Increasing evidence exists of a correlation between zidovudine-resistant HIV and disease progression in patients treated with zidovudine monotherapy [3-8]. It has been shown that HIV can also develop resistance to didanosine [9-11]. As with the resistance of HIV to zidovudine, the decrease in susceptibility of HIV to didanosine has been shown to be caused by specific mutations in the HIV reverse-transcriptase gene. St. Clair and colleagues [9] identified the first mutation in the reverse-transcriptase gene to confer resistance to didanosine, a mutation at codon 74 that results in an amino acid change from leucine to valine. The codon 74 mutation in an HIV construct can induce an eightfold decrease in susceptibility to didanosine [9]. Although other mutations have been reported to confer didanosine resistance [12], most data to date suggest that the codon 74 mutation is the primary mutation responsible for didanosine resistance in patients receiving didanosine monotherapy [9, 13-16].

Researchers have postulated [17, 18] that genotypic assays that can detect HIV reverse-transcriptase mutations may be used in place of drug susceptibility testing when a proven association exists between a specific mutation and drug resistance. Previously we reported [5] that in patients who are receiving zidovudine monotherapy, the development of a serum HIV RNA mutation at codon 215 (which confers a 16-fold decrease in susceptibility to zidovudine [18]) was strongly associated with and predictive of decreases in CD4⁺ T cells in these patients. In the current study, we obtained serum HIV RNA from patients who were switched from zidovudine to didanosine therapy, and we examined the relation of the codon 74 mutation in patient serum HIV RNA to changes in CD4⁺ T-cell levels and HIV virus burden. The development of the codon 74 mutation in relation to the genotype at codon 215 was also examined because researchers have shown that the combination of codon 215 and codon 74 mutations can restore HIV susceptibility to zidovudine [9] and that the preexistence of zidovudine resistance mutations may actually augment didanosine resistance [9, 13, 14].

Methods

Top Methods Results Discussion Author & Article Info References

Patients

Sixty-four patients infected with HIV were enrolled in three protocols at Stanford University Medical Center: 1) Eight patients were enrolled in Stanford University/San Mateo County Didanosine Protocol [19], an open-labeled study involving patients who had received zidovudine [Retrovir; Burroughs Wellcome, Research Triangle Park, North Carolina] for more than 16 months who were switched to didanosine; 2) 33 patients were enrolled in AIDS Clinical Trial Group Protocols [2] 116b/117 who had tolerated zidovudine for at least 16 weeks and were then switched to didanosine; and 3) 23 patients were enrolled in AIDS Clinical Trial Group Protocol 118 who were intolerant to zidovudine and who had been switched to didanosine. The baseline CD4⁺ T-cell counts of the patients ranged from 6 to 400 CD4⁺ T cells/mm³ (median, 105 CD4⁺ cells/mm³), and patients had either AIDS, AIDS-related complex, or were asymptomatic. No patient had an active opportunistic infection at the time of enrollment in the study, and patients received Pneumocystis carinii prophylaxis as defined by each protocol and were allowed to continue suppressive therapy for previously diagnosed opportunistic infections [2, 19].

Patients received didanosine (Videx; Bristol Laboratories, Princeton, New Jersey) monotherapy at one of three possible dosages: 200 mg/d (6 patients), 500 mg/d (31 patients), or 750 mg/d (27 patients). Serial serum samples were saved at week 0, 2, 8, 12, 16, 24, 32, 40, and 48; at these same time points, CD4⁺ T-cell determinations were done. All patients were followed until the close of study or until death.

Reverse-Transcriptase Gene Mutational Analysis

Cryopreserved ( –70°C) serum was thawed and 200 µL was ultracentrifuged at 125 000 g for 10 minutes. The resulting pellet was dissolved in 400 µL of 5-M guanidium thiocyanate. Serum HIV RNA was then extracted as previously described [20-22]. Extracted viral RNA was then reverse-transcribed to cDNA using 500 ng of primer 35-NE1 [17] and 5 units of murine leukemia virus reverse transcriptase (Bethesda Research Laboratories, Gaithersburg, Maryland) with reaction conditions and controls as previously described [5, 16]. The cDNA was amplified by PCR using 250 ng of primer 35-A [17] with the reaction conditions described by Larder and Boucher and colleagues [17, 18]. For selective PCR, 5 µL of the 805-basepair product from the first PCR was used in the second series of reactions. Primers 3W (wild type), 3M (mutant), and B [17, 18] were used to determine the sequence at codon 215; primers X2 and 74WT (wild type) and 74M (mutant) [9, 16] were used to determine the sequence at codon 74. Master mix, negative, wild-type, and mutant sequenced controls were amplified in each sample run. Non-reverse-transcribed control samples subjected to the PCR procedure yielded negative results, and patient RNA samples treated with deoxyribonuclease yielded identical results to patient RNA samples using our extraction procedure, validating that the reverse-transcriptase PCR product was the result of HIV RNA and not DNA. The purity of the extracted RNA using the guanidium thiocyanate-phenol-chloroform extraction technique has been previously reported by Chomczynski and Sacchi [22]. Fifteen HIV isolates that had been tested in our laboratory by nested PCR for the codon 74 and 215 mutations were also sequenced and confirmed the presence or absence of these mutations.

Products from PCR were analyzed on a 3% agarose gel with ethidium bromide staining. All samples had been blinded by a code number from the start and thus the evaluator who scored the PCR product had no knowledge of sample origin (that is, the corresponding patient). Samples yielding a product with only the wild-type primers were considered wild type. Samples that yielded product with only the mutant primers were considered mutant. If a sample yielded product with wild-type and mutant primers, the second PCR step was repeated with serial dilutions of the first round PCR (at dilutions of 1:20, 1:400, and 1:8000); if a mixture of wild type and mutant was still present after serial dilutions, the sample was considered a mixture of wild-type and mutant sequences at the codon of interest. Samples with a mixture of wild-type and mutant sequences at the codon were included in the mutant group in our statistical analysis.

Serum HIV RNA Preparation for Virus Burden

Duplicate serum samples were ultracentrifuged, and the pellet was purified by phenol-chloroform extraction and alcohol precipitation, as previously described for plasma [20]. Polymerase chain reaction quantification of viral RNA was done using reverse-transcriptase PCR, and the PCR product was detected using a nonisotopic enzyme hybridization assay, as previously described [20, 21]. Results were then expressed as HIV RNA copies per milliliter of serum.

HIV Biological Phenotype

For the 25 patients for whom cryopreserved peripheral blood mononuclear cells were available, viral stocks were created from these mononuclear cells by cocultivation with peripheral blood mononuclear cells from patients who were seronegative for HIV. Viral stock supernatant (the tissue culture infective dose₅₀ was about 2000), 200 µL, was cultured with 8 mL of MT-2 cells (0.5 x 10⁶ cells/mL) in duplicate. Cultures were maintained for 3 weeks and were examined for syncytia twice a week, as described by Koot and colleagues [23].

CD4⁺ Cell Counts

CD4⁺ cell counts were done at weeks 0, 2, 4, 8, 12, 16, 24, 32, 40, and 48 by the Stanford University Blood Bank (a certified member of the National Institute of Allergy and Infectious Diseases, Division of AIDS, the CD4⁺ T-cell quality assurance program for the AIDS Clinical Trial Group).

Statistical Analysis

A two-sided t-test was used to compare changes in virus burden between patient groups. A standard one-sample, two-sided Wilcoxon test was used to compare differences in the change of slopes of CD4⁺ T cells (slopes of CD4⁺ cells before the mutation compared with slopes of CD4⁺ cells after the mutation) for patients developing a mutation at codon 74. The slope difference test analysis pertained to the slope difference statistic, defined to be the difference in fitted slope of the CD4⁺ counts after compared with before the mutation at codon 74. Thirty-four of 38 patients in the mutation group had sufficient data to allow computation of the slope difference statistic.

We compared the slope difference statistic values for the 34 mutants with the corresponding values from the wild-type group in order to see if they were more negative in the mutant group. This is not possible directly because the wild-type group by definition has no mutation time from which to define "before" and "after." Instead, artificial slope difference values were constructed as follows: 1) a random participant was chosen from the 25 in the wild-type group; 2) a random mutation time was chosen from the 38 such times in the mutation group; and 3) if the mutation time was less than the last observation time for the randomly selected participant, then the wild-type slope difference value was set equal to the slope difference for that participant using that mutation time to define before and after. If the mutation time was not less, then we started over at step 1. One hundred fifty-two values were constructed in this way. A two-sample, two-sided Wilcoxon test was used to show that the 34 mutant values were statistically significantly more negative than the 152 wild-type values. A two-sided chi-square test was used to analyze the differences in the frequency of the mutations in the patients receiving different dosages of didanosine.

