now.jpg (1640 bytes)

January 1998 Volume 1, Number 3-4, Section 10

NATAP Resistance Supplement

Part I

In October ‘97, NATAP held a Resistance Forum at NYU Medical Center. The morning session was a discussion of drug resistance. The afternoon session was a discussion on genotypic and phenotypic resistance testing. This first report reviews the discussions by 4 expert speakers in the morning session on drug resistance. A second report on the afternoon session is being prepared.

1. Introduction by Dr. Roy Gulick, NYU/Bellevue Medical Center
2. Nucleoside Resistance by Dr. Daniel Kuritzkes, University of Colorado
3. Protease Inhibitor Resistance by Dr. Jody Lawrence, Stanford University
4. 1592U89 Resistance by Dr. Marty St Clair, Glaxo Wellcome

The NATAP Resistance Forum was organized to discuss resistance and cross-resistance to drugs used for treatment of HIV, because resistance can be the most important reason for failure of a therapy; therefore, it is crucial to try and prevent it, and in many cases it is preventable. We will discuss ways in which you can act to help prevent resistance.

In previous years when treatment consisted of AZT, ddI or d4T monotherapy treatment choices were so limited that the development of resistance was not important because you could not prevent it for a long period of time. Although using two nucleosides in combination such as AZT/3TC or d4T/3TC could delay resistance better than monotherapy, it was still limited. Now, by using potent 3 or 4 drug combinations, viral load can be suppressed to below detection, and resistance can be prevented or at least significantly delayed. The suppression of HIV viral load is the key to preventing or delaying resistance; and, the prevention of resistance is the key to keeping viral load suppressed. The suppression of viral load can delay disease progression, that is the delay or prevention of the development of opportunistic infections and the preservation of your immune system. Following is a detailed reproduction of each speaker’s discussion, which can be very helpful for you in understanding and preventing resistance.

Introduction by Dr. Roy Gulick

Dr. Gulick said the new goal of antiretroviral therapy is to suppress HIV replication as much as possible for as long as possible. The achilles heal of therapy, up until the development of the new treatments, has been the development of drug resistance. A simple definition of resistance is a change in the virus such that it is allowed to grow in the presence of an antiretroviral drug. A loss of drug sensitivity or susceptibility defines resistance. An increase in the viruses’ fitness occurs. Fitness means the ability of the virus to duplicate itself in the presence of drug defines drug resistance.

Although many people are doing very well with the new treatments why is it failing others? There are multiple reasons why not everyone is responding well:

• Baseline resistance and cross resistance which will be the focus of today’s discussion.
• The use of less potent regimens.
• Sequential monotherapy, which can mean starting one drug, adding a second drug weeks or months later and then adding a third weeks or months after that; adding them in piecemeal does not constitute triple drug therapy.
• Non-adherence or non-compliance
• Drug levels in blood can vary between individuals; drug interactions can effect blood levels of a drug.
• Tissue reservoir penetration by a drug such as the brain and genital tract; a drug may not suppress viral load in the brain as well as the blood, and consequent resistance could develop in the brain.
• Other unknown reasons.

Despite this long list of reasons, many of them lead to the development of resistance; and so, resistance is the main culprit. Resistance is the reason that many people are not doing well on today’s new treatments. Today we’ll be talking about baseline resistance, which refers to having resistance before even trying drugs, and cross resistance. Cross resistance refers to when a person takes one drug and fails, has a rebound in viral load, may not be very sensitive or susceptible to a second drug even if they’ve never taken the second drug before. Taking one drug will diminish the effect of other drugs in the same class taken down the road. Resistance to a nucleoside will not cause crossresistance to a protease inhibitor. But resistance to one protease inhibitor can cause cross resistance to another protease inhibitor.

It helps to understand resistance by understanding how HIV duplicates itself. In the HIV Life Cycle diagram, HIV is over on the left and the human T-cell to the right of it. HIV specifically recognizes the T-cell and binds to it through a couple of different proteins, which have been relatively recently identified. It loses its outer protein coat and its inner protein coat and exposes its genes in the form of RNA inside the host T-cell. It brings with it and makes its own protein called reverse transcriptase (RT). RT changes the viral RNA into viral DNA which develops a complex which enters into the nucleus of the cell and integrates or combines with the cell DNA. At the point of integration the cell is infected for the life of the cell. Although research is addressing this issue, we know of no way yet of removing the viral DNA once it has integrated.

After a possible period of being quiet the process turns back on and DNA is used to make RNA. RNA is used to make proteins which all assemble and are specifically cut by an enzyme or protein called the HIV protease. After that occurs the virus is fully mature and buds off from the cell. Each infected T-cell can make up to a thousand or perhaps many thousands of new viral particles which have bud off from the cell. Many of these will be fully capable of starting the process all over again and infecting other T-cells. After enough of these viral particles have budded off from the cell, the membrane or the outer coating of the T-cell is compromised and the cell dies. After enough cells die the immune system becomes compromised and people get sick.

A goal of understanding the process by which HIV reproduces itself is to try and figure ways to interfere with it. Most of the drugs that we have either interfere with the reverse transcriptase step, and we call them reverse transcriptase or RT inhibitors, or the step very late in the life cycle where the proteins are cut by the HIV protease, and drugs which work at that step, and we call them protease inhibitors.

There are research efforts ongoing to develop other drugs that are targeted to work at the initial step of recognition and binding by a virus particle and the T-cell, and at the step of integration (integrase inhibitors). Research for these drugs is not very far along in development. Today’s discussion will focus on the available RTIs and protease inhibitors.

Why Does Resistance Occur?

It has to do with factors that we’ve come to understand only in the last couple of years. Most individuals have a high viral burden (load). It is believed that a person has billions of viral particles (viruses) being produced and cleared every day. So, there is an enormous amount of duplication (replication) of the virus on a daily basis throughout the whole course of HIV infection. There is also a high rate of mutation. Every time HIV duplicates itself it makes errors or mistakes in the genes. On average, it makes one or several mistakes every time it duplicates its own genes. And, many T-cells are made and lost on a daily basis.

