| |
Effect of HAART Interruption on mtDNA
|
| |
| |
"Effect of treatment interruption monitored by CD4 cell count on mitochondrial DNA content in HIV-infected patients: a prospective study"
"....Discontinuation of treatment provoked no changes in the amount of mtDNA per cell in CD4 T lymphocytes, but a significant increase in mtDNA copies in CD8 T cells [18.20 copies/cell per month (P = 0.0018) in the first year of therapy suspension, such an increase started only 6 months after therapy discontinuation....it is important to understand if the increase in mtDNA content shown in this paper reflects a recovery from drug-related toxicity or a direct negative effect of the virus.... our results are more likely to reflect a recovery from drug-related toxicity.... further follow-up of this study is needed to answer some of these questions.... break in treatment of at least 6 months is required to initiate an observable increase..."
AIDS: Volume 19(15) 14 October 2005 p 1627-1633
Mussini, Cristinaa,c; Pinti, Marcellob; Bugarini, Robertob; Borghi, Vannia,c; Nasi, Milenab; Nemes, Elisab; Troiano, Leonardab; Guaraldi, Giovannia,c; Bedini, Andreaa,c; Sabin, Carolined; Esposito, Robertoa,c; Cossarizza, Andreab
From the aClinic of Infectious and Tropical Diseases
bDepartment of Biomedical Sciences, University of Modena and Reggio Emilia, Modena
cAzienda Policlinico, Modena, Italy
dRoyal Free Centre for HIV Medicine and Department of Primary Care and Population Sciences, Royal Free and University College Medical School, London, UK.
Introduction
Treatment with HAART, in particular using regimens containing nucleoside reverse transcriptase inhibitors (NRTI), is associated with a variety of side effects [1,2], many of which can be ascribed to direct mitochondrial toxicity of the drugs [3-5]. Moreover, HIV infection per se can also cause mitochondrial damage [6], and it has been hypothesized that the negative effect of HIV on mitochondria could render HIV-positive patients more susceptible to NRTI-induced mitochondrial toxicities.
Such effects, the most severe of which are hyperlactataemia and lactic acidosis [7-10], are mainly caused by the ability of NRTI drugs to bind and alter the functionality of gamma-polymerase, the enzyme present in the mitochondrial matrix that replicates mitochondrial DNA (mtDNA). A decrease in mtDNA content leads to impairment in the production of respiratory chain enzymes, with consequent alteration of cellular metabolism and the production of ATP.
Although the interpretation of data obtained ex vivo on blood cells is difficult because of possible contamination by platelets, which contain mtDNA but not nuclear DNA [11], there is general agreement that the decrease in mtDNA content induced by NRTI can be considered to be a marker of more generalized mtDNA depletion. While the mechanism by which HIV leads to mitochondrial damage is elusive, it has been shown that mtDNA depletion occurs in several regions of the body, including skeletal muscle, adipose tissue, the liver, peripheral nerves and the placenta [4,12-18].
In an attempt to decrease drug exposure and, therefore, toxicity, two main strategies have been considered: the first is the maintenance of HAART but with the substitution of certain drugs [19], and the second is the use of treatment interruptions. Structured treatment interruptions comprise different periods of alternating on and off cycles of HAART, the length of the cycles either being fixed in advance [20-23] or being determined by monitoring for CD4 cell count [24-26] with the aim of decreasing drug exposure while preserving the CD4 T lymphocyte level.
The aim of the present study was to evaluate the effect of treatment interruptions monitored by CD4 cell count on mtDNA in highly purified, platelet-free peripheral blood CD4 and CD8 T lymphocytes.
Abstract
Background: HIV infection per se and HAART can alter mitochondrial functionality, leading to a decrease in mitochondrial DNA content.
Objective: To evaluate whether treatment interruption monitored by CD4 cell count can restore mitochondrial DNA content in peripheral blood lymphocytes.
