This is a report of abstract presented at the 3^rd International Workshop on Drug Resistance & treatment Strategies. These were selected because I feel they are the most clinically relevant. Additional reports may follow after further review of the presented abstracts.

John Erickson, a well-known researcher at the National Cancer Institute has been in protease inhibitor research from early years and has been reporting for several years at AIDS conferences that he wanted to develop a protease inhibitor specifically designed to be active against viruses containing primary PI resistance mutations. At the Resistance Workshop he discussed RS-344, a PI designed using structure-based approaches with a novel flexible core to target some of the active site mutants. Erickson reported that RS-344 showed similar potency to currently available protease inhibitors in both enzymatic and cell culture assays. He showed a table of clinical isolates with high-level phenotypic resistance to multiple protease inhibitors. In every case these clinical isolates had either <4-fold or 4-10 fold resistance to RS-344. These clinical isolates had from 8 to 18 mutational changes. Erickson said he thought RS-344 looked good against these highly resistant viruses.

Bharat Ramratnam, with the Aaron Diamond AIDS Research Center, reported levels of replication competent HIV in the latent reservoir of resting memory CD4 cells decayed in 17/25 individuals infection with an overall mean half-life of 49 weeks (range 17-194 weeks). Individuals had chronic (n=20) and acute (n=13) infection, and individuals had been infected from 4 months to 10 years. They were treated with a 3 or 4 drug PI HAART regimen for at least 12 months. The baseline average CD4 was 421 (117-954 range) and viral load was 190,000 copies/ml (4,023 to 1.6 million copies/ml). No decay was observed in 7/32. Intermittent viremia was seen in 17/25 with decay and 7/7 without decay suggesting that individuals with intermittent viremia may be less likely to show decay in this compartment. He said individuals with intermittent viremia had slower decay compared to individuals consistently <50 copies/ml (half-life 59 weeks vs 26 weeks). Ramratnam said replication competent

As you may know some researchers have said HIV in this compartment will not decay leaving a potentially permanent pool of HIV which, they think, could create an insurmountable obstacle to eradication. However, researchers from Aaron Diamond have reported seeing decay in this compartment in a highly selected group who were treated during acute infection and consistently remained below 50 copies/ml for 2-3 years without interruption. A number of researchers have commented that this reservoir is established shortly after infection and remains stable, although they admit ongoing replication may feed this reservoir. Some researchers think that the reservoir decays but is replenished by ongoing replication. They think that more potent treatment might stop the ongoing replication.

Joe Wong, from this research group at UCSD, has previously reported that HIV RNA persists in lymph nodes at low levels from patients on potent anti-retroviral therapy with undetectable plasma viremia (<50 copies/ml for up to 2 years). But, Gunthard said it is not known whether such residual RNA represents continuous replenishment by chronically infected cells or whether it represents trapped virus from the time of initiation of therapy.

Lymph node biopsies from 6 patients with the San Diego Cohort of the Merck 035 study were obtained at year 1 (n=3) and at year 2 (n=6) after initiation of therapy. Evolution of V3 sequences was examined by comparing pairwise distances between baseline plasma and year 1 and 2 LN sequences.

Huldrych Gunthard reported that he found from this small subset of patients that for patients who had rapid initial RNA decay and undetectable (<50 copies/ml plasma viremia) for up to 2 years, there was no evolution of the V3 env region detectable in sequential RNA extracts from lymphatic tissue. Gunthard said this suggests that such residual RNA represent trapped virus rather than the product of ongoing replication. However, he reported that in patients with slower RNA decay and intermittent low level plasma viremia, viral evolution in the lymph node can be detected and correlates with the magnitude of viral replication.

Sabine Yerly, with the Geneva University Hospital, reported on a study to look at whether the time of initiation of HAART in the course of HIV infection affects the levels of plasma viremia. She found that initiation of anti-retroviral therapy during primary or acute infection and while CD4 cell counts are high leads to a stronger suppression of residual viral replication. But the clinical benefits of this achievement are uncertain.

