Altered myeloid & lymphoid composition of tumor microenvironment following anti-SEMA4D and immunotherapies Elizabeth E. Evans, Terrence L. Fisher, Holm Bussler, Crystal Mallow, Christine Reilly, Sebold Torno, Desa Rae Pastore, Maria Scrivens, Alan Howell, Leslie Balch, , John E. Leonard, Terrence L. Fisher, Clint Allen*, Paul E. Clavijo*, Gregory Lesinski**, Christina Wu**, Conor Steuer**, Nabil Saba**, Brian Olson**, Siwen Hu-Lieskovan***, Antoni Ribas***, Emily G. Greengard**** Ernest S. Smith, Maurice Zauderer. NIH/NIDCD Head and Neck Surgery Branch*, Winship Cancer Institute of Emory University**, David Geffen School of Medicine at UCLA***, University of Minnesota Division of Pediatric Hematology/Oncology ****, and Vaccinex, Rochester, New York

Summary PRECLINICAL: Anti-SEMA4D Mab neutralizes SEMA4D barrier at tumor CLASSICAL-Lung Phase 1/2b Trial: Combination with Avelumab

Anti- 4D (SEMA4D, CD100) blocking antibody promotes immune infiltration, margin and shifts the balance of tumor immunity This phase 1b/2, open label, single arm, first-in-human combination study is designed to evaluate the safety, Exploratory Biomarker Analysis: shift in balance of CD8:Treg reduces immunosuppressive myeloid cells, and enhances T cell activity and tumor growth tolerability and efficacy of pepinemab in combination with avelumab in 62 subjects with advanced (IIIB/IV) NSCLC. A. Sema4D inhibition in combination with various immunotherapies in preclinical animal models. B. Increased infiltration of pro-inflammatory APC C. Shift in balance of tumor immunity in TME • Pre-treatment tissue shows gradient and CD8+ T cells infiltrating malignant cells (green), SEMA4D represents a novel target to regulate immune infiltration and mesenchymal Dose Escalation 5 mg/kg VX15 10 mg/kg VX15 20 mg/kg VX15 Recruitment suppression, sources of resistance to current immunotherapies. Phase 1b stained with pan-cytokeratin. (n=12) (n=3) (n=6) (n=3) Closed PRECLINICAL MECHANISM OF ACTION STUDIES: Blocking antibody to SEMA4D directly Vessel • Post-treatment biopsy demonstrates reactive alveolar parenchyma and enhanced M1/M2 ratio and reduced both expression of chemokines that recruit MDSC and M2-TAM MDSC + 10 mg/kg Avelumab Q2W inflammation. According to the ability of MDSC to suppress T cell activity. Antibody blockade simultaneously restored the pathology report, “No evidence of ability of dendritic cells and cytotoxic T cells to infiltrate the TME, increasing ratio of Teffector Tumor Phase 2 Dose Expansion Pre-treatment Post-treatment tumor.” to Tregulatory cells, in syngeneic tumor models. IO Failure • CD8+ to FoxP3 T cell ratio increases Control Ig Control (n=50) Recruitment + + 10 mg/kg VX15 CD8 T FoxP3 Tregulatory Tumor PRECLINICAL COMBINATION THERAPIES: anti-SEMA4D MAb enhanced the activity of co- Ongoing in the tumor microenvironment post administered immunotherapies, including antibodies to PD1, CTLA-4, and LAG3, and (n=28) treatment SEMA4D CD11c F4/80 CD8 Tumor-specific Cytotoxic T Cells Teff:Treg See POSTER CD8+ and FoxP3+ T-Cells CD8 / FoxP3 T-Cell Ratio epigenetic modulators including entinostat. Tumor IHC samples are from one subject and + 10 mg/kg Avelumab Q2W CLASSICAL CLASSICAL 67 #CT086 represent core biopsies that were isolated from the TRANSLATIONAL BIOMARKER ANALYSIS: Methods to assess immune phenotype and %CD3+CD8+ IFNg ELISPOT 150 15 4-fold same lung lesion both at 2 days pre-treatment and 33 biomarkers in translational and clinical studies include flow cytometry and whole slide scans IO Naïve days post-treatment with pepinemab + avelumab. This Recruitment 2 increase

MAb patient was a PR according to RECIST, with a -46% of multiplex IHC panels to examine MDSC, M1/M2 macrophage, monocytes, activated DC, B 10 mg/kg VX15 100 Ongoing 10 lesion decrease at first scan (day 50) and overall lesion cells, exhausted, activated and stem-like populations of T cells in clinical samples. Tumor (n=22) shrinkage at every scan after that. 5 micron FFPE sections were stained sequentially with CLINICAL TRANSLATION: Clinical trials of immune checkpoint inhibitors (ICI) in combination 50 5