Methods: Mitochondrial DNA content was measured in platelet-free CD4 and CD8 T cells by real-time polymerase chain reaction; flow cytometry was used to identify and quantify activated CD4 and CD8 T lymphocytes.
Results: The 30 patients had been treated for a mean of 107 months (range, 27-197). Median CD4 cell count at discontinuation was 702 cells/ml (range, 547-798). Median observational time from HAART discontinuation was 11.3 months (range, 4-26). Discontinuation of treatment provoked significant increases in mitochondrial DNA in CD8 T cells, which started only 6 months after therapy discontinuation [5.12 copies/cell per month from 0 to 6 months (P = 0.37) and 26.96 copies/cell per month from 6 to 12 months (P < 0.0001)].
Conclusions: This study is the first showing that mitochondrial DNA content can increase in peripheral blood lymphocytes during treatment interruption, but only after at least 6 months of interruption. Consequently, interruptions of shorter periods, whether by clinician or patient decision, are unlikely to allow restoration of mitochondrial DNA and so decrease HAART-related toxicity.
Results
Between December 2001 and December 2003, 30 patients entered the study and discontinued HAART. No HIV-related or clinical event occurred during the period of discontinuation. No transmission of HIV to sexual partners was recorded. Epidemiological and immuno-virological characteristics of the patients at start of HAART and at discontinuation are shown in Table 1. At discontinuation, HIV plasma viral load was > 50 copies/ml in nine patients (range, 53-5802). Viral load had been detectable in these nine patients for a median period of 4 months (range, 1-12) before stopping treatment. Patients discontinued treatment for a median of 11.3 months (range, 4-26); 15 patients (50%) restarted treatment after a mean period of 10 months (range, 4-20) and nine of these had HIV plasma viraemia <50 copies/ml after 2 months of treatment.
At baseline and during discontinuation, mtDNA content was analysed in highly purified, platelet-free CD4 and CD8 T lymphocytes. Discontinuation of treatment provoked no changes in the amount of mtDNA per cell in CD4 T lymphocytes, but a significant increase in mtDNA copies in CD8 T cells [18.20 copies/cell per month (P = 0.0018) in the first year of therapy suspension]. However, it is noteworthy that such an increase started only 6 months after therapy discontinuation [5.12 copies/cell per month from 0 to 6 months (P = 0.37) and 26.96 copies/cell per month from 6 to 12 months (P < 0.0001)].
Interestingly, mtDNA increase among CD8 T cells was higher in subjects who had a larger increase (slope) in CD4 T lymphocytes during the treatment period (0.34 copies/cell for each CD4 T lymphocyte cell increase; P = 0.038). No significant association was found for the other covariates included in the statistical model (including CD4 T lymphocyte absolute count, CD8 lymphocyte count, HIV plasma viraemia, CD4 T lymphocyte nadir value before the beginning of therapy, antiretroviral regimens and previous duration of exposure to antiretroviral treatment).
The mean CD4 T lymphocyte count decreased by 52.30 cells/ml per month [95% confidence interval (CI), 39.38-80.27; P < 0.001] over the first 6 months of discontinuation, then decreased by 10.38 cells/ml per month (95% CI, 1.66-17.28; P < 0.001) in subsequent months. A mean increase in HIV RNA of 0.32 log10 copies/ml per month (95% CI, 0.096-0.76; P = 0.00695) was apparent over the first 6 months of discontinuation, followed by a mean increase of 0.18 log10 copies/ml per month (95% CI, 0.054-0.23; P = 0.008) thereafter. The rate of change of CD4 T lymphocytes and HIV RNA were not associated with the CD4 T lymphocyte value at the time of discontinuation. After stopping therapy, the monthly decrease in CD4 T lymphocyte count was reduced by 11.8 cells/ml for each 10 cells/ml higher CD4 T lymphocyte pretreatment nadir value. While some patients experienced a decline in CD4 T lymphocyte count during HAART to levels below their pretreatment nadir, this on-treatment nadir was not associated with the CD4 cell decrease (P = 0.76).
