0

10

20

30

40

50

60

Immune Design Corp. Page 1 of 1Protocol Number IMDZ-G142 Date of Datacut: 03Nov2017

Figure 2.2.3 Spider Plot of CD8 Positive Cells - G100 (10 ug) + anti-PD-1

All Dosed Patients

Program Name: figure-2-2-3-spider-cd8-c.sas Executed: 24NOV2017

Immune Design Corp. Page 1 of 1Protocol Number IMDZ-G142 Date of Datacut: 03Nov2017

Figure 2.2.2 Spider Plot of CD8 Positive Cells - Part 2 G100 (10 ug)

All Dosed Patients

Program Name: figure-2-2-2-spider-cd8-c.sas Executed: 24NOV2017

**

*

*

*

** **

**

N= 10 N= 11

*R/R (previously treated) patient

I. RATIONALE / BACKGROUNDLocal Treatment è Systemic Response

II. G142 STUDY DESIGN

Table 3. Overall Response

Intratumoral G100 Induces Systemic Immunity and Abscopal Tumor Regression in Patients with Follicular Lymphoma: Results of a Phase 1/ 2 Study Examining G100 Alone and in Combination with Pembrolizumab Christopher Flowers, MD1, Carlos Panizo, MD, PhD2, Iris Isufi, MD3, Alex F. Herrera, MD4, Craig Okada, MD, PhD5, Elizabeth H. Cull, MD6, Bela Kis, MD7, Jorge M. Chaves, MD8, Nancy L. Bartlett, MD9, Weiyun Ai, MD, PhD10, Luis de la Cruz-Merino, MD11, Locke J. Bryan, MD12, Roch Houot, MD, PhD13, Kim Linton, MBChB PhD14, Javier Briones, MD15, Ian Chau, MD16, Gottfried R. von Keudell, MD, PhD3, Hailing Lu, PhD17, Frank J Hsu, MD18 and Ahmad S. Halwani, MD19

1Emory University, Atlanta, GA; 2Clínica Universidad de Navarra, Pamplona, Spain; 3Yale Cancer Center, New Haven, CT; 4City of Hope, Duarte, CA; 5OHSU, Portland, OR; 6GHS Cancer Institute, Greenville, SC; 7Moffitt Cancer Center, Tampa, FL; 8Northwest Medical Specialties, PLLC, Tacoma, WA; 9Washington University, St Louis, MO; 10UCSF, San Francisco, CA; 11Hospital Universitario Virgen Macarena, Seville, Spain; 12Augusta University, Augusta, GA; 13University Hospital Ponchaillou, Rennes, France; 14Christie Hospital, Manchester, UK; 15Hospital Santa Creu i Sant Pau, Barcelona, Spain; 16Royal Marsden, London, UK; 17Immune Design, Seattle, WA; and 18South San Francisco, CA; 19University of Utah, Salt Lake City, UT

Abstract #2771

1 2

3 ReleasedTumor Antigens

A. B.

• Background: Follicular lymphoma (FL) tumor microenvironment (TME) can suppress anti-tumor immune activity by physical barriers (fibrosis), blockade of attractant chemokines, production suppressive factors, and activation of negative regulatory checkpoint pathways. Studies of FL TME have demonstrated a favorable survival outcome in patients with higher tumor CD8 T cells or a profile enriched for T cell restricted genes (an immunologically “hot” tumor) compared to a patients without these features (a T cell “cold” tumor) and indicates a potential area for therapeutic intervention. (Sugimoto JClinExpHem2016; Dave NEJM2004)

• Hypothesis: Intratumoral (IT) G100 with low dose radiation can overcome immune suppression and cause immune mediated tumor shrinkage by transforming the TME from an immunologically "cold" to a "hot” one, thereby enabling effective presentation of endogenous neoantigens from apoptotic tumor and inducing systemic anti-tumor immunity. Furthermore, the addition of an anti-PD1 checkpoint inhibitor (pembrolizumab) would be synergistic by preventing the negative immune regulatory effects.