Study Objectives Hematoxylin, pan-cytokeratin, CD8 and FoxP3. Images # Cells per mm # Cells

with pepinemab (VX15/2503), humanized anti-SEMA4D antibody, are currently underway in CD8 /Ratio FoxP3 were taken at 10x magnification with CD8 (red) and • The primary objective of dose escalation is safety, tolerability, and identification of the RP2D for dose expansion. FoxP3 (blue) overlays of pre and post treatment several cancer indications. Pepinemab treatment was well tolerated in a Phase I oncology 0 αSEMA4D αSEMA4D 0 • Secondary objectives include evaluation of efficacy, immunogenicity, and PK/PD, and an exploratory objective is biopsies. Scans were co-registered for each stain; trial (NCT01313065) and is currently being evaluated as single agent or in ICI combinations Pre-treatment Post-treatment Pre-treatment Post-treatment to identify candidate biomarkers of activity total number of CD8+ and FoxP3+ cells were quantified in: (i) a Phase 1b/2a combination trial of pepinemab with avelumab in ICI naïve or ICI SEMA4D is strongly expressed at the invasive margin of tumors. Antibody blockade of SEMA4D facilitates migration of APCs and T cells into the tumor. (A) CD8+ T-Cells from entire section and normalized by sample area • Pre and post treatment biopsies are collected when possible FoxP3+ T-Cells using Visiopharm software. refractory NSCLC (CLASSICAL-Lung) (NCT03268057); (ii) a phase 1 combination trial of SEMA4D expression at invasive margin of murine Colon26 tumors restricts infiltration of PLXNB1+ DC into TME. Brackets indicate area of SEMA4D gradient. (B) Anti- pepinemab with nivolumab or ipilimumab in melanoma patients who have progressed on SEMA4D MAb promotes infiltration of pro-inflammatory CD11c+/F4-80+ antigen presenting cells, while reducing CD206+ M2 macrophage. Pro-inflammatory APC any anti-PD-1/PD- (NCT03425461); (iii) a neoadjuvant integrated biomarker trial in patients recruit and activate CD8+ T cells within TME. Colon26 tumor-bearing mice were treated with Control Ig or anti-SEMA4D/MAb67 antibodies (50 mg/kg, weekly IP) and tumors were harvested on day 27 and FFPE sections were stained by IHC or tumors were dissociated and assessed for immune cell markers by flow cytometry. with metastatic colorectal, pancreatic (NCT03373188) and head and neck (NCT03690986) Leukocytes were enriched from whole tumor digests using lympholyte-M and cultured for 2-days and supernatants were assessed for T cell activity by ELISPOT (C), Integrated Biomarker Window of Opportunity Study: MSS CRC with resectable liver metastases, PANC, HSNCC, Melanoma cancers treated with pepinemab in combination with nivolumab or ipilimumab; and (iv) a n=8-12 mice/group. Phase 1/2 trial of pepinemab in children with solid tumors and children and young adults See POSTER #CT016 Exploratory Biomarker Analysis: Immune infiltration and drug penetration in TME with osteosarcoma (NCT03320330). Clinical trials will evaluate safety, tolerability, efficacy, and biological endpoints, including immunophenotyping tumors and blood. Resectable PDAC PRECLINICAL: Combination Immunotherapy Shift in balance of infiltrating CD8+ T cells and MDSC T cells in TME are saturated with VX15 Resectable melanoma Resectable HNSCC (n=16) VX15/2503 Anti-SEMA4D (n=36) Anti-SEMA4D Mab blocks Colon26: anti-CTLA-4 Combination Colon26: anti-LAG3 Combination (n=36) CRC resectable liver EMORY CRC M-MDSC's within Total Tumor M-MDSC CD8+ T Cell R-Ras, RhoA and A. D. A. TUMOR MAKEUP PERCENTAGE B. C. D. (D) Representative data binding to receptor & Post-treatment Resections Center of Tumor Invasive Margin SEMA4D ROCK-kinase Total enrolled: 8 Total enrolled: 3 mets (n=16) Post-Treatment from melanoma patient PLXNB1/ 30 Normal liver 06A Expressing Immune signaling activity Total enrolled: 3 100 dosed on Days 1 and 21 PLXNB2 • Inhibits differentiation/ with pepinemab (15mg/kg) 80 Cell: and and nivolumab (240mg); Cytoskeletal function of MDSC, M2 VX15/2503 (20 mg/kg) 20 M-MDSC Infiltrating Myeloid No treatment (SOC) 60 tumors were harvested 15- Crosslinking reorganization, TAM and Treg No treatment (SOC) Ipilimumab (1 mg/kg) Bed Tumor 28 days later. TIL were leukocytes, and Cells 40 2 migration, • Promotes infiltration of 10 isolated and CD3+ T cells Tumor signaling differentiation,

SEMA4D potent APC and T cells VX15/2503 (15 mg/kg) 20 were assayed for saturation

per mm

No Treatment No % Bed total Tumor of Cytokine secretion into TME VX15/2503 (20 mg/kg) 10% cells # S100A9+CD33+ Tumor Bed of the cellular SEMA4D VX15/2503 (20 mg/kg) 0 0 target. Nivolumab (480 mg) no treatment pepinemab pepinemab + no pepinemab pepinemab+ normal treatment ipilimumab adjacent Normal liver Isotype, Anti Hu IgG4, VX15/2503 (15 mg/kg) ipilimumab Ipilimumab (3 mg/kg) Anti Hu IgG4 + pepinemab PRECLINICAL: SEMA4D blockade reverses myeloid suppression % Tumor Nodules VX15/2503 (20 mg/kg) % Stroma Nivolumab (480 mg) VX15/2503 (20 mg/kg) Ipilimumab (1 mg/kg) % Necrotic Area MDSC Function Nivolumab (480 mg) A. M1/M2 ratio B. B. MOC1 HNSCC: anti-CTLA-4 Combo E. MeColon26:an Col26 anti15on-PDc18-21 PCombinationD The key observations relate to distribution of T cells and MDSC. FLOW: TIL AND PBMC EVALUATION Human Multiple Myeloma Surgical resections were analyzed from one CRC patient/treatment arm Pepinemab 3 • target (SEMA4D) saturation Control VX15/2503 (15 mg/kg) VX15/2503 (20 mg/kg) Tumor Bed m Nivolumab (3 mg/kg) Ipilimumab (1 mg/kg) following one cycle (3-5 weeks) of treatment and one patient who did

m 1000 • CD3+ CD4+/CD8+ T cells aSEMA4D 0.3423 0.0307 Nivolumab (480 mg) not receive antibody treatment. No conclusion can be drawn regarding e Ipilimumab (1 mg/kg) Normal liver

m tumor necrosis because patients received neoadjuvant chemotherapy • CD19+ B cells

u aPD-1 + CTRL 0.0599

l +

o 2 Doses 1 Dose 1 Dose prior to surgery. • CD3+/- CD56+ NK-T and NK v

0.0030

r aPD-1 + aSEMA4D

o Pre-treatment Surgical 5 micron FFPE sections were stained sequentially for each marker and 500 • CD3+ CD4/8+ HLA-DR+ CD38+ m * u Biopsy Resection scanned at 40X. Scans were co-registered for each stain in multiplex

T Ki67+ stem-like PD-1 responsive T

n A) Percent of total tumor bed area for each component was quantified a cell population

e and averaged from 2 sections/patient. B) 10x images with

Ipiilimumab M 0 Biomarker Analysis: S100A9+/CD33+ MDSC (blue) overlays on cytokeratin stain (grey Pepinemab 10x Tumor Bed 3.3x • PD1/Tim3,CD26 expression - + + + M1 = CD14 CD16 M2 = CD14 CD16 0 20 40 60 80 scale) are shown. Total number of S100A9+/CD33+ cells were • Treg: CD3+ CD4+ CD25+ GITR+ Safety: • Compare relative to baseline biopsy and Day of Study quantified from entire tumor bed area, normalized by area of tumor bed CD127- T cells; EVALUATE PK/PD among Rx cohorts using Visiopharm software, and 2 sections/patient were averaged in bar • immune profile in tumor and peripheral blood graphs. C) CD8+ T cells (red) overlays on cytokeratin stain • Foxp3 & SEMA4D MDSC Recruitment: YUMM2.1 Melanoma: anti-CTLA-4 Combo Colon26: Epigenetic Modulator (black/white) at tumor/normal liver margin are shown (3.3x). • SEMA4D, Plexin B1/B2, and PD-L1 In vitro treatment of MOC1 tumor cells Entinostat Combination In vivo treatment: MOC1 TME C. F. * Total enrolled, as of March 15, 2019. C. CXCL1 CXCL2 CXCL5 D. CXCL1 CXCL2 CXCL5 Melanoma – PD-1 or PD-L1 Refractory Ph 1/2b Trial: TRANSATIONAL Biomarker analysis: Serial Multiplex IHC