Results

Top Methods Results Discussion Author & Article Info References

Baseline Characteristics

Sixty-four patients who had received zidovudine and were switched to didanosine monotherapy were evaluated for the presence of codon 74 and 215 reverse-transcriptase mutations at serial time points (baseline through week 48). The mean (±SD) CD4⁺ T-cell count at the start of didanosine monotherapy for all patients was 129 ±88 CD4⁺ cells/mm³ (median, 105 CD4⁺ cells/mm³). The mean (±SD) time on didanosine monotherapy for all patients was 10.5 ±5.7 months (median, 9 months).

Reverse-Transcriptase Mutations

At baseline, 54 of 64 (84%) patients had a mutation at codon 215, whereas 0 of 64 (0%) patients had a mutation at codon 74. By week 24 of didanosine monotherapy, 36 of 64 (56%) patients had developed a mutation at codon 74. At week 24, the number of patients who had a mutation at codon 215 in serum HIV RNA had decreased from 84% to 59% (38 of 64 patients); 16 patients previously mutant at codon 215 in their serum HIV RNA now only had wild-type sequences (a 25% difference, 95% lower CI, 15%; P < 0.0001 by the Fisher exact test). The rate of mutation at codon 74 and codon 215 at weeks 0, 8, 12, 16, and 24 is shown (Table 1).

A greater proportion of patients receiving the highest dose of didanosine developed the codon 74 mutation, but this difference was not significant; thus, 54% (20 of 37) of patients receiving didanosine, 200 mg/d or 500 mg/d, developed the mutation, whereas 70% (19 of 27) of patients receiving didanosine, 750 mg/d, developed the mutation (P = 0.18, two-sided chi-squared test). As shown in Table 1, the most common genotype in the 64 patients at the final time point for each patient before they went off study was that of a mutation at codon 74 with a persisting codon 215 mutation (74 MUT/215 MUT), followed by patients remaining wild type at codon 74 and wild type at codon 215 (74 WT/215 WT), and then patients with combinations of wild type and mutant genotypes (74 MUT/215 WT and 74 WT/215 MUT). No significant difference was found between patients with syncytium-inducing and non-syncytium-inducing isolates with respect to the presence of the codon 74 mutation in HIV serum RNA: In patients with syncytium-inducing isolates, 67% (6 of 9) had the codon 74 mutation, whereas 56% (9 of 16) of patients with non-syncytium-inducing isolates had the codon 74 mutation.

CD4⁺ T-Cell Changes Related to the HIV Reverse-Transcriptase Genotype

Changes in patients CD4⁺ T-cell counts were analyzed to determine the relation to the codon 74 mutation. The starting CD4⁺ T-cell counts were not significantly different for the 25 patients who remained wild type at codon 74 compared with the 39 patients who developed a mutation at codon 74 (mean [±SD], 134 ±102 CD4⁺ cells/mm³ compared with 125 ±78 CD4⁺ cells/mm³; P > 0.5). Almost all patients (63 of 64) had follow-up CD4⁺ T cell counts available for analysis; the remaining patient was noncompliant and did not have CD4⁺ T-cell counts available. The patients who had mutations as early as 8 to 16 weeks after starting didanosine had lower starting CD4⁺ T-cell counts than the patients who did not have mutations until week 24 or after (mean [±SD], 85 ±58 CD4⁺ cells/mm³ compared with 153 ±79 CD4⁺ cells/mm³; P < 0.01).

Figure 1 shows the pattern of CD4⁺ T-cell changes before and after the appearance of the codon 74 mutation in the 38 patients receiving didanosine who developed this mutation. Most of the patients (34 of 38) who developed the codon 74 mutation had CD4⁺ T-cell values available after developing the mutation; 4 patients developed the codon 74 mutation at the last evaluable time point before they stopped taking didanosine and had no further CD4⁺ T-cell values. The slopes of the CD4⁺ T-cell changes were significantly more negative after the codon 74 mutation compared with before the mutation (a median greater decline of –1.54 CD4⁺ cells/wk; CI, –2.68 to –0.71;P < 0.002). These slope values were also compared with randomized slope differences generated by superimposing a mutation time from the mutant group on a randomly chosen case history from the codon 74 wild-type group; 152 values were compared. Patients with the mutation had greater decreases in their CD4⁺ T-cell counts after developing the mutation (significantly more negative slopes) than did patients who remained wild type during the same time period (P < 0.001).

View larger version (15K): [in this window] [in a new window]   Figure 1. Mean CD4+ T-cell changes before the appearance of the HIV reverse-transcriptase mutation at codon 74 and CD4+ T-cell changes after the appearance of the mutation in 38 patients switched from zidovudine to didanosine. The point of the codon 74 mutation (measured in the serum HIV RNA of the patients) is at time 0 and is the reference point for all other CD4+ T-cell counts to be compared; x-axis, –24 to 0 represents the time receiving didanosine before the codon 74 mutation and 0 to 24 represents the time receiving didanosine after the mutation. The y-axis represents the mean CD4+ T-cell change in relation to the CD4+ T-cell level at the time of the mutation. The number of patients evaluable are above each time point.

HIV Genotype Related to Changes in Serum Virus Burden

Serum virus burden determined by quantitative PCR was done at week 0 and week 24 for 58 of 64 patients (Table 2). The remaining 6 patients did not have enough serum left after mutational analysis for quantitative testing. Patients who developed the codon 74 mutation had a higher mean virus burden at initiation of didanosine therapy (83 000 HIV RNA copies/mL serum [CI, 51 900 to 133 000 copies/mL serum] compared with patients who remained wild type at codon 74 (50 100 HIV RNA copies/mL serum; CI, 24 200 to 103 000), a 1.6-fold greater virus burden, although the difference was not significant (CI, 0.7 to 3.94; P = 0.20). Patients who developed a codon 74 mutation had a greater serum virus burden after 24 weeks of didanosine than did patients who remained wild type (a 2.7-fold greater virus burden; CI, 1.26 to 5.89; P = 0.01). Although patients with the codon 74 mutant had a greater increase in their serum virus burden during the first 24 weeks of didanosine than did patients remaining wild type at codon 74, the magnitude of the difference was not significant (Table 2).

Discussion

Top Methods Results Discussion Author & Article Info References

It has been established that five HIV reverse-transcriptase gene mutations (at codons 41, 67, 70, 215, and 219) can confer resistance to zidovudine [24, 25]. Data from several groups [25-27] indicate the relative frequency and pattern of development of these mutations in patients receiving zidovudine monotherapy. It has become important to investigate the frequency of didanosine resistance because more patients who have taken zidovudine are being switched to didanosine. Previous research has established that the mutation at codon 74 appears to confer the greatest degree of resistance to didanosine (a 5- to 26-fold decrease in susceptibility [9]), and this mutation has been shown to provide a replication advantage to HIV-1 compared with that of wild-type viruses in vivo and in vitro (cell culture experiments) in the presence of didanosine [13, 14].

Because of these reports, we investigated the frequency and pattern of development of the codon 74 mutation in 64 patients who had previously taken zidovudine and who had been switched to didanosine monotherapy. The codon 74 mutation was detected in serum HIV RNA from these patients as early as 8 weeks after starting didanosine; by week 24 of didanosine treatment, 56% of the patients had developed a mutation at codon 74. All of the patients had previous zidovudine therapy; at entry into the study, 84% of them had a mutation at codon 215 (a mutation known to confer a 16-fold decrease in zidovudine susceptibility [18]). After 24 weeks of didanosine treatment, the number of patients with a codon 215 mutation had decreased to 59% (that is, 16 patients who had had mutant sequences at codon 215 in their serum HIV RNA were now wild type at this codon). This transformation from mutant to wild type may have been the result of a reversion of the reverse-transcriptase gene mutation back to the wild-type sequence or a more likely possibility is that once the pressure of zidovudine was removed, there was re-expression and release of the wild-type virus into the serum from the large reservoir of different viral quasi-species infecting a given patient.

Zidovudine resistance occurs more rapidly in patients with decreased CD4⁺ T-cell levels [28], and the mutations that cause zidovudine resistance are associated with a greater virus burden [5]. Both of these findings for zidovudine resistance also apply to codon 74 mutations for patients receiving didanosine. Patients with mutations at codon 74 after just 8 to 16 weeks of didanosine therapy had decreased starting levels of CD4⁺ T cells when compared with patients who did not develop a codon 74 mutation until week 24 or later. All patients who developed the codon 74 mutation tended to have a higher serum virus burden at the start of therapy and had a greater serum virus burden (statistically significant) after 24 weeks of didanosine therapy than did the patients who were wild type at codon 74. This finding may reflect that a patient with a greater virus burden is likely to generate mutations faster given the chance for more replicative events than does a patient with a low virus burden.

The most common genotype found in the 64 patients during the 48 weeks of didanosine treatment was the combination of a codon 74 mutation with a preexisting zidovudine resistance mutation at codon 215 (in 44% of patients). This subgroup of patients also had the highest serum virus burden after 24 weeks of didanosine therapy (mean, approximately 275 400 HIV RNA copies/mL serum). The high frequency of a HIV genotype with codon 74 and 215 mutations suggests that although this combination would not be favored by the virus in the presence of zidovudine because it can restore zidovudine susceptibility [9], it is readily selected for patients receiving didanosine monotherapy. The selection of this combination by the virus and its association with a high virus burden may be related to the findings of Eron and colleagues [14] and St. Clair and colleagues [9]. These researchers, using site-directed mutagenesis, showed that the presence of zidovudine resistance mutations in the HIV reverse transcriptase in combination with the codon 74 mutation may confer an even greater degree of didanosine resistance than does the codon 74 mutation alone. Thus, viruses with this combination of mutations are likely to have a selection advantage in a patient receiving didanosine monotherapy.

The development of the zidovudine resistance mutation at codon 215 has been reported [5, 8] to be predictive of decreases in CD4⁺ T cell counts in asymptomatic patients receiving zidovudine. In this study, we evaluated the effect of a codon 74 mutation on decreases in CD4⁺ T cells and virus burden in patients with advanced disease who were receiving didanosine monotherapy. Patients with the codon 74 mutation were found to have a statistically significant greater decrease in their CD4⁺ T cell counts after the development of the codon 74 mutation when compared with patients who remained wild type at codon 74 during the same time period. Thus, a temporal relation appears to exist between the development of the codon 74 mutation in the serum HIV RNA of these patients and a more severe subsequent decrease in CD4⁺ T cells when compared with patients who remain wild type at this codon while receiving didanosine monotherapy.

In patients with very advanced immunodeficiency, the development of a new AIDS diagnosis or the occurrence of death may be more a function of opportunistic infection and chance than of drug resistance [29]. Perhaps a better population in which to study the effects of drug resistance would be patients who still have at least partial immune function; thus, when the suppressive effect on viral replication is lost because of drug resistance, the resultant effect on the immune system can be better studied and fewer confounding factors would be present. Although the codon 74 mutation was associated with a greater rate of decrease in CD4⁺ T cells, some patients who remained wild type at this codon also had a decrease in CD4⁺ T cells. In patients remaining wild type at codon 74, this decrease in CD4⁺ T cells may have been caused by didanosine resistance conferred by other HIV mutations not looked for in this study or caused by other virologic characteristics known to be associated with disease progression, such as the syncytium-inducing phenotype [30, 31]. We could not control for the syncytium-inducing phenotype in our study because only 25 of 64 patients had peripheral blood mononuclear cells available for HIV biological phenotype testing. Patients with syncytium-inducing isolates had similar rates of the codon 74 mutation in their serum HIV RNA (67%) when compared with patients with non-syncytium-inducing isolates (56%).

In patients infected with HIV, zidovudine resistance [3-8] and HIV resistance to non-nucleoside reverse-transcriptase inhibitors [29, 32-34] have been shown to be associated with and predictive of decreases in CD4⁺ T cells and disease progression. Despite the size of this study and the advanced disease stage of the patients at the start of didanosine therapy, an association was found between the development of the didanosine resistance mutation at codon 74 in the serum HIV RNA of these patients and a subsequent accelerated decrease of CD4⁺ T cells when compared with patients without the mutation who were receiving didanosine monotherapy. In addition, patients with the mutation were found to have a greater serum virus burden (statistically significant) than the patients without the mutation after 24 weeks of didanosine. This mutation should be studied in a larger number of patients with less advanced disease to determine if it is a marker for disease progression in patients receiving didanosine monotherapy. The rapid development of the codon 74 mutation and its occurrence in most patients receiving didanosine monotherapy in this study support previous findings [9, 13, 14] that this mutation confers a survival advantage for HIV in vivo.