Full length article

In vivo administration of ritonavir worsens intestinal damage causedby cyclooxygease inhibitors

Barbara Renga a,n, Andrea Mencarelli a,n, Sabrina Cipriani a, Claudio D’Amore a,Daniela Francisci b, Luca Santucci c, Franco Baldelli b, Eleonora Distrutti c, Stefano Fiorucci a

a Dipartimento di Medicina Clinica e Sperimentale, Università di Perugia, Nuova Facoltà di Medicina e Chirurgia, Via Gerardo Dottori 1S. Andrea delle Fratte,Perugia 06132, Italyb Dipartimento di Medicina e Scienze Biochimice, Università di Perugia, Nuova Facoltà di Medicina e Chirurgia, Via Gerardo Dottori 1S. Andrea delle Fratte,Perugia 06132, Italyc Azienda Ospedaliera di Perugia, Ospedale Santa Maria della Misericordia, S. Andrea delle Fratte, 06132 Perugia, Italy

a r t i c l e i n f o

Article history:Received 26 September 2013Received in revised form27 November 2013Accepted 27 November 2013Available online 4 December 2013

Keywords:Acetylsalicylic acidNaproxenKetoprofenRitonavirMisoprostolNon-steroidal anti-inflammatory drugsIntestinal damage

a b s t r a c t

The protease inhibitor ritonavir is part of the highly active anti-retroviral therapy (HAART) successfullyused in the treatment of human immunodeficiency virus (HIV)-1 infection. There is evidence that ritonaviralters intestinal permeability and induces damage to the small intestine. Because HIV infected patientstaking HAART are at high risk for developing cardiovascular complications, there might be a need for theuse of low dose of aspirin (ASA) to prevent ischemic events. Similarly, long term survival exposes HIVinfected persons to detrimental interactions of ritonavir with non-steroidal anti-inflammatory drugs(NSAIDs). In the present work we tested whether ritonavir worsens intestinal injury caused by NSAIDs andASA. C57BL6 mice were treated for 25 days with ritonavir and for a further 5 days with the combination ofritonavir plus ASA or ritonavir plus naproxen. In a second set of experiments C57BL6 mice were cotreatedwith ritonavir plus misoprostol, a PGE1 analog. We found that ritonavir administration caused intestinaldamage and its co-administration with naproxen or ASA exacerbated the severity of injury and intestinalinflammation, as assessed by measuring haematocrit, MPO, mucosal levels of PGE2 and mRNA levels ofiNOS, MCP-1 and VLA-1. Co-administration of misoprostol protected against intestinal damage induced bynaproxen and ritonavir. In conclusion we demonstrated that ritonavir causes intestinal damage and that itsassociation with NSAIDs or ASA worsens the damage caused by COX-inhibitors. Misoprostol rescues fromintestinal damage caused by ritonavir. Further studies are need to clarify whether this observation has aclinical readout.

& 2013 Elsevier B.V. All rights reserved.

1. Introduction

Gastrointestinal ulcerations and bleedings caused by acetyl-salicylic acid (ASA) and traditional non steroidal anti-inflammatorydrugs (tNASIDs) are well recognized complication related to the use ofthese agents occurring at a rate of E3% (Abraham et al., 2010;Santucci et al., 1995). Despite the inhibition of gastric acid secretionby proton pump inhibitor (IPP) has been demonstrated effective inreducing the incidence of gastro-duodenal ulcers and gastroduodenalbleeding, this approach has been shown to be poorly effective inprotecting against intestinal injury. The basic mode of action of ASAand non selective NSAIDs lies in the inhibition of cyclo-oxygenases(COX), with COX-2 generating prostaglandins at site of inflammationand COX-1 being responsible for generation of prostaglandins involved

in protecting the gastrointestinal mucosa (Catella-Lawson et al., 2001;Fiorucci et al., 2007, 2003). Thus, while tNSAIDs are thought to injurythe gastrointestinal tract through a COX-1 related mechanism, selec-tive COX-2 inhibitors effectively reduce both gastro-duodenal andintestinal complications, but their use associates with an increased riskto develop cardiovascular ischemic events (Catella-Lawson et al., 2001;Fiorucci et al., 2007, 2003). Because the above limitations, there is stilla need for the discovery of safer NSAIDs (Fiorucci et al., 2007, 2003;Fiorucci, 2001) or effective co-treatments that could limit intestinaldamage in specific subset of NSAID-taking patients.

