IMMUNE RECONSTITUTION IN HIV-1 INFECTED PATIENTS TREATED FOR TWO YEARS WITH HIGHLY ACTIVE ANTIRETROVIRAL THERAPY

The aim of this paper was to evaluate the immune reconstitution of HIV-1 patients subjected to highly active antiretroviral therapy (HAART) for two years or more according to CD45RA and CD45RO cell count; determination of IL-2, IFN-γ, IL-4, IL-10 and TNF-α serum levels; CD4 T and CD8 T lymphocyte count; and plasma viral load (VL) determination. For this purpose, a cross sectional study was carried out in the Tropical Diseases Area, Botucatu School of Medicine, São Paulo State University, UNESP, Botucatu, São Paulo, Brazil. Between June 2001 and April 2002, 37 HIV-1 infected patients were evaluated, 13 with treatment indication but untreated (G1), 9 subjected to HAART for 5-7 months (G2), and 15 treated for two years or more (G3); both treated groups used medication regularly and without failure. Forty-nine normal individuals were studied as controls (GC-1 and GC-2). There was a tendency (p<0.10) for the predominance of two nucleoside reverse transcriptase inhibitors (NRTI) associated with one non-nucleoside reverse transcriptase inhibitor (NNRTI) regimen in G2; and two NRTI associated with a protease inhibitor (PI) in G3. Statistical differences between groups were seen for CD45RA (G1<[G3=GC-2]; p<0.05) and CD45RO (G1[G2=G3]; p<0.001), TNF-α serum determination ([G1>G3; G2=intermediate]>GC-1; p<0.001), IL-2 (G1<[G2=G3=GC-1]; p<0.01), IFN-γ ([G1=GC1]<[GC-2=G3]; p<0.001), IL-4 and IL-10 ([G1=G2=G3]>GC-1; p<0.001), serum cytokine profiles, with a higher proportion of subtype 2 in G1 and mature subtype 0 in G2 and G3 (p<0.005). There was no statistical difference for CD8 T lymphocyte counts (G1=G2=G3; p<0.50). Consistency was seen between positive correlations of profile 1 definer cytokines (IL-2 and IFN-γ), CD45RA and CD45RO cells, and CD4 T lymphocyte counts and between positive correlations of profile 2 definer cytokines (IL-4 and IL-10) with TNF-α, and VL. The negative correlations were also consistent as they expressed the inverse of the positives. The variables with the highest number of correlations were IL-2, IFN-γ, and VL, followed by CD45RA and CD45RO cells, and IL-10. The variables with the lowest number of correlations were CD4 T and CD8 T lymphocytes. The results express the partial but important immune reconstitution in HIV-1 infected individuals with the interference of HAART and the importance of cytokines especially IL-2 and IFN-γ, and CD45RA and CD45RO cells as surrogate markers of this reconstitution.


INTRODUCTION
The introduction of highly active antiretroviral therapy (HAART) was a major milestone in the attempt to recuperate HIV-1 infected individuals' immune systems.Despite this recuperation being slow, partial, and variable, the suppression of HIV-1 replication has permitted a higher level of control over associated diseases with a consequent decrease in mortality and improvement in quality of life.(10) The main surrogate markers of natural history and therapeutic efficacy currently used in HIV infected patient follow-up are clinical condition, CD 4 + T lymphocyte count, and plasma viral load (VL) determination (21,23).However, after VL drops below detectable levels, it does not allow us to consistently express the behavior of immune response reconstitution.Also, CD 4 + T lymphocyte count can show irregular behavior without significant elevation even in individuals with major clinical improvement (4) and not demonstrate whether this elevation expresses the real immune system reconstitution through the production of naive cells, which have CD 45 RA surface molecules, or just the recirculation of memory cells, also known as CD 45 RO (10,13,20).In addition, CD4 + T lymphocyte count does not characterize the quality of the immunomodulatory substances they produce.

PARTICIPANTS
Between June 2001 and April 2002, 37 HIV-1 patients on or not on HAART were studied.They were treated at the Special Outpatient Clinic and Infirmary of Tropical Diseases, Botucatu School of Medicine, UNESP.Twenty-five were male and twelve female, aged between 18 and 62 years (X=37.6years).
The assay detection limit varied from 3 to 5 pg/ml, depending on the cytokine.

