Introduction Peptide-based cancer vaccines require presentation by a specific HLA molecule. HLA-A2 is the most common class I allele in the US (40-50% of the population) and, therefore, the most commonly targeted. Our group has been investigating HER2-derived peptide vaccines administered in the adjuvant setting to high risk breast cancer patients in order to prevent disease recurrence. This study was undertaken to compare known prognostic factors and disease-free survival (DFS) in control HLA-A2+ and HLA-A2- patients in order to better define these populations for future trial design. Methods Our group is currently enrolling patients in a phase II trial evaluating the HER2-derived peptide vaccines, AE37 (MHC Class II, HLA-non-restricted epitope) and GP2 (MHC Class I, HLA-A2+ restricted epitope). The studies are enrolling high-risk, disease-free breast cancer patients with any level of HER2 expression (IHC 1+, 2+ or 3+) after completion of standard of care therapy. Patients are HLA-typed. HLA-A2+ patients are randomized to GP2+GM-CSF or GM-CSF alone. HLA-A2- patients are randomized to AE37+GM-CSF or GM-CSF alone. Demographics between groups are compared using chi squared or fisher exact as appropriate. DFS is compared using log rank. Results Thus far, 407 patients have been enrolled to the study (181 HLA-A2+ and 226 HLA-A2-). Demographics are shown in Table 1. Table 1. Demographics (all) A2+A2-pn181226 Age (median)51500.47Node Positive62%66%0.46Grade 354%54%0.99Tumor >/ = 2 cm59%64%0.23ER/PR Negative36%38%0.59HER2 Overexpression55%51%0.39Triple Negative13%15%0.52 There are no differences between groups with respect to age, node positivity, grade, tumor size, ER/PR status, HER2 over-expression, or triple negative breast cancer. Of those enrolled, 83 HLA-A2+ patients and 109 HLA-A2- patients have been randomized to the control groups. Within the control group, there are no differences between the HLA-A2+ and HLA-A2- patients regarding age, node positivity, grade, tumor size, ER/PR status, HER2 over-expression, or triple negative breast cancer (Table 2). With a median follow-up of 30 months, DFS is similar between A2+ and A2- control patients (83% v. 80%, p = 0.93). Conclusions Baseline clinico-pathologic factors are similar between HLA-A2+ and HLA-A2- breast cancer patients with no correlations to known prognostic factors. Well-matched blinded control patients treated only with GM-CSF demonstrate no differences in DFS between HLA-A2+ and HLA-A2- patients. Therefore, it does not appear that HLA-A2 status is a prognostic factor in breast cancer, and HLA-A2+ and HLA-A2- patients should be comparable in peptide-based breast cancer vaccine trials. Table 2. Demographics (Control Group Only) A2+A2-pn83109 Age (median)51510.75Node Positive65%64%0.99Grade 359%57%0.76Tumor >/ =59%71%0.08ER/PR Negative36%38%0.84HER2 Overexpression55%46%0.19Triple Negative11%16%0.34 Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-13-05.