It is of note that activated CD4 T cells showed a slight but significant increase (of the order of 40%) in the first 4 months of therapy interruption then decreased. The increase in activated CD8 T lymphocytes was much more marked (about six times) and rapid (2 months); the percentage of these cells then decreased and remained stable until the end of the analysis.
Discussion
This study is the first to show that mtDNA content increases in CD8 T lymphocytes in individuals who discontinue HAART, but that this process is not immediate. Indeed, a break in treatment of at least 6 months is required to initiate an observable increase. The longitudinal design of our study meant that we were able to follow changes in mtDNA content (along with several other viro-immunological parameters) in the same patients over prolonged periods of time during treatment interruption.
Treatment discontinuation is generally followed by several clinical and immunological phenomena that often resemble acute HIV infection [31,32]. We show here that activated CD8 T lymphocytes increased shortly after the interruption, as did CD4 T cells, even if the magnitude and kinetics of these increases differed among T cell subpopulations. Therefore, it is important to understand if the increase in mtDNA content shown in this paper reflects a recovery from drug-related toxicity or a direct negative effect of the virus.
Our results suggest that there is no direct correlation between the level of HIV plasma viraemia and mtDNA, in contrast to a previous report [6]. Given the viral rebound that is usually seen in the first few weeks after treatment discontinuation, such an association could imply that any increase in mtDNA content can be explained by mitochondrial proliferation in response to a direct toxic effect of HIV. However, if this was the explanation for our findings, then the increase in mtDNA should have been immediate rather than delayed until 6 months after treatment discontinuation. The absence of correlation with the intensity of HIV infection, the fact that all these patients were treated for a long period of time with NRTI and the latency in mtDNA content increase suggest that our results are more likely to reflect a recovery from drug-related toxicity.
There is considerable concern regarding the suitable cells for studies on NRTI-induced changes in mtDNA. To overcome the obvious difficulty in obtaining adipose tissue, where changes in mtDNA or mtRNA have been clearly shown [19,33-35], many investigators are using a more accessible tissue, such as blood, and then analyse mtDNA content in peripheral blood mononuclear cells or in peripheral blood lymphocytes [2]. However, peripheral blood mononuclear cells and blood lymphocytes are heterogeneous populations of different cells (B and T lymphocytes, natural killer cells, monocytes, dendritic cells, among others) at different stages of activation or differentiation. In these cell types, the kinetics of mtDNA and mitochondrial proteins production are unknown, and the simple measure of mtDNA content in peripheral blood mononuclear cells or blood lymphocytes, often without considering contamination by platelets and without performing the appropriate checks for homogeneity, is leading to conflicting results [36]. Using platelet-free, highly purified lymphocyte populations, we found an increase in mtDNA content among CD8, but not CD4, T lymphocytes that started 6 months after treatment discontinuation. A possible explanation for the different behaviour of T cell subtypes could be that during HIV infection the half-life of CD4 T cells may be much shorter than that of CD8 lymphocytes. As a result, there may be more newly formed lymphocytes present among CD4 T cells than among CD8 T cells, and this may result in a shorter period of drug exposure for these cells. Since the increase only began to appear 6 months after discontinuation, when the number of activated T cells was significantly lower than in the first period of discontinuation, we can exclude the possibility that T cell activation per se is responsible for mtDNA changes. CD4 and CD8 T lymphocytes, however, contain subpopulations that express different surface markers (such as CD45RA, CD62L, CCR7), which allow their classification as naive, central and effector memory, or terminally differentiated cells. It cannot be excluded that changes in the relative percentages of these subpopulations could occur during treatment interruption and so influence mtDNA analysis. Regardless, it is of note that we could detect changes in mtDNA content in cells outside adipose tissue and in patients who did not interrupt therapy because of the onset of side effects.