The many different mutations that occur every time the virus duplicates itself causes genetic diversity. That is, a person is infected with many different viruses, different because they contain different combinations of mutations. Drugs select certain viruses to suppress. The fittest viruses may survive and go on to duplicate themselves, unless you can fully suppress replication including all or almost all viruses. It appears as though an effective triple drug regimen when it is successful is suppressing all or almost all viruses. The fit viruses can more easily evade suppression when therapy is not potent enough to fully suppress viral replication and sustain it for a relatively long period of time. Certainly, that is why viral load will usually rebound relatively quickly after using therapy consisting of one RT inhibitor.

Eleven approved antiretroviral drugs:

• Nucleoside RT inhibitors:

        AZT, zidovudine (ZDV), Retrovir
        ddI, didanosine, Videx
        ddC, zalcitabine, HIVID
        d4T, stavudine, Zerit
        3TC, lamivudine, Epivir

• Non-Nucleoside RT inhibitors:

         nevirapine, Viramune
        delavirdine, Rescriptor

• Protease Inhibitors:

        indinavir, Crixivan
        ritonavir, Norvir
        Fortovase, saquinavir soft gel capsule
        Invirase, hard gel capsule saquinavir
        nelfinavir, Viracept

New drugs in development:

• Nucleoside Analogue RT Inhibitors:

            1592U89 (abacavir)
            adefovir dipivoxil

• Non-Nucleoside RT Inhibitors:

            DMP-266, efavirenz, Sustiva

• Protease Inhibitors:

            142W94, VX-478

Currently, there are 3 ways to test for HIV resistance: genotypic resistance, phenotypic resistance and clinical resistance. In a genotypic resistance test, you are looking at the genes or genetic material of the virus itself. Such a test can be performed before starting a drug(s), while taking a drug(s), or after stopping the drug(s). You are looking to identify changes in the genes (mutations). In using a phenotypic resistance test, a person’s virus can be taken into the lab. They are not simply looking at the genes. The virus is grown in the lab in a culture in the presence of drug to see how much drug may be necessary to suppress the virus. If more drug is required than the amount used in recommended doses, that means resistance has developed. A measure of how much more drug is required than the recommended dose reflects the amount of phenotypic resistance. Phenotypic resistance can occur as a result of changes or mutations in the genes (genotypic mutations).

Clinical resistance is reflected by viral load test results. Most doctors who take care of patients use this approach. After starting a drug regimen viral load can go down below detection and stay down. This might mean there is no resistance developing. After starting therapy, viral load might not go down much or at all, or it might go down and then come back up. This could mean resistance has developed. If a person is completely not taking the drugs prescribed, of course no change in viral load should occur, which means there is no resistance. Clinical resistance may not correlate with genotypic or phenotypic resistance; you may not be able to detect geno- or phenotypic resistance, although clinical resistance or failure occurs.

The use of geno and phenotypic resistance tests may be effective tools in the future, but some people are using them now to try and figure out which drugs may or may not work for a person and which drugs they could successfully switch to.

How do you prevent resistance? You can maximize antiviral activity of the drugs you are taking by selecting the most efficacious or potent combination regimen. These days a number of drugs are available to put together such a regimen. Another way is to maximize the genetic barrier to resistance. Combine drugs which have no overlapping resistance mutations. Generally, choose drugs that need more than one mutation for resistance to develop. Choose drugs that the person has never taken before. Do not treat sequentially. That is, do not add drugs to a regimen one at a time spaced out over a period of weeks or months. When switching a regimen because a person is failing the regimen they are taking, the US Public Health Service Treatment Guidelines recommends changing to at least two new drugs. It is preferable, if possible, to start therapy with 3 new drugs.

Reverse Transcriptase Inhibitor (RTI, nucleoside) Resistance by Dr. Daniel Kuritzkes

There are many causes for failure of drug therapy, as Dr Gulick outlined. Possibly the most important cause today is problems of adherence to therapy which leads to suboptimal treatment which leads inevitably to resistance to drug therapies. Its important to realize the presence of reservoirs (e.g., brain) where drugs may not penetrate well where viruses may find themselves uninhibited or only partially inhibited by the drugs. As well, differences of metabolism within a cell may play an important role in whether a drug is effective or not. The nucleosides are actually prodrugs. The drugs you actually take are not themselves effective against the reverse transcriptase. They must be activated by the cells. They are activated by a process called phosphorylation, in which phosphate groups are added one at a time until there are three of them and the drugs are then in their triphosphate form. It is the triphosphate which is now a nucleotide that is recognized by reverse transcriptase and is able to inhibit reverse transcriptase by competing with the naturally occurring triphosphate nucleotides inside the cell.

Different kinds of cells and cells at different stages in their lifecycle have a different ability to activate these drugs because the enzymes that are responsible for activating the nucleosides are expressed differently in cells that are resting or cells that are activated. For example, we know that AZT is activated or phosphorylated best in T-cells that are in an active state and is phosphorylated less extensively in resting T-cells. Whereas, ddI appears to be phosphorylated equally well in resting and active cells. It is unclear what the clinical relevance of this information is, but it is important to keep in mind that the differences in the ability of the cells to phosphorylate and activate the nucleosides could potentially have some implications.

Perhaps the clearest example we have to date is that drugs that compete for the same enzymes for phosphorylation may antagonize each other in their ability to inhibit HIV. D4T has a very similar structure to AZT and uses the same enzymes as AZT. Preliminary data from ACTG study #290 suggests that using AZT and d4T together does not work well clinically, supporting research from lab studies suggesting antagonizism between these two drugs. Individuals in study 290 who had been on long term AZT and added d4T to AZT had a more rapid decline in CD4 count than those who simply switched to d4T or those who stayed on AZT. The study was stopped as a result and more complete data will be presented at ‘98 Human Retroviruses Conference. This is an example potentially of intracellular metabolism being the cause not of resistance as we understand it, but of metabolic failure.

