García, Felipe*; Plana, Montserrat†; Arnedo, Mireia‡; Brunet, Merçe§; Castro, Pedro*; Gil, Cristina‡; Vidal, Elena§; Millán, Olga†; López, Anna†; Martorell, Jaume†; Fumero, Emilio*; Miró, José M.*; Alcamí, José‖; Pumarola, Tomás‡; Gallart, Teresa†; Gatell, José M.*
Mycophenolate mofetil (MMF), a morpholinoethyl ester of the active compound mycophenolic acid, is a well-characterized drug widely used in renal transplantation because of its ability to selectively inhibit lymphocyte division. 1 MMF may inhibit HIV type 1 (HIV-1) replication by a dual mechanism: an antiviral mechanism involving depletion of the substrate for the reverse transcriptase, and an immunologic mechanism arising from its ability to reduce the pool of activated CD4+ T lymphocytes, which may in turn support productive HIV-1 infection. 2 MMF selectively inhibits synthesis of guanosine nucleotides by competing with inosine monophosphate dehydrogenase. Because there is no alternative enzymatic way for producing guanosine nucleotides in lymphocytes, MMF appears to exert a cytostatic and potential antiviral effect by depletion of this substrate. The effect of mycophenolic acid on cell activation and HIV-1 infection has been investigated both in vitro and in experimental animal studies. 3 In vitro data showed that MMF inhibited proliferation of activated T cells, especially cells with intermediate and low CD4 expression, driven apoptosis, and cellular death even in the presence of interleukin-2. 2,4,5 From these results, the activity of mycophenolic acid on T-cell proliferation and HIV-1 was tested in a pilot clinical study. 4 Overall, this study suggested that an immune modulation strategy using MMF may be effective in controlling viral replication and may reduce the size of the viral reservoir.
The primary aim of this study was to determine the potential efficacy of MMF either as an adjunct to highly active antiretroviral therapy (HAART) or during a structured therapy interruption period. The study was based on 2 hypotheses. First, if MMF blocks cellular activation, the likelihood of viral load (VL) rebound after HAART interruption will be lower in patients administered the drug. Second, it is known that the effect of MMF on the enzyme substrate is rapidly reversed 6; therefore, its efficacy will depend on its ability to inhibit lymphocyte proliferation throughout the dose interval. We also hypothesized that control of viral replication would be poorer in those patients who did not achieve inhibition of lymphocyte proliferation with MMF throughout the dose interval.
Study Design and Patients
Seventeen patients with chronic HIV-1 infection at very early stages (baseline VL, 200–5000 copies/mL; CD4 cell count, >500/mm3) were treated with abacavir (300 mg twice a day), efavirenz (600 mg every day), and nelfinavir (1250 mg twice a day) for 12 months (from day −365 to day 0). They were then randomized (day 0) to receive the same HAART plus MMF (0.25 g twice a day) (HAART–MMF group, n = 9) versus the same HAART alone (HAART group, n = 8) for 120 additional days. At day 120, HAART was discontinued, and MMF administration was continued in the HAART–MMF group for at least 6 months. Because of the longer half-life of efavirenz, this drug was discontinued 4 days before interruption of the other 2 drugs. Two patients from the HAART group were lost to follow-up immediately after randomization and were not evaluated. This left 6 patients in this group for further study.
Plasma VL was assessed, and an immunologic study (including CD4+ and CD8+ T-cell subset assessment, annexin V [Pharmingen] staining, and Ki67+ T-cell measurement) was performed at baseline (day −365), day 0 (day of randomization), day 120 (when HAART was discontinued), and every 2 weeks thereafter. Tonsillar tissue VL was determined at baseline, day 0, day 120, and day 150 (4 weeks after interruption of HAART) as previously described. 7 The pharmacokinetic profile of mycophenolic acid (area under the curve [AUC0–12h]) and pharmacodynamic studies (CEM response assay, which basically determines the ability of patients' plasma to inhibit proliferation of the CEM cell line) were carried out 7, 28, 120, and 150 days after MMF treatment.