Protease inhibitors (PI) as a part of highly active anti-retroviraltherapy (HAART) have been used successfully in the treatment ofhuman immunodeficiency virus (HIV)-1 infection. Incorporation ofHIV protease inhibitors in the HAART causes profound and sustainedsuppression of viral replication, significantly reduces the morbidityand mortality, and prolongs the lifespan of patients with HIV infection(Riddle et al., 2001; Moyle and Carr, 2002; Spector, 2003). HAART haschanged the clinical profile of HIV infection from a sub-acute lethal

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/ejphar

European Journal of Pharmacology

0014-2999/$ - see front matter & 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.ejphar.2013.11.035

n Corresponding authors. Tel./fax: þ39 75 585 8120.E-mail address: barbara.renga@unipg.it (B. Renga).

European Journal of Pharmacology 723 (2014) 194–201

disease to a chronic ambulatory disease (Clevenbergh et al., 2003).Despite its efficacy in controlling the disease progression, the use of PItherapy associates with an increased risk of development of prematureatherosclerosis. An increasing body of evidence suggests that treat-ment of HIV-infected patients with HIV PIs causes a dyslipidemiawhich contributes to the development of cardiovascular diseases(Spector, 2003; Clevenbergh et al., 2003; Cohen, 2005). A significantincrease in plasma triacylglycerols and total cholesterol concentrations,often associated with abnormal body fat distribution and peripheralinsulin resistance (hyperinsulinemia, hyperglycemia and diabetesmellitus), has been detected in HIV PIs-treated patients (Spector,2003; Clevenbergh et al., 2003; Cohen, 2005; Dubé, 2000) makingthese subjects putative candidates for the treatment with ASA and/oranti platelet drugs. HIV PI causes gastrointestinal adverse effectsincluding abdominal pain and diarrhea (Wallace and Brann, 2000).Recent studies have shown that the two of most commonly used HIVPIs, ritonavir and lopinavir, trigger apoptosis of intestinal cells byactivating endoplasmic reticulum (ER) stress, thus disrupting theintestinal epithelial barrier integrity both in in vitro cell cultures andin mice (Wu et al., 2010).

Because tNSAIDs and ASA might be needed in long-term survivalHIV infected persons we have designed a study to investigatewhether ritonavir exacerbates intestinal injury caused by tNSAIDsand ASA (Zhang et al., 2013; Wallace et al., 2011) and definemechanisms involved in these interactions.

2. Materials and methods

2.1. Materials

Aspirin (ASA), naproxen, ketoprofen, misoprostol and ritonavirwere purchased from Sigma. Prostaglandins E1 and E2 (PGE1 andPGE2) were purchased from Cayman Chemical.

2.2. Cells

HT29 cells, a human colon adenocarcinoma cell line, werecultured in RPMI-1640 medium containing 10% FCS, 1% glutamineand antibiotics.

2.3. In vitro detection of apoptosis

To test whether ritonavir, TNFα and naproxene drive colon cellsto apoptosis HT29 (300.000 cells/wells) were incubated 36 h withmedium alone or with the following reagents: ritonavir (15 μM),naproxen (50 μM), TNFα (100 ng/ml), the combination of ritonavirplus naproxen or ritonavir plus TNFα or ritonavir plus naproxeneplus TNFα. To test whether ritonavir, TNFα and ketoprofen drivecolon cells to apoptosis HT29 (300.000 cells/wells) were incubated12, 24, 36 and 48 h with medium alone or with the followingreagents: ritonavir (15 μM), ketoprofen (50 μM), TNFα (100 ng/ml), the combination of ritonavir plus ketoprofen or ritonavir plusTNFα or ritonavir plus ketoprofen plus TNFα. To investigate theeffect of PGE1 and PGE2 on Ritonarir induced apoptosis HT29 cellswere incubated 36 h with ritonavir (15 μM) alone or in thepresence of prostaglandins PGE1 (1 μM) or PGE2 (5 μM).