As Dr. Gulick explained earlier, HIV has a high replication and mutation rate which leads to genetic diversity; because of that, there are mutations pre-existing even before starting any drug therapy. These mutations may by themselves cause resistance to a drug, but they appear rarely. Treatment by a single drug may inhibit virus that do not have mutations while those with mutations may not be inhibited. These mutated viruses may continue to replicate and resistance to the drug may develop over a period of time causing drug failure. The longer this drug continues to be used the more resistance will develop as the mutant viruses replicate more.

If you use two drugs, some of the viruses may be resistant to one drug but not the other and some viruses may be resistant to both drugs. But you’ve created a higher genetic barrier to the development of resistance. That is, resistance to the two drug regimen should develop, but it is more difficult to develop and should occur more slowly. So, you have mutant viruses that may have been pre-existing or might emerge on therapy. Where only a single mutation is needed for resistance to one of the drugs, it is highly likely for the mutation to be present. If more mutations are necessary for resistance to a drug it is much less likely that there are pre-existing viruses containing these multiple mutations.

Using a potent triple drug combination has been shown to be more successful at inhibiting virus replication and therefore the development of mutations causing resistance. An extremely potent two drug combination which requires multiple mutations for resistance to develop may be just as effective. By potent inhibition of virus replication the opportunities for the emergence of virus with mutations is much more limited. The 1 and 2 year results from several studies show that with proper adherence 80-90% of the study participants were able to sustain viral load below detection by standard commercially available tests.

An extremely potent triple drug combination has created a high genetic barrier. There may be rare pre-existing viruses resistant to one or even two of the drugs, but there are no rare mutant viruses that are resistant to all the drugs. So, in the short term (1-2 years) we do not yet see viruses escaping inhibition by the triple drug therapy.

(Commentary - Some researchers believe that low level virus replication is ongoing even in the face of so-called complete viral suppression. Some believe eventually resistance will develop to the potent triple drug regimens. They believe that lowering viral load to as low as possible may be the best approach to dealing with that potential problem. At the December ‘97 ACTG meeting, Doug Richman, MD, a leading AIDS researcher from UCSD, said the goal of therapy should be to lower viral load to <1 copy/ml. The results of two studies suggest lowering viral load to <20 copies/ml may increase durability of viral suppression. However, some researchers suggest that it may be possible to delay disease progression if you can maintain viral load at a low but still detectable level. One study suggested that keeping viral load below 5,000 copies/ml using just two nucleosides as an initial therapy could delay disease progression for a year. The criticism of that approach is that resistance will develop to the drugs used possibly within a year or less, and that could limit subsequent choices for constructing effective combinations. After developing resistance to nucleoside(s), it appears that benefit will diminish from other nucleosides used afterwards.)

Dr. Kuritzkes went on to discuss two concrete examples of resistance. First he discussed nevirapine (Viramune) which is a NNRTI; this was followed by a discussion using AZT and 3TC as an example of a NRTI. A single mutation can be enough for development of resistance to nevirapine. Resistance is the loss of ability to suppress virus. Although it’s rare, the nevirapine single mutation resistant virus can pre-exist in a person even if they’ve never before taken the drug. If nevirapine is used alone, resistance would usually develop within weeks. The overgrowth of pre-existing mutants, viruses already resistant to nevirapine, emerge very rapidly leading to a complete loss of antiviral activity because only one drug was used. However, in Boerhinger Ingelheim study #1046 which studied individuals with no prior drug therapy experience, nevirapine was effective in inhibiting or suppressing viral replication in a triple regimen with two nucleosides, but a two drug regimen consisting of nevirapine plus just 1 nucleoside did not adequately inhibit virus replication. So, resistance can develop quickly if you use a drug alone which needs only one resistance mutation for a loss of viral load suppression.

A different and more complicated situation can occur with the use of nucleoside RT inhibitors (NRTI). For example, although a drug such as 3TC can develop high level resistance very quickly with a single mutation, it can interact in a complicated way with AZT. In Glaxo Wellcome study NUCA 3001, when 3TC was used alone a potent inhibition of viral replication occurred but resistance developed fairly quickly accompanied by a rebound in viral load. The development of a single mutation, such as with nevirapine, caused high level 3TC resistance and a loss of antiviral activity. But, in the NUCA 3001 study viral load did not rebound to its level just prior to starting 3TC. Although high level resistance was detected viral load did not go back to baseline, there was a little persisting antiviral activity. When AZT and 3TC were combined you can observe a persisting greater additive or possibly synergistic antiviral effect, which may in part be due to 3TC preventing AZT resistance mutations.

It has been found that in some patients 3TC reverses and delays AZT resistance until possibly a large number of AZT mutations have accumulated, and some other so far poorly defined mutations accumulated. But, in a greater number of patients, in whom 3TC does not have that effect, it is not understood why viral load reduction persist despite the presence of the 3TC 184 resistance mutation. And 3TC resistance develops within 6 weeks even when taken with AZT. At week 52, viral load inhibition was sustained in the study, although after one year some study participants started to see a rise in viral load as viruses resistant to both drugs emerged. In this study for those taking AZT/3TC, it can take many months for enough AZT mutations to accumulate for viruses to become highly AZT resistant. Over the course of the first year despite the accumulation of one or just a few mutations in about a 1/3 of patients their viruses by and large remained highly sensitive to AZT although they were resistant to 3TC.

Geno and Phenotype Testing. You can measure in the lab how much drug is needed to inhibit HIV by either 50% or 90%. When you talk about the amount of drug required to do this you are talking about the inhibitory concentration (IC), so you’ll hear terms such as IC50 or IC90. This is called phenotypic resistance testing. If you need more drug to suppress the virus than is normally needed that means there is phenotypic resistance. If the amount of drug necessary to inhibit virus replication is 8 times more, then you have 8 fold phenotypic resistance. A different way to look at resistance is to perform a genotypic analysis, where you are looking at the sequence of the genes in the virus itself. You are looking for changes in the genes (mutations); but you have to look for mutations that have been proven to result in drug resistance, to make your findings useful for the patient.

