Background:  Good markers for clinical progression are highly warranted. Programmed cell death 1 (PD-1) has recently been shown to be directly involved in T cell exhaustion during HIV infection and appears less expressed in slow progressors. We here explored the predictive properties of quantitative PD-1 expression in relation to the actual CD4 T cell loss rate over a 1-year period in patients with untreated chronic HIV infection. These results were compared to markers which previously have been best related to progression, namely expression of CD38 and HIV RNA.

Methods:  Cross-sectional study including 50 consecutive, healthy HIV-infected patients off ART; 43 had the required observation times >12 months. PD-1 (MFI) and CD38 (molecules/cell) were determined on various T cell subsets by fluorescence activated cell sorting (FACS) analyses in fresh and later in parallel cryopreserved samples at end of observation period. FACS acquisition was performed after rigorous adjustments of PD-1 and CD38 negative populations; CD38 was quantified by PE-labelled beads (Quantibrite). Rapid progressors were defined by CD4 loss rates < cohort median at -45.7/µL/year. Non-parametric methods were used to characterize and compare groups, whereas logistic regression analysis was used to estimate predictive strength.

Results:  PD-1 and CD38 densities in fresh blood were lower (p <0.001) in patients on ART (n = 14) and seronegative controls (n = 8). CD4 loss rates correlated significantly to current HIV RNA (R = –0.30), CD38 (R = –0.33), and PD-1 densities (R = –0.38) on CD8⁺ T cells, and best to D-CD38, i.e. the difference in CD38 between the PD-1⁺CD8⁺ and CD8⁺ subsets (R = –0.51). PD-1 was highest on the CD27⁺CD28^–CD8⁺ intermediately differentiated subset with best correlation to progression (R = –0.54) in rapid progressors. Logistic regression models from HIV RNA, CD38, and PD-1 predicting rapid progression included PD-1 as best independent variable in combination with D-CD38 or CD38. PD-1 did not correlate with any of the other candidate variables. Cryopreservation reduced the CD38⁺ and PD-1⁺ fractions but corresponding densities became more suppressed through a non-linear loss most pronounced in CD38^high/PD-1^high cells with loss of predictive power.

Conclusions:  PD-1 was best independent predictor for CD4 loss rates in fresh blood compared with CD38 and HIV RNA, but not so after cryopreservation.