Apoptosis was detected by staining the cells with propidiumiodide (PI). Briefly, cell pellets were washed twice in PBS, resus-pended in hypotonic fluorochrome solution (50 μg/ml PI in 0.1%sodium citrateþ0.1% Triton X-100), kept 4–8 h at 4 1C in the dark,and analyzed using a Epics XL flow cytometer (Beckman-Coulter,Miami, FL). The percentage of apoptotic cells was determined byevaluating hypodiploid nuclei after proper gating on DNA content.

2.4. Assessment of caspase 3 and 8 activity

After incubation with appropriate agents HT-29 cells were recov-ered into lysis buffer (10 mM Tris–HCl (pH 7.3), 25 mM NaCl, 0.25%Triton X-100, 1 mM EDTA). After centrifugation at 13,000 rpm for30 min at 4 1C, the resulting supernatants were adjusted to 1 mg/mlwith lysis buffer and 25 μg total proteins were incubated in 100 μl ofcaspase buffer (50 mM HEPES (pH 7.2), 100 mM NaCl, 1 mM EDTA(pH 8.0), 10% sucrose, 0.1% 3-[(3-cholamidopropyl)dimethylammo-nio]-1-propanesulfonate (CHAPS), and 1 mM DTT) with variousfluorogenic substrate peptides (100 μM) including acetyl-Asp-Glu-Val-Asp-(7-amino-4-trifluoromethyl-coumarin) (Ac-DEVD-AFC) forcaspase-3 and Ac-Ile-Glu-Thr-Asp-AFC (Ac-IETD-AFC) for caspase-8.Caspase-3 and -8 activities were assayed using a fluorimeter platereader in kinetic mode with excitation and emission wavelengths of405 and 519 nm, respectively, continuously measuring release of AFCfrom substrate peptides (Fiorucci et al., 2003).

2.5. Animal protocols

C57BL6 were from Harlan Nossan (Udine, Italy). Mice were housedunder controlled temperatures (22 1C) and photoperiods (12:12-hlight/dark cycle), allowed unrestricted access to standard mouse chowand tap water and allowed to acclimate to these conditions for at least5 days before inclusion in an experiment. Protocols were approved bythe University of Perugia Animal Care Committee. The ID for thisproject is #98/2010-B. The authorization was released to Prof. StefanoFiorucci, as a principal investigator, on May 19, 2010.

2.6. Intestinal damage, intestinal mucosal prostanoids, haematocritand MPO

Groups of six mice (not fasted) were randomized and treated asfollows: control group: mice received vehicle alone (a solution of 1%methylcellulose); ritonavir group: mice were treated daily for 25 dayswith ritonavir (50 mg/kg/day/OS); ASA group: mice were treated forfive days with ASA (50 mg/kg/day/OS); naproxene group: mice weretreated for five days with naproxen (100mg/kg/OS); ritonavir plus ASAgroup: mice were administered 25 days with ritonavir (50 mg/kg/day/OS) and for other five days with the combination of ritonavir plus ASA(50 mg/kg/day/OS); ritonavir plus naproxen group: mice were admi-nistered 25 days with ritonavir (50 mg/kg/day/OS) and for other fivedays with the combination of ritonavir plus naproxene (100 mg/kg/OS). In a second set of experiment C57BL6 mice were randomized andtreated as follows: control group: mice received vehicle alone(a solution of 1% methylcellulose); ritonavir group: mice were treatedfor 25 days with ritonavir (50 mg/kg/day/OS); naproxene group: micewere treated for five days with naproxen (100 mg/kg/OS); ritonavirplus naproxene group: mice received 25 days ritonavir (50 mg/kg/day/OS) and for other five days with the combination of ritonavir plusnaproxen (100 mg/kg/OS); ritonavir plus misoprostol: mice received25 days ritonavir (50 mg/kg/day/OS) and for other five days with thecombination of misoprostol (100 μg/kg/OS) plus ritonavir. Six hourafter the final administration of drugs, the mice were anesthetizedwith sodium pentobarbital (100 mg/kg) and blood sample was drawn,by intracardiac puncture, for determination of haematocrit (Fiorucciet al., 2003). The small intestine was excised, and the extent ofhaemorrhagic damage to the small intestine blindly quantified, undera microscope, by measuring the lengths of each lesion in mm and thensumming these to obtain a damage score for each mice (Fiorucci et al.,2003). Generation of PGE2 by gastric and intestinal mucosa wasmeasured according to previously published methods (Fiorucci et al.,2003), using a specific ELISA kit (Cayman Chemical Company, AnnArbor, MI). Gastric MPO activity was measured using a spectrophoto-metric assay with tri-methylbenzidine (TMB) as a substrate. Activity isexpressed as mU per mg protein.