Statistical Analysis
Group comparisons were made with the non-parametric Kruskal-Wallis test and analysis of variance for entirely randomized experiments (ANOVA-ERE).The χ 2 test was used for serum cytokine profile proportion comparisons in the three HIV-1 groups, and for PI proportion comparisons in the two HIV-1 groups under treatment.When the proportion was zero, Yates correction was applied.For the χ 2 test, statistical significance was when p<0.05 and the value equal or higher than 3.84 for one degree of freedom, and equal or higher than 5.99 for two degrees of freedom.In all other analyses, statistical significance was p<0.05.For the correlations between pairs of variables, Spearman coefficient of rank correlation was calculated with significance α=0.05 and α=0.01, (30).
This study was approved by the Research Ethics Committee of Botucatu School of Medicine -UNESP.

RESULTS
Table 3  slightly higher than that of controls.The prevailing mode of transmission in all infected groups was heterosexual.
Groups were not homogeneous to ARV treatment regimens as they were grouped using this parameter.G1 patients had not yet been treated; six G2 patients received two NRTI associated with one NNRTI, and three G2 patients, two NRTI associated with one PI; 11 G3 patients received two NRTI associated with one PI, and four G3 received two NRTI and one NNRTI.Although there was no significant difference between groups under treatment, PI tended to predominate in G3 (χ 2 1 =3.70; p<0.10).At exams collection, G1 had the highest number of individuals (six) with AIDS defining diseases; there were also 2 oligosymptomatic and 5 asymptomatic patients.In G2, asymptomatic was predominant; there were no oligosymptomatic, and only one individual with AIDS, who showed paradoxical reaction to tuberculosis.All G3 were asymptomatic.
Analyzing data from Table 4, statistical difference was seen between G1, and G3 and GC-2 for CD 45 RA cell count, with patients in G1 showing lower counts than those in G3 and GC-2 which were similar, and between G1, G3, and GC-2 for CD 45 RO cell count, with G1 showing lower counts than GC-2, and GC-2 patients showing lower counts than G3.

Table 4 shows patient characterization in relation to serum cytokine level determination.
There was a progressive TNF-α serum level decrease from G1 to G3, and all showed higher TNF-α than GC-1.Statistical difference was seen between G1 and G2, G3, and GC-1 for IL-2, with serum levels in G1 being less elevated than in groups G2, G3, and GC-1, which had no statistical difference between them.There was no difference in IFNγ serum levels between G1 and GC-1, but they were less elevated than in G2 and G3, which were not statistically different either.There was no statistical difference between G1, G2 and G3 for IL-10 and IL-4 serum levels, however these were higher than in GC-1 individuals.Statistical difference was seen for serum cytokine profile with predominance of profile 2 in G1, and pronounced predominance of mature profile 0 in G2 and G3 (χ 2 2 =13.65; p<0.005).In G3, only mature profile 0 was seen.No patient in the HIV-1 groups showed profile 1.According to data in Table 4, there was statistical difference between groups for CD 4 + T lymphocyte count with progressive increase from G1 to G3; this did not occur with CD 8 + T lymphocyte count.
In G1, all individuals had detectable VL and in G2 there were seven undetectable and two detectable VL patients; in G3 there were nine undetectable and six were greater than 80 copies/ml; statistical difference was seen between groups, with G1 being statistically greater than both G2 and G3 which had no difference between them (Table 4).
G1: HIV-1 infected individuals before the start of HAART.
G2: HIV-1 infected individuals under HAART for between five and seven months.
G3: HIV-1 infected individuals under HAART for more than 24 months. GC-

