of therapy selection substantially1, 2; for example, approximately 24% of the therapy selections reported in the 2006 version of our Arevir database turned out to be ineffective. Using THEO could have helped reduce the error rate in selecting effective therapies below 15%.

The EuResist project (http://www.eu-
resists.org) adds two qualities to the re-
search we discuss here: Data collection
includes data from several European
countries; and, on the EuResist predic-
tion server, three independently devel-
oped prediction engines are executed
and return individual results and a con-
sensus prediction. 18

new Drugs

Using sophisticated methods to administer antiviral combination drug therapies does not obviate demand for

figure 7. theo applet.

gp41

hIV gp120

Coreceptor
CD4

host Cell figure 8. Proteins participating in hIV cell entry (courtesy Pfizer).

continually developing new drugs. For an individual patient, administering a drug provokes resistance mutations that accumulate within the virus genome. Eventually, only new drugs with new modes of action or even new target proteins will deliver additional effective drug therapies. Moreover, AIDS drugs age as resistance mutations accumulate in the global viral population, necessitating continuous development of new drugs. And clinical side effects enforce the development of new drugs with the same mode of action as existing “old” drugs. Such new drugs might replace the “old” drugs but might also provoke slightly different resistance mutations.

Drugs targeting RT and PR were the basis of AIDS therapy until the early 2000s. Since 2003, drugs targeting other proteins have come onto the market. Especially attractive targets for anti-HIV drugs are proteins facilitating cell entry of HIV. Such targets are chosen because blocking viral cell entry helps prevent integration of the viral DNA into the cellular genome. To understand how we block viral cell entry we must look at the process of HIV entering the cell in more detail (see Figure 8). First, the viral surface protein gp120 binds to the cellular receptor protein CD4. This leads to a conformational change in gp120 so it can then bind to an additional cellular protein, the so-called co-receptor. The binding of gp120 to the cellular co-receptor triggers the actual viral cell entry, during which the helical ( cork-screw-like) viral surface protein gp41 penetrates the cellular membrane, and the hull of HIV fuses with the cellular membrane. HIV can use one of two cellular surface proteins—CCR5 or CXCR4—as a co-receptor; some viral strains use either. The co-receptor specificity of HIV is also called viral tropism. A virus using CCR5 is called R5-virus. Analogously, a virus using CXCR4 is called X4-virus. A virus using either co-receptor is called dual-trop-ic, or R5/X4-virus.

Viral tropism has important clinical consequences. For example, the initial infection results almost exclusively in an R5-virus population; we assume that X4-viruses may infect the patient but can be controlled initially by the immune system. Approximately 1% of the