B. Renga et al. / European Journal of Pharmacology 723 (2014) 194–201 195

2.7. RT-PCR

The method and conditions for RT-PCR are the same describedelsewhere (Fiorucci et al., 2005). Forward and reverse primers usedwere as follow: mGAPDH: ctgagtatgtcgtggagtctac and gttggtggtgcag-gatgcattg; mTNFα: acggcatggatctcaaagac and gtgggtgaggagcacgtagt;mCOX1: tgccctctgtacccaaagac and tgtgcaaagaaggcaaacag; mCOX2:agaaggaaatggctgcagaa and gctcggcttccagtattgag; meNOS: agaagagtc-cagcgaacagc and tgggtgctgaactgacagag; miNOS: acgagacggataggcagagaand cacatgcaaggaagggaact; mMCP1: cccaatgagtaggctggaga and tctgga-cccattccttcttg; mVLA1: tgtacccaattggatggtca and cagtcctggattgtgcctct.

2.8. Statistical analysis

All data are presented as the mean7S.E.M. Comparisons ofgroups of data were performed using a 1-way analysis of variancefollowed by the Student–Newman–Keuls post hoc test. An associatedprobability (P value) of less than 5% was considered significant.

3. Results

3.1. Ritonavir enhances the pro-inflammatory effects of ASA in thesmall intestine

To investigate the role of ritonavir in the regulation of intestinalhomeostasis C57BL6 mice were treated 25 days with ritonavir and

then challenged for others 5 days with the combination ofritonavir plus ASA. As illustrated in Fig. 1A, the intestinal mucosaldamage caused by administering mice with ritonavir was less thanthat caused by ASA. However, the mucosal damage was almostdouble in mice treated with the combination of ritonavir plus ASAin comparison with mice treated only with ASA. However, ritona-vir did not increase gastric mucosal injury when co-administeredwith ASA or naproxen (Supplementary Fig. 1).

The mechanism by which ritonavir induces intestinal damagedoes not seem to be due to a different inhibition of prostanoids,because as shown in Fig. 1B, ritonavir per se do not caused asuppression PGE2 formation and the treatment of ritonavir plusASA caused a same suppression of PGE2 formation in comparisonwith the administration of ASA alone. We have therefore assessedthe intestinal levels of MPO, a classical marker of inflammationwhich reflects a neutrophil infiltration. As shown in Fig. 1C, micechallenged with ritonavir had higher mucosal levels of MPO thancontrol mice. The same levels were found in mice treated with ASAalone while the combination of ritonavir plus ASA furtherincreases the levels of MPO. We found no significant decrease inhaematocrit in response to ASA or ritonavir but the combination ofthe two significantly decreased the volume percentage of redblood cells in blood (Fig. 1D). The observation of macroscopicappearance of intestinal injury confirmed a very little damage inthe intestine of mice treated only with ritonavir or ASA, whilethe combination of ritonavir plus ASA resulted in ulcerationsand bleeding of the small intestine (Fig. 1L). Finally, we applied a

NAIVE

ASARIT

RIT+ ASA

0

10

20

30

Inte

stin

al d

amag

e(m

m2 )

