# Levels of age‐related blood CD8 T cells binding a non‐self peptide/HLA epitope track with Alzheimer's disease status in HLA‐A2+ and HLA‐A2‐ cohorts: a T cell biomarker assay applicable to all patients

## TL;DR

A new T cell biomarker for Alzheimer's disease works in both HLA-A2 positive and negative patients, improving diagnostic accuracy.

## Contribution

A novel T cell biomarker assay for Alzheimer's disease that is applicable to all patients, regardless of HLA-A2 status.

## Key findings

- T cells binding to a non-self pHLA multimer show reduced levels in Alzheimer's patients across HLA subtypes.
- The assay achieves high diagnostic accuracy (AUC 0.878-0.912) with strong sensitivity and specificity.
- The biomarker's performance is comparable to APP-specific T cell assays in HLA-A2-positive individuals.

## Abstract

We recently developed a CD8 T cell‐induced mouse model that recapitulates definitive hallmarks of Alzheimer's disease (AD) and allowed prediction of previously unknown properties of human sporadic AD (DOI: 10.1073/pnas.2401420121). Identification of a peptide antigen to which the inducing T cells respond in this model (a non‐Aβ epitope on Amyloid Precursor Protein [APP]) allowed us to quantify analogous T cells in human brain and blood, using an APP471‐479/Human Leukocyte Antigen(HLA)‐A2 multimer (pHLA) and flow cytometry on blood. Blood levels of these T cells distinguished AD and related MCI from normal aging controls with high accuracy (AUC: 0.883‐0.892), but the assay is currently suitable only for HLA‐A2‐positive (about half of all) patients due to HLA subtype‐restricted binding.

CD8 T cell binding to a non‐self (ALIAPVHAV/HLA‐A0201) multimer correlated with APP multimer binding in HLA‐A2‐positive patients (0.42‐0.58x relative to APP/HLA‐A2; r = 0.81; P < 0.001). Surprisingly, the same non‐self multimer bound at similar levels to HLA‐A2‐negative patients’ Tcells. We thus quantified levels of age‐related, KLRG1+ CD8 T cells binding to this multimer in HLA‐A2‐positive (n = 79) and ‐negative (n = 44) subjects by flow cytometry from Alzheimer's disease (AD) and normal aging control cohorts, and assessed their correspondence to disease status.

As with the APP pHLA reagent, levels of T cells binding to the non‐self pHLA multimer were significantly diminished in AD irrespective of HLA subtype (P < 0.015 by 2‐sided T‐Test). In receiver operating characteristic (ROC) analysis of the HLA‐negative cohort, area under the curve (AUC) was 0.878 for AD (P < 0.00001) with 79.4% sensitivity and 98.6% specificity, nearly identical to the APP‐specific CD8 T cells in HLA‐A2‐positive individuals (AUC 0.88; 72.2% sensitivity, 97.2 specificity). Moreover, disease correspondence of the non‐self pHLA‐reactive T cells was improved in HLA‐A2‐positive patients as well (AUC 0.912).

Subtype‐independent pHLA binding by CD8 T cells strongly corresponds to AD status in HLA‐A2‐negative and HLA‐positive individuals, providing a working basis for expanding our AD biomarker assay to all patients. Ongoing work will focus on increasing the sensitivity of HLA‐independent multimer binding to further increase assay utility.