Introduction: We are conducting a phase II clinical trial of the GP2+GM-CSF vaccine for the prevention of breast cancer recurrence in disease-free, node-positive or high-risk node negative patients (pts). GP2 is an HLA-A2+-restricted immunogenic peptide from the HER2 protein (aa: 654-662) capable of stimulating CD8+ cytotoxic T-lymphocytes (CTLs) to recognize and destroy HER2-expressing tumor cells. Here, we present the final pre-specified analysis of our study, disease free survival (DFS) at one year from last enrolled pt. Methods: The trial is a prospective, randomized, multi-center, placebo-controlled, single-blinded, phase I/II trial designed to evaluate the safety and clinical efficacy of GP2 in breast cancer patients. Pts with any level of HER2 (immunohistochemistry [IHC] 1-3+) expression were enrolled after standard of care therapy. Blinded HLA-A2+ pts were randomized 1:1 to receive GP2 + GM-CSF or GM-CSF alone. Pts in both treatment arms received 6 monthly intradermal inoculations during the primary vaccine series (PVS) followed by 4 booster inoculations administered every 6 months. Data was collected for demographic, safety, immunologic, and clinical recurrence. Per the statistical plan, DFS was analyzed for both intention-to-treat population (ITT; all randomized pts) and per-treatment population (PT; excluding recurrences during the PVS and second malignancies) using Kaplan-Meir methods. Continuous variables are compared using analysis of variance techniques and proportions compared with Fisher's exact test. Results: This trial enrolled 180 pts (vaccine group n=89 and control group n=91). There were no differences in age, grade, hormone receptor status, tumor size or nodal status between groups (all p>0.2). With median follow up of 36.2 months, the five-year DFS estimates did not demonstrate a difference between vaccinated (VG) and control groups (CG) with ITT (82.6% v 80.4%, p=0.96, respectively) or PT (88.9% v 84.3%, p=0.52, respectively). On further subset analysis, in HER2 3+ by IHC (or positive via FISH) pts, ITT analysis showed a DFS of 93.8% v 87.2% (VG, n=50 v CG, n=51; p=0.67, respectively) while the PT analysis revealed a DFS of 100% v 87.2% (VG, n=48 v CG, n=50; p=0.052, respectively). In ER/PR+ pts, ITT analysis showed a DFS of 88.1% v 80.6% (VG, n=54 v CG, n=60; p=0.41, respectively) while PT analysis demonstrated a DFS of 91.5% v 85.2% (VG, n=52 v CG, n=57; p=0.42, respectively). Conclusions: The final pre-specified analysis of this randomized phase II trial of the GP2+GM-CSF adjuvant breast cancer vaccine shows no statistical differences between treatment arms. However, it does identify a particular patient population where the vaccine may have efficacy; HER2 3+ patients appear to derive the greatest clinical benefit from GP2 vaccination after trastuzumab treatment with the PT analysis nearing statistical significance. Further studies are required to confirm these findings and investigate the potential link between hormone receptor status and the induction of clinically beneficial anti-tumor immunity with this vaccine. Citation Format: Greene JM, Schneble EJ, Perez S, Murray JL, Berry JS, Trappey AF, Hale DF, Vreeland TJ, Clifton GT, Ardavanis A, Litton JK, Shumway NM, Papamichail M, Peoples GE, Mittendorf EA. Final pre-specified analysis of the phase II trial of the GP2+GM-CSF peptide vaccine in high risk breast cancer patients to prevent recurrence. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P2-11-01.

Background: HER2 over-expression is associated with more aggressive malignant disease. The introduction of trastuzumab and other HER2-directed therapies, however, has led to improved prognosis for patients (pts) with HER2 over-expressing (OE) tumors. Currently, no HER2-targeted therapies are available for patients with HER2 low-expressing (LE) (1+, 2+ by IHC) tumors. We are conducting a randomized, controlled Phase II trial of multiple peptide vaccines enrolling patients with any level of HER2 expression (1+, 2+ and 3+). Here, we report survival data based on levels of HER2 expression in our unvaccinated, control pts. Methods: After standard of care therapy, disease-free, high-risk BCa pts were randomized to receive either peptide+GM-CSF (Vaccine Group, VG) or GM-CSF alone (Control Group, CG) in six, monthly doses followed by four boosters every six months. Pts were prospectively followed for recurrence. Demographic information was available for all pts and was compared between groups using chi square or fisher exact tests. Disease-Free Survival (DFS) was compared using log rank. Results: To date, we have enrolled 196 pts in the CG. 96 pts had HER2 OE tumors, 100 had LE tumors. The only significant demographic difference between the CG OE and LE groups was more ER/PR positive patients in LE (LE 72% vs OE 51%, p = 0.008). 83% of CG OE pts received trastuzumab, 3% of CG LE pts received trastuzumab. At a median f/u of 30 mo, DFS was significantly higher for CG OE vs CG LE (92.5% v 65.5%, p = 0.001). Conclusions: In the cohort of control pts from our ongoing vaccine trial, conducted in an era when Tz has been standard of care therapy for patients with HER2 OE tumors, we have shown that HER2 LE pts are at higher risk of recurrence than OE pts, despite having more ER/PR positive. This calls for increased efforts to develop novel therapies for patients with HER2 LE disease. We have previously shown a trend towards increased DFS with the HER2 vaccines, AE37 (p = 0.13, median f/u 22 mo) and E75 (p = 0.16, median f/u 60mo) in HER2 LE pts, suggesting that these vaccines may represent one such novel therapeutic approach. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P2-14-01.