0

25

50

75

MP

O(m

Uni

ts/m

g of

pro

tein

)

25

40

55

*H

aem

atoc

rit (%

)

0

25

50

75

muc

osal

PG

E-2

(pg/

mg

of p

rote

in)

0.0

2.5

5.0

7.5

10.0

iNO

S(m

RN

A le

vels

)

0

25

50

75

MC

P-1

(mR

NA

leve

ls)

NAIVE

ASARIT

RIT+ ASA

NAIVE

ASARIT

RIT+ ASA

NAIVE

ASARIT

RIT+ ASA

Naive ASA Ritonavir + ASARitonavir

0.0

0.5

1.0

1.5

2.0

2.5

TNF

(mR

NA

leve

ls)

0

1

2

3

eNO

S(m

RN

A le

vels

)

0

1

2

3

VLA

-1(m

RN

A le

vels

)

Fig. 1. Effect of aspirin (ASA) and ritonavir on small intestine mucosa after the administration of ASA, ritonavir or the combination of both drugs in C57BL6 mice.(A) Intestinal mucosal damage. (B) Mucosal content of prostaglandin E2 (PGE2). (C) Mucosal myeloperoxidase activity (MPO). (D) Haematocrit. ((E)–(I)) Relative mRNAexpression of iNOS, eNOS, TNFα, VLA-1 and MCP-1 was assayed by quantitative Real-Time PCR. Values are normalized relatively to GAPDH mRNA. (J) Macroscopic analysis ofsmall intestine. Data are expressed as means7S.E.M. of 6 mice/group. nPo0.05 vs not treated mice. #Po0.05 vs ritonavir administered mice.

B. Renga et al. / European Journal of Pharmacology 723 (2014) 194–201196

RT-PCR investigation to analyze the expression profile of severalpro-inflammatory mediators. Fig. 1E–I demonstrates that micetreated with ritonavir were characterized by an enhanced expres-sion of eNOS and MCP-1. The treatment with ASA significantlyincreased the mRNA relative expression of iNOS, TNFα, and MCP-1while did not changed the mRNA levels of eNOS and VLA-1. Thecombination of ritonavir plus ASA exacerbated the induction ofMCP-1 and iNOS and significantly up-regulates VLA-1.

3.2. Ritonavir worsens intestinal injury caused by naproxen

We have next investigated whether exposure to ritonavir exacer-bates intestinal damage induced by naproxen, a widely used tNSAIDs(Fiorucci et al., 2007). As shown in Fig. 2, we have confirmed that,in comparison with mice treated exclusively with naproxen, thecombination of ritonavir plus naproxen worsened the intestinaldamage score and boosted the MPO levels (Fig. 2A and B). Astatistically significant reduction of the haematocrit (Fig. 2C) wasobserved. In addition, we confirmed that the treatment with ritonavir,per se, did not change the PGE2 levels and that the co-treatment ofritonavir plus naproxen caused a same suppression of PGE2 formationwhen compared with mice administrated only with naproxen alone(Fig. 2D). Finally, the macroscopic analyses of the small intestinerevealed that administration of naproxen resulted in extensive

intestinal ulceration and bleeding. However, while ritonavir induceda very minor damage, the combination of these two agents stronglyexacerbated the intestinal damage induced by naproxen (Fig. 2E).

3.3. Ritonavir induces apoptosis of colon epithelial cells

To gain insights on the mechanisms of damage involved in theabove mentioned effects, we have investigated whether ritonavirinduces apoptosis of intestinal epithelial cells. For this purposeHT29 cells, a human colon adenocarcinoma cell line, were used.As illustrated in Fig. 3A, treatment of HT29 cells with ritonavirresulted in a E10% increase of the percentage of cells on sub G1phase (apoptotic cells). By contrast, incubation of HT29 cells withnaproxen or TNFα failed to increase the apoptotic (Fig. 3A).Compared with cells exposed to ritonavir alone, the combinationof ritonavir plus naproxen failed to ncrease HT29 apoptosis, whilethe combinations of ritonavir plus TNFα or ritonavir plus TNFαand naproxen resulted in a 25 and 35% increase of apoptosis rate(Po0.05 vs ritonavir alone). These results associated with anincreased activity of caspase-3 and caspase-8 by (Fig. 3B and C).Once again, we observed that the combinations of ritonavir plusTNFα or ritonavir plus TNFα and naproxen resulted in a stronginduction of caspase 3 and 8 activities in comparison withritonavir alone (Po0.05 vs ritonavir alone). Similar results were