Our results may have important clinical implications; indeed, in theory, structured treatment interruptions of fixed time intervals could reduce costs and possibly increase adherence among highly motivated patients, who may prefer to take their medications for only 5 days a week. However, the use of such a strategy may encourage the development of drug-resistance mutations [23] and, on the basis of the present study, it is unlikely to restore mtDNA content in CD8 T lymphocytes unless the periods of interruption are prolonged. Treatment interruptions that are guided by CD4 cell monitoring may allow the patient to remain off therapy for longer periods of time [24-26], thus allowing for restoration of mtDNA content in CD8 T lymphocytes. Since the factors that influence CD4 T lymphocyte decay have generally been similar in all studies, in particular the nadir pretreatment CD4 T lymphocyte count and the duration of undetectable HIV plasma viraemia, they should be considered when identifying patients who can safely discontinue treatment for periods of longer than 6 months in an attempt to restore mtDNA content. Moreover, although all prospective studies have suggested that this strategy is safe [24-26,37], patients should be carefully monitored for any sudden decrease in CD4 cell counts.
There are a number of limitations to the present study. Since frozen samples from earlier periods are not available, we are unable to say whether the level of mtDNA returned to pretreatment levels. We also did not evaluate changes in body shape using standardized methods, so we cannot comment on the clinical relevance of these biological data. We do not know if an increase in mtDNA content will prevent or delay the onset of mitochondrial-related toxicities (such as lactataemia) once HAART is restarted in these patients. Finally, while the statistical methods chosen utilized all longitudinal data to measure changes in mtDNA content within each individual, they may be subject to informative censoring, for example if individuals with the most rapidly decreasing CD4 T lymphocyte cells restarted therapy quickly and, therefore, had the shortest periods off-treatment. If this was the case, estimates of mtDNA content recovery over the longer term may be biased towards patterns seen in patients who have slower CD4 decreases and who are off therapy for longer periods of time.
In conclusion, while further follow-up of this study is needed to answer some of these questions, our results do provide useful information for clinicians and patients who wish to use treatment interruption strategies to decrease drug-related toxicities. Candidates should be carefully selected for such a strategy. On the basis of several studies including the present one, in order to achieve an off-treatment period as long as possible, these persons should start HAART relatively early in order to preserve a CD4 cell count nadir > 300 cells/ml, spend 1 year with a plasma viral load < 50 copies/ml and discontinue treatment with a CD4 cell count > 500 cells/ml. Finally, patients should be advised that self-decided short treatment interruption periods (so-called drug holidays) are unlikely to be sufficient to reduce any previously established mitochondrial damage.
Methods
Patients
This study includes 30 patients with HIV infection followed as outpatients at the Clinic of Infectious and Tropical Diseases, Policlinico of Modena, University of Modena and Reggio Emilia, Italy, who were treated with HAART. Patients, after consultation with the physician, interrupted HAART if they satisfied the following inclusion criteria: > 18 years of age, taking HAART for at least 12 months, and with a CD4 T lymphocyte count pre-interruption > 500 cells/ml. HIV plasma viral load at the time of discontinuation could be below or above 50 copies/ml. Patients treated with immunomodulatory agents and hydroxyurea were excluded. The criteria for restarting treatment were a CD4 T lymphocyte count < 350 cells/ml on two different occasions 1 month apart, a clinical manifestation of AIDS or the patient's desire to resume HAART.
Patients were advised of a possible higher risk of HIV transmission to sexual partners during the period of treatment discontinuation. Patients were followed prospectively and viro-immunological assays were performed at baseline and then bimonthly. Plasma HIV RNA was measured by a branched DNA assay (Chiron, Emeryville, California, USA) and a value < 50 copies/ml was considered undetectable. At the time of blood collection, 27 ml were drawn to perform viro-immunological analyses including analysis of mtDNA.
Flow cytometry
Cytofluorimetric analysis for the detection of activated CD4 and CD8 T cells followed standard methods using a CyFlow ML flow cytometer (Partec, Muenster, Germany) [27]. Activated T cells (expressing the CD3 molecule) were considered those with a high (bright) expression of CD38 on the plasma membrane, as described [28].