Viremia was measured using an assay with a detection limit of 3 copies/ml in 3 groups of patients. HAART was initiated in drug-naïve patients during primary infection (n=10), during chronic infection (n=10) "without immune suppression" (CD4s >500, median 577 CD4s), and "after immune suppression" developed (CD4s <500; median 113; n=21). Individuals received HAART for at least 72 weeks. Adherence was estimated to be >90%. Median baseline CD4 was 418 (range 131-806) and plasma viral load was about 26,000 copies/ml (range about 600 to 537,000 copies/ml) in the primary infection group. In the >500 CD4 group, the median baseline CD4 was 577 (range 500-743), and the plasma viral load was about 10,470 copies/ml (range about 117 to 33,800 copies/ml). In the <500 Cd4 group, the median baseline CD4 was 113 and viral load was about 117,000 copies/ml. In the >500 CD4 group the median time between infection and initiation of treatment was about 3 years (range about 1 to 6 years). In the <500 CD4 group, the median time from infection to initiation of treatment was about 5.7 years (range about 2.5 yrs to 11.5 yrs).

Overall, 250 blood samples were collected and analyzed (median of 6 samples per patient) between 24 and 120 weeks after therapy started. In the primary infection group 75% of the samples for the 10 patients were <3 copies/ml, all samples were <100 copies/ml, and 5/10 maintained viremia <3 copies/ml during the follow-up. Many more samples in this group were <3 copies/ml than in the other two groups.

In the >500 CD4 group, viremia was >3 copies/ml in the majority of samples, the highest viral load was 50 copies/ml, only 32% of the samples were <3 copies/ml (compared to 75% in primary infection group). None of the patients persistently maintained viremia <3 copies/ml during the follow-up.

Higher viremia was measured in the group with <500 CD4s with the highest viremia being 400 copies/ml. Eighty percent of the samples were <50 copies/ml and 8% of samples had <3 copies/ml, but none of the patients had persistently maintained <3 copies/ml during the follow-up.

The primary infection group of patients had lower residual viremia than chronically infected patients with >500 CD4s, and patients with >500 CD4s had lower viremia than those with <500 CD4s.

These preliminary data show the initiation of HAART during primary infection, high baseline CD4 count (per 100 cell increase), low baseline plasma viral load, and shorter duration of delay between infection and starting therapy were all statistically significant predictors of low residual viremia. Only starting therapy during primary infection and baseline CD4 count (per 100 cells increase) were independent predictors of mean residual viremia. I think she said that when looking at the association between residual viremia and baseline viremia, there was a difference between those treated during primary infection and those treated when CD4s were <500, but not between those treated during primary infection and when CD4s were >500.

Yerly assessed cell associated RNA and DNA in the 5/10 patients treated during primary infection who had <3 copies/ml after 18 months of treatment. The detection limits for RNA were 3 copies/ml and 5 copies/ml for DNA. Two patients had cell-associated undetectable RNA after 6 months of treatment. One other patient’s cell associated RNA declined to undetectable prior to 6 months, rose to detectable at 12 months and then slowly declined again to undetectable by about 20 months. Only the 2 patients with persistent undetectable cell associated RNA from 6 months to end of follow-up- one patient’s follow-up was 24 months and the other was 45 months-- had a clear decline in call associated DNA resulting in an undetectable level. One of the 2 patients reached undetectable DNA at month 24. The other patient was treated within 1 month of infection, and had a low viral load when starting therapy. Plasma viral load, cell associated RNA and DNA were all below detection for more than 30 months. The DNA and plasma viral load went undetectable at about 5 months, and the cell associated RNA went undetectable at about 15 months after starting therapy. After 45 months the patient decided to stop therapy. Four weeks later plasma viral load was detected at 30 copies/ml, and 1 week later it was 80 copies/ml. Therapy was re-started and viral load was undetectable within 1 week. For the other 3 patients only a slow 2^nd phase decay in cell associated DNA was observed between 3 and 24 months, without reaching undetectable.

Despite cell associated RNA and DNA declining to undetectable in 2/5 patients the other 3 patients were not able to reach undetectable for these parameters.

Author’s summary: These preliminary results show that initiation of therapy at the time of primary infection and while CD4 counts are high results in lower residual viremia. But we don't know if such lower viremia predicts clinical benefit. In patients treated during primary infection, a persistence of cell associated RNA and DNA was generally observed despite sustained plasma viral load below 3 copies/ml for more than 18 months. In these patients a slow 2^nd phase decay of cell associated DNA was seen during 3-36 months after starting therapy.

