Key Points Tumor IDO inhibits CD19-CART activity, likely via induction of the kynurenine pathway, whose metabolites directly inhibit T cells. Fludarabine and cyclophosphamide, frequently used before CART administration, downregulate IDO expression in lymphoma cells.

Second-generation (2G) chimeric antigen receptors (CARs) targeting CD19 are highly active against B cell malignancies, but it is unknown whether any of the costimulatory domains incorporated in the CAR have superior activity to others. Because CD28 and 4-1BB signaling activate different pathways, combining them in a single third-generation (3G) CAR may overcome the limitations of each individual costimulatory domain. We designed a clinical trial in which two autologous CD19-specific CAR-transduced T cell products (CD19.CARTs), 2G (with CD28 only) and 3G (CD28 and 4-1BB), were infused simultaneously in 16 patients with relapsed or refractory non-Hodgkin's lymphoma. 3G CD19.CARTs had superior expansion and longer persistence than 2G CD19.CARTs. This difference was most striking in the five patients with low disease burden and few circulating normal B cells, in whom 2G CD19.CARTs had limited expansion and persistence and correspondingly reduced area under the curve. Of the 11 patients with measurable disease, three achieved complete responses and three had partial responses. Cytokine release syndrome occurred in six patients but was mild, and no patient required anti-IL-6 therapy. Hence, 3G CD19.CARTs combining 4-1BB with CD28 produce superior CART expansion and may be of particular value when treating low disease burden in patients whose normal B cells are depleted by prior therapy.

Vaccines prevent HPV-associated cancer but, although these tumors express foreign, viral antigens (E6 and E7 proteins), they have little benefit in established malignancies, likely due to negative environmental cues that block tumor recognition and induce T cell anergy in vivo. We postulated that we could identify mechanisms by which ex vivo stimulation of T cells could reactivate and expand tumor-directed T-cell lines from HPV-positive cancer patients for subsequent adoptive immunotherapy. A total of 68 patients with HPV-associated cancers were studied. Peripheral blood T cells were stimulated with monocyte-derived dendritic cells loaded with pepmixes (peptide libraries of 15-mers overlapping by 11 amino-acids) spanning E6/E7, in the presence or absence of specific accessory cytokines. The resulting T-cell lines were further expanded with pepmix-loaded activated B-cell blasts. IFNγ release and cytotoxic responses to E6/E7 were assessed. We successfully reactivated and expanded (>1200-fold) E6/E7-specific T cells from 8/16 cervical and 33/52 oropharyngeal cancer patients. The presence of the cytokines IL-6, -7, -12 and -15 is critical for this process. These T cell lines possess the desirable characteristics of polyclonality, multiple T-cell subset representation (including the memory compartment) and a TH1 bias, and may eliminate E6/E7-positive targets. In conclusion, we have shown it is possible to robustly generate HPV16 E6/E7-directed T-cell lines from patients with HPV16-associated cancers. Because our technique is scalable and good-manufacturing-procedures compliant, these lines could be used for adoptive cellular immunotherapy of patients with HPV16-positive cancers.