• Study Agent: G100 is glucopyranosyl lipid A (Figure 1A) formulated in a stable oil-in-water emulsion– Potent TLR4 agonist that activates the MyD88, TRIF/TRAM, and non-canonical inflammasome to stimulate

innate and adaptive immunity– TLR4 is an accessible cell surface receptor as opposed to TLR9 (CpG) receptors, which are endosomal– In preclinical studies, IT G100 activated DCs, T cells/NK cells, changed tumor-associated macrophages (TAM)

phenotype from M2 to M1, mediated leukocyte infiltration/inflammation and led to abscopal responses and increased survival. Radiation Treatment (RT) designed to release tumor antigens increased G100’s effects. In addition, B-cell lymphomas express TLR4 and exposure to G100 directly increased co-stimulatory molecules on these tumors (e.g. CD40, CD80, CD86) that enhanced their ability to present antigens and be effective immunologic targets (Figure 1B, Sagiv-Barfi, ASH 2015).

– In animal models of NHL, the combination of G100/anti-PD1 improved survival, whereas anti-PD1 inhibitors alone were not effective. • G100 was safe and well-tolerated: All AEs considered possibly related to

G100 were grade 1 or 2 and no related SAEs were observed.• The addition of Pembrolizumab to G100 did not result in unexpected or

worsening toxicity

G100 (n=13) G100 + P (n=13)Best Objective Response Rate (all lesions)^

CR, pts (%) 0 (0%) 0 (0%)

PR, pts (%) (-50% to < -100%) 2 (15%) 5 (38.5%)

SD Overall, pts (%)(+25% to < -50%) 9 (69%) 7 (54%)

MR, pts (%) (-25% to < -50%) 5 (39%) 4 (31%)

PD, pts (%) 2 (15%) 1 (8%)

Abscopal Tumor Reduction (10-100% -non-injected distal lesions)Overall, pts (%) 6 (46%) 9 (69%)

Range of shrinkage, % 19% to 61% 10% to 83%

Shrinkage ≥10cm2, pts (%) 1 (8%) 5 (39%)

Progression Free Survival (PFS)Median PFS (range), mos 7.4 (1.7 to 11.5+) 11.1 (1.9 to 11.1+)Median duration of observation (range), mos 7.6 (4.9 – 11.5) 8.3 (5.8 – 14.3)

Data cut-off 04Nov2017; ^responses are best responses in all index lesions combined (injected and non injected) with or without follow-up confirmation based on irRC (Wolchok, et al., 2009)

Figure 3. Concordance Between Overall Response (All Lesions) and Abscopal (Untreated Lesions) Shrinkage

G100 (n=13) G100 + P (n= 13)Treatment

NaïveRelapsed / Refractory

Treatment Naive

Relapsed / Refractory

N= 6 7 8 5Best Objective Response Rate (All lesions)

CR, pts (%) 0 (0%) 0 (0%) 0 (0%) 0 (0%)PR, pts (%) (-50 to < -100%) 0 (0%) 2 (29%) 2 (25%) 3 (60%)SD, pts (%) (+25 to < -50%) 6 (100%) 3 (43%) 6 (75%) 1 (20%)PD, pts (%) 0 (0%) 2 (29%) 0 (0%) 1 (20%)

Abscopal Tumor Reduction (10-100%- Non injected lesions)Overall, pts (%) 4 (67%) 2 (29%) 6 (75%) 3 (60%)Range of shrinkage, % -19% to -49% -55% to -61% -10% to -83% -33% to -82%Shrinkage ≥10cm2, pts 0 1 3 2

Progression-Free SurvivalMedian PFS (range), mos)

NA(1.9 to 11.2+)

3.7(1.7 to 11.5+)

11.1(6.5 to 11.1+)

5.4(1.9 to 9.8+)

Table 4. Overall Response By Prior Treatment Status

GELF High Tumor Burden§ Failed R-Chemo <2y¶

G100 (n= 6) G100 + P (n= 9) G100 (n=5) G100 + P (n= 3)PR, pts (%) 0 (0%) 3 (33%) 1 (20%) 2 (67%)Abscopal reduction, pts (%) 3 (50%) 6 (67%) 3 (60%) 1 (33%)

Abscopal reduction >10cm2, pts 1 5 1 1

§GELF Criteria: Reflects tumor burden and identifies pts with poorer survival who may need to begin treatment (Brice, et al. JCO 1997)¶Failure within 2 years of R-Chemo defines a poorer survival outcome group and unmet need: Originally described in first-line R-CHOP but becoming recognized as a broader group (failure <2 years at any line of therapy) with an unmet need for new agents (Kahl et al., Blood 2016)

Table 5. Responses In Previously Treated High Risk Patients

• Phase 1 Dose Escalation (5,10, 20ug in n=9) data of IT G100/RT (G100) in FL patients (pts) was presented at ASCO ‘17.