Commentary - There are some concerns I have about genotypic testing. Will your genotypic analysis detect a MDR (multi-drug resistant) mutation which is discussed below? You can fail d4T without any mutation being detectable (discussed below)? If you stopped taking a drug, after a while a resistance mutation may not be detectable although it might have been detectable while on therapy.

Nucleoside Resistance Mutations Primer. This is a review of different kinds of mutations that are responsible for resistance to various drugs with a discussion of how resistance to one drug can overlap with or cause resistance to another drug (cross-resistance). The principle AZT mutations that lead to AZT resistance are 41, 67, 70, 215, and 219. But, we’ve recently learned that there are additional minor mutations that continue to accumulate even after the first 4 or 5 have accumulated. By themselves these mutations do not give rise to broad cross-resistance. But clearly patients that have failed AZT and have highly resistant viruses do not respond as well to subsequent (nucleoside) therapy. What we don’t know for sure is if its due to true cross-resistance at the level of the virus or is it because people who are failing therapy have been selected in a sense and are going to do poorly whichever (nucleoside) therapy they are given. We do not yet have adequate data for individuals who were treated first with other nucleosides and subsequently switched to AZT+3TC, to know if individuals who failed ddI or d4T would have the same or worse subsequent outcome.

The principal mutation that can occur from ddI therapy is 74. This mutation is shared by ddC and is also potentially shared by 1592U89. Although we don’t know yet what happens in people who fail ddI (how well will they respond to other nucleosides?), it’s clear that people who are resistant to ddI are not going to respond to ddC; and, people who are ddC resistant are not going to respond to ddI. (Commentary- there is some suspicion that 3TC and ddI may have some cross-resistance).

D4T is a real puzzle. Although a mutation at 75 has been found in the lab, there is little convincing evidence that patients treated with d4T show resistance to d4T. Clearly, prolonged d4T therapy leads eventually to d4T failure, but there are no virologic markers that we can measure identifying viruses that are no longer responding to d4T. But, there is a curious mutation that gives rise to resistance to many of the nucleosides. This is a mutation, called the multi-drug resistant mutation (MDR), that can emerge from using the combination of AZT+ddI. There are several that can emerge (62, 75, 77, 115), but the principal one is 151. When a virus accumulates a MDR mutation plus others you can get virus that is highly resistant to AZT, ddI, ddC+d4T and perhaps to 3TC (if 184 is also present), and we don’t know yet about 1592. About 10% of patients using the combination of AZT+ddI can get the MDR mutation, so you should be concerned about using this combination certainly as an initial therapy.

Selecting Treatment Based on Resistance. Each treatment option carries with it the risk of certain resistance and cross-resistance patterns (and certain side effects). At the present time it may be difficult to say that any one specific set of treatments is clearly better than an equally potent set of treatments in terms of what the downstream options are going to be. But, whichever choices are made initially it carries important implications for what options remain because of the patterns of resistance and cross-resistance described above. (commentary - The results of several studies revealed in the past year suggested that d4T+3TC had similar effects as AZT+3TC, in terms of CD4 increases and viral load reductions for individuals who had never before taken therapy. Some doctors are using d4T/3TC as a first line nucleoside option).

Several studies have shown that patients who failed AZT do not respond as well to subsequent (nucleoside) treatments. But, to be fair we are not sure if this is due to cross-resistance due specifically to AZT resistance, or people who fail an initial nucleoside therapy are not going to respond well to any subsequent nucleoside therapy.

We don’t have data yet on individuals failing d4T or ddI first and then switching to AZT. As mentioned above ddI failure clearly leads to ddC failure. It is yet uncertain of the effect of a ddI failure on subsequent 1592 treatment in patients. Although the MDR mutation that can result from AZT+ddI combination therapy can be uncommon, the potential problem is severe enough that you should avoid using it as an initial regimen. 3TC may have some degree of cross-resistance with ddI and ddC, but there is little data from patients on the use of ddI after 3TC failure. It seems less likely that 3TC resistance will play a major role in cross-resistance to 1592. (Commentary- having dual resistance to AZT and 3TC may play a more significant role in causing cross-resistance to 1592. See the 1592 resistance report later in this section).

Strategies for Preventing Resistance

• Raise the genetic barrier to resistance; that is, select drugs for your regimen wisely in order to choose drugs that require complex patterns of mutations; you don’t want to use many drugs together each of which needs only one mutation that will lead to resistance; that would make it easier to select from the pre-existing pool of mutations that would cause a drug to fail more quickly
• You want to use drugs with non-overlapping patterns of resistance such as AZT/3TC or d4T/3TC.
• We need better drugs that are easier to take and better tolerated; by making drugs that have characteristics making it easier to take, compliance will improve; for example, a once-a-day regimen should be easier to take for many individuals. The current crop of available drugs do have limitations; increasing the amount of a certain drug may increase its effectiveness, but you can’t do that because using higher concentrations of the current drugs would cause safety problems.

Protease Inhibitor Resistance by Dr. Jody Lawrence

Additional protease inhibitors in development:

• Two protease inhibitors from Bristol Myers Squibb, one is in a human study
• PNU 140690 from Upjohn & Pharmacia in phase I human study
• DuPont Merck 850 & 851 about to enter phase 1 human study.
• Parke Davis has several protease inhibitors in early development.

Although protease inhibitors are potent and can lead to success, they must be taken as prescribed (compliance) to have prolonged suppression. Drug levels in the blood must be kept above a certain level. If by missing a dose or not taking full doses (poor adherence), or due to high metabolism you drop down below that level needed to suppress the virus, it allows the virus to mutate. If the mutated virus replicates enough and resistance develops, the drug will not be potent enough to suppress viral replication any longer.