1913 Vaccines effectively prevent HPV-associated cancers, but their use as therapy for established neoplasms has been disappointing. Although the target tumors express the viral E6 and E7 antigens, patients' immune responses against virally infected cells are limited, and active immunization fails to generate effective immunity to these antigens. Ineffective endogenous immunity to E6/E7 is likely due to negative environmental cues that block initial tumor cell recognition and subsequent T cell activation, expansion or persistence in vivo. We postulated that ex vivo stimulation of patient T cells in an immunologically favorable milieu would allow us to reactivate sufficient numbers of active tumor-directed CTLs for adoptive transfer to patients with HPV-associated tumors. We studied 67 patients with HPV-associated neoplasms (16 with cervical and 51 with oropharyngeal carcinoma). To investigate the presence of HPV16 E6- and E7-specific T cells in peripheral blood, we performed a γ-IFN ELISpot assay targeting those antigens, measuring both baseline immunity detectable in peripheral blood mononuclear cells (PBMCs) and the responses of T cells stimulated for 9 days by monocyte-derived dendritic cells (DCs) loaded with pepmixes (peptide libraries of 15-mers overlapping by 11 aa) spanning E6 and E7. We anticipated finding HPV-reactive T cells in a significant proportion of patients with oropharyngeal cancer, given that approximately 75% of these tumors are associated with HPV16, and a lesser rate of responsiveness in subjects with cervical neoplasia, since only half of subjects have HPV16-positive tumors. However, we found no PBMC or T cell reactivity against E6 or E7 under the tested conditions. Because HPV-specific T cells from patients with HPV-associated cancers may be anergized by their tumors, we therefore postulated that potent antigen presenting strategies might be required to reactivate them. We have previously shown that in vitro stimulation in the presence of the accessory cytokines IL-6, -7, -12 and -15 is able to reactivate T cell responses even to poorly immunogenic non-viral tumor antigens. Indeed, in combination with these cytokines, DCs successfully reactivated and expanded HPV16 E6- and/or E7-specific T cells detectable by ELISpot in 8/16 cervical and 32/51 oropharyngeal carcinoma patients. To develop a technology that would allow us to consistently produce and expand these HPV-specific T cells to the number needed for clinical use, given that sufficient DCs may be difficult to obtain when sourced from the peripheral blood of patients with cancer, we prepared substitute antigen presenting cells, B-blasts (BBs), derived from the subjects' autologous B cells. These BBs were generated by culturing patient PBMCs in the presence of IL-4 and cyclosporine on a CD40 ligand-positive fibroblast feeder layer. Stimulation of DC-stimulated HPV-specific T cells by E6/E7 pepmix-loaded BBs allowed further expansion of HPV-specific T cells (3.8 ± 1.5 fold per round of stimulation), with successive increase in antigen specificity. Analysis of the Vβ-chain T cell receptor (TCR) repertoire established that the cell lines were polyclonal. Phenotypic analysis of the cell lines showed them to be almost exclusively composed of T cells (97.5 ± 3.4% CD3+), with a variable proportion of CD4+ and CD8+ cells (36.7 ± 28.2% and 49.4 ± 27.0%, respectively) displaying a predominantly effector memory phenotype (CD45RA–, CD45RO+, CD62L– and CCR7–); and minimal CD3– CD56+ NK cells (1.5 ± 0.9%). ELISpot assays using mini-pools of peptides from each antigen allowed identification of the epitopes recognized, which mapped to aa 49–71, 77–91 and 125–143 for E6, and aa 1–19 and 73–87 for E7. Cytotoxicity assays using these HPV-specific cell lines as effectors (E) against autologous pepmix-loaded PHA blasts as targets (T) demonstrated for several of these lines dose-dependent specific killing of antigen-loaded targets (specific lysis range 45–61% versus 0–8% in controls at 40:1 E:T ratio). Hence, these cells are true CTLs. In summary, we have developed a system that for the first time allows the robust generation of HPV-directed CTLs from the peripheral blood of patients with an HPV16-associated cancer, and shown that these lines recognize specific epitopes in tumor-associated molecules. Our cell lines have the potential to be used for adoptive cellular immunotherapy of HPV-associated carcinomas. Disclosures: No relevant conflicts of interest to declare.