BACKGROUND: HER2 is a commonly expressed tumor-associated antigen in breast (BrCa) and ovarian cancer (OvCa) and, therefore, an attractive target for immunotherapy. We have investigated HER2-derived peptides as vaccines mixed with GM-CSF to include GP2 (a HLA-A2 and HLA-A3 restricted, CD8+ eliciting epitope) and AE37 (a HLA unrestricted, MHC class II, CD4+ eliciting epitope). Both peptide vaccines (PV) have shown clinical promise individually. There is clear rationale for combining GP2 and AE37 to elicit a more robust immune response (IR) of both CD4+ and CD8+ T cells. Here, we summarize initial toxicity (tox) and in vivo IR data from a phase 1 trial of the combined PV. METHODS: The trial is being performed as a five cohort, 3+3 dose-escalation, safety trial. Clinically disease-free, HLA-A2+ and A3+, BrCa and OvCa patients with tumors expressing any level of HER2 (IHC 1-3+) and who have completed standard-of-care therapy are accrued. In the first cohort, three patients received six, monthly intradermal inoculations (R1-R6) of 100mcg of AE37, 100mcg of GP2, and 125mcg of GM-CSF or 100:100:125. The second cohort received 250mcg of AE37, 100mcg of GP2, and 125mcg of GM-CSF or 250:100:125. Three additional cohorts were vaccinated: 250:250:125, 500:250:125, and 500:500:125. Toxicity was graded 48-72 hours post vaccination using NCI Toxicity Criteria v4.0. After each inoculation, local reactions (LR) are measured via the sensitive ballpoint pen method and reported as the orthogonal mean (OM). IR is assessed in vivo by delayed type hypersensitivity (DTH) reactions with separate intradermal inoculations of AE37, AE36, and GP2 antigens, measured both pre-vaccination (R0) and after the vaccine series (R6) via the sensitive ballpoint pen method, and reported as the OM. Means were compared using paired t-tests. RESULTS: 28 patients enrolled; 8 withdrew consent, 1 recurred prior to completing R6, 3 had an intercurrent illness, 14 patients completed R1-R6, and the vaccine series is ongoing in 2 patients. Six patients did not receive any inoculations and, therefore, are not included in this safety analysis. In 22 patients, the vaccine was well tolerated (max local tox: 23% Grade (Gr) 1, 73% Gr 2, 4% Gr 3; max systemic tox: 14% Gr 0, 50% Gr 1, 36% Gr 2). No dose-limiting toxicity was observed. For the 14 patients who completed the VS, the median age was 51(35-83). Breast tumor size was 3.3±1.1cm and ovarian tumor size was 10.0±2.3cm. Compared to GP2 LR at R1 (15.5±4.1mm), LR increased at R2 (31.7±5.9mm), R3 (42.9±7.4mm), R4 (35.3±7.3mm), R5 (45.0±9.9mm), and R6 (25.9±6.7mm, p = 0.17). Compared to the AE37 LR at R1 (18.5±3.8mm), LR increased at R2 (37.3±6.7mm), R3 (36.4±4.6mm), R4 (42.2±5.9mm), R5 (46.0±8.9mm), and R6 (36.2±6.6mm). Unless stated, all LR p-values < 0.05. After the VS, AE37 DTH increased from 0.0±0.0mm to 19.6±6.7mm (p<0.01), AE36 DTH increased from 0.0±0.0mm to 10.3±3.9mm (p<0.01), and GP2 DTH reactions increased from 0.3±0.2mm to 4.1±2.0mm (p = 0.056). CONCLUSIONS: Initial results from a phase I trial of a vaccine combining GP2 and AE37 peptides show that dual administration of the peptides is well tolerated at all tested dosing levels. Additionally, the combination is capable of stimulating strong peptide-specific in vivo immune responses. Continued testing of this vaccination strategy is underway. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P4-13-02.