The approved protease inhibitors have overlapping resistance mutations. That is, they have some of the same mutations in common for resistance to that drug to develop. The major or primary mutations for indinavir and ritonavir for patients taking the drugs are almost identical: 82, 84, 54, 46. The major mutations reported for saquinavir are 48 and 90. It’s been reported that the major mutation for nelfinavir is 30; and for 141W94, 50 appears to be the major mutation. If you concentrate only on the major mutations it looks simpler than it is. It may appear as though there aren’t as many overlapping mutations and possibly you could switch from one protease inhibitor in some situations to another. But a number of other mutations develop with time on treatment and many of them are overlapping. There appears to be much more cross-resistance between these drugs than we originally thought.

(Commentary - For example, additional indinavir mutations found in humans include 71, 90, 10, 20, 32, 64, and 63. Additional ritonavir mutations found in humans during ritonavir therapy include 20, 33, 36, 71, 63 and 90. Additional mutations that have been identified with saquinavir include 10, 46, 54, 63, 71, 82 and 84. Although Agouron says that only a mutation at 30 causes resistance and that these other mutations observed are not relevant to resistance to nelfinavir, these have been observed - 71, 77, 88, 46, and 84. For 141W94, 46 and 47 are additional mutations that have been observed. These additional mutations are usually referred to as secondary mutations. Some combination of the major mutations and/or the secondary mutations can cause resistance to a specific protease inhibitor.)

As you can see, although the major mutations may not be as overlapping, the secondary mutations can be. This is a major factor in the development of cross-resistance.

For both ritonavir and indinavir it has been uniformly reported by both manufacturers, Merck and Abbott, that an accumulation of several mutations is required for resistance to emerge. Merck has reported it takes an accumulation of three or more mutations for resistance to begin to develop to indinavir. Merck has reported that mutations at 46+63+82 produced a 4-fold resistance when tested in the test tube or lab.

They also reported that mutations at 46+63+82+84 produced an 8-fold increase in resistance to indinavir in the lab. Merck has stated that the accumulation of mutations causing resistance to indinavir do not necessarily occur in an ordered fashion. In other words, it may be a random accumulation of several different mutations. However, Abbott has said that the mutations for ritonavir occur in a step wise and ordered way, possibly with 82 occurring first followed by specific other mutations.

Abbott has reported that a mutation at 82 has been observed in the lab to confer 2.5 fold resistance; that mutations of 82+54+71+20+36 observed in the lab caused an 8 fold increase in resistance to ritonavir. These examples stated above for ritonavir and indinavir are only isolated examples for specific situations. Individ- uals can experience different genotypic resistance mutation profiles.

Dr. Lawrence showed an example of one person’s experience in developing resistance to indinavir. This person at baseline, before taking indinavir, had a few mutations that sometimes are referred to as polymorphisms or as we described earlier pre-existing mutations. At week 12, the person picked up a couple more genotypic mutations but no phenotypic resistance was detected. In this Resistance Supplement, Dr. Gulick defines phenotypic resistance. At week 24, an 82 mutation was observed along with a 63 mutation, and now a 4 fold increase in phenotypic resistance was detected. With further therapy additional mutations (84, 71, 24) were seen at week 40 accompanied by about an additional 4 fold increase in phenotypic resistance. By week 52, additional mutations were observed (20, 54) and phenotypic resistance at least doubled from its level at week 40. A rebound in an individual’s viral load may or may not occur at the same time as genotypic or phenotypic resistance is observed. A person’s viral load may rebound before these changes are seen.

A more detailed discussion about protease inhibitor resistance and mutations accompanied with tables is available on the NATAP web site "Protease Inhibitors: Success and Resistance".

Dr. Lawrence then went on to discuss some of the studies presented at recent conferences. The study presented by Dr. Steven Deeks caught the headlines in the mainstream press. He reported that 53% of patients in his clinic had failed protease inhibitor therapy when he looked back at their charts. But, his purpose was to show how protease inhibitors should not be used. The mainstream press misinterpreted his point. The high failure rate was due to using protease inhibitors improperly. In fact, he reported that individuals using them properly had 80-90% success rates. Deeks pointed out that only 1/13 individuals who were drug naive had drug failure. Non-adherence was a problem among the group with a high failure rate. They had high baseline viral load and low baseline CD4 counts meaning that therapy should be started before an individual reaches such an advanced stage. Some of them merely added a protease inhibitor to there current nucleoside therapy. As well, some had previously taken saquinavir so may have had some cross-resistance.

Protease Inhibitor Salvage Therapy

Dr. Lawrence reviewed several studies for individuals who failed indinavir and then were switched to a ritonavir+saquinavir regimen. Although initial viral load reductions may have been as much as 1.5 log, which is still not as much as seen when an individual is naive to protease inhibitors, the reductions were generally short lived. By week 24 or week 16, the percentage of individuals who were undetectable (<500 copies/ml) was low ranging from 10% and 25% to 37%.

She reviewed several studies of individuals who failed saquinavir therapy and then were switched to an indinavir regimen. The results were a little bit better but not very much. She reviewed one study of individuals who failed saquinavir and were switched to ritonavir+saquinavir. Individuals received an initial 1.52 log reduction in viral load but by week 16 viral load rebounded.

She discussed a study she is conducting where individuals who failed saquinavir were first switched to a nelfinavir regimen. NNRTIs were not used in this study because it was before we knew how to use NNRTIs with protease inhibitors. The individuals were mostly heavily pretreated with nucleosides so when they switched from saquinavir to nelfinavir only 30% were able to change to new nucleosides. Some individuals recycled old nucleosides. After an initial nice viral load reduction for some individuals, by week 12 only 2/16 were still undetectable. Subsequently, individuals stopped nelfinavir and added indinavir+nevirapine to their unchanged nucleoside therapy. There was a nice initial viral load reduction that lasted a little longer but over time the durability appears to be fading.

(Commentary - One strategy to deal with protease cross-resistance is discussed in more detail in the Protease Inhibitor Update in the section called "Monthly Monitoring of Viral Load." Essentially, if you monitor your viral load monthly and change to a potent therapy possibly consisting of 4 or even 5 drugs as soon as you detect a rebound in viral load, you may be able to limit cross-resistance).

