Long-term changes in circulating CD4 T lymphocytes in virologically suppressed patients after 6 years of highly active antiretroviral therapy

Highly active antiretroviral therapy (HAART) initiated in advanced HIV disease is associated with CD4 lymphocyte increases (200–300 cells/mm3after 2–4 years), although longer-term cellular dynamics have not been studied. We observed a significant median CD4 lymphocyte increase of 126 cells/mm3and 54 naive CD4 lymphocytes from year 3 to 6 of HAART among 20 individuals with pre-HAART CD4 cell counts of 100–300 cells/mm3. This cohort represents the longest prospective immunological follow-up of virologically suppressed patients on HAART. Untreated HIV infection leads to a progressive loss of peripheral CD4 lymphocytes, immunological impairment and an increased risk of opportunistic illnesses. Several studies have examined CD4 cell responses to potent antiretroviral therapy, and have demonstrated that prolonged successful viral suppression is associated with substantial increases in CD4 lymphocyte counts and clinical benefit [1]. After 2–4 years of antiretroviral therapy, mean (or median) CD4 increases are approximately 200–300 cells/mm3 [2–5]. However, the rate of the CD4 cell increase has been shown to diminish over time with little additional increase observed after 3–4 years of therapy [2–5]. AIDS Clinical Trials Group (ACTG) 315 was a study designed to evaluate the immunological effect of highly active antiretroviral therapy (HAART). HIV-infected individuals with CD4 cell counts between 100 and 300 cells/mm3 and a history of therapy with zidovudine for at least 3 months were enrolled in this three-center study. Patients with previous treatment with lamivudine or any HIV-1 protease inhibitor were excluded. After treatment with zidovudine, lamivudine and ritonavir for a minimum of 48 weeks, study participants rolled over into ACTG 375, in which changes in antiretroviral treatment could be made if plasma HIV-1-RNA levels were more than 100 copies/ml, and changes were required if plasma HIV-1-RNA levels were repeatedly more than 3000 copies/ml. Twenty patients with plasma HIV-1-RNA levels less than 1000 copies/ml at 3 years and for at least 75% of the determinations between week 12 and year 3 have been described in detail through 3 years of follow-up [4]. The same 20 patients continued in follow-up, and CD4 and CD8 lymphocyte counts and subsets are available through 6 years of HAART. Measurements were averaged within a window from week 266 to week 322 (one subject went off-study at 4.5 years, and this timepoint was carried forward in all the following analyses). None of the subjects had confirmed HIV-1-RNA levels greater than 1000 copies/ml between years 3 and 6. At year 6, 17 subjects remained on zidovudine, lamivudine and ritonavir, one subject was on zidovudine, lamivudine and indinavir, one subject was on zidovudine, lamivudine, ritonavir and saquinavir, and one subject was on stavudine, nevirapine, ritonavir and saquinavir. No subjects took IL-2 or other adjunctive immune therapies. Changes in lymphocyte subsets are presented as medians and are summarized in Fig. 1. The Wilcoxon signed-rank test was used to evaluate whether median changes from year 3 to year 6 were different from zero, and the Wilcoxon rank-sum test compared changes between subjects above versus below the median thymus thickness. There were no corrections for multiple comparisons. CD4 cell counts continued to increase from year 3 to year 6, and the median increase of 126 cells/mm3 was statistically significant (P < 0.001, 42 cells/mm3 per year). From baseline to year 3, CD4 cell counts increased an average of 71 cells/mm3 per year [4]. Notably, the median change in CD4 cell count from baseline to year 1 was +159, whereas the average yearly change from year 3 to year 6 was +42 cells/mm3. Surprisingly, 55% of subjects demonstrated decreases in CD4 cell counts (ranging from −20 to −178 cells) from year 5 to year 6. During the study period, no subject had a confirmed CD4 cell count above 900 cells/mm3; one subject had a confirmed CD4 cell count above 800 cells/mm3 and six subjects had confirmed CD4 cell counts above 700 cells/mm3.Fig. 1. Median CD4 and CD8 lymphocyte counts and subsets over time (top two panels) and median CD4 and CD8 activation percentages over time (bottom panel).: Each timepoint represents 20 subjects.The number of peripheral blood memory (CD45RO+/CD45RA−) CD4 lymphocytes increased significantly from baseline (127 cells/mm3) to year 3 (241 cells/mm3) and to year 6 (321 cells/mm3) (P < 0.001). In addition, the number of naive (CD45RA/62L) CD4 and CD8 T lymphocytes increased significantly (P < 0.001); the median year 3 to year 6 increases were 54 and 113 cells/mm3, respectively (Fig. 1). Unexpectedly, the percentages of activated (CD38/HLA-DR) CD4 and CD8 lymphocytes increased from year 3 to year 6, in contrast to the declines observed from baseline to year 3 (Fig. 1). This pattern was observed in the majority of subjects; 75 and 85% of subjects increased activated CD4 and CD8 cell percentages over this period, respectively. The thymus size was measured approximately one year after the initiation of antiretroviral therapy in 16 of the subjects in this study using methods described elsewhere [6]. We have previously described a correlation between thymus size as measured by chest computed tomography scan and the number and percentage of naive CD4 cells in these subjects after one year of antiretroviral therapy [7]. Comparing naive CD4 cell increases from baseline to year 6 among the seven subjects with year one thymus size (thickness) above the median (5 mm) to the nine at or below the median, a marginally significant difference was noted (median 204 versus 97 naive cells/mm3, P = 0.071, Wilcoxon test). No differences were seen in total CD4 cell increases (319 versus 306 cells/mm3). This cohort represents the longest prospective immunological follow-up of virologically suppressed patients on potent antiretroviral therapy. We have demonstrated sustained CD4 cell increases in subjects who had moderately advanced HIV disease at the start of therapy. It is notable, however, that despite improvements in absolute, naive and memory CD4 cell counts and percentages, the majority of subjects have CD4 cell subsets that are lower than seen among HIV-negative subjects. We have expanded upon our previous findings, and confirmed the findings of others regarding the importance of thymus size in naive CD4 cell restoration after HAART [8,9]. In addition, we have demonstrated that despite long-term viral suppression for more than 5 years, subjects have persistently elevated markers of immune activation, and the majority of subjects demonstrated small increases in immune activation in the final 3 years of treatment. The mechanism for this is unclear. It is possible that the increase in CD4 cells has provided new targets for viral replication. Low-level increases in viral replication or early viral escape could thus be driving heightened immune activation. We could not find a correlation between plasma HIV RNA and immune activation in these subjects; however, as most subjects had HIV levels below the level of detection of the assay used in this study, a more sensitive measure of viral replication may be needed to answer this question. We will address this and other questions in the follow-up study to ACTG 375, ACTG 5136, which will examine the effect of HAART intensification in these subjects. Sponsorship: This work was supported by the AIDS Clinical Trials Group and by grants AI 38855, AI 38858 and AI 25879 from the National Institutes of Health and by GCRC grant RR-0051 and Cancer Center grant CA-46934 at the University of Colorado Health Sciences Center. HIV-RNA testing was performed by Glaxo-Wellcome Inc., who also supplied zidovudine and lamivudine; Abbott Laboratories supplied ritonavir; Bristol–Myers Squibb supplied stavudine and didanosine; Hoffman La Roche supplied saquinavir; and Agouron Pharmaceuticals supplied delavirdine.