The HIV Protease Inhibitors Nelfinavir and Saquinavir, but Not a Variety Of

Home , Abacavir, Nelfinavir, Saquinavir

Antiviral Therapy 10:215–223 The HIV protease inhibitors nelfinavir and saquinavir, but not a variety of HIV reverse transcriptase inhibitors, adversely affect human proteasome function Marco Piccinini1, Maria T Rinaudo1*, Annalisa Anselmino1, Barbara Buccinnà1, Cristina Ramondetti1, Antonio Dematteis1, Emanuela Ricotti 2, Lucia Palmisano3, Michael Mostert2 and Pier-Angelo Tovo2

1Department of Medicine and Experimental Oncology and 2Department of Paediatric Sciences, University of Turin, Turin, Italy 3Istituto Superiore di Sanità, Rome, Italy

*Corresponding author: Tel: +39 011 670 5303 Fax: +39 011 670 5311; E-mail: [email protected]

Background: In HIV-infected patients some clinical and drugs at different concentrations. Intracellular protea- immunological benefits of antiretroviral therapy, which some proteolytic activity was evaluated by searching for frequently include a combination of HIV protease inhibitors ubiquitin-tagged proteins in HL60 cells incubated with (PIs) and reverse transcriptase inhibitors (RTIs), cannot be and without the drugs. solely explained by the drugs’ action on viral enzymes. Results: At therapeutic dosages, nelfinavir and saquinavir Proteasomes constitute the central protease of the ubiq- inhibited proteasome peptidase activity and caused uitin ATP-dependent pathway involved in many cellular intracellular accumulation of polyubiquitinated proteins, processes, as well as in HIV maturation and aggressiveness. a hallmark of proteasome proteolytic inhibition in vivo; Objective: To explore whether the PIs nelfinavir and the RTIs failed to evoke either effect. saquinavir and the RTIs abacavir, nevirapine, delavirdine, Conclusion: Proteasomes are targeted by the two PIs but stavudine and didanosine affect proteasome function in not the RTIs. Therefore, in HIV-infected patients the vitro and in vivo. beneficial effect of a therapy including one of the two Methods: Peptidase activity of purified human 26S and PIs should partly rely on inhibition of host proteasome 20S proteasomes was assayed with and without the function.

Introduction

HIV-related morbidity and mortality have dramatically degradation of foreign or damaged proteins, as well as decreased since the introduction of highly active anti- of short half-life proteins playing a key role in a retroviral therapy (HAART), commonly based on the variety of primary cell functions. These include cell combined administration of specific HIV protease cycle progression, processing of transcriptional factors inhibitors (PIs) and reverse transcriptase inhibitors such as the nuclear factor κB (NF-κB) and its inhibitor (RTIs). HAART usually leads to suppression of viral IκB, production of pro-inflammatory cytokines, replication, reversal of HIV-driven immune alterations expression of surface protein receptors, apoptosis, and slower disease progression. Failure of antiretro- T-cell reactivity and degradation of intracellular antiviral therapy is ascribed to the development of resistant gens and their presentation onto MHC class I viral mutants. However, about 40% of patients treated molecules [15–18]. In addition, this proteolytic for more than 1 year exhibit a steady increase in CD4+ pathway regulates the maturation and release of infec- cell counts and do well clinically, despite a plasma viral tious HIV particles [19]. RNA level rebound [1,2]. The reason for this discrep- Overall, the ubiquitin ATP-dependent pathway ancy remains unclear. involves two discrete steps: covalent, ATP-dependent An increasing body of evidence highlights direct attachment of oligomers of ubiquitin molecules to the immunomodulatory effects of antiretroviral agents protein targeted for proteolysis and degradation of [3–11], and a few of these agents have been shown to the marked protein to small peptides by the 26S affect proteasome function [8,12–14]. Proteasomes proteasome [20]. constitute the central protease of the ubiquitin ATP- The 26S proteasome is a large multi-subunit particle dependent proteolytic pathway involved in the comprised of a catalytic core, the 20S proteasome,