Dr. Lawrence then went on to discuss two reports from ICAAC in October ‘97 where individuals who failed nelfinavir were switched to other therapies to assess their response. The prospective study conducted and reported by Keith Henry showed more optimistic results. Individuals were divided into two small but different group in this study. The first group of 12 individuals had limited prior nucleoside experience, CD4s of about 200, and viral load of about 60,000 copies/ml. They failed the regimen they were taking in nelfinavir study #511 of nelfinavir+AZT+3TC; they were drug naive prior to enrolling in 511. They were switched to ritonavir+saqui-navir+d4T/3TC. Initially, 12/12 (100%) were <500 copies/ml and at week 16 6/7 (85%) were still <500 copies/ml, which is encouraging. But this is still only 16 weeks. The key is to see how durable the effect will be. A follow-up report will be presented at the ‘98 Human Retroviruses Conference in February.

Henry took a group of 7 individuals from nelfinavir study #525 who had much more nucleoside experience, protease naive, and had significantly lower CD4s (65) and higher viral load (233,000 copies/ml). Only 3/7 were able to achieve a viral load <500 copies/ml. The other study was a retrospective study (looked back at patient charts). 6 went on indinavir therapy and 6 went on ritonavir+saquinavir therapy, after failing nelfinavir therapy. Only 3/12 individuals had a good response.

Dr. Lawrence suggested that if you can detect viral load rebound before it develops too much (as discussed in the article, Monthly Monitoring of Viral Load), a quick switch to another potent regimen may work. If you delay switching therapy, over time the number of mutations will build up for all the protease inhibitors with cross-resistance developing.

(Commentary - She suggested one way to increase the efficacy and durability of protease inhibitor therapy is to increase the trough levels of a drug. This is the subject of ongoing research. The trough level of a drug is the amount of drug in the blood at the end of the dosing period: for example, 8 hours for indinavir and 12 hours for ritonavir. Some individuals fail therapy because they have low trough levels due to high metabolism or taking a dose too late. Combining 2 protease inhibitors or a protease inhibitor with a NNRTI may raise the trough level if the drugs are dosed properly and if the right drugs are selected. This may prevent an individual from having inadequate blood levels of a drug. However, sometimes when you raise blood levels you can increase side effects. Several protease-protease combinations are being explored with this idea in mind: indinavir-ritonavir, indinavir-nelfinavir, ABT-378/ritonavir, ritonavir-nelfinavir. As well, delavirdine used in combination with other protease inhibitors may be useful to apply this concept because delavirdine is the only NNRTI that significantly increases protease inhibitor blood levels. However, the proper dosing regimens must be identified before experimenting with these combinations, and ongoing studies are exploring this now. With the exception of ritonavir+saquinavir, where dosing regimens have been identified.)

Dr. Lawrence was asked a question from the audience, is nelfinavir a preferable choice for a person’s first protease therapy? She said, that a key question now is should a person start with nelfinavir or indinavir? On the one hand, the preliminary Keith Henry data may suggest, if the data holds up beyond 4 months, that you may have more leeway for bringing on other drugs for salvage therapy if you fail nelfinavir as a first line therapy. Although individuals with more extensive prior nucleoside experience did not do as well. However she said, we don’t know if you might develop resistance to nelfinavir more rapidly if you started on that drug, than for indinavir if you started on indinavir first. She said we need head to head comparisons of the two drugs to determine which may be a better first line protease inhibitor.

1592U89 Resistance by Dr. Marty St Clair

New data and resistance information about the use of 1592U89 has been recently reported at the 6th European Conference on Clinical Aspects and Treatment of HIV Infection in Hamburg, Germany (Oct 11-15,1997), and at the NATAP Resistance community education forum at NYU Medical Center (Oct. 25, 1997). Some of the information was previously reported, but much of it is brand new. If you do not want to read all the information in detail, first is a brief summary of the report.

Treatment-naive individuals in studies have achieved about a 2 log reduction in viral load due to 1592. For individuals who are nucleoside experienced the benefit can be less. Based on in vitro data and preliminary data from 22 week results of a small human study (CNAA 2003, page 35), individuals with extensive experience with nucleosides and/or resistance to a broad number of nucleosides may or may not respond well to 1592. The participants have viral load reductions at week 22 ranging from none to -2.44 log. It appears too early to predict exactly how an individual may respond, but Table 3 below lists HIV RNA responses for the 18 study participants and other related information, and the data in the table may offer some suggestions on how 1592 will effect treatment experienced individuals. Glaxo Wellcome believes, and the preliminary data supports this thinking, that individuals with extensive experience and/or resistance to a number of nucleosides will be less likely to respond well to 1592. However, from examination of the data in Table 3 of the 18 study participants, some individuals who you might expect not to respond well to 1592 do in fact get a good response to 1592; as well some individuals you might expect to respond well do not. As study results become available over time, analysis will reveal more helpful information. In an animal study, 1592 penetrated the brain and CSF well. An ongoing dementia study using 1592 should be more revealing. For more information on the CSF animal data, see the article "The CSF and Indinavir, Ritonavir+Saquinavir, DMP-266, 1592U89, AZT vs d4T".

Emerging Resistance Patterns. A more comprehensive review and discussion of the in vitro resistance data is available in an article posted to the NATAP web site several months ago called 1592U89 Cross-Resistance.

One of the first things researchers do with a new drug which is expected to be used by humans is to passage virus in the presence of the new drug to see if mutations could be found that might render the virus resistant to the drug. Researchers took a wild type virus (no resistance mutations) and put it into the test tube with 1592; 4 mutations emerged. The first to emerge was the M184V mutation which is the 3TC mutation followed by 65R, 74V and 115F upon further passage. Each of these 4 mutations individually result in 2, 3 or 4 -fold increase in resistance. In more scientific terminology, each of these single mutations resulted in no more than a 4-fold increase in the IC50. The IC50 is the inhibitory concentration of drug that will block replication of virus by 50%. This is a standard test used by researchers.