Felipe Garcia, with the Hospital Clinic at the University of Barcelona, reported follow-up data on this group of patients first reported on at the ’99 Human Retroviral Conference in February, which was reviewed in the April NATAP Reports newsletter available on the NATAP web site. Interrupting and stopping therapy has become an interesting therapeutic approach because of the difficulties in maintaining long-term adherence to HAART, and because eradication from HAART appears elusive. With the waning of hope for eradication, there has been interest that an immune system response to HIV could be elicited by interrupting therapy, which might lead to control of HIV by the immune system. There was a lot of discussion about this at the Retrovirus Conference. The April NATAP Reports contains comprehensive coverage of the discussions at Retrovirus on interrupting therapy. A hard copy of the newsletter is available by calling NATAP at (212) 219-0106.

Garcia’s data is interesting but there are a couple of cautions. He warns as others do that individuals who want to consider interrupting therapy do so only in a study-controlled situation. The ACTG is planning such a study now. Other researchers are planning such studies. Second, the individuals who interrupted therapy in the study described below all started HAART in the earlier stages of their HIV disease when their viral load was <20 copies/ml and their CD4s were above 500. Another serious concern about stopping or interrupting therapy is if a person’s HAART regimen contains efavirenz, nevirapine or any antiviral with a long half-life. Efavirenz has a long half-life of 40-50 hours. This means it remains in the blood for several days after you stop taking it. This is a reason for its once daily dosing and likely a reason for its potency. If you stop all drugs in a regimen that includes efavirenz, the other drugs will leave your blood quickly but efavirenz levels will remain in your blood as its blood levels decline more slowly. This creates a concern that resistance to efavirenz could develop during that period of declining efavirenz blood levels, and viral replication while there is little viral suppression from other drugs.

Interrupting therapy may eventually be a viable approach but a number of questions remain unanswered including how to do it, what is the effect of re-seeding tissue where viral load was lowered due to HAART, and what happens to the immune system over the longer term after interrupting therapy.

Garcia reported data after two therapy interruptions. Ten patients were initially included in the study with CD4 counts >500. They started therapy with D4T+3TC and either ritonavir or indinavir. After 1 year of treatment patients had a plasma viral load <20 copies/ml, volunteered to stop therapy and had been below 20 copies/ml for at least 8 months.

The study methods are as follows. Patients initially stopped therapy after 1 year of therapy (day 0) when viral load was <20 copies and re-started therapy after 4 weeks if VL >200 copies/ml. They received treatment for 6 more months. Then if viral load was <20 copies/ml the patient was offered to stop therapy again (week 28). Treatment was re-introduced (week 32) after 4 weeks if viral load was >200 copies/ml and if viral load did not drop spontaneously (without HAART). The follow-up is just beyond the 32-week point. They are planning to give patients 6 months more therapy followed by a third interruption.

Measurements of plasma and CSF and lymphatic tissue viral load, CD4 lymphocyte proliferative responses to mitogens and specific antigens (including HIV-1 Ag), and lymphocyte immunophenotyping were performed weekly. Genotypic resistance was assessed after the rebounds of viral load and at baseline.

Just prior to the first stop (week 0), plasma viral load was <20 copies in all cases and below 5 copies/ml in 7/10 patients. CSF viral load was <20 copies/ml in all patients. Among the 6 patients with accessible tonsils, tonsillar tissue VL was <40 copies/ml of tissue in 5 and 485 copies/mg in the other patient. So this was a well suppressed group of patients. Nine patients reached the time point of the second stop. At week 28 (2^nd stop), plasma VL was <20 copies/ml in all cases and below 5 copies/ml in 7/9.

A rebound in plasma VL occurred in all cases, with a mean doubling time (DT) of 2.32 days in the first stop and 3.38 days in the second stop. Doubling time is the amount of time for viral load to increase 100%. In 2 patients the DT increased from 2.48 to 8.68 and from 3.85 to 8.65 days, respectively. DT did not change in 7/9 patients who completed both stops. Four patients’ viral load rose higher than their baseline before starting therapy, both at the first and second stop.