which is placed between two caps formed by the 19S UK); antibodies to ubiquitin and actin were from Santa regulatory component. The 20S proteasome, which Cruz Biotechnology (Heidelberg, Germany) and also exists as a free particle, has a hollow cylindrical Sigma-Aldrich (Milan, Italy), respectively. The other shape; in mammals it consists of 14 pairs of α and β products were from Sigma-Aldrich and Roche- subunits staked in four rings, the external ones Boehringer (Monza, Italy). Red blood cell units and composed of α subunits and the internal ones of β fresh blood samples were from Associazione Volontari subunits. Only three subunits (β1, β2 and β5) are Italiani del Sangue (AVIS) (Turin, Italy). catalytically active and are responsible for splitting peptide bonds after acidic, basic and hydrophobic Proteasome purification amino acid residues. These activities are referred to as The 26S and 20S proteasomes were purified from peptidyl glutamyl peptide hydrolytic (PGPH) activity, human red blood cell units as previously reported trypsin-like activity and chymotrypsin-like activity, [13,24] and stored at –20ºC. respectively, and are considered to be the major peptidase activities of the proteasome [21,22]. Substrates Proteasome peptidase activity assay enter the cylinder through the gated channel defined by Chymotrypsin-like, trypsin-like and PGPH activities of the α subunits. In the free 20S proteasome, the channel purified 26S and 20S proteasomes were evaluated by a is too narrow to be crossed by folded proteins, a spectrofluorometric assay in which the fluorogenic feature that explains why this proteasome cannot peptides Suc-LLVY-AMC, Z-LLR-AMC and Z-LLE- degrade native or ubiquitinated proteins [21]. βNA were used as substrates [24,25]. Peptides However, when the 20S proteasome binds with the 19S (130 µM) in dimethylsulphoxide (DMSO) were incu- regulator, its subunits undergo structural modifications bated in a final volume of 100 µl with: (i) 50 µM ATP, µ inducing in the newly formed particle – the 26S protea- 5 mM MgCl2 and 0.125 or 0.250 g of 26S proteasome – widening of the channel, acquisition of some for chymotrypsin-like and trypsin-like activities proteolytic activity on ubiquitinated proteins, enhance- or PGPH activity, respectively; the reaction lasted ment of peptidase activity and ATP dependence of 30 min for the former two activities and 120 min for µ catalytic properties [21–23]. the latter activity; (ii) 5 mM MgCl2 and 2 or 4 g of Here, we explore whether the PIs nelfinavir and 20S proteasome for chymotrypsin-like and trypsin-like saquinavir, and the RTIs abacavir, nevirapine, delavir- activities or PGPH activity, respectively; the reaction dine, stavudine and didanosine, at clinically relevant lasted 30 min for the former two activities and dosages, affect proteasome function in vitro and in 120 min for the latter activity. vivo. In this respect, peptidase activity of purified human 20S and 26S proteasomes, as well as the levels Chymotrypsin-like activity of the 20S proteasome in of ubiquitin-tagged proteins in cultured cell lines, were lysates from human red blood cells evaluated in the presence or absence of the drugs. In order to assess proteasome peptidase activity in a Intracellular accumulation of ubiquitin-tagged proteins context close to the intracellular milieu, red blood cells is considered to be a hallmark of proteasome inhibition from healthy donors were pelleted at 600 ×g at 15ºC, in vivo [8]. washed twice with PBS, and then lysed by hypotonic shock at 7ºC in 20 mM Tris-HCl pH 7.3, supple- Materials and methods mented with 10 mM NaCl, 1 mM dithiothreitol (DTT) and 1 mM EDTA. After 45 min, lysates were added Reagents with 20% glycerol and centrifuged at 4ºC for 60 min The fluorogenic peptides, Suc-LLVY-AMC, MeOSuc- at 30 000 ×g. The clear supernatant was aliquoted and GLF-AMC, Z-LLR-AMC and Z-LLE-βNA (in which stored at –20ºC. Thawed lysate aliquots (6 µg of single-letter amino acid codes are used and AMC is 7- protein) were incubated at 37ºC for 3 h in a final amido-4-methylcoumarin, βNA is β-naphthylamide, Suc volume of 100 µl with 50 mM HEPES pH 7.8, 65 µM is succinyl, MeOSuc is 3-methoxysuccinyl and Z is Suc-LLVY-AMC and 5 mM MgCl2, with and without benzyloxycarbonyl) were purchased from Bachem 50 µM ATP. The activity tested in these conditions was Feinchemikalien AG (Bubendorf, Switzerland). almost completely aspecific because it was 99% Chromatographic supports, ECL Plus and Hyperfilm unaffected by 10 µM lactacystin or epoxomicin, which autoradiography films were from Amersham are irreversible and selective proteasome peptidase and Pharmacia Biotech (Milan, Italy), and Centricon YM- proteolytic activity inhibitors [27]. To potentiate 30 (30 000 MW cut-off) from Amicon-Millipore Corp proteasome-specific activity, the reaction mixture (Milan, Italy). Antibodies to human 20S proteasome α4 described above was supplemented with 0.05% and 19S complex Rpn8 (S12) subunits, as well as lacta- sodium dodecyl sulphate (SDS), which strongly cystin and epoxomicin, were from Affiniti (Exeter, enhances 20S proteasome chymotrypsin-like activity

[21]. The activity was then evaluated in the presence of immunoblotting with antibody to ubiquitin. Detection either Suc-LLVY-AMC or MeOSuc-GLF-AMC [28], of immunodecorated protein bands was performed as with or without lactacystin or epoxomicin. The proce- described elsewhere [25]. At the end of incubation, cell dure could not be used to quantify trypsin-like and viability was assessed by trypan blue exclusion. PGPH activities of the 20S proteasome because both were below the detection limit. The ATP-dependent Results peptidase activities were undetectable, in agreement with the described disassembly of the 26S proteasome Identification of purified proteasomes caused by SDS [21]. The purified 20S and 26S proteasomes were identified by: (i) migration velocity in non-denaturing gel elec- Inhibition assay trophoresis (Figure 1); (ii) subunit composition and Each drug was dissolved in DMSO at a final concen- immunoreactivity to antibodies specific for α4 or Rpn8 tration of 20 mM, aliquoted and stored at –20°C. (S12) subunit of the 20S proteasome and the 19S regu- Immediately before use, drugs were suitably diluted in lator, respectively (Figure 2); (iii) peptidase activity DMSO to 2.5 mM and then in bidistilled water assay (Table 1). As expected, compared with the