The Glaxo Wellcome researchers believe that a greater than approximately 8-fold increase in resistance is necessary to cause a concerning reduction in sensitivity to 1592. In the research reported here, a 10 fold or higher resistance to 1592 occurred only with the presence of three mutations; except, one double mutation caused an 8.5 fold increase in resistance to 1592.

The Fold Increase in IC50 (or, increase in resistance) The following increases in IC50 have been reported for these listed single, double and triple mutations: See Table 1

Table 1

wild type virus
























As you can see in table 1, one double mutation (74V/184V) and two triple mutations have >8 fold increase in IC50 (or, increase in resistance) to 1592. It is important to remember that this is an in vitro experiment. That is, it was conducted in the laboratory. Lab experiments do not always predict responses to a drug by an individual.

Researchers conducted an additional in vitro experiment. See Table 2. They took viruses which were resistant to nucleosides and measured the fold increase in IC50 for 1592. In other words, the amount of resistance to 1592 a virus had if it were resistant to another nucleoside.

Table 2.

Virus Known resistance

Fold increase in IC50 for 1592

Wild type none


mutations-67N/70R/215F/219Q AZT (120 fold)


mutation- 74V ddI (5-10 fold)


  ddC (18 fold)  
mutation- 181C NNRTI (>100 fold)


mutation- 184V 3TC (500 fold),


  ddI/ddC (2-10 fold)  

The wild type virus had no resistance to any nucleoside and was not resistant to 1592. The next virus had 4 AZT associated mutations and was very resistant to AZT (120 fold). The third virus only had one mutation (74V) but was 5-10 fold resistant to ddI and 18 fold resistant to ddC. The fourth virus had a NNRTI mutation and possesses some NNRTI resistance. The fifth virus was extensively 3TC resistant (500 fold) with 2-10 fold resistance for both ddI and ddC. In each case for each of the viruses listed in the table, the amount of resistance to 1592 (or the increase in IC50) was less than 4 fold.

The question remains, how will 1592 perform for individuals who have had varying degrees of experience with different nucleosides; and, have varying degrees of levels of resistance associated with those nucleosides. Keep reading.

Glaxo Wellcome researchers have referred to one patient in the treatment-naive study (CNAA 2001). This patient entered the study with a 184V mutation although 3TC experience was prohibited by study design. The person achieved about a 1 log drop in viral load after 4 weeks of 1592 monotherapy. For weeks 4 through 12 of the study the person received 1592+AZT and achieved a viral load reduction at week 12 of about 2 logs. Despite their 3TC resistance the person responded to 1592, but of course this is only one person.

CNAA 2003: Treatment Experienced. In this study 18 individuals with greater than 6 months nucleoside experience with AZT, AZT/3TC, d4T, and ddI, but who had not taken protease inhibitors received 1592U89. The therapies were all failing because all the study participants were required to have >10,000 copies/ml of viral load. In this study, 1592 was merely added to their current therapy; no other component of the therapy was changed. By merely adding one drug onto a failing regimen, you may be limited in the capacity to capture the individual’s full ability to respond to the newly added drug.

The average reduction in viral load from baseline for the 18 individuals was at:

        • week 4: -1.11 log
        • week 22: -1.30 log

For those who did not have the 3TC M184V mutation at baseline (prior to receiving 1592), their reduction in viral load from baseline was:

        • week 4: -1.49 log
        • week 22: -1.48 log

Other individuals who did have the M184V mutation at baseline, some of whom also had additional mutation(s), their reduction in viral load was:

        • week 4: -0.52 log
        • week 22: -0.97 log

See below - More on Emerging Resistance Patterns. Generally, it appears as if 3TC resistance should not cause cross-resistance to 1592; but, dual AZT/3TC resistance may or may not cause cross-resistance to 1592.

These averages can be misleading because each of the 18 study participants had varying baseline characteristics and varying prior nucleoside experience. Phenotypic resistance and genotypic mutations to 1592, 3TC and AZT were measured at baseline. As well, the changes in viral load were measured at week 4 and week 22. The 18 participants had varying baseline phenotypic resistance to 1592, 3TC and AZT, varying genotypic mutations and varying viral load responses. The viral load responses at week 22 did not always correlate in a way you might expect with the baseline genotype or phenotype data. At the NATAP community forum, Dr. St Clair presented all the data for each of the 18 participants which follows in Table 3. A closer review of the individual data may be more helpful in understanding the treatment effect of 1592 for treatment experienced individuals. But, data from a larger set of individuals is needed in order to better understand how to use 1592 in treatment experienced individuals. Glaxo has said they will be compiling data on treatment response to 1592 from their ongoing trials for treatment experienced individuals. This data will be helpful in deciding how to best use 1592 for a given individual.

Glaxo researchers, Lanier at Hamburg and St Clair at the NATAP meeting, concluded treatment experienced individuals would generally be less responsive to 1592, if:

        • their phenotypic resistance to 1592 at baseline tended to be at a higher level
        • the more nucleoside associated genotypic mutations one has, the less responsive to 1592 you should be
        • Dr St Clair said they believe phenotypic resistance >8 fold to 1592 may be the cutoff above which individuals may not respond well or at all to 1592.

The following table was presented by Dr. Marty St. Clair, of Glaxo Wellcome, at the NATAP community education forum on October 25, 1997 at NYU Medical Center. See Table 3.

Table 3. Viral Load Changes and Baseline Genotypic and Phenotypic 1592 Resistance for the 18 CNAA 2003 (treat. exp.) Study Participants

Pt ID # Prior Th mos. 1592 STC AZT wk 4 wk22 bsl geno mt
442 AZT 18 0.41 0.68 0.93 -2.34 -2.44 wt
445 AZT 12 1.03 1.33 1.11 -3.15 -1.81 wt
465 AZT 15 0.83 1.52 0.76 -2.33 -2.03 wt
482 AZT 38 0.70 1.21 0.67 -0.28 +0.07 wt
490 AZT 18 47.00 100.72 7.46 +0.42 -0.72 41, 74, 184, 210, 215
491 AZT 14 7.01 100.72 1.30 -1.19 -2.26 74, 184
495 AZT 30 5.39 13.12 98.56 -0.34 -0.92 67, 70, 215, 219
441 AZT/3TC 52/10 3.22 109.72 1.07 -2.35 - 184
473 AZT/3TC 19/10 9.41 109.72 11.50 -0.08 +0.25 41, 70, 184
497 AZT/3TC 12/11 3.11 109.72 9.56 -0.00 -0.56 184
515 AZT/3TC 12/12 4.68 3.27 15.26 -0.38 -2.34 67, 184, 210, 215
516 AZT/3TC 12/12 9.82 109.72 28.92 -0.05 +0.17 41, 67, 74, 184, 210, 215
517 AZT/3TC 19/12 1.59 1.01 0.97 +0.07 -0.55 wt
447 ddI/AZT 48/0 1.46 4.55 0.51 -1.01 -0.74 wt
443 d4T 9 0.58 0.34 0.95 -2.16 -1.72 65
459 d4T 6 1.02 1.10 0.79 -2.53 -1.99 wt
496 d4T 11 1.18 1.62 0.85 -1.11 -2.33 wt
513 d4T 10 7.32 12.45 125.77 -1.19 -1.86 41, 210, 215

Analysis of Table 3. The mean viral load reduction at week 22 for the 8 individuals with no baseline genotypic mutations (wild type) is -1.49 log. Surprising to me, PT# 482 had no genotypic mutations at baseline and had virtually no phenotypic resistance to 1592, AZT and 3TC at baseline; but had 38 months of prior AZT experience. PT# 482 had virtually no response to 1592 at week 22. In contrast, pts 442, 445, and 465 also had no genotypic mutations at baseline, virtually no phenotypic resistance to 1592, AZT and 3TC at baseline, but had less prior experience with AZT (18, 12 and 15 months, respectively). The viral load reductions for pts 442, 445 and 465 at week 22 are -2.44 log, -1.81 log, and -2.03 log, respectively.

PT# 447 had 48 months of prior ddI experience, also had no genotypic mutations identifiable at baseline, and had relatively low phenotypic resistance to 1592, AZT and 3TC; this person had a limited response at week 22 with a -0.74 reduction in viral load. A possible explanation for the results discussed here are that the actual length of time of prior nucleoside experience may be a factor regardless of a lack of identified geno- or phenotypic resistance.

Pts 473, 497 and 516 had 109-fold baseline phenotypic 3TC resistance, and >9 fold AZT resistance; their viral load responses were limited or none at week 22 ( +0.25 log, -0.56 log, and +0.17 log, respectively).

4/6 individuals with 3 or more genotypic mutations at baseline had a limited response to 1592 at week 22 (pt 513, -1.86; pt 490, -0.72 log; pt 495, -0.92 log; pt 473, +0.25 log; pt 515, -2.34 log; pt 516, +0.07 log).

More on Emerging Resistance Patterns. In Hamburg, Randall Lanier reported the findings from an in vitro experiment looking at viruses with phenotypic resistance to 3TC but phenotypically AZT sensitive, and separately to viruses with dual phenotypic resistance to AZT/3TC. The fold increase in 1592 IC50 was measured.

3TC Resistance. All 25 clinical isolates (individual’s blood samples) with 3TC resistance but AZT sensitivity had less than 8-fold phenotypic resistance to I592. 10 of the 25 isolates had between 5 and 7 fold resistance. 10/25 had approximately 3 fold or less phenotypic resistance. The median was 4.4 fold phenotypic resistance to 1592. In other words, 3TC phenotypic resistance caused a median 4.4 fold increase in 1592 resistance which means that 3TC resistance by itself may not cause resistance to 1592.

AZT/3TC dual resistance. 13 clinical isolates had dual phenotypic resistance to AZT/3TC. 5/13 had > 8 fold phenotypic resistance to 1592. 8/13 had <8 fold resistance to 1592. Remember, the Glaxo researchers believe that >8-fold phenotypic resistance to 1592 may predict a poor response to 1592 or no response at all. The median response for the 13 isolates was a 6.7 fold increase in phenotypic resistance. The phenotypic resistance for 6 of the isolates was approximately equal to or less than 5-fold. 2 of the isolates had between 8 and 10 fold resistance, and 3 of the isolates were about 12 fold or higher.

This in vitro data suggests that some individuals with AZT/3TC resistance may not respond well to 1592, but that some individuals with AZT/3TC resistance may be able to respond well to 1592. In fact, the Glaxo Wellcome researchers said they believe all the data compiled above suggests that failure to respond to 1592 is more likely to occur when an individual has broad and/or extensive resistance to nucleosides (AZT, 3TC, ddI, d4T, ddC). A close look at Table 3 shows how 18 individuals with a variety of different nucleoside experience and resistance actually responded to 1592.

Expanded Access Tips. Here are a few words about how to best use the 3 drugs available through expanded access. If you have extensive nucleoside experience including use of AZT/3TC, ddI, and d4T, you may or may not get a good response to 1592U89. As you can see from the results of CNAA 2003, there were a variety of viral load responses from no response at all to a reduction of 2.44 log. Anecdotally, I have heard from several individuals who have received good responses to 1592 despite extensive nucleoside experience. However, when planning your regimen you should consider the possibility that you may not respond well to 1592. If you have not yet tried a NNRTI you should respond well to efavirenz which is a potent NNRTI, but resistance to a NNRTI can develop quickly. A NNRTI must be used in a regimen well designed to adequately suppress viral load in order to prevent resistance. Based on preliminary data a 0.5 log reduction in viral load is average for PMEA. However, the preliminary data suggests those with extensive AZT experience may not respond as well. If possible, your goal should be suppression of viral load to below detection. You may need additional therapy to accomplish this. It is a personal choice and everyone’s situation is different, but sometimes it may be preferable to delay changing your regimen until you can assemble a regimen you feel more confident in.

Back to top