Vaccines prevent HPV-associated cancer (Ca), but their use as therapy for established Ca has been disappointing. Although the target tumor cells express the viral E6 and E7 antigens (Ag), patients' immune responses against virally infected cells are limited, even after active immunization, likely due to negative environmental cues that block initial tumor cell recognition and subsequent T cell (TC) activation and expansion in vivo. We postulated that ex vivo stimulation of patient TCs in an immunologically favorable milieu would allow us to reactivate tumor-directed CTLs.We studied 67 patients with HPV+ Ca (16 cervical and 51 oropharyngeal, OPCa). To investigate the presence of HPV16 E6-and E7specific TCs (HPV-TCs) in blood, we measured the g-IFN ELISpot responses of TCs stimulated by monocyte-derived dendritic cells (DCs) loaded with pepmixes (peptide libraries of 15-mers overlapping by 11 aa) spanning E6 and E7. Although $75% of OPCa are HPV16+, we initially found no evidence of E6/E7-reactive T cells in the patients tested. Because HPV-TCs from these patients may be anergized by their tumors, we postulated that potent Ag presenting strategies might be required for reactivation. In other studies, we have found that in vitro stimulation of T cells in the presence of DCs and IL-6, -7, -12 and -15 can induce responses to poorly immunogenic Ag. We therefore tested this approach in patients with HPV+ Ca, and found we could successfully reactivate HPV-TCs in 8/16 cervical and 32/51 OPCa patients.Given it is difficult to obtain large numbers of DCs, we expanded these HPV-TCs to clinically useful numbers by substituting patient B-cell blasts (BBs) as APCs, which we made by culturing autologous PBMCs with IL-4 on a CD40L+ feeder layer. Stimulation of DCstimulated HPV-TCs by E6/E7 pepmix-loaded BBs further expanded HPV-TC lines (3.8 6 1.5Â/round), whose phenotype is summarized in the Table. The epitopes recognized by the HPV-TCs mapped to E6 aa 49-71, 77-91 and 125-143, and E7 aa 1-19 and 73-87. These cells achieved dose-dependent specific killing of E6/E7+ target cells (specific lysis up to 45-61% vs. 0-8% in controls, 40:1 E:T ratio). Thus, we have generated true CTLs. Table. Phenotypic analysis of HPV-specific cell lines Marker % ± SD CD3 97.5 ± 3.4 CD4 36.7 ± 28.2 CD8 49.4 ± 27.0 CD56 (CD3 negative)1.5 ± 0.9All cell lines are almost exclusively composed of T cells, with a variable proportion of CD4 + and CD8 + cells, displaying a predominantly effector memory phenotype (CD45RA -, CD45RO + , CD62Land CCR7 -). There are minimal CD3 -NK cells. Vb-chain TCR repertoire analysis established polyclonality.In summary, we have developed a system that allows the robust generation of HPV-directed CTLs from the blood of patients with HPV16+ Ca, and shown that they recognize specific epitopes in tumor-associated Ag. Our lines have the potential to be used for adoptive cellular immunotherapy of HPV+ Ca. Objective: In order to improve the graft versus leukemia (GVL) effect of DLI, we investigated the efficacy and safety of combinin...

2558 Background: Vaccines prevent HPV-associated Ca, but their benefits in established Ca are disappointing: although these tumors express viral E6 and E7 antigens (Ags) immune responses are limited even after vaccination, likely due to negative environmental cues that block tumor recognition and T cell (TC) activation in vivo. We postulated that ex vivo stimulation of TCs in an immunologically favorable milieu would allow us to reactivate tumor-directed CTLs from Ca patients and benefit the recipients on reinfusion. Methods: We studied 68 patients with HPV+ Ca. To detect HPV16 E6/E7-specific TCs (HPV-TCs) in blood, we measured the IFNγ ELISpot responses of TCs stimulated by monocyte-derived dendritic cells (DCs) loaded with pepmixes (peptide libraries of 15-mers overlapping by 11 aa) spanning E6/E7. Because HPV-TCs from these patients may be anergized by their tumors, potent Ag presenting strategies might be required for reactivation, and thus we stimulated these cells in the presence of IL-6, -7, -12 and -15, as we have shown that these can induce responses to poorly immunogenic Ags. Results: We successfully reactivated HPV-TCs from 8/16 cervical and 33/52 oropharyngeal Ca patients. We could expand these HPV-TCs to clinically useful numbers by using patient B-cell blasts (BBs) as APCs. Stimulation of DC-stimulated HPV-TCs by E6/E7 pepmix-loaded BBs further expanded (3.8 ± 1.5×/round) HPV-TC lines, which phenotypically are almost exclusively composed of TCs (98 ± 3% CD3+), with a variable proportion of mostly effector memory (CD45RA–, CD45RO+, CD62L– and CCR7–) CD4+ and CD8+ cells (37 ± 28% and 49 ± 27%, respectively). The viral/tumor associated epitopes recognized mapped to E6 aa 49-71, 77-91 and 125-143, and E7 aa 1-19 and 73-87. The expanded cells from patients killed E6/E7+ targets (specific lysis up to 45-61% vs. 0-8% in controls, 40:1 E:T ratio). Conclusions: We have developed a system that allows the robust generation of HPV-directed CTLs from patients with HPV16+ Ca, which recognize specific epitopes in tumor-associated Ags. Our lines have the potential to be used for adoptive cellular immunotherapy of HPV+ Ca.