NAIVE

NAPROXRIT

RIT+NAPROX

0

10

20

30

40

50

inte

stin

al d

amag

e(m

m2 )

25

40

55

Hae

mat

ocrit

(%)

0

50

100

150

MP

O(m

Uni

ts/m

g of

pro

tein

)

0

25

50

75

muc

osal

PG

E-2

(pg/

mg

of p

rote

in)

NAIVE

NAPROXRIT

RIT+NAPROX

Naive Naproxen Ritonavir Ritonavir + Naproxen

Fig. 2. Effect of naproxen and ritonavir on small intestine mucosa after the administration of naproxen, ritonavir or the combination of both drugs in C57BL6 mice.(A) Intestinal mucosal damage. (B) Mucosal myeloperoxidase activity (MPO) (C). Haematocrit. (D) Mucosal content of prostaglandin E2 (PGE2). (E) Macroscopic analysis ofsmall intestine. Data are expressed as means7S.E.M. of 6 mice/group. nPo0.05 vs not treated mice. #Po0.05 vs ritonavir administered mice.

B. Renga et al. / European Journal of Pharmacology 723 (2014) 194–201 197

obtained using ketoprofen another tNSAID. Indeed, as shown inSupplementary Fig. 2, an increased rate of apoptosis was observedin HT29 cells exposed to TNFα or ketoprofen in the presence ofritonavir.

3.4. Misoprostol protects against intestinal damage induced byritonavir and naproxen

We have then investigated the hypothesis that the PGE1analogue misoprostol may prevent the intestinal damage inducedby ritonavir. As shown in Fig. 4, the severity of intestinal damagewas significantly reduced in animals receiving misoprostol com-pared to animals administered with ritonavir alone (Fig. 4A).The “clinical” readout of this protection translates in arobust reduc-tion of intestinal bleeding as demonstrated by the analysis thehaematocrit (Fig. 4B, Po0.05 verus ritonavir plus naproxen). Addi-tionally mucosal levels of MPO were reduced in mice administeredwith ritonavir plusthe PGE1 analogue compared with control animalsadministered only with ritonavir (Fig. 4C; Po0.05 vs ritonavir plusnaproxen). Finally, Real-Time PCR analysis of intestinal samplesdemonstrated that ritonavir significantly up-regulated the expressionof both COX-1 and COX-2 enzymes and that these changes werereversed by co-treating mice with misoprostol (Fig. 4D and E), alongwith other markers of intestinal inflammation such as MCP-1 andTNFα (Fig. 4G).

3.5. Apoptosis induced by ritonavir is reversed PGE1 and PGE2

Finally, we have investigated whether prostaglandins PGE1 andPGE2 protects HT29 cells against apoptosis induced by ritonavir andnaproxen or TNFα. As shown in Fig. 5A and B, incubation of HT29cells with PGE1 or PGE2 resulted in a robust reduction of apoptosisinduced by ritonavir and naproxen or ritonavir and TNFα.

4. Discussion

Gastrointestinal injury, ranging from acute erosions to ulcerbleeding and perforation, represents a well recognized complicationof the use of ASA and NSAIDs (Wolfe et al., 1999). The risk ofgastrointestinal bleeding in patients taking low doses of ASA(o100 mg/day) or medium doses of any tNSAID increases approxi-mately 4 folds in comparison with untreated populations (Wolfeet al., 1999). The pathogenesis of injury caused by ASA and NSAIDsis largely related to the suppression of COX-1 derived prostanoids inthe gastro-intestinal tract, however, additional mechanisms seem tobe involved as exemplified by the demonstration that COX-1 geneablation results in mutant mice that survive well, have no gastricpathology, and show less indomethacin-induced gastric ulcerationthan wild-type mice, even though their gastric PGE2 levels are onlyE1% of control mice (Langenbach et al., 1995).