After the first stop, one month after re-starting therapy VL dropped below 20 copies/ml again in all patients and remained <20 copies/ml after 6 months until the 2^nd stop.

However, after the second stop 4 patients’ viral load rebounded to similar levels of their baseline level but then dropped spontaneously (without therapy). The drops ranged from -0.80 log to -2.09 log. A proliferative p24 response was detected in 2 of these 4 only after the second stop.

Although viral loads rebounded after both stops they did not detect RT and protease mutations known to be associated with resistance to RT inhibitors or protease inhibitors. They used the Visible Genetics System for genotyping.

Changes in Immune System. At the beginning of the first stop, the patients’ CD4s, naïve CD4s and CD8s decreased. There was an increase in CD38+CD8 cells and of both memory CD4s and memory CD8s. There was no proliferative p24 Ag response detected in any patients. Garcia reported that 6 months after restarting therapy patients regained these losses in the immune system back to where they were prior to the stop. After the second stop CD4s decreased but CD38+ CD8 cells did not change and the naïve and memory cells did not change. However, a proliferative p24 response was detected in 2/10 patients. Again, these 2 were 2 of the 4 patients whose viral load dropped spontaneously after the second stop.

In summary, Garcia concluded by saying that in a cohort of well controlled (virologically) asymptomatic patients in very early stages of disease after 1 year of HAART—

1. viral load rebounded in all cases
2. the amount of time it took for VL to rebound increased 3-fold in 2/9 patients
3. following 2 consecutive planned interruptions VL dropped spontaneously in 4 of 9 patients
4. 2 of these 4 developed a CD4 T-cell proliferative response to p24 antigen
5. he characterized interruption as being relatively safe—lack of resistance, quick virological response after re-introducing HAART, and complete recovery of immune deterioration after 6 months of HAART
6. lesser immune deterioration after second stop
7. these data could be a proof of concept about the likelihood of inducing an immune response against HIV antigens in chronic infection with intermittent therapy
8. Garcia recommended not to experiment with therapy interruption outside of a study

However, Garcia does not address the concern about re-seeding tissue or areas of the body where HIV may have been greatly reduced by HAART. These areas may include the CSF, lymph tissue, genital tract and secretions, and the brain.

I think this is an important study because the study data suggests that the longer a person remains undetectable the chances of viral rebound decrease. With all the headlines in the press screaming about the likelihood of patients losing their viral load suppression many patients get discouraged. This study suggests that if you keep your viral load undetectable for a year and remain adherent long term viral suppression may occur.

Andrew Phillips, with the Royal Free and University College Medical School in London, reported on this study of 406 drug-naïve patients starting a HAART regimen including protease inhibitor or NNRTI therapy with 2 NRTIs. The median baseline viral load was 250,000 copies/ml (range 830-5 million). The median CD4 was 259. Using a sensitive viral load assay the median VL decline was 4.8 logs by week 48. This method permitted characterizing the full VL decline, which is normally cut off by detection limits of the assays usually used. Ninety-one percent reached VL <500 copies/ml by 24 weeks. Of those who achieved a VL <500 (n=342), 69 had a viral rebound (2 consecutive values >500 copies/ml) in 344 person years of follow-up. There was a fairly low overall rate of rebound. By week 48 21% experienced a VL rebound after initial suppression below 500 copies/ml. However, the rate of VL rebound decreased as only 26% experienced rebound at the 96 week time-point.

Ninety-seven percent of patients who reduced their VL <500 copies/ml had at least 1 HIV-RNA value <50 copies/ml. Statistical analysis (Cox regression analysis) indicated that there was a significant decrease in the rate of viral rebound with increasing length of viral suppression. That is, the longer you stay undetectable the longer you are more likely to remain undetectable, if adherent. There was a rate of only 0.8 rebounds per 10 years (8 rebounds in 99 person-years; i.e. average of 1 rebound per 12.4 years) in people who had experienced VL <500 copies/ml for over 1 year.

In conclusion-- Phillips said that although we have not followed patients for more than a maximum of about 3 years this low and decreasing rate of viral rebound in patients with at least 1 year viral suppression implies that, if prolonged complete drug adherence were possible, long-term viral suppression for 10 years and more seems within the capacity of presently available regimens.