• We now present a Randomized Phase 2 (n= 26) (Figure 2A) examining G100 (n=13) vs. G100 + pembrolizumab (P) (n=13) in treatment (Tx) naïve or relapsed/refractory (R/R) FL.

• Study Treatment: Local low dose RT (2Gy x2) followed by intratumoral/peritumoral 10µg/dose G100 injection into tumor mass(es) x6–9 by palpation or by image guidance (e.g. ultrasound) (Figure 2B). Pts randomized to receive pembrolizumab received 200mg IV q3wk beginning on D14 for up to 2 yrs, as tolerated. Pts were required to have one tumor site for IT injection and at least one additional untreated tumor site outside the radiation field evaluable for distal (abscopal) response.

• Endpoints: Safety, response of injected lesion and abscopal (distal) sites, PFS, exploratory biomarkers

Safety of G100 and G100 + PembrolizumabTable 2. Summary of Treatment Related TEAEs

• G100 induces systemic clinical responses in FL pts (Table 3) • Addition of pembrolizumab resulted in higher rate of clinical response, deeper

abscopal (untreated sites) reduction (Table 3, Figure 3) and prolonged progression free survival in FL pts (Figure 4)

• G100 + P combination induced systemic clinical responses in FL pts with significant medical need:− Patients with recurrent/refractory disease− Patients with high risk disease based on GELF criteria or disease

recurrence within 2 years after rituximab containing chemotherapy regimen

<50% ≥50% <60% ≥60% <70% ≥70%

0/9 (0%)

9/20 (45%)

0/11 (0%)

9/18 (50%)

0/12 (0%)

9/17 (53%)

OR

R

p= 0.018 p= 0.005 p=0.003

Baseline TumorTLR4 expression

PR/total pts

ORR in Pts with >50% Baseline Tumor TLR4 Expression

n # pts with PR % ORR

G100 13 5 39%G100+P 7 4 57%

PR ptSD pt (including MR) with abscopal shrinkage

• Addition of pembrolizumab to G100 demonstrated a trend to a higher number of CD8+ TILs following treatment.

Immune Infiltrate Responders Non-Responders Odds Ratio p-valueCD8 (ratio of post/pre) 1.61 0.9 30 0.032

CD8/CD4 (ratio of post/pre) 2.99 1.14 5.1 0.054

CD8/Foxp3 (ratio of post/pre) 3.16 0.88 34 0.06IHC analysis of all matched pre and post tumors from patients in the Phase 1 dose escalation and this randomized phase 2. [Group (#matched sets post/pre): CD8 (22); CD8/CD4 (17); CD8/FoxP3 (20)]

Table 6. Increased CD8+ TILs Are Associated With PR*

• Increase in CD8+ Post/Pre ratio significantly correlated with higher odds of having a PR.

G100 (n= 10) G100 + P (n= 11)

Figure 7. Association Of Baseline Tumor TLR4 Expression And PRs

Figure 5. G100 + P Treatment Resulted in Increased CD8+ Tumor Infiltrating Lymphocytes (TILs)

Figure 1. G100 Structure (A) and Mechanism of Action (B)

Day:01257142128354249 56180XRT

After EOS visit, patients enter

long-term follow-up with restaging

q8 wks

G100 x 6

Anti-PD-1 à q3wk

If tumor mass has not regressed completely

Optional: G100 x 3 B.