(ddH2O, Milli-Q grade water) to 0.5 mM, and then particle identified as 20S proteasome, the particle iden- added to the reaction mixture to obtain the desired tified as 26S proteasome showed ATP dependence and final concentrations. enhanced peptidase activity (Table 1), typical features of the 26S proteasome [22,24,33,34]. Electrophoresis Non-denaturing and denaturing gel electrophoreses Effects of PIs on 26S proteasome peptidase activities (SDS-PAGE) were carried out as previously described The three major peptidase activities of the 26S [24]. Following gel fixation by 10% v/v trichloroacetic proteasome were inhibited by nelfinavir or saquinavir acid, proteins were visualized by Coomassie blue or at 5 µM and more so at 10 µM. At 20 µM, only silver staining [29]. Alternatively, gels were subjected trypsin-like activity was further depressed, and only to Western immunoblotting as specified below. by nelfinavir (Figure 3). When the two drugs were used in combination, each at 10 µM, inhibition was Western immunoblotting intermediate to that elicited by the single drugs (data Proteins resolved in SDS-PAGE were electrotransferred not shown). onto polyvinylidene difluoride (PVDF) membranes according to Towbin et al. [30] with minor modifications [31]. Detection of immunodecorated protein Figure 1. Resolution of human 20S and 26S proteasomes by bands was performed as reported [25]. non-denaturing gel electrophoresis A B Protein quantitation 12 12 The protein content of samples separated in each purification step was determined according to Bradford et al. [32] using human serum albumin as a standard.

Cell culture 26S The effects of the drugs on proteasome proteolytic activity in intact cells were evaluated on HL60 cells (a 20S human leukaemic cell line) cultured at a density of 2×106/ml on 6-well plates in RPMI-1640 supplemented with 10% fetal bovine serum.

Detection of ubiquitin protein conjugates Ubiquitin protein conjugates were assayed in cells exposed overnight to the single drugs at the concentrations reported in Figure 7, or to drug vehicle (negative Aliquots of fractions from glycerol gradient ultracentrifugation step: 28–30% control), or for 3 h to 10 µM lactacystin (positive glycerol [(A), lanes 1 and 2] and 32–34% glycerol [(B), lanes 1 and 2] were control). After incubation, cells were harvested, loaded on a 4.5% non-denaturing polyacrylamide gel and electrophoresed for 250 V·h. At the end of the procedure, 20S and 26S proteasomes were localized washed with PBS, lysed and centrifuged [25]. Aliquots on the gel by substrate overlay [(A) and (B), lane 1], and Coomassie blue staining [(A) and (B), lane 2] as previously reported [24]. (A) and (B) of the clear supernatant were diluted with reducing Electrophoresed fractions were used for the evaluation of the effects of the sample buffer and subjected to SDS-PAGE and Western antiretroviral drugs on proteasome peptidase activity.

Antiviral Therapy 10:2 217 M Piccinini et al.

Figure 2. SDS-PAGE and Western immunoblotting of purified Figure 3. Effects of (A) nelfinavir and (B) saquinavir on human 26S and 20S proteasomes chymotrypsin-like, trypsin-like and PGPH activities of human 26S proteasome kDa 123 4 A 205 - 110 CTL 116 - TL 98 - 100 PGPH

67 - 90 80

45 - 70 60 50 29 - 40 30 0 5 20 15 20 µM Purified 26S and 20S proteasomes from glycerol gradient were subjected to SDS-PAGE followed by silver staining (lanes 1 and 2) or Western immuno- B blotting in the presence of a monoclonal antibody against the 19S regulator non-ATPase subunit Rpn8 (S12 in humans) (lane 3), and the 26S and 20S 110 proteasome α4 subunit (lane 4). Molecular markers were: Escherichia coli β- galactosidase (116 kDa), rabbit muscle phosphorylase b (98 kDa), bovine serum 100 albumin (67 kDa), ovalbumin (45 kDa) and bovine carbonic anhydrase (29 kDa). 90 80 70 Table 1. Chymotrypsin-like, trypsin-like and PGPH activities of 26S and 20S proteasomes purified from human red blood 60 cells 50 Activity (pmoles/µg protein/min) 26S proteasome 20S proteasome 40 Chymotrypsin-like 41.22 ±1.84 6.30 ±0.25 30 ± ± Trypsin-like 33.50 0.68 4.50 0.10 0 5 20 15 20 PGPH 24.10 ±0.76 1.75 ±0.05 µM The results are the mean of four independent experiments ±SD. PGPH, peptidyl glutamyl peptide hydrolytic. Activity is expressed as percentage of controls (proteasomes exposed to the drug vehicle). Values are the mean of four independent experiments ±SD. CTL, chymotrypsin-like; TL, trypsin-like; PGPH, peptidyl glutamyl peptide hydrolytic.