Plasma HIV-1 RNA levels were determined using the Amplicor HIV-1 Monitor Ultra Sensitive Specimen Preparation Protocol Ultra Direct Assay (Roche Molecular Systems, Inc., Somerville, NJ) with a limit of quantification of 20 copies/mL. Samples with findings below the detection limits of this test were retested with a lower level of 5 HIV-1 RNA copies/mL. 8 Subpopulations of CD4 and CD8 cells were determined by 3-color flow cytometry. For apoptosis studies, annexin V-PE (Pharmingen) was used. Annexin V staining was performed in the presence of 5 mM CaCl2 after 18 hours of cultivation of peripheral blood mononuclear cells (PBMCs) at 1 × 106 cells/mL in 24-well plates in the presence or absence of 1% phytohemagglutinin. Ki67 staining was performed after fixing and permeabilizing PBMCs. Mouse Ig isotypes conjugated with Per-CP, PE, or FITC were always used as negative controls for nonspecific binding. The stained cells were analyzed on a FacScalibur flow cytometer (Becton Dickinson, Mountain View, CA).
The primary end point of the study was the number of individuals maintaining a plasma VL set point of <200 copies/mL after at least 6 months off HAART. The study was explained to all patients in detail, and all patients gave written informed consent. The institutional ethical review board approved the study.
Substudy of Inhibition of Lymphocyte Proliferation and Correlation With Control of VL Replication: Pharmacokinetic and Pharmacodynamic Assays
The pharmacokinetic profile of mycophenolic acid (AUC0–12h) was carried out with blood samples (EDTA) collected 0 (predose), 20, and 40 minutes and 1, 2, 4, 6, 8, 10, and 12 hours after administration of the morning MMF dose. Plasma concentrations of mycophenolic acid were analyzed by a validated and previously described HPLC method 1 with a minor modification concerning the internal standard used. Briefly, 100 μL of carboxy butoxy ether mycophenolic acid solution (10 mg/L in methanol) was added to 0.2 mL of human plasma as an internal standard and 1 mL of 80% vol/vol methanol in 0.1 M acetate buffer (pH 4.4). The mixture was vortexed and centrifuged for 10 minutes at 3500 rpm. Chromatographic analysis of mycophenolic acid and the internal standard was achieved with a C-18 Novapak HPLC column (4.6 mm × 25 cm; Waters, Milford, MA) connected to a reverse-phase microguard column (Waters). Chromatography was carried out at 25°C with a flow rate of 1.0 mL/min and monitored at a UV wavelength of 215 nm. The isocratic mobile phase was 0.05% aqueous phosphoric acid:acetonitrile in a ratio of 55:45. The total run time was 12 minutes. The working range for mycophenolic acid was 0.1 to 50 mg/L. Within-run and between-run imprecision ranged from 4.5% to 9.7%. The concentrations of mycophenolic acid were determined by the ratio of its peak height in relation to that of the internal standard.
The efficacy of MMF treatment was evaluated by the capacity of patients' sera to inhibit the response of a T-cell line (CEM). To study the CEM response, blood samples were collected in tubes without anticoagulant 0, 1, 2, 4, and 12 hours after MMF administration. The CEM response assay determines the ability of patients' plasma to inhibit proliferation of the CEM cell line. 1,5,9 The results were expressed as follows: CEM response (% CEM) = (CEM proliferation [cpm] in the presence of patients' sera) × 100/(CEM proliferation [cpm] in the presence of normal human sera). On the basis of previous results, 10 we considered that a patient had good control of the proliferation response in the dose interval if the predose CEM response was <40%. When our transplant patients received standard dosages of MMF (2–3 g/d), all presented with a pre-dose % CEM of <40%, but when they received doses below the standard some presented with a predose % CEM above this figure. We classified the population into 2 groups according to the results of this assay: patients who maintained an inhibitory capacity of >60% (<40% of CEM response) throughout the dose interval and patients who maintained an inhibitory capacity only 0 to 4 hours after dosing. We expected the GTP depletion and control of proliferative response to be more constant in the first group than in the second group, because the MMF effect on enzyme substrate is rapidly reversed. 6 Six of 9 patients in the HAART–MMF group appeared to maintain an inhibitory capacity of >60% (<40% of CEM response) (inhibition group), and none of the patients in the HAART alone group inhibited proliferation of the CEM cell line, giving a total of 9 nonresponders (all 6 patients from the HAART group plus 3 patients from the HAART–MMF group) (no inhibition group).