Combination with Ipilimumab or nivolumab PD-L1 PanCK FoxP3 CD33 CD11c CD163 Evaluation of spatial and cell-specific expression of SEMA4D and its cognate CTRL αS4D CTRL αS4D CTRL αS4D CTRL αS4D CTRL αS4D CTRL αS4D receptors: Liver metastases from Phase 1b Dose Escalation 5 mg/kg VX15 Q2W 10 mg/kg VX15 Q2W colorectal cancer patient (NCT03373188) SEMA4D blockade increases ratio of M1/M2 when exposed to SEMA4D+ tumors. (A) Human PBMC (n=up to 30) (n=3) (n=6) Recruitment was assessed for various cell types using were cultured with conditioned media from co-culture of multiple myeloma RPMI 8226 with human Immunomodulatory effects of SEMA4D blockade can enhance other immunotherapies. A, D, E. Colon26 (500,000 cells) were subcutaneously implanted into Ongoing sequential serial stains on the same bone marrow stroma (mock) or with Control Ig or anti-SEMA4D/2503 (αS4D). Balb/c mice, that were then treated with αSEMA4D / MAb67 (10 mg/kg, weekly IP X4), αLAG3/C9B7W (10 mg/kg 2x/week X4; n=20); αCTLA-4 / MAb UC10-4F10 with nivolumab 240 mg, Q2W section to allow multiplex IHC and SEMA4D directly promotes function of myeloid derived suppressor cells (MDSC); suppression is (100/50/50 µg, q3 days; n=20), αPD-1 / MAb RMP1-14 (10 mg/kg, twice/week, starting on day10, n=20), B. MOC1 HNSCC (5x106 cells) were subcutaneously or ipilimumab 1 mg/kg, Q6W x4 colocalization of markers for various reversed by antibody blockade. (B) MDSC were isolated from HNSCC MOC1 in vivo tumors and co- implanted into C57Bl/6 mice, that were then treated with αSEMA4D/MAb67 (10 mg/kg, weekly IP), αCTLA-4 / MAb 9H-10 (5 mg/kg, q5D); n=10. C. YUMM2.1 immune cell subsets. cultured ex vivo with rSEMA4D (10 μg/ml) or antibodies, in presence of naïve T cells labeled with CFSE melanoma were implanted into C57Bl/6 mice and treated with αSEMA4D/MAb67 (10 mg/kg, weekly IP), αCTLA-4 / MAb UC10-4F10 (5 mg/kg 2x/wk X3 doses), IHC MULTIPLEX 3 VX15/2503, RP2D, Q2W in a T cell suppression assay (right panel). gMDSC were assessed for ARG1 expression via qRT-PCR after αPD-1 / MAb RMP1-14 (10 mg/kg 3x/week); n=8. F. Treatment of established tumors with Entinostat (ENT, 20 mg/kg 3x/wk, at TV ~250mm , n=20). Phase 2 Dose Expansion Similar analysis will be applied to clinical ex vivo exposure to recombinant or antibody for 3 hours (n=5/group). Arginase production in (n=36) + nivolumab, 240 mg, Q2W Sema4D CD45 CD8 S100A9 CD20 CD68 biopsies from all trials. TME suppresses T cell function (right panel) (C) MOC1 cells in vitro were exposed to Sema4D mAb (10 µg/mL) or isotype for 24 hours and analyzed for myeloid chemokine expression by qRT-PCR. (D) Mice FORWARD LOOKING STATEMENT: To the extent that statements contained in this presentation are not descriptions of historical facts regarding Vaccinex, Inc. (“Vaccinex,” “we,” “us,” or “our”), they VX15/2503, RP2D, Q2W 1.Clavijo PE et al. Cancer Immunol Res. 2019 Feb;7(2):282-291 bearing MOC1 tumors were treated in vivo with isotype control or anti-Sema4D Ab (αS4D) (n=5/group). are forward-looking statements reflecting management’s current beliefs and expectations. Words such as “may,” “will,” “expect,” “anticipate,” “estimate,” “intend” and similar expressions or their 2.Evans, EE et al 2015. Cancer Immunol Res. 3(6):689-701. Whole tumor digests were analyzed for myeloid chemokine expression via qRT-PCR. negatives (as well as other words and expressions referencing future events, conditions, or circumstances) are intended to identify forward-looking statements. No representations or warranties are + ipilimumab, 1 mg/kg, Q6W x4 offered in connection with the data or information provided herein. This presentation is intended for informational purposes only and may not be relied on in connection with the purchase or sale of 3.Evans EE, Paris M, Smith ES & Zauderer M. 2015): OncoImmunology [email protected] any security. Any offering of our securities will be made, if at all, only upon the registration of such securities under applicable securities laws or pursuant to an exemption from such requirements. 4.Patnaik A et al. Clin Cancer Res. 2016 Feb 15;22(4):827-36. 5.Fisher et al, 2016. MAbs. 8(1): 150-162.