DISCUSSION
For a better evaluation of the qualitative aspects of the immune system behavior in patients on HAART, this paper used three other markers including serum cytokine determination and CD 45 RA and CD 45 RO cell count.
In relation to patient homogeneity, the sex behavior, age and transmission mechanism are in agreement with the most recent epidemiological condition of HIV-1 infected individuals in Brazil (18).
Predominance of infected individuals with AIDS defining diseases among untreated patients is strong evidence for the benefit of the treatment (11).
CD 45 RA cell count increased significantly with the treatment in patients treated for at least 2 years, with predominance of PI; levels reached the same values as in control group individuals, which did not occur with CD 4 + cells.According to literature, as treatment goes on, there is a progressive increase in the number of naive CD 4 + T lymphocytes (10).These findings suggest that CD 45 RA was more significant than CD 4 + and VL in the evaluation of immune reconstitution.They perhaps suggest the partial preserved capacity of thymus regeneration (4,10).There was a significant increase in CD 45 RO cells in this study for patients treated for over two years.It should be mentioned that the mean of values from this group was higher than that from controls.This is in agreement with Mezzaroma et al. (20), Haase (10), and Powderly et al. (24).This count, similar to that of CD 45 RA, was more sensitive than CD showing that plasma and in vitro TNF-α levels fall with treatment but not to normal levels.These authors (2) associate the persistent activation of TNF-α system components to therapeutic, virological, and immunological failure.
Serum IL-2 increased in the two under-treatment groups as treatment went on, reaching levels statistically equal to those of normal individuals.In absolute values, the G3 mean was higher than that of control group.This increase indicates a recuperation of the infected individual's immune system.Kaufmann et al. (13) report an increase in IL-2 secretion and Imami et al. (12), an elevation in specific RNA expression by RT-PCR in the first weeks of treatment.Weiss et al. (29) found increased levels of CD 4 + T cells, producers of IL-2; in most studied patients, IL-2 production capacity under stimulation was similar to that of seronegative control individuals.
At pretreatment phase, IFN-γ levels were statistically equal to those of normal individuals, increasing significantly with treatment, which is in agreement with Imami et al. (12).This increase may be explained by the maintenance of large quantities of CD 8 + T lymphocytes, which are the main IFN-γ producers.( 8) IL-4 and IL-10 were not statistically different between infected groups as treatment went on, but were always higher than in controls.However, in absolute values, in the pretreatment group, they were higher than in the other infected groups; this is in agreement with Imami et al. (12).
Characterization of serum cytokine profiles in patients from the three study groups showed none with profile 1 (Th-1).Most studied patients were considered as having mature 0 profile (Th-0), or 2 (Th-2  (14,25) have also reported a very small number of profile 1 patients, agreeing with this study.These authors (14,25), however, did not mention profile 0 in their findings.
When only patients under treatment are considered, absolute serum cytokine median values show patients treated for over 2 years with higher IFN-γ and lower IL-4.These variations, although small, are important because they may be interpreted as an evolution of the mature 0 related to the duration of treatment and predominance of PI.
There was a progressive increase in CD 4 + T lymphocyte counts in patients on HAART for over 2 years, with PI predominance; the median passed 300 cells/mm 3 by a small margin.Therefore, according to Haase (10), the increase in CD 4 + T lymphocyte count with treatment was progressive but partial.
There was no difference in CD 8 + T lymphocyte count behavior with treatment.This therefore did not offer any support for the evaluation of immune recuperation.The best performance was in the group of patients treated for over 2 years, and the analysis of all parameters used perhaps suggests that their immune reconstitution may be attributed to longer treatment time and participation of PIs.Hence there is still the need to perform further studies with these variables with more patients and longer treatment duration.
Autran et    al.(3) and Kaufmann et al.(13) found similar results.HIV-1 VL showed a marked decrease with treatment duration, with some individuals remaining with VL at detectable levels (>80 copies/ml).These results agree with literature(3,20).The relationships of pairs of variables showed coherence of positive correlations between profile 1 defining cytokines (IL-2 and IFN-γ), CD 45 RA and CD 45 RO cells, and CD 4 + T lymphocyte count, as well as for the association between them.These correlations also show coherence of correlations between profile 2 defining cytokines (IL-4 and IL-10) and TNF-α, and VL, as well as for the association between them.The negative correlations are also coherent as they express the reverse of positive associations.The variables with the highest number of positive or negative correlations were IL-2, IFN-γ, and VL, followed by CD 45 RA and CD 45 RO cells, and IL-10.The variables with the lowest number of correlations were CD 4 + T and CD 8 + T lymphocytes.These results show the importance of cytokines especially IL-2 and IFN-γ, VL, and CD 45 RA and CD 45 RO cells as surrogate markers of HIV-1 infection with HAART interference.In literature (21), however, most authors consider CD 4 + T lymphocyte count R. A. M. B. Almeida et al.IMMUNE RECONSTITUTION IN HIV-1 INFECTED PATIENTS TREATED FOR TWO YEARS WITH HIGHLY ACTIVE ANTIRETROVIRAL THERAPY.J. Venom.Anim.Toxins incl.Trop.Dis., 2006, 12, 1, p.105 and VL are the markers of natural history, biological activity, and therapeutic efficacy in HIV-1 individuals.

Table 2 )
. All 19 GC-2 individuals were subjected to CD 45 RA and CD 45 RO cell count.