CEM was measured as follows. The CEM cell line (American Type Culture Collection, Rockville, MD) is a human T-lymphoblastoid cell line that was obtained from peripheral blood buffy coat of a 4-year-old white girl with acute lymphoblastic leukemia. We used the subclone, which we called CEM.2b. This clone will be referred to henceforth as CEM. The CEM cells were cultured in RPMI-1640 culture medium (Bio-Whittaker, OptiPhase “HiSafe” 2, Wallac Scintillation Products) supplemented with 10% FCS and gentamicin. The cells were grown at 37°C under 5% CO2. Standards of mycophenolic acid and carbamazepine were obtained from Sigma-Aldrich (Madrid, Spain). Methanol, acetonitrile (HPLC grade), and acetic acid (glacial) were purchased from Scharlau (La Jota, Barcelona, Spain), and the C-18 solid-phase extraction column was from Supelco (Bellefonte, PA). In the presence of patients' sera, CEM cells were resuspended in RPMI-1640 medium containing 10% heat-inactivated FCS; 2 × 108 cells/L were seeded in 96-well microtiter plates, and 100 μL of patient sera was added. 3 H-Thymidine was added 24 hours after culture, and incorporated thymidine was measured in a Beckman Scintillation Counter 24 hours later.
To test whether the mycophenolic acid effect on inosine monophosphate dehydrogenase activity and proliferation was reversible in our model, PBMCs from a healthy donor were activated with phytohemagglutinin plus interleukin-2, and spontaneously dividing CEM cells were cultured for 2–8 hours with mycophenolic acid. Mycophenolic acid removal in in vitro cultures was done as follows: 5 × 105 PBMCs in RPMI-1640 medium containing 10% heat-inactivated FCS were seeded in 96-well microtiter plates, and the cells were activated with phytohemagglutinin (5 μg/mL) plus interleukin-2 (20 IU/mL). At 0 hours, 3 mg/L mycophenolic acid was added. At 2, 4, and 8 hours, mycophenolic acid was removed by washing, and cells were resuspended in media containing phytohemagglutinin plus interleukin-2. 3 H-Thymidine was added at 24 hours and measured 24 hours later. Spontaneously dividing CEM cells were treated in a similar manner, with the exception that neither phytohemagglutinin nor interleukin-2 was added. All proliferation cultures were done in triplicate.
For the purpose of analysis, undetectable RNA levels (<5 copies/mL) were considered equivalent to 5 copies/mL. The HIV RNA values were log10 transformed before analysis. The doubling time of VL was calculated as described elsewhere. 11 The baseline VL and baseline CD4 T-cell count before any antiretroviral therapy were defined as the average of the screening findings within 0 to 3 months before study enrollment (day −365). The VL set point was determined by the average of the last 2 stable measurements (with a difference of <0.3 log10) separated by at least 1 month after 24 weeks off therapy. The proportion of patients with undetectable VL after interruption of HAART in the groups was compared using the Fisher exact test. Quantitative data of VL and lymphocyte subtypes were compared between groups using the Student t test for paired samples for variables with normal distribution and similar variances or with the Wilcoxon matched pairs test for variables without normal distribution. Changes in RNA viremia, Ki67+ T cells, and annexin level over a period of 16 weeks after interruption of HAART were analyzed by an AUC measurement that incorporated the baseline value. Spearman rank order correlations were performed for inhibition of CEM proliferation and mycophenolic acid blood levels.
Baseline characteristics of the patients are shown in Table 1. Two patients in the HAART group were lost to follow-up immediately after randomization and before interruption of HAART and were not included in the analysis. There were no statistically significant differences at baseline between the HAART–MMF group and the HAART group. At day 0 (day of randomization) and day 120 (day of interruption of HAART), all patients had an undetectable plasma VL (<5 copies/mL). Tonsillar tissue VL was detectable at day 0 in 3 of 9 patients in the HAART–MMF group and in 2 of 6 patients in the HAART group, and all patients had an undetectable VL (<40 copies/mg of tissue) at day 120.