BACKGROUND FBP is overexpressed in 20-50% of breast(B) cancers(Ca) and roughly 90% of endometrial(E) and ovarian (Ov) Ca. E39 (FBP191-199, EIWTHSYKV)+GM-CSF is an HLA-A2 restricted FBP peptide vaccine, which has been shown to generate significant in vivo immunologic response(IR) in a phase I/IIa trial in E Ca and Ov Ca patients (pts). There is a risk of inducing immunologic tolerance after multiple inoculations with a highly immunogenic vaccine. Thus, we are investigating a novel vaccination series using combinations of E39 and E39' (EIWTFSTKV, an attenuated version of E39) in a phase Ib, randomized, single-center trial. We are assessing short and long-term IR. Here, we present the initial IR analysis to the primary vaccination series (PVS) within B Ca pts. METHODS HLA-A2 positive B or Ov Ca pts were enrolled after completion of standard of care therapy and randomized into three arms: EE (6 inoculations of E39); EE'(3 inoculations of E39, then 3 of E39'); or E'E(3 of E39', then 3 of E39). Theoretically, due to lower FBP expression and less aggressive chemotherapy regimens, B Ca pts are more antigen naïve and have a less suppressed immune system. Thus, only B Ca pts were included in this analysis. The PVS includes 6 inoculations total (R1-R6), one every 3-4 weeks, and containing 250mcg GM-CSF+500mcg peptide in the first 5 pts per arm and 1000mcg of peptide in second 5 pts. To assess the in vivo IR, local reaction(LR) was measured 48 hours after each inoculation (R1-R6), and delayed type hypersensitivity(DTH) was measured pre-PVS (R0), 1, and 6-months post-PVS (RC1, RC6). Ex vivo IR was measured via dextramer assay for E39-specific CD8+ T-cells at R0, RC1, and RC6. Statistical analyses were completed using appropriate tests. RESULTS Thirty-five B Ca pts were enrolled, with 27 completing the PVS (EE n=10, EE' n=8, E'E n=9). No clinicopathologic differences between groups or significant toxicities > grade 2 were appreciated. LR increased from R1 to R6 in all groups (ΔEE= 24.80mm, p=0.14; ΔEE'=38.13mm, p=0.07; ΔE'E=8.05mm, p=0.38), the greatest increase approaching statistical significance in the EE' arm. The only arm with a statistically significant increase for in vivo DTH from R0-RC1-RC6 was in the EE' arm (ΔEE=-6.17mm, p=0.27; ΔEE'= 44.58mm, p<0.05; ΔE'E=-1.42, p=0.37). Ex vivo analysis of IR revealed no significant difference between groups at R0(p=0.45) or RC6(p=0.72), nor within groups over time (EE p=0.32, EE' p=0.47, E'E p=0.30). CONCLUSION In this phase Ib trial analyzing the IR of B Ca pts receiving a different vaccination strategy, both peptides were noted to be safe and immunogenic. While no difference was seen in E39-specific CD8+ T cells between groups, the in vivo response was enhanced with the use of E39' after E39; this may indicate expansion of more effective clonal populations of CD8+ T cells with this strategy. These results may be specific to B Ca pts who are relatively antigen-naïve with relatively intact immune systems. Further analysis of these pts as this trial continues will determine the optimal vaccination strategy capable of stimulating and maintaining an IR to prevent B Ca recurrence. Citation Format: Jackson DO, Qiao N, Peace KM, Hale DF, Vreeland TJ, Greene JM, Berry JS, Trappey AF, Clifton GT, Ibrahim N, Toms A, Peoples GE, Mittendorf EA. Determining the optimal vaccination strategy using a combination of the folate binding protein (FBP) peptide vaccine (E39+GM-CSF) and an attenuated version (E39') to maximize the immunologic response in breast cancer patients [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-10-04.