0

25

50

75

Cas

pase

3 a

ctiiv

ity(m

UF/

g of

pro

tein

)

crtl

Rit

Naprox Tnf

Rit+Nap

rox

Rit +TNF

Rit+TNF

+Nap

rox0

10

20

30

40

% A

popt

otic

cel

ls

0

100

200

300

400

Cas

pase

8 a

ctiv

ity(m

UF/

g pr

otei

n)

crtl

Rit

Naprox

Rit+Nap

rox

Rit +TNF

Rit+TNF

+Nap

rox crtl

Rit

Naprox

Rit+Nap

rox

Rit +TNF

Rit+TNF

+Nap

rox

Fig. 3. Ritonavir induces apoptosis in colon adenocarcinoma cell line HT29. HT29 cells were left untreated (ctrl) or treated 36 h with ritonavir, naproxen, TNFα, ritonavir plusnaproxen, ritonavir plus TNFα and ritonavir plus naproxen plus TNFα. (A) Apoptosis was determined by staining the cells with propidium iodide and assaying the number ofapoptotic/necrotic cells at flow cytometry. (B) Caspase 3 activity. (C) Caspase 8 activity. Data are means7S.E.M. of 4 experiments. nPo0.01 vs control cells. #Po0.05 vs cellsstimulated with ritonavir.

B. Renga et al. / European Journal of Pharmacology 723 (2014) 194–201198

The introduction of HAART has deeply changed the clinicaloutcome of HIV-1 infection from an acute disease to a chronicillness. HIV-1 infected persons taking HAART survive decades andare therefore exposed to other chronic disorders (Mugavero et al.,2011; Yiannoutsos et al., 2012). In virologically suppressed personstaking HAART, a state of low grade chronic inflammation has beenassociated with an increased risk of developing cardiovascularcomplications (Lipshultz et al., 2012). In addition, HIV-PI inhibitors,specifically ritonavir, carry on a specific risk for causing cardiovas-cular diseases, at least in part due to their ability to induce insulinresistance and dyslipidemia (Cipriani et al., 2013a; Mencarelli et al.,2012, 2010). In addition, several studies have provided evidencethat ritonavir might increase intestinal permeability thus contribut-ing to a sub-chronic inflammation (Wu et al., 2010). In the presentstudy we have shown that treating mice with ritonavir exacerbatesintestinal injury caused by ASA, naproxen and ketoprofen, a widelyused tNSAIDs (Fiorucci et al., 2007). Co-treating mice taking ASAwith ritonavir increased the severity of intestinal injury and theexpression of markers of intestinal damage including iNOS, eNOS,TNF, VLA-1 and MCP-1 (Fiorucci et al., 2003). Importantly, thecombination of ritonavir and ASA resulted in a significant reductionof the hematocrit, a robust indicator of small intestinal injury.Similarly, intestinal injury caused by naproxen was worsened by

co-treating mice with ritonavir. The mechanism of injury caused byritonavir was prostaglandin-independent because in contrast toASA and naproxen, ritonavir failed to reduce the levels of intestinalPGE2. However, the prostaglandin-depleted environment facilitatesthe intestinal injury caused by ritonavir, as demonstrated by the factthat in vivo administration of misoprostol, a PGE1 analogue, rescuesfrom injury caused by naproxen and ritonavir.

In the present study we have shown that ritonavir causes a directinjury to epithelial cells. In contrast to naproxen, exposure to riton-avir directly increased the apoptotic rate of HT29 cells. This effect wasnot additive to that of naproxen, but was additive to TNFα, a cytokinethat is released directly in vivo in response to exposure to ASA andtNSAIDs (Fiorucci, 2009; Wallace, 2008; Fiorucci et al., 1999; Santucciet al., 1996, 1994). TNFα is well recognized for its role in regulatingthe activity of pro-apoptotic signals in the stomach and intestinalcells including caspase-3 and -8 (Fