Characterization of Viral Epitopes and the HLA Restriction That Govern Anti-Adenoviral Response to Viral Specific T-Lymphocyte Therapy in a Pediatric Cohort

Abstract

<h3>Introduction</h3> Adenoviral (AdV) infections are common after HSCT and are a major cause of complications. Antiviral medications are frequently ineffective and toxic. At CCHMC, viral specific T cells (VST) from 3^rd party donors are generated by pulsing peripheral blood mononuclear cells (PBMCs) with viral peptide pools (pepmix). The choice of VST product for each patient is based on pre-clinical evidence of anti-AdV activity and the degree of HLA match as the viral epitopes and HLA restriction driving this response remain poorly understood. <h3>Objective</h3> The focus of this study is to characterize the HLA restriction of the AdV response in our VST bank. <h3>Methods</h3> Epitope prediction for AdV hexon and penton was done with free software (HLA restrictor 1.2 and IEDB) in conjunction with published data. 20 9mer to 15mer peptides were synthesized. VSTs were pulsed with peptides based on the HLA haplotype of the product and activity was evaluated using FACS for intracellular interferon gamma (IFNg). Post-infusion analysis was performed using peptide pulsed PBMC followed by IFNg ELISpot. HLA restriction was confirmed using co-culture of VST with a set of mouse fibroblasts expressing a single HLA allele (SALs) loaded with AdV pepmix. <h3>Results</h3> We have characterized 93 VST products and identified that 60/93 (65%) products have AdV activity <i>in vitro</i>. 20 patients have received VST for AdV with a 75% response rate in evaluable patients. A response against at least one AdV peptide has been demonstrated in 12/13 assessed products (4 of which have been infused into patients), and a polyclonal response was seen in 6/12 products (Fig 1). PBMC from 4 VST recipients were probed with the same AdV peptides that had stimulated a response in their respective VST products. In 3/4 recipients there was a response to the same peptides post-infusion as observed in the VST product, confirming <i>in vivo</i> activity (Fig 2). To discern HLA restriction, AdV pepmix was loaded onto class II HLA SALs and then co-cultured with 4 VST products expressing those same class II alleles. Anti-ADV activity was identified for specific class II HLA for all products, confirming the antiviral response in each product is restricted differentially for distinct HLA alleles. In retrospective analysis of one case, a patient with poor clinical response was found to have received a product restricted largely through DRB1*03:01, which was not a site of HLA match, rather than at the matched site of DRB1*04:01 (Fig 3). <h3>Conclusions</h3> Robust methods can be used to characterize the viral epitopes driving anti-AdV activity of VST products and the subsequent HLA restriction of that response. Response is frequently polyclonal and to some degree restricted through HLA class II antigens. While our sample is small, we propose that prospective adoption of these methods, especially the use of SALs as an initial screen, may allow for more precise VST product selection with improved response rates.