Figure 2. G142 Study Design (A) and Study Treatment (B)

Table 1. Patient CharacteristicsG100 (n=13) G100 + P (n=13)

Age: median (range) 64 (39-72) 58 (36-80)

Gender, M/F 10M/3F 10M/3F

ECOG, 0 / 1 10 / 3 11 / 2

Grade 1 / 23

Unknown/Missing

11 (85%)2 (15%)

9 (69%)3 (23%)1 (8%)

Stage At Entry

IIA/BIIIA/BIVA/B

Unknown/Missing

1552

0652

PriorTreatment

Naïve R / R

1-2>3

Median (Range) (for R/R)

6 (46%)7 (54%)

2 (15%)5 (39%)3 (1 to 5)

8 (62%)5 (38%)

2 (15%)3 (23%)3 (1 to 6)

Prior Auto-SCT 3 2

Growing Tumor at Entry (PI assessment) 6 (46%) 8 (62%)

G100 Doses, median (range) 8 (3 - 9) 8 (3 -9)

III. CLINICAL RESULTS IV. EXPLORATORY BIOMARKER RESULTS

Pts with at least 1 TEAE G100 (n=13) G100 + P (n=13)All 9 (69%) 11 (85%)

Grade 1 6 (46%) 6 (46%)

Grade 2 3 (23%) 4 (31%)

Grade 3 0 1 (8%)1

Grade 4 0 0

Grade 5 0 0

SAE 0 1 (8%)2

TEAE – treatment emergent adverse event1 colitis, adrenal insufficiency, hyponatremia, hypocalcemia (all related to pembrolizumab)2 adrenal insufficiency (related to pembrolizumab)Most common treatment related TEAEs by treatment arm:•G100: injection site conditions (39%), GI disorders (1 pt each nausea, vomiting, diarrhea)•G100+P: asthenia (31%), nausea (31%), diarrhea (23%), injection site conditions (23%), rash (23%)

Efficacy of G100 and G100 + Pembrolizumab

Figure 4. Progression-Free Survival

2-03

2-08

2-01

2-11

2-20

2-24

2-07

2-23

2-15

2-12

2-19

2-14

2-27

*Censored data point

G100 (n=13) 2-05

2-06

2-10

2-02

2-04

2-18

2-25

2-09

2-22

2-26

2-21

2-16

2-17

G100 + P (n=13)

Total Best % Change from Baseline of Sum of Product Diameters (SPD)Abscopal Best % Change from Baseline of SPD

G100 (n= 13)

All lesions

Abscopallesions

Clinical Conclusions• Intratumoral G100 with low dose radiation demonstrated significant

systemic clinical responses in FL patients. • Combination of G100 with pembrolizumab was well-tolerated

without new or unexpected toxicities. • Addition of pembrolizumab resulted in more PRs than G100 alone

(38.5%), deeper abscopal tumor shrinkage (≥10cm2), and a trend to a better PFS. Pembrolizumab monotherapy in relapsed/refractory FL showed 11% PR (Ding, ASH 2017).

• Clinical responses are observed in FL patients with high unmet medical need, including recurrent/refractory disease, treatment failure <2 yrs after rituximab containing chemotherapy and high risk pts based on GELF criteria.

Figure 6. Baseline Tumor TLR4 Expression And Tumor Response

Baseline tumor TLR4 expression was analyzed by IHC assay in patients from the Phase1 dose escalation and this randomized phase 2 (n=29). All patients with PR had baseline TLR4 expression at or above 50% of tumor cells (Figure 7). The four pts who had baseline tumor TLR4 expression >=50%, but no significant tumor shrinkage (orange color) were heavily pre-treated (2 post stem cell transplant) or had tumor masses >7cm2.

• A strong association between baseline tumor TLR4 expression by IHC and objective clinical response (irRC PR) was observed (Figure 7) indicating the potential of patient selection and enriching for responders.

Biomarker & Overall Conclusions• Addition of pembrolizumab to G100 increases the CD8 T-cell

infiltration of the tumor microenvironment (TME), as an indication of the pro-inflammatory activity of G100.

• Baseline tumor TLR4 expression may serve as an enrichment marker for clinical response.

• Systemic clinical responses supported by immune biomarker data, and association of baseline tumor TLR4 expression with clinical responses warrant further investigation of G100 in combination with pembrolizumab as a novel immunotherapy in pts with recurrent and refractory high risk indolent lymphomas.

Treatment naïve and recurrent/refractory FL pts

with injectable tumors

Local Radiation Therapy followed by G100N=13

Local Radiation Therapy followed by G100 and pembrolizumab (P)

N=13

A.

G100 + P (n= 13)

* R/R patients

PR

*Single baseline abscopal lesion was 4.5x3 cm

*

Progression-Free Survival (Months) Progression-Free Survival (Months)