Effects of PIs on 20S proteasome peptidase activities The three peptidase activities of the 20S proteasome were depressed by nelfinavir at 5–20 µM. In contrast, Effects of RTIs on peptidase activities of 26S and 20S saquinavir inhibited chymotrypsin-like and, to a lesser proteasomes extent, trypsin-like activities, leaving PGPH activity Abacavir, nevirapine, delavirdine, stavudine and unaltered (Figure 4) even at 40 µM (not shown). With didanosine, even at 40 µM, did not affect the three the two drugs in combination, each at 10 µM, major peptidase activities of either proteasome, inhibition of chymotrypsin-like or trypsin-like activity whether used alone or in various combinations. was intermediate to that elicited by the single drugs, Furthermore, at 10 µM, each drug did not modify the while inhibition of PGPH activity was close to that effects elicited by 10 µM nelfinavir or saquinavir (data elicited by nelfinavir alone (data not shown). not shown). These results contrast with what was Drug-dependent inhibition of the peptidase activity previously observed using the two RTIs zidovudine of both proteasomes was reversed by dialysis. and lamivudine, which, although only at concentra-

Figure 4. Effects of (A) nelfinavir and (B) saquinavir on Table 2. Effects of zidovudine and lamivudine on chymotrypsin-like, trypsin-like and PGPH activities of human chymotrypsin-like and trypsin-like activities of human 26S 20S proteasome proteasome

Zidovudine, µM Lamivudine, µM

A Activity 10 20 10 20

110 CTL Chymotrypsin- 84.8 ±14.3 68.3 ±8.7 83.3 ±12.0 67.1 ±8.8 TL like 100 PGPH Trypsin-like 90.8 ±12.6 79.8 ±15.0 88.6 ±16.1 72.8 ±5.6 90 Activity is expressed as percentage of control (proteasome exposed to the drug 80 vehicle). The values are the mean of four independent experiments ±SD. 70 60 Effects of PIs and RTIs on chymotrypsin-like activity 50 of the 20S proteasome in lysed red blood cells 40 In lysates exposed to SDS, only the chymotrypsin-like activity was appreciable and was entirely proteasome- 30 specific. In fact, it was detected not only with 0 5 20 15 20 Suc-LLVY-AMC, but also with MeOSuc-GLF-AMC, µM which has been described as not cleavable by other B proteases [28] (Figure 5), although with the latter the activity was eightfold lower than that assayed with the 110 former (data not shown). In addition, the activity was 100 almost completely suppressed by lactacystin, and even more so by epoxomicin, with both substrates (Figure 5). 90 The peptidase activity was inhibited by both nelfi- 80 navir and saquinavir, irrespective of the fluorogenic 70 peptide used, although the inhibition was less with MeOSuc-GLF-AMC (Figure 5). Using the two drugs in 60 combination, each at 10 µM, inhibition was similar to 50 that elicited by each drug (data not shown). 40 The RTIs examined were all ineffective even at 40 µM. µ 30 At 10 M, they did not modify the inhibition elicited by 10 µM nelfinavir or saquinavir (data not shown). 0 5 20 15 20 µM Ubiquitin protein conjugates in cells incubated with PIs and RTIs Activity is expressed as percentage of controls (proteasomes exposed to the In lysates from HL60 cells exposed to nelfinavir or drug vehicle). The values are the mean of four independent experiments ±SD. CTL, chymotrypsin-like; TL, trypsin-like; PGPH, peptidyl glutamyl peptide saquinavir, accumulation of ubiquitin conjugated hydrolytic. proteins was clearly appreciable with the drugs at 5–10 µM. The ladder of ubiquitinated bands in cells exposed to each drug was reproduced in cells exposed to 10 µM lactacystin (Figure 6). tions in the high therapeutic range, inhibited the No accumulation of ubiquitin conjugated proteins chymotrypsin-like and the trypsin-like activities of was appreciable in cells exposed to the RTIs considered human 26S proteasome (leaving the human 20S up to 20 µM (data not shown). proteasome unaffected) [13]. To establish whether the Cell viability in all samples at the end of incubation contrast relies on the different characteristics of the procedures was close to 97%. RTIs used, a new set of tests was performed in which the effects of zidovudine and lamivudine on peptidase Discussion activity of the two proteasomes were evaluated in parallel with the effects of the other RTIs. The results RTIs and PIs were designed to inhibit HIV reverse tran- confirmed the previous and present observations scriptase and to block HIV protease, as well as resist (Table 2). cellular proteases, respectively. This study reveals that

Antiviral Therapy 10:2 219 M Piccinini et al.