Evaluation of MMF Efficacy: Comparison of MMF–HAART and MMF Groups
Comparison of the 2 randomized groups (HAART–MMF and HAART groups) showed no differences in the dynamics of VL rebound during interruption of HAART (Table 2). Moreover, 5 of 9 patients in the HAART–MMF group versus 1 of 6 patients in the HAART group maintained a VL set point of <200 copies/mL after at least 6 months off HAART after interruption of HAART (Table 2) (P = 0.28). At day 120 (before HAART interruption), differences between the groups in the lymphocyte subset tested (CD4+ or CD8+), Ki67+ T cells, or annexin V staining levels were not statistically significant (data not shown). There was a significant correlation between VL after interruption of HAART and CD4+ and CD8+ Ki67+ cells (r = 0.3, P = 0.0001; and r = 0.5, P = 0.0001; respectively). After interruption of HAART, there was an increase in Ki67+ CD4+ T cells (P = 0.05), and this increase was similar in the HAART–MMF (mean AUC ± SE, 1.34 ± 0.31) and HAART (mean AUC ± SE, 1.52 ± 0.35) groups (P = 0.69) (Table 2).
Annexin V binding levels did not change over all the time points tested (P = 0.28), and there were no differences between groups (Table 2). From the day of randomization (day 0) to the day of interruption of HAART (day 120), CD4+ T-cell counts did not change significantly in either group (P > 0.05). Moreover, the change in the CD4+ T-cell counts did not differ significantly between the groups (mean ± SE of change of the CD4+ T-cell count, −61 ± 99/mm3 and 110 ± 65/mm3 in HAART–MMF and HAART groups, respectively; P = 0.27).
Substudy of Inhibition of Lymphocyte Proliferation and Correlation With Control of VL Replication: Pharmacokinetics and Pharmacodynamics
The median AUC0–12h of MMF was 15 mg xh/L (data not shown). There were no differences in pharmacokinetic or pharmacodynamic studies when comparing the different time points analyzed (7, 28, 120, and 150 days after MMF treatment). To assess whether CEM response inhibition could be predicted by mycophenolic acid blood levels, the capacity of sera to inhibit the spontaneous CEM proliferation at 0, 1, 2, 4, and 12 hours was correlated with mycophenolic acid blood levels (mg/L) at the same times pooling together data for all patients obtained on days 7, 28, 120, and 150 of MMF treatment. Mycophenolic acid blood levels had a significant negative relationship with the CEM response (P < 0.01 in all cases). The discriminatory power of mycophenolic acid plasma levels was especially evident when mycophenolic acid plasma levels were >1 mg/L (1 and 2 hours). However, for plasma levels of <1 mg/L, which are predominant at 0, 4, and 12 hours, a high variability in the CEM response was found, even in patients with similar mycophenolic acid plasma levels (0, 4, and 12 hours) (data not shown). Moreover, in the assay of the reversibility of the mycophenolic acid effect, we observed that incubation of both activated PBMCs and CEM cells with MMF for up to 8 hours did not modify the future proliferative capacity of those cells when MMF was removed (Table 3). When all the postdose curves obtained for patients in the MMF group were divided into those belonging to patients who had increased VL at day 150 and those who did not, no differences were observed for minimum concentration, maximum concentration, or AUC, suggesting that mycophenolic acid levels were not good predictors of virologic control. However, predose % CEM showed an acceptable correlation with the capacity of patients to control viral rebound.
We used the CEM response as the criterion for classifying the population into 2 groups according to the results of the CEM response assay: patients maintaining an inhibitory capacity of >60% (<40% of CEM response) throughout the dose interval (inhibition group) and patients maintaining an inhibitory capacity only 0 to 4 hours after dosing (no inhibition group). We expected that GTP depletion and control of proliferative response to be more constant in the first group than in the second group, because the MMF effect on enzyme substrate is rapidly reversed. Six of 9 subjects in the HAART–MMF group appeared to maintain an inhibitory capacity of >60% (<40% of CEM response) (inhibition group), and none of the patients in the HAART alone group inhibited proliferation of the CEM cell line, giving a total of 9 nonresponders (all 6 patients from the HAART group plus 3 patients from the HAART–MMF group) (no inhibition group). Because there were no differences in pharmacokinetic or pharmacodynamic studies when comparing the different time points analyzed (7, 28, 120, and 150 days after MMF treatment), classification of the inhibition and no inhibition groups was based on the results of all these time points.
A second analysis reclassifying the patients according to their CEM response (<40% at all time points) and comparing the group with the no inhibition group was performed. At baseline (day −365), there were no differences in VL or CD4+ T-cell count between the groups (Table 1). One month after interruption of HAART, plasma VL and tonsillar tissue VL were not above the detectable level in any of the 6 patients in the inhibition group but were detectable in all 9 patients in the no inhibition group (P = 0.0001). The mean doubling time ± SE was significantly higher in the inhibition group (10.22 ± 1.3) than in the no inhibition group (4.6 ± 1.6) (P = 0.03) (Table 2). The mean AUC ± SE of viral rebound after interruption of HAART was significantly lower in the inhibition group (0.82 ± 0.23) than in the no inhibition group (1.92 ± 0.2) (P = 0.003) (Table 2). There were no significant differences between the 2 groups in peak VL after interruption of HAART (P = 0.28) (Table 2). After interruption of HAART, 5 of 6 patients in the inhibition group maintained a VL set point of <200 copies/mL versus 1 of 9 patients in the no inhibition group after at least 6 months off HAART (P = 0.01) (Table 2, Fig. 1).
At day 120, there were no differences between the groups in lymphocyte subset tested (CD4+ or CD8+), Ki67+ T cells, or annexin V staining levels (data not shown). After interruption of HAART, the increase in Ki67+ CD4+ T cells was significantly lower in the inhibition group (mean AUC ± SE, 0.98 ± 0.27) than in the no inhibition group (mean AUC ± SE, 1.94 ± 0.35) (P = 0.05) (Table 2, Fig. 2A). There was a significant correlation between VL after interruption of HAART and CD4+ and CD8+ Ki67+ cells (r = 0.3, P = 0.0001; and r = 0.5, P = 0.0001; respectively). Annexin V binding levels were similar in the 2 groups at all time points tested (Table 2, Fig. 2B).
The mean CD4+ T-cell counts ± SE were 912 ± 143/mm3 and 891 ±130/mm3 at day 0 (before randomization) and 808 ± 166/mm3 and 973 ± 196/mm3 at day 120 (before discontinuing HAART) for the inhibition and no inhibition groups, respectively (Fig. 3). The change in the CD4+ T-cell count between day 0 and day 120 did not differ significantly between the groups (mean ± SE of change of the CD4+ T-cell count, −103 ± 140/mm3 and 82 ± 76/mm3 for the inhibition group and no inhibition group, respectively; P = 0.22). Nonsignificant drops in CD4+ T-cell counts were observed after interruption of HAART (mean ± SE of difference of the CD4+ T-cell count between set point and day 0, −191 ± 72/mm3 and −2 ± 162/mm3 for the inhibition group and no inhibition group, respectively; P > 0.05). CD4+ T-lymphocyte counts did not drop below the baseline value (>500/mm3) in any patient at any time.
No patients had to discontinue HAART or MMF because of adverse effects, although 4 patients had to change therapy for this reason (all of them during the period between day 0 and day 120 before the cycles of interruption of HAART). In 2 cases with diarrhea, nelfinavir was replaced by indinavir; in 1 case with rash, abacavir was changed to stavudine; and in 1 case with psychiatric disorders, efavirenz was switched to lamivudine. Mean fasting cholesterol, triglyceride, and glucose levels did not change significantly during the follow-up. There were no cases of clinical lipodystrophy. None of the patients for whom MMF treatment was maintained reported any adverse effects during the period off HAART. The proportion of patients with adverse effects did not differ between the HAART–MMF group and the HAART group (33% in both groups).
