Di Mascio's study looked carefully at the frequency and duration of blips above 50 copies/mL as recorded in 123 treatment naive patients from eight different research cohorts starting a PI-containing regimen. The mean CD4 count at treatment initiation was 474 (+/- 254) cells/mm3. Overall, the analysis looked at an average of 26 viral load tests per subject over as many months, finding a wide variation in blip frequencies, with 41 patients showing no blips and one patient blipping at every other determination. The average number of blips per sample was 0.09.
The study found that blips were not due simply to assay variation or to chance alone but that different people inherently have different tendencies to blip. They next showed that, within the limits of monthly testing, having one blip does not predict having another and that blip arrival is substantially random. Furthermore, in the patients studied, neither the frequency nor amplitude of blips seemed to increase with time on therapy, which suggests that poor adherence was not responsible for these viremic episodes. There was a relationship, however, between blip frequency and baseline CD4 count, with those having more advanced HIV disease at the time of starting therapy being more likely to become blippers. The significance of this is not clear, although during the period of observation reported here no increase in blip frequency was seen.
The different rates and amplitudes of blipping suggest that there is a great deal of individual variability in the replication rate of HIV, even when mostly suppressed by drug pressure. Another study reported in the November Journal of Virology investigated the relation between viral load and replication rate in individuals who are not taking antiretroviral drugs.
It's long been recognized that viral genetics plays a role in how aggressively HIV behaves in a host. The X4 coreceptor-using variant is particularly famous for kicking HIV immune damage into high gear. More recently it's been recognized that for people who have been on therapy and have developed drug-resistance, their mutant virus may be "less fit" than a wild type drug-susceptible virus. If so, then staying on a failing regimen may be clinically protective despite loss of viral control. Growth competition experiments have also shown that viruses from several long-term non-progressors were inherently less replication competent than viruses from people with normal rates of disease progression.
On the host side, the best known genetic trait that affects susceptibility to HIV infection and subsequent disease progression is a mutation found in a small segment of the population that limits or eliminates the CCR5 cell surface protein, an essential co-receptor for HIV entry. But this flaw in the CCR5 gene is not the only source of CCR5-dependent variability in HIV replication. Even in persons with two functional copies of the CCR5 gene there may considerable inter-patient variability in levels of CCR5 expression at the cell surface. Individuals may also express different amounts of RANTES, a messenger protein that competes with HIV for using CCR5, with elevated levels of RANTES associated with slower disease progression. Different degrees of innate and acquired immunity to HIV may also play a large role in keeping HIV replication under control during the years of slowly progressing disease that follows primary infection. HIV-specific CD8 cells in particular are thought to help in controlling runaway HIV disease and it is hoped that one day a vaccine can be made to boost these protective cells.
The amount of virus found in the blood (viral load) is likely determined by a balance between the elimination of virus and the production of new virus. HIV-specific CD8 cells are generally considered the leading candidate for effecting viral elimination. But this theory remains shaky because most studies haven't found the expected correlation between the strength and specificity of CD8 T-cell response and lowered viral load. If CD8s are mainly responsible for clearing out unwanted HIV, then why don't people with the most qualified CD8s always have the lowest viral loads?
Thomas Campbell and colleagues from the University of Colorado, Denver, sought to establish if replication rate was correlated with plasma viral load levels by performing two different kinds of replication rate assays on the viruses of 12 individuals with chronic HIV infection who were not receiving treatment. Eight of the 12 were treatment naive and none of the participants had detectable drug resistance mutations.
Each individual's virus was cultivated in cell cultures for up to ten days with assessments of HIV p24 protein production performed daily. Changes in the amount of p24 detected from one assessment to the next produced a growth curve that revealed each virus' particular replication dynamics. Typically, each virus had a daylong lag before any p24 production was seen. After p24 was detected, growth proceeded exponentially for the next six days or so. Finally, a plateau phase appeared after the sixth day when additional p24 production tapered off, probably due to saturation of infectable cells after day four.
In addition to the growth curves, the replication capacity of each virus' reverse transcriptase and protease enzymes were determined by genetic recombination techniques using a modified version of the Phenosense drug susceptibility assay.
The investigators found a strong linear relationship between replication rate and viral load that held true from 1000 copies to 100,000 copies/mL. Furthermore, they established that, among these 12 individuals, there was significant natural variation in rates of viral replication due entirely to viral qualities. Another interesting finding was that RT and PR replication capacity were related to the cell-based replication rate. This suggests that genetic variations in these wild type enzymes may be responsible for the different replication rates of different viruses, even in the absence of drug exposure.
One limitation to the study is that in cell systems the role of the host's genetics and immune system are removed, so an individual's actual response to their virus can not be predicted from these results. This issue aside, however, the authors make a provocative suggestion that different viral replication rates may be obscuring measurements of immune-based factors that influence HIV viral load in the body. In particular, they suggest that CD8 cell responses, which have previously not correlated well with viral load, should be reexamined after controlling for replication rate. It's possible that the expected CD8 impact on viral load may only become clear after the "noise" of variation in replication rate has been reduced. If so, then this could help unlock one of the central mysteries of immune control of HIV and remove one of the stubborn stumbling blocks in the way of finding a vaccine.
Back to the GMHC Treatment Issues November 2003 contents page.