Table 5
. A. M. B. Almeida et al.IMMUNE RECONSTITUTION IN HIV-1 INFECTED PATIENTS TREATED FOR TWO YEARS WITH HIGHLY ACTIVE ANTIRETROVIRAL THERAPY.J. Venom.Anim.Toxins incl.Trop.Dis., 2006, 12, 1, p.98 + T lymphocytes and CD 45 RA cells.There was a significantly strong positive correlation (α=0.01) between TNF-α and IL-10; TNF-α and VL; IL-2 and IFN-γ; IL-2 and CD 45 RA cells; IL-10 and IL-4; CD 4 + T lymphocytes and CD 45 RA cells; CD 4 + T lymphocytes and CD 45 RO cells; and CD 45 RA and CD 45 RO cells.Analysis of the table also shows that there was a significantly negative correlation (α=0.05) between IL-10 and CD 45 RA cells; IL-4 and CD 45 RO cells; VL and CD 45 RA cells; and VL and CD 45 RO cells.A significantly strong negative correlation was seen (α=0.01) between TNF-α and IL-2; TNF-α and IFNγ; IL-2 and IL-10; IL-2 and IL-4; IL-2 and VL; IFN-γ and IL-10; IFN-γ and IL-4; IFN-γ and CV; and CD 4 + T lymphocytes and VL.The highest number of significant correlations was for IL-2, IFN-γ, and VL, with seven correlations each; this was followed by IL-10, CD 45 RA and CD 45 RO cells with six, TNF-α R

Table 5 :
IMMUNE RECONSTITUTION IN HIV-1 INFECTED PATIENTS TREATED FOR TWO YEARS WITH HIGHLY ACTIVE ANTIRETROVIRAL THERAPY.J. Venom.Anim.Toxins incl.Trop.Dis., 2006, 12, 1, p.101 Linear correlation between pairs of variables: TNF-α, IL-2, IFN-γ, IL-10, IL-4, CD 4 + T and CD 8 + T lymphocytes, plasma viral load (VL), and CD 45 RA and CD 45 RO cell count for six G1, three G2, and eight G3 individuals, all with HIV-1, sick or not. .A. M. B. Almeida et al.IMMUNE RECONSTITUTION IN HIV-1 INFECTED PATIENTS TREATED FOR TWO YEARS WITH HIGHLY ACTIVE ANTIRETROVIRAL THERAPY.J. Venom.Anim.Toxins incl.Trop.Dis., 2006, 12, 1, p.102 * U = Below detection limit (< 80 copies/ml).†I = Intermediate.R. A. M. B. Almeida et al.R 4 + count and VL values as indicators of immune reconstitution of patients on HAART.Another result of this study, which in a way reinforces these statements, was the strongly positive correlation between CD 4 + T lymphocytes and CD 45 RA and CD 45 RO cells, at the same time that there was a weak negative correlation between both CD 45 RA and CD 45 RO and VL. .A. M. B. Almeida et al.IMMUNE RECONSTITUTION IN HIV-1 INFECTED PATIENTS TREATED FOR TWO YEARS WITH HIGHLY ACTIVE ANTIRETROVIRAL THERAPY.J. Venom.Anim.Toxins incl.Trop.Dis., 2006, 12, 1, p.103 Serum TNF-α values decreased as treatment went on with lower levels in the over-twoyears-treatment group, with PI predominance.This decrease, even though at levels about four times higher than in controls, seems to suggest a compatibility with clinical improvement since all patients in this group were asymptomatic.Ledru et al. (15) have reported that the decrease in TNF-α levels is compatible with the decrease of viral replication and apoptosis once this cytokine induces these phenomena (1, 15).Lew et al. (16) have reported a decrease in the number of T cells producers of TNF-α in the early weeks of HAART.Kaufmann et al. (13) showed a decrease in TNF-α secretion with the same treatment.Aukrust et al. (2) have reported data agreeing with this study, R . A. M. B. Almeida et al.IMMUNE RECONSTITUTION IN HIV-1 INFECTED PATIENTS TREATED FOR TWO YEARS WITH HIGHLY ACTIVE ANTIRETROVIRAL THERAPY.J. Venom.Anim.Toxins incl.Trop.Dis., 2006, 12, 1, p.104 predominance of Th-2 (53.1%) and Th-0 (38.7%) profiles in their patients.Although the study group formation was different, results agreed.Other authors ). Meira et al. (19) called attention to the R