Although many studies have investigated the influence of HLA on the risk of Alzheimer's Disease (AD), there have been inconsistent results. Part of this problem has been attributed to limitations of a clinical assessment in the absence of histopathological confirmation or other quantitative assessments. This study employed a subset of the AD Sequencing Project, representing 2663 cases with histopathological confirmation of AD and confirmation of 881 cognitively normal cases. Two HLA allelic subtypes, DQB1*02:01 and DQB1*02:02, were associated with lower Braak staging, a measure of tauopathy in the brain. These HLA subtypes were also associated with a later age of onset. There was a lower occurrence of HLA-DQB1*02:01 and HLA-DQB1*02:02 in AD cases compared to cognitively normal cases. For all of the above results, replicative sets were confirmatory. The above results were also maintained for both HLA-DQB1*02:01 and HLA-DQB1*02:02 when removing the effect of APOE4 or APOE2. Interestingly, the HLA-DQB1*02 allele binds tau better than all other HLA-DQB1 alleles tested, per an in silico assessment, raising the question of whether deletion of tau binding, auto-reactive T-cells in the thymus could reduce the likelihood of the onset of AD?