Our pilot study suggests that interrupting HAART and maintaining MMF treatment may substantially reduce the size of the pool of dividing CD4+ T cells and that this reduction is strongly correlated with viral replication but only in patients who maintain the capacity to inhibit lymphocyte proliferation for at least 4 hours after dosing of MMF (as measured by the CEM response assay). Moreover, in our cohort of patients with early-stage chronic HIV-1 infection, subjects who maintained CEM inhibition throughout the MMF dosing interval maintained VLs of <200 copies/mL. Conversely, we found that MMF levels were not predictors of virologic control and therefore could not be used as predictors of response.
In the light of these results, the question that arises is how MMF exerts its effect on viral replication. MMF may inhibit HIV-1 infection by an antiviral mechanism exerted by depleting intracellular deoxyguanosine triphosphate and enhancing the activity of abacavir and other nucleoside analog reverse transcriptase inhibitors as demonstrated by Margolis et al 2 for patients with advanced-stage disease. It has also been suggested that MMF might inhibit viral replication by an immunologic mechanism able to reduce the pool of activated CD4+ T lymphocytes, which may in turn support productive HIV-1 infection. 4 However, this immunologic mechanism has been difficult to differentiate in vivo from virologic mechanisms, because MMF has always been used in association with HAART. 4 Our pilot study suggests that interrupting HAART and maintaining MMF may substantially reduce the size of the pool of dividing CD4+ T cells and that this reduction is correlated with viral replication. Moreover, a considerable reduction in the VL set point was observed by maintaining MMF treatment during interruption of HAART, at least in those patients who maintained inhibitory capacity of >60% (<40% of CEM response) throughout the dose interval. This finding lends further support to the hypothesis that an immunologic mechanism alone is able to partially control viral replication in vivo. We previously reported that maintaining HU during discontinuation of antiretroviral drugs may significantly increase the percentage of patients able to achieve control of viral replication probably by limiting the subsequent wave of viral replication among activated T lymphocytes by virtue of its cytostatic effects. 12 This strategy of using drugs that interfere with the HIV life cycle—acting on the target cells of HIV rather than on viral enzymes—has the advantage of avoiding the development of antiretroviral drug–resistant HIV mutants and could be used as a complementary strategy to HAART. Nevertheless, these data should be considered with caution. This is a pilot study, and it may have been viral replication that was driving the Ki67+ T-cell expression; the decreased levels we observed in the inhibition group could be a consequence (and not a determinant) of lesser viral rebound in these patients.
Chapuis et al 4 reported that MMF might exert an effect on the pool of resting latently infected CD4+ T cells. They observed that HAART plus MMF reduced the ability to isolate virus from the CD4+ T-cell population. It is likely that MMF has no effect on a resting cell and, thus, does not affect directly the size of this pool of cells. However, once these cells are activated in the presence of MMF, MMF may in turn induce apoptosis and therefore cell death. In this pilot study, we were unable to demonstrate an increase in apoptosis either before or after interrupting HAART in patients with maintained inhibition of lymphocyte proliferation (measured by the CEM response). It may be that measurement of annexin V binding in PBMCs is not an accurate method for demonstrating this increase in apoptosis; as a result, alternative explanations should be explored.
One of the main drawbacks of our study was the early loss to follow-up of 2 patients randomized to the control group (25% of the total group). Moreover, the initial selection criteria for study participants (chronic infection but with low VL and high CD4 cell count) may have produced a group already biased toward “spontaneous” viral control and slow progression. This may limit the generalization of these findings. Further research should be performed with a higher number of patients in different stages of infection to confirm these results, to test whether those patients without a sustained CEM inhibitory response might have benefited from a higher dose of MMF and to address other issues such as the optimal duration of MMF therapy. The combination of HAART and an immunosuppressive drug such as MMF for HIV-1 infection should be applied with caution until the long-term impact and safety have been further investigated in larger clinical trials.
The authors are indebted to all the participants of the study, laboratory technicians (M.J. Maleno, A. Capón, M. García, and A. García), and Roche for supplying MMF.
© 2004 Lippincott Williams & Wilkins, Inc.