The second enzyme of HIV to be successfully targeted was the viral aspartyl protease. HIV produces the structural proteins of the viral core, as well as the integral viral enzymes, as Gag and Gag-Pol polyproteins that subsequently must be cleaved by the HIV protease. This cleavage is essential to the production of mature virions; and if HIV protease is inhibited, the virions produced are not infectious. The three-dimensional structure of HIV protease has been elucidated, and this information, as well as previous studies elucidating its target sequence, permitted the rational development of inhibitors of this enzyme. Unlike the reverse transcriptase inhibitors, PIs block the production of infectious virions by chronically infected cells. PIs can thus suppress viral production by macrophages and other cells that can produce virus over relatively long periods of time.

Although HIV protease was an attractive target, a number of obstacles had to be overcome to develop clinically useful inhibitors. Many of the initial candidate compounds were highly potent in vitro but had poor bioavailability and relatively short plasma half-lives. In addition, the compounds are quite complex and difficult to synthesize. Nonetheless, by March 1997, four HIV PIs had been approved and several others are presently in clinical development. The PIs developed to date fall into several broad structural categories. Saquinavir and indinavir are peptide-based inhibitors with substitutions in the dipeptidic cleavage site. Ritonavir is a twofold (C₂ ) symmetrical inhibitor, designed to take advantage of the symmetry of the protease active site. Finally, nelfinavir is a non-peptidic inhibitor. All of these are highly selective for viral protease and have little or no activity against human proteases.

Used alone, PIs can induce substantial decreases in the HIV viral load of 100-fold (2 log₁₀ ) or more. However, their usefulness as single agents is limited by the relatively rapid development of resistance, often within 3 months if used alone. This resistance is associated with the emergence of HIV strains with mutations at key sites in the HIV protease gene. Usually two or more mutations are required for high-level resistance to emerge. Cross-resistance among these drugs is frequently found, but because the resistance generally involves several mutations, the issue of cross-resistance in this class of compounds is a complex one. It is worth noting that there is no cross-resistance between PIs and either NRTIs or NNRTIs. As noted earlier, the development of resistance to PIs can be substantially slowed or even prevented by their use in highly potent combinations with NRTIs, and they should generally only be used in such combinations.

The first PI to be developed was saquinavir. Although this drug was found to be active and reasonably well tolerated, its usefulness was somewhat limited by its poor oral bioavailability, which is 4% when given with food and less in the fasting state. However, when saquinavir is given with an inhibitor of cytochrome P-450 (such as ritonavir), the area under the time-plasma concentration curve is substantially increased. When used in this manner, saquinavir has quite effective anti-HIV activity. The other approved PIs are better absorbed by mouth, although each must be taken in a defined way with relation to food. The serum half-life of these drugs ranges from 1 to 2 hours (saquinavir) to as much as 3½ to 5 hours (nelfinavir). One potential drawback to this class of drugs is that in general they penetrate but poorly into the central nervous system.

All of the presently available PIs are metabolized by the cytochrome P-450 system, and all of them inhibit cytochrome P-450 (particularly 3A4) to various degrees. Because of this, a large number of drug interactions occur with this class of drugs, some of which can be quite serious or even fatal. Ritonavir is a potent inhibitor of cytochrome P-450, and thus has a greater potential than the other available PIs for affecting the metabolism of other drugs. Even so, all the members of this class of drugs can cause drug interactions, and they can be quite complex to use, especially in sick patients who may require a number of other drugs. Some drugs that are recommended to be avoided with one or more of the PIs are listed in

Table 418-3. In addition, a number of drug interactions may require adjusting the dose levels. It is beyond the scope of this section to list all the drug interactions involving the HIV PIs, and physicians prescribing any of these drugs should pay particular attention to the information on drug interactions found in the package inserts and in pharmaceutical reference works such as that published by the American Hospital Formulary Service.

The toxicity profile of the PI is generally different from that of the NRTI. All the members of this class of drugs can cause gastrointestinal intolerance, ranging from nausea to diarrhea. Ritonavir and saquinavir can cause elevated hepatic transaminase levels, and some cases of hepatitis have been observed in patients on ritonavir. Indinavir can cause a clinically inconsequential elevation of the indirect bilirubin level. The most frequent dose-limiting toxicity seen with indinavir is nephrolithiasis with drug crystals, and it is important to keep patients receiving this drug well hydrated. There have also been reports of increased bleeding episodes in hemophiliac patients receiving PIs, and diabetes mellitus has also been reported to occur in patients receiving these drugs. Also, a number of patients who have been on combination therapy that included a PI have developed a form of lipodystrophy. These patients characteristically have wasting of the face and limbs along with adipose tissue accumulations in the abdomen and the back of the neck, the latter giving a “buffalo hump” appearance. The pathogenesis ofthis condition is not clear at present and there is some evidence to suggest that it may result from the substantial suppression of HIV in advanced patients rather than from a direct drug toxicity.

A substantial body of knowledge has been accumulated about the three-dimensional structure of the HIV protease, about its interactions with its peptide substrate and available PIs, and on the patterns of mutations that develop. Some of this information had been used to develop the present drugs, and new generations of PIs are being developed using sophisticated computer modeling techniques along with iterative rounds of synthesis and in vitro testing to optimize desirable properties. Some PIs now in clinical testing include VX-478, a non-peptide-based PI that may penetrate better into the central nervous system than some other members of this class; 141W94 (amprenavir), a highly active C₂ symmetry-based inhibitor; PNU-140690; and DMP-450. Amprenavir is now available under an expanded access program.