J. Lewis (1,2), J. F. Standing (1), A. S. Walker (3), N. Klein (1) and R. Callard (1,2)
(1) UCL Institute of Child Health, (2) UCL CoMPLEX, (3) MRC Clinical Trials Unit
Objectives: A two-compartment model of CD4 memory T-cell homeostasis has already provided a compelling explanation of the slow decline in CD4 count during chronic HIV infection, as a set-point modified by changing population dynamics[1]. We aimed to use a similar model to investigate recovery of CD4 count in HIV-infected children starting antiretroviral therapy (ART), and to understand mechanisms and determinants of long-term reconstitution.
Methods: We studied 9166 CD4 counts collected over 2 years in 914 HIV-infected children (median(IQR) age 5.6(2.3,8.9) years; pre-ART CD4 count 342(165,643) cells/uL) showing typical, asymptotic CD4 recovery following ART initiation[2]. Using NONMEM[3] and nonlinear mixed-effects models based on ordinary differential equations, we fitted fixed and random effects for thymic activity, stimulation of T-cells to divide and death of resting T-cells.
Results: Diagnostic plots and simulated datasets demonstrated the model’s fit to the data and good predictive properties. Examining estimated child-specific parameters we found that children with poor recovery had low thymic output, faster stimulation of T-cells to divide and increased death of resting cells (all p<0.002). Unexpectedly, children older at ART initiation (and therefore HIV-infected and ART-naïve for longer) had a progressively higher thymic output, relative to that expected for their age (p<0.0001). We found elevated expression of division and activation markers in children with faster transition of T-cells to the dividing pool and higher division markers in children with faster death of resting cells (all p<0.002). Finally, we found a poorer fit of the model where higher activation and division had been recorded (both p<0.0006), suggesting that qualitatively different population dynamics may be at work in patients with significant bystander activation of T-cells.
Conclusions: A homeostatic model of T-cell dynamics can be used to understand CD4 T-cell dynamics and recovery in HIV-infected children starting ART. Faster T-cell death, faster activation to divide and lower thymic output are all candidate causes for incomplete T-cell reconstitution in children. Children starting ART at older ages have higher thymic output, relative to expected level in a healthy child. We found correlations between random effects and markers of T-cell activation and division, suggesting that between-child differences in recovery are related to altered T-cell homeostasis.
References:
[1] Yates A, Stark J, Klein N, Antia R, Callard R (2007) Understanding the slow depletion of memory CD4+ T cells in HIV infection. PLoS Med. 4:e177.
[2] Picat M-Q, Lewis J, Musiime V, Prendergast A, Nathoo K, Kekitiinwa A, Nahirya Ntege P, Gibb DM, Thiebaut R, Walker AS, Klein N, Callard R and the ARROW Trial Team (2013) Predicting Patterns of Long-term CD4 Reconstitution in HIV-infected Children Starting Antiretroviral Therapy in Sub-Saharan Africa: an Observational Study. Under review by PLoS Med.
[3] Beal S, Sheiner L, Boeckmann A, Bauer R (1989-2009) NONMEM User’s Guides. Ellicott City, MD, USA.
Reference: PAGE 22 () Abstr 2785 [www.page-meeting.org/?abstract=2785]
Poster: New Modelling Approaches