Figure 5. Effects of saquinavir and nelfinavir, as well as of Figure 6. Ubiquitin protein conjugates in HL60 cells exposed lactacystin and epoxomicin, on chymotrypsin-like activity of to nelfinavir or saquinavir 20S proteasome in human red blood cell lysates, evaluated in the presence of the fluorogenic peptides (A) Suc-LLVY-AMC and (B) MeOSuc-GLF-AMC A 12 3 4

A kDa 205 SQV 98 100 NFV LAC 67 80 EPO 45 60

29 40

20 actin - 0 0 5 10 15 20 µM B 12 34 B SQV NFV kDa 205 100 LAC 98 EPO

80 67

60 45 40 29 20

0 actin - 0 5 10 15 20

µM HL60 cells (2×106) were incubated overnight with 5–10 µM nelfinavir, saquinavir or drug vehicle (negative control), or for 3 h with 10 µM lactacystin, Activity is expressed as percentage of controls (lysates exposed to the drug elective proteasome inhibitor [25] (positive control). Cells were then lysed, vehicle). The values are the mean of four independent experiments ±SD. SQV, centrifuged and stored at –80°C. To evaluate the amount of ubiquitin- saquinavir; NFV, nelfinavir; LAC, lactacystin; EPO, epoxomicin. conjugated proteins, one volume of the whole cell extract (approximately 20 µg of protein) was diluted with one volume of Laemmli sample buffer, electrophoresed on 12% polyacrylamide gel, electrotransferred onto PVDF and probed with a monoclonal antibody directed to ubiquitin. For equal protein loading, the same blot was stripped and reprobed with an anti-actin monoclonal antibody. (A) Cells exposed to nelfinavir. (B) Cells exposed to saquinavir. the two PIs nelfinavir and saquinavir, at clinically rele- (A) and (B): lane 1: lysates from cells supplemented with drug vehicle; lane 2: vant dosages (5–10 µM), are inhibitors of human 20S lysates from cells supplemented with lactacystin; lanes 3 and 4: lysates from cells supplemented with the drugs at 5–10 µM. Molecular mass markers were: and 26S proteasome peptidase activity in vitro, as well rabbit muscle myosin (205 kDa), rabbit muscle phosphorylase b (98 kDa), bovine as of intracellular proteasome proteolytic activity. In serum albumin (67 kDa), ovalbumin (45 kDa) and bovine carbonic anhydrase (29 kDa). contrast, the RTIs examined, with the exception of lamivudine and zidovudine, are ineffective at therapeutic dosages. Thus, these two RTIs and the two PIs have an additional target in a range of concentrations effects in vitro should offer clues on how they work in higher than the one for which they were designed vivo in HIV-infected patients on treatment. [35,36]. A further finding of the study is that the two With the exception of PGPH activity of the 20S PIs, in vitro and in intact cells, inhibit proteasome proteasome, which is not affected by saquinavir only, activity at the same dosages. An understanding of drug all the other activities are inhibited by the two drugs in