A Yvon, with the Pitie Salpetriere Hospital in Paris, reported data from this study which gives some clues in using resistance testing to improve the predictability of response to amprenavir for individuals with protease inhibitor experience. Fifty individuals experienced with protease inhibitors but amprenavir-naïve patients were studied for genotypic and phenotypic resistance. The ABI System was used for genotypic analysis. Phenotypic analysis was performed using recombinant virus assays. Patients were divided into 2 groups based on their susceptibility to amprenavir—sensitive (S), <4-fold increase in IC50 over wild-type; and, resistant R, >4-fold increase in IC50.

The key mutations involved in PI resistance were compared between the 2 amprenavir susceptibility groups. Mutations at 54, 84, and 90 were more frequent in the R group. While, mutations at codons 46, 48, and 82 were present at the same frequency in both groups. None of the 50 patients had an I50V mutation, which is known to reflect resistance to amprenavir. In amprenavir resistant patients, the number of protease gene mutations was higher. This stands to reason to occur in most cases of protease inhibitor resistance. In general, you would want to minimize how many mutations you accumulate. This suggests again, that you should monitor your viral load closely (every 4-6 weeks) and if you are failing your first PI regimen, you should strongly consider stopping the PI and changing the regimen so you don’t accumulate more PI mutations. Ongoing replication while taking a drug results in mutations to that drug accumulating. And, results in cross-resistance to other drugs in the same class of drugs.

The study author concluded that in this study mutations at codons 54 and 84 are predictive for amprenavir resistance in the case of intensely pretreated patients.

R Zierman and others with ViroLogic, located in South San Francisco, reported this study data. ViroLogic developed a phenotypic resistance test which recently became available to the public. The phenotypic and genotypic resistance of 200 viruses from patients failing therapy with indinavir or nelfinavir were evaluated. An increase in susceptibility (2.5 to 10-fold) to amprenavir was found in 20 viruses. The most pronounced increases in susceptibility were strongly associated with the presence of the N88S protease mutation. I think this information may be a little preliminary to rely upon but is worthwhile knowing.

K Soderbarg, with the Professional Genetics Lab in Uppsala, reported on this study whose objective was to determine whether drug resistance mutations in the RT and protease genes could be found in 7 patients which, during ongoing drug therapy, show minor increases in viral load.

Viral RNA was isolated from samples with as few as 50 copies/ml, after which PCR amplification was performed. Mutations that are likely to cause drug resistance were detected in all patients. These mutations were detected early in the rebound process. Viral loads were as low as between 50 to 500 copies/ml. But in samples with as few as 50 copies/ml, it was possible to detect mutant viruses. The authors concluded mutations causing drug resistance are possible to detect in samples with very low viral load. Therefore, ongoing replication may result in further development of resistance. Some researchers question the reliability of mutations observed in individuals with viral load <50 copies/ml.

Rob Lloyd is with Applied Sciences Lab and Visible Genetics Inc. Visible Genetics Inc is located in Toronto and manufactures a genotypic resistance test. Applied Sciences is in Norcross, Georgia. It has generally been accepted that if a person’s viral load were above 1000 copies/ml a lab would not be able to test their blood sample for resistance. In fact if a person’s viral load is above 1000 copies/ml some labs will not accept or discourage submission of the blood sample. But many individuals with <1000 copies/ml of VL would I think justifiably like to test themselves for resistance.

Lloyd reported he detected genotypic mutations from actual samples of individuals in this study. He said sequence results demonstrated reproducible and accurate genotypes between duplicate samples at low viral loads down to 60 copies/ml. In speaking with Lloyd he has told me he can detect genotypic resistance mutations when viral load is below 50 copies/ml. In all fairness this data ought to be considered preliminary and should be subjected to confirmatory studies. A few other researchers expressed doubt or uncertainty that you could in fact reliably and reproducibly detect genotypic mutations when viral load is very low (<50 copies/ml). If you have a low viral load and request genotypic resistance testing, I would be cautious in interpreting and relying on the results. One way to address this question would be to submit 2 or more samples to the same lab and see if the results are the same. If your viral load is <50 copies it may be irrelevant to perform resistance testing because you are not going to change your regimen. But if your viral load is several hundred to 1000 copies/ml detecting resistance may help in deciding to modify your regimen and what drugs to use.