Abstract Body

Background:

Immunological features predicting time-to-rebound during analytical treatment interruption (ATI) may differ between cohorts, and be impacted by whether or not therapeutic interventions were used. Here, we used CyTOF to identify shared and unique pre-ATI features of T cells associated with time-to-rebound in the non-interventional ACTG A5345 cohort and 3 interventional Danish cohorts.

Methods:

A 40-parameter CyTOF T cell phenotyping panel was applied on pre-ATI blood specimens from 33 chronic-treated and 11 acute-treated individuals from the non-interventional ACTG A5345 cohort, and 3 chronic-treated Danish cohorts receiving therapy prior to ATI: CLEAR (Panobinostat intervention, n=9), TEACH (TLR9 agonist intervention, n=9), and REDUC (therapeutic vaccination + romidepsin intervention, n=16). Clustering analysis was performed to identify T cell phenotypes associated with time-to-rebound upon ATI.

Results:

Using cluster-resolution optimization, nine and five clusters were identified from the A5345 and the 3 collective Danish cohorts, respectively. Cluster 9 from A5345 and Cluster 3 from TEACH were both significantly (p<0.01) positively associated with longer time-to-rebound. Cluster 9 consisted of memory T cells expressing high levels of the T resident memory (Trm) marker CD103, and high levels of the homeostatic proliferation marker CD127 and the Th17 marker CCR6. As Cluster 9 consisted of both CD4+ and CD8+ T cells, we split it into subclusters 9a (CD8+) and 9b (CD4+). Cluster 9a expressed high levels of the pro-survival factor BIRC5, and low levels of the integrin component CD29. By contrast, Cluster 9b expressed low levels of BIRC5 and high levels of the co-stimulatory molecule CD28. Cluster 3 from the Danish cohorts consisted of CD8+ T cells expressing high levels of BIRC5 and low levels of CD29. These cells also expressed high levels of CXCR4 and CCR7, suggesting homing to inflamed and lymphoid tissues. Our observation that CD8+ T cells from both A5345 Cluster 9a and Danish Cluster 3 expressed high levels of BIRC5 and low levels of CD29 suggest that phenotypic features predicting longer time-to-rebound can be shared among non-interventional and interventional ATI cohorts.

Conclusions:

CD8+ T cells expressing high levels of BIRC5 and low levels of CD29 predict longer time-to-rebound in both interventional and non-interventional ATI cohorts. These ‘pro-survival’ CD8+ T cells may be better able to survive during the early post-ART period and thereby slow down viral rebound upon ATI.