Peripheral blood lymphocyte isolation
Peripheral blood lymphocytes were obtained by standard methods (density gradient centrifugation on Ficoll-Hypaque and depletion of monocytes by plastic adhesion). CD4 or CD8 T cells were then positively isolated by magnetic sorting (Miltenyi, Bergisch Gladbach, Germany) to eliminate the contamination of platelets that is usually present after density gradient centrifugation [29]. The purity of cell populations was always > 95%, with a platelet contamination < 1% as assessed by parallel cytofluorimetric analysis.
Quantification of mitochondrial DNA
The cell content of mtDNA was measured by an original method that has been recently developed [30]. The assay used two parallel polymerase chain reactions (PCR) that quantify mtDNA or nuclear DNA using the same construct, known amounts of which were amplified as reference molecules. Briefly, DNA was extracted from isolated CD4 or CD8 T cells according to standard methods. In the first reaction, mtDNA was measured using a mix of 1× PCR buffer (Promega, Madison, Wisconsin, USA), 3.0 mmol/l magnesium chloride, 400 pmol primers for mtDNA, 0.2 mmol/l dNTP and 2 U Taq polymerase (Promega). The mtDNA primers used were: mtDir (5'-CACAGAAGCTGCCATCAAGTA-3') and mtRev (5'-CCGGAGAGTATATTGTTGAAGAG-3'). The TaqMan probe used for mtDNA (5'-FAM-CCTCACGCAAGCAACCGCATCC-BlackHole Quencher1-3') was included in the reaction mixture throughout PCR as a real-time detector for the amplified product. One cycle of denaturation (95°C for 6 min) was performed, followed by 45 cycles of amplification (94°C for 30 s, 60°C for 60 s). PCR was carried out in an iCycler Thermal cycler (BioRad, Hercules, California, USA), and all samples were analysed in triplicate. The same approach was used to quantify nuclear DNA, which was required to obtain the number of copies of mtDNA per cell. In this case, primers GenDir (5' GGCTCTGTGAGGGATATAAAGACA-3') and GenRev (5'-AAACCACCCGAGCAACTAATCT-3'), designed on the gene sequence for FasL (present in two copies in the human genome, one in each chromosome), were used. The TaqMan probe Genprobe (5'-Texasred-CTGTTCCGTTTCCTGCCGGTGC-BlackHole Quencher2-3') was included in the reaction mixture throughout PCR as a real-time detector for the amplified product. To optimize the precision of the assay and express the results as absolute number of mtDNA copies per cell, the regions used as template for the two amplifications (i.e., those of mtDNA and nuclear DNA) were cloned tail to tail in a vector (pGEM-11Z; Promega) to have a ratio of 1:1 of the molecules used as reference. Then, serial known dilutions of this vector, amplified in triplicate, were included in each PCR run to generate a standard curve from which the relative copy number of either mtDNA or nuclear DNA present in the unknown samples was determined. The measured values for mtDNA and nuclear DNA were always within the range of the standard curve and the correlation coefficient was always > 0.995. Absolute values for the copies of mtDNA per cell were then simply obtained from the ratio between the relative values of mtDNA and nuclear DNA (obtained versus the same vector), multiplied by 2 (as two copies of the nuclear gene are present in a cell).
Statistical analysis
Using a mixed model with a random effect for slopes and an autoregressive variance-covariance structure for each patient, changes over time were studied for mtDNA content in CD4 and CD8 T lymphocytes, CD4 T lymphocyte count, plasma HIV RNA and percentage of activated CD4 and CD8 T lymphocytes. This type of model was chosen because there was substantial intra- and interpatient variation in the timing and number of measurements. To model different trends for levels of response variables over time, a spline regression was used. The knot for the splines was chosen as the time point that maximized the differences in slopes over time. For the analysis of mtDNA content in CD4 and CD8 T lymphocytes, the covariates were duration of therapy, length of period of undetectable viral load, CD4 cell nadir value, CD4 cell count and HIV RNA at discontinuation.
|
|
| |
| |
|
|
|
