The critical difference is in a virus protein called Nef, which increases virus infectivity and replication. But even more important than what HIV-1 Nef does may be what it does not do.
The German group tested 30 different nef alleles in an HIV-1 vector. Most SIV Nef proteins down-regulated expression of the T cell receptor protein CD3 and thus blocked activation of infected T cells. Nef from HIV-1 and some closely related SIV relatives had neither of these effects.
The T cell activation matters because of activation-induced cell death (AICD). A normal immune reaction starts with T cell proliferation. Once an invader is cleared, the reaction is self-limited by AICD—the suicide of most of the recently proliferating T cells. During HIV-1 infection, however, the virus is not cleared and still remains visible to the immune system. Repeated cycles of activation lead to excessive proliferation and excessive AICD. The vast turnover of T cells eventually exhausts the immune system.
This AICD was high for HIV-1 and the close SIV relatives. But for SIV strains that down-regulated CD3, there was no activation and thus little cell death. This presumably provides a longer-lived cell within which SIV takes shelter.
Kirchhoff plans two further tests of the hypothesis. SIV engineered to express Nef that doesn't down-regulate CD3 should be more pathogenic in monkeys. And less pathogenic HIV-2 strains may express Nef proteins that down-regulate CD3 less.
A more long-term goal will be to reduce AICD, probably by reducing the activation step. This has been tried in patients with full-blown AIDS, “but this is too late,” says Kirchhoff. “You need to start this early when the immune system is still functional.” Kirchhoff confirms that using immune suppression to fight an immunosuppresive disease will be difficult: “It won't be easy to find the right balance. The [simian] virus can; the question is whether humans can mimic this situation.”