both proteasomes. It emerges that in affecting peptidase last only as long as the drug is bound to the subunit, activity of the 20S and 26S proteasomes, the two drugs since drug inhibition is reversible. Plasticity of subunit target the active subunits of either proteasome [12] and conformation should confer the ability on proteasomes not the 19S regulator, which is a component of the 26S to respond to changing environmental conditions to proteasome only. Furthermore, inhibition exerted by match new challenges [38]. With the exception of nelfinavir on trypsin-like activity of the 26S proteasome zidovudine and lamivudine, inability of the RTIs, is stronger than that exerted on the same activity of the whether alone or in combination, to affect even one of 20S proteasome; however, inhibition of chymotrypsin- the peptidase activities of either proteasome or to inter- like activity of the 26S proteasome is weaker than that fere with the effects of PIs, suggests that these drugs do exerted on the same activity of the 20S proteasome. not target the subunits because either they do not enter One explanation for this comes from the fact that the the catalytic core, or do so but find an unreceptive subunits of the 20S and 26S proteasomes have the same subunit conformation. primary structure but different tertiary structure due to Preparation of SDS-treated lysates is a rapid proce- interaction of the 20S proteasome with the 19S regu- dure applicable to small samples, different from the one lator [21–24,33,34]. These structural differences are needed to produce purified proteasomes. Therefore, this responsible for the different activity levels that the procedure may offer preliminary and reliable informa- catalytically active subunits have in the two protea- tion on 20S proteasome chymotrypsin-like sensitivity to somes [21–24,33,34]. The respective tertiary structures pharmacological agents when only small samples are should explain not only the reported differences in drug available. Of note, in lysates of human cancer cells, effects on the same peptidase activity of the two protea- inhibition of 20S proteasome chymotrypsin-like activity somes, but also ineffectiveness of saquinavir on PGPH by saquinavir has been observed in comparable experi- activity of the 20S proteasome, as opposed to its effec- mental conditions [14]. tiveness on the same activity of the 26S proteasome. In In previous studies, the two PIs, ritonavir [8] and indi- this view, proteasome subunit dynamics is central to navir [13], were seen to affect proteasome proteolytic drug effectiveness. An alternative explanation is offered activity in agreement with what has been observed in by the ‘two-site modifier’ model that assumes that the this study with the two PIs nelfinavir and saquinavir. effector (in our case the drug) binds to an active site or Concerning the RTIs in a previous study [13], as well as to a second non-catalytic modifier site that in turn in this study (Table 2), zidovudine and lamivudine were affects the active site [37]. seen to inhibit proteasome activity; however, in this In red blood cell lysates exposed to SDS, only study five RTIs were shown to be ineffective. The non- chymotrypsin-like activity of the 20S proteasome is appre- uniform effects that the RTIs have on proteasome ciable. Specificity of this activity is underscored firstly by function compared with the examined PIs, show that the the strong inhibition exerted by lactacystin – almost RTIs are not a uniform category of drugs with respect to complete in the presence of epoxomicin – and secondly by effects on proteasomes. The question arises whether the cleavage of MeOSuc-GLF-AMC the analogues of drugs affecting proteasome function share a common which are resistant to the action of other intracellular target on the proteasome and whether they have proteases [28]. Here too, irrespective of the peptide common molecular features not shared with the ineffec- substrate used, chymotrypsin-like activity is affected by tive drugs, irrespective of whether they are RTIs or PIs. nelfinavir and saquinavir, and inhibition occurs with the Inhibition of proteasome function in intact cells by drugs at therapeutic dosages, highlighting that drug effi- nelfinavir and saquinavir at clinically relevant dosages ciency is unaffected by the intracellular milieu. strongly supports the view that, in individuals on anti- In the presence of SDS, inhibition in lysates is retroviral therapy including these two PIs, at least a part stronger than that of purified 20S proteasome. Of note, of the large array of cellular processes in which protea- SDS modifies structural properties of 20S proteasome somes are implicated should be down-regulated. Virus subunits with consequent modifications of the relative aggressiveness should also be attenuated in the pres- baseline peptidase activities [21]. Therefore, besides ence of the two drugs, since the virus exploits host cell proteasome commutation, external agents, including proteasomes for its maturation [19,39–41]. However, drugs, which succeed in modifying proteasome subunit ineffectiveness of the majority of the RTIs on protea- structural conformation, will also affect peptidase some peptidase and proteolytic activity suggests that activity. Consequently, inhibition of chymotrypsin- and these drugs do not mediate their antiretroviral action trypsin-like activities by saquinavir and nelfinavir in through a modification of proteasome-controlled combination, which in either proteasome is intermediate cellular processes. Competition between the two PIs on to that evoked by each drug alone, should signal compe- proteasome peptidase activity in vitro is an indication tition between the drugs in determining a new subunit that the use of these two drugs in combination in the structural conformation. The new conformation should treatment of HIV-infected patients should be avoided.

Antiviral Therapy 10:2 221 M Piccinini et al.

