanoma cell line that is genetically engineered to produce the cytokine granulocyte macrophage colony stimulating factor (GM-CSF). This cytokine is imperative in the growth, function, and maturation of myeloid derived granulocytes including macrophages and dendritic cells. The use of GVAX as a vaccination regime would mean the presence of GM-CSF as an innate adjuvant. Phagocytosis was performed with fresh BMDMs. 510.sup.4 macrophages were co-cultured with 110.sup.5 CFSE-labeled GVAX and B16F10 tumor cells for 2 hours at 37 C. and 5% CO.sub.2 in the presence of 0.5 g opsonizing antibodies, anti-CD47 miap301 (BD Biosciences) and anti-gp75 TA99 (BioXCell). Macrophages were then stained with APC-tagged F4/80. Phagocytosis analysis was done on a BD FACS Aria II flow cytometer.

(107) Phenotyping tumor cells. gp75 profiling for use of TA99 antibody as opsonizing agent in in vitro phagocytosis assay. The B16 WT and CD47.sup./ 3BD9 cells were treated with purified anti-mouse TA99 primary antibody (BioXCell), and then with a rat anti-mouse Alexa Fluor 488-tagged secondary antibody (Sigma). As a negative control, the same cells were treated with only the secondary antibody. Cells were analyzed by flow cytometry on the BD FACS Aria II.

(108) CD47 profiling of irradiated cells for use as vaccines in vivo. The gamma irradiated cells used as vaccinesB16 WI and CD47.sup./ 3BD9 cells were stained for cell surface CD47 expression using an Alexa Fluor 647-tagged anti-CD47 antibody, clone miap301 (Biolegend), and analyzed on the BD FACS Aria II. Isotype control for the CD47 antibody, a rat IgG-k Alexa Fluor 647 (Biolegend), was used as a negative control.

(109) Comparison of CD47.sup./ and WT melanoma tumor growth in vivo. Tumor implants. 510.sup.5 B16F10 tumor cells (CD47.sup./ and WT) were implanted subcutaneously into the left flanks of 7-week-old female C57BL/6 mice. Tumor growth was observed every alternate day and tumors were measured using a Vernier caliper. Two separate experiments with 4 mice per group were performed to compare the tumor growth pattern. Another cohort of 15 mice per group was used to study tumor infiltrating lymphocytes and TDLN immunophenotypes at three different stages of tumor growthsmall (200-300 mm.sup.3 tumors), medium (500-600 mm.sup.3 tumors), and large (800-900 mm.sup.3 tumors).

(110) Organ Preparation and Immuno-Phenotyping.

(111) Tumors. Tumors were harvested at the three stages mentioned above, stored and transported in media containing RPMI 1640 (Gibco), 2% FBS (Sigma), and 1 Pen-Strep antibiotic (Gibco) on ice. Tumors were then digested using 1 mg/ml Collagenase and 10 mg/ml DNase (Sigma) at 37 C. for one 20-minute cycle with intermittent high-speed vortexing, and another 20-minute cycle with intermittent vigorous pipetting. Homogenized suspensions were filtered using a 70 m filter and suspended in FACS Buffer (1 PBS, 2% FBS, 2 mM EDTA (Gibco), and 25 mM HEPES (Gibco)) until use.

(112) Lymph nodes. TDLNs were harvested from mice along with the tumors and transported separately in media containing RPMI 1640 (Gibco), 2% FBS (Sigma), and 1 Pen-Strep antibiotic (Gibco) on ice. TDLNs were cut into small pieces using a sharp scalpel and 26G needle (BD Biosciences). Fragments were placed in 2 ml warm Digest Buffer containing 1 mg/ml Collagenase and 10 mg/ml DNase and digested at 37 C. for 30 minutes with intermittent high-speed vortexing. Fragments were filtered using a 70 m filter and suspended in FACS Buffer until use.

(113) Immunostaining. Approximately 2-4 million cells from tumors and 0.5-1 million cells from TDLNs were transferred to non-treated U-bottom 96-well plates (Costar). Two multi-color panels were used for tumors and TDLNsPanel T1 for tumors and Panel SL1 for TDLNs (containing all the antibodies for phenotyping the APC compartment, and tumor antigens), and Panel P2 for both (containing an antibody cocktail for phenotyping the effector cell compartment). All the antibodies used in this experiment were titrated using mouse melanoma tumors and the dilutions were optimized for use in the multi-color panels. CD16/CD32 Fc Block (Biolegend, CA) was used to block immune cell receptors along with the antibody cocktails. Cells were stained for extracellular antigens in FACS Buffer for 30 minutes on ice, and then fixed and permeabilized using the Transcription Factor Buffer Set (BD Biosciences) for 30 minutes at 4 C. The intracellular antigen staining was performed using the Perm/Wash Buffer from the set for 30 minutes at 4 C. Stained cells were stored in the dark at 4 C. in FACS Buffer until analysis by flow cytometry. Multi-color compensation was performed by staining polymer beads from the AbC Antibody Compensation Kit (Invitrogen) with the appropriate amounts of all antibodies used in the staining panels.