Saquinavir and ritonavir have been described as dell’Adolescenza, University of Turin, Turin, Italy) for affecting a number of cellular processes and responses technical assistance. Sponsorship: Ministero della including antigen processing and presentation, activa- Sanità, Istituto Superiore di Sanità, IV AIDS project, tion of the transcriptional nuclear factor NF-κB Italy. associated with degradation of its inhibitory factor IκB, and production and release of inflammatory cytokines References and chemokines such as tumor necrosis factor (TNF)-α, 1. Perrin L & Telenti A. HIV treatment failure: testing for interleukins 6 and 8 from peripheral mononuclear HIV resistance in clinical practice. Science 1998; blood cells (PBMCs) and endothelial cells. Furthermore, 280:1871–1873. ritonavir, saquinavir, indinavir and nelfinavir are 2. Ledergerber B, Egger M, Opravil M, Telenti A, Hirschel B, Battegay M, Vernazza P, Sudre P, Flepp M, Furrer H, capable of blocking activation of endothelial cells and T Francioli P & Weber R. Clinical progression and virologic cells, as well as maturation and function of dendritic failure on highly active antiretroviral therapy in HIV-1 patients: a prospective cohort study. Swiss IV Cohort cells. In addition, these drugs interfere with prolifera- Study. Lancet 1999; 13:863–868. tion, apoptosis and differentiation in a variety of cells 3. Heagy W, Crumpacker C, Lopez PA & Finberg RW. including PBMCs, T cells, adipocytes and osteoblasts. A Inhibition of immune function by antiviral drugs. Journal number of these actions are now thought to rely on a of Clinical Investigation 1991; 87:1916–1924. 4. Andersson J, Fehniger TE, Patterson BK, Pottage J, Agnoli drug-mediated modulation of proteasome function (see M, Jones P, Behbahani H & Landay A. Early reduction of [42] for a review). Therefore, although assessed in HIV- immune activation in lymphoid tissue following highly free experimental systems, these effects may all active HIV therapy. AIDS 1998; 30:F123–F129. 5. Chavan S, Kodoth S, Pawha R & Pawha S. The HIV contribute to the effectiveness of HAART in HIV-infected protease inhibitor indinavir inhibits cell-cycle progression patients by down-regulating otherwise up-regulated in vitro in lymphocytes of HIV-infected and uninfected cellular functions due to the virus stimuli or to oppor- individuals. Blood 2001; 98:383–389. 6. Sloand EM, Kumar PN, Kim S, Chaudhuri A, Weichold FF tunistic infections frequently associated with AIDS. In & Young NS. Human immunodeficiency virus type 1 this respect, it is of note that in subjects undergoing PI- protease inhibitor modulates activation of peripheral blood based HAART as prophylactic therapy without acquired CD4(+) T cells and decreases their susceptibility to apoptosis in vitro and in vivo. Blood 1999; 94:1021–1027. HIV infection, production of inflammatory cytokines, 7. Sloand EM, Maciejewski J, Kumar P, Kim S, Chaudhuri A, including TNF-α, interferon-γ and interleukin 2, is Weichold FF & Young N. Protease inhibitors stimulate defective [11], an indication that, in triggering this hematopoiesis and decrease apoptosis and ICE expression in CD4(+) cells. Blood 2000; 96:2735–2739. reduction in vivo, the drugs are working independently 8. Andre P, Groettrup M, Klenerman P, de Giuli R, Booth BL of their inhibitory action on virus replication. Jr, Cerundolo V, Bonneville M, Jotereau F, Zinkernagel RM Final confirmation of the regulatory role of HIV & Lotteau V. An inhibitor of HIV-1 protease modulates proteasome activity, antigen presentation, and T cell enzyme inhibitors on host cell proteasome function responses. Proceedings of the National Academy of will be achieved by the assessment of proteasome Sciences, USA 1998; 95:13120–13124. peptidase and proteolytic activities in PBMCs of HIV- 9. Lu W & Andrieu J-M. HIV protease inhibitors restore impaired T-cell proliferative response in vivo and in vitro: a infected patients before and during treatment. Such a viral-suppression-independent mechanism. Blood 2000; study would be of great relevance were it performed on 96:250–258. primary CD4+ cells from a cohort of HIV-infected 10. Hosseini H, Andre P, Lefevre N, Viala L, Walzer T, Peschanski M & Lotteau V. Protection again experimental patients in which these cells are numerically stable autoimmune encephalomyelitis by a proteasome modulator. despite virological failure, or alternatively in which Journal of Neuroimmunology 2001; 118:233–244. counts of these cells decrease following therapy failure. 11. Tovo P-A. Highly active antiretroviral therapy inhibits cytokine production in HIV uninfected subjects. AIDS However, based on the experience acquired with 2000; 14:743–744. human red blood cells, there are difficulties in 12. Schmidtke G, Holzhutter HG, Bogyo M, Kairies N, Groll collecting PBMCs and/or CD4+ cells from such indi- M, de Giuli R, Emch S & Groettrup M. How inhibitor of the HIV-I protease modulates proteasome activity. Journal viduals in sufficient amounts to evaluate specific of Biological Chemistry 1999; 274:35734–35740. proteasome peptidase activity; furthermore, variability 13. Piccinini M, Rinaudo MT, Capello N, Ricotti E, Mostert of proteasome peptidase activity in CD4+ cells from M & Tovo PA. The human 26S proteasome is a target of antiretroviral agents. AIDS 2002; 16:693–700. patients is likely to be considerable; thus, a suitable 14. Pajonk F, Himmelsbach J, Riess K, Sommer A & McBride number of cases would have to be examined. WH. The human immunodeficiency virus (HIV)-1 protease inhibitor saquinavir inhibits proteasome function and causes apoptosis and radiosensitization in non-HIV-associated Acknowledgements human cancer cells. Cancer Research 2002; 62:5230–5235. 15. Baldwin AS Jr. The NF-kappaB and IkappaB proteins: new The authors would like to thank Dr S Vella (Istituto discoveries and insights. Annual Review of Immunology Superiore di Sanità, Rome, Italy) for the generous gift 1996; 14:649–681. 16. Schwartz AL & Ciechanover A. The ubiquitin-proteasome of drugs and critical revision of the manuscript, and Dr pathway and pathogenesis of human diseases. Annual G Bertolotto (Dipartimento di Scienze Pediatriche e Review of Medicine 1999; 50:57–74.

17. Kisselev A & Goldberg AL. Proteasome inhibitors: from procedure and some applications. Proceedings of the research tools to drug candidates. Chemistry & Biology National Academy of Sciences, USA 1979; 76:4350–4354. 2001; 8:739–758. 31. Piccinini M, Merighi A, Bruno R, Cascio P, Curto M, 18. Glickmann MH & Ciechanover A. The ubiquitin-protea- Mioletti S, Ceruti C & Rinaudo MT. Affinity purification some proteolytic pathway: destruction for the sake of and characterization of protein gene product 9.5 (PGP 9.5) construction. Physiological Reviews 2002; 82:373–428. from retina. Biochemical Journal 1996; 318:711–716. 19. Schubert U, Ott DE, Chertova EN, Welker R, Tessmer U, 32. Bradford MM. A rapid and sensitive method for the quan- Princiotta MF, Bennink JR, Krausslich HG & Yewdell JW. titation of microgram quantities of protein utilizing the Proteasome inhibition interferes with gag polyprotein principle of protein-dye binding. Analytical Biochemistry processing, release, and maturation of HIV-1 and HIV-2. 1976; 72:248–254. Proceedings of the National Academy of Sciences, USA 2000; 97:13057–13062. 33. Braun BC, Glickman M, Kraft R, Dahlmann B, Kloetzel PM, Finley D & Schmidt M. The base of the proteasome 20. Hershko A & Ciechanover A. The ubiquitin system. regulatory particle exhibits chaperone-like activity. Nature Annual Review of Biochemistry 1998; 67:425–479. Cell Biology 1999; 1:221–226. 21. Coux O, Tanaka K & Goldberg AL. Structure and func- 34. Bochtler M, Ditzel L, Groll M, Hartmann C & Huber R. tions of the 20S and 26S proteasomes. Annual Review of The proteasome. Annual Review of Biophysics & Biochemistry 1996; 65:801–847. Biomolecular Structure 1999; 28:295–317. 22. Baumeister W, Walz J, Zühl F & Seemüller E. The proteasome: paradigm of a self-compartimentalizing protease. 35. Gaucher B, Rouquayrol M, Roche D, Greiner J, Aubertin Cell 1998; 92:367–380. AM & Vierling P. Prodrugs of HIV protease inhibitors- saquinavir, indinavir and nelfinavir-derived from 23. Ferrell K, Wilkinson CRM, Dubiel W & Gordon C. diglycerides or amino acids: synthesis, stability and anti- Regulatory subunit interactions of the 26S proteasome, a HIV activity. Organic & Biomolecular Chemistry 2004; complex problem. Trends in Biochemical Sciences 2000; 2:345–357. 25:83–88. 36. Miller JF, Furfine ES, Hanlon MH, Hazen RJ, Ray JA, 24. Piccinini M, Tazartes O, Mostert M, Musso A, DeMarchi Robinson L, Samano V & Spaltenstein A. Novel arylsul- M & Rinaudo MT. Structural and functional characteriza- fonamides possessing sub-picomolar HIV protease activities tion of 20S and 26S proteasomes from bovine brain. Brain and potent anti-HIV activity against wild-type and drug- Research. Molecular Brain Research 2000; 76:103–112. resistant viral strains. Bioorganic & Medicinal Chemistry 25. Piccinini M, Tazartes O, Mezzatesta C, Ricotti E, Bedino S, Letters 2004; 14:959–963. Grosso F, Dianzani U, Tovo PA, Mostert M, Musso A & 37. Schmidtke G, Emch S, Groettrup M & Holzhutter HG. Rinaudo MT. Proteasomes are a target of the antitumor Evidence for the existence of a non-catalytic modifier site drug vinblastine. Biochemical Journal 2001; 356:835–841. of peptide hydrolysis by the 20S proteasome. Journal of 26. Fenteany G, Staendaert RF, Lane WS, Choi S, Corey EJ & Biological Chemistry 2000; 275:22056–22063. Schreiber SL. Inhibition of the proteasome activities and subunit-specific amino-terminal threonine modification by 38. Husom AD, Peters EA, Kolling EA, Fugere NA, Thompson lactacystin. Science 1995; 268:726–731. LV & Ferrington DA. Altered proteasome function and subunit composition in aged muscle. Archives of 27. Meng L, Mohan R, Kwok BH, Elofsson M, Sin N & Biochemistry & Biophysics 2004; 421:67–76. Crews CM. Epoxomicin, a potent and selective proteasome inhibitor, exhibits in vivo antiinflammatory activity. 39. Bultmann A, Eberle J & Haas J. Ubiquitination of human Proceedings of the National Academy of Sciences, USA immunodeficiency virus type 1 env glycoprotein. Journal of 1999; 96:10403–10408. Virology 2000; 74:5373–5376. 28. Harding CV, France J, Song R, Farah JM, Chatterjee S, 40. Mulder LC & Muesing MA. Degradation of HIV-1 inte- Iqbal M & Siman R. Novel dipeptide aldehydes are protea- grase by the N-end rule pathway. Journal of Biological some inhibitors and block the MHC-I antigen-processing Chemistry 2000; 275:29749–29753. pathway. Journal of Immunology 1995; 155:1767–1775. 41. Patnaik A, Chau V & Wills JW. Ubiquitin is part of the 29. Rabilloud T. Silver staining of 2-D electrophoresis gels. In retrovirus budding machinery. Proceedings of the National 2-D Proteome Analysis Protocol, 1999; pp. 297–305. Academy of Sciences, USA 2000; 97:13069–13074. Edited by AJ Link. Totowa, NJ: Humana Press. 42. Sgadari C, Monini P, Barillari G & Ensoli B. Use of HIV 30. Towbin H, Staehelin T & Gordon J. Electrophoretic transfer protease inhibitors to block Kaposi’s sarcoma and tumour of proteins from polyacrylamide gels to nitrocellulose sheets: growth. Lancet Oncology 2003; 4:537–547.

Received 7 June 2004, accepted 24 December 2004

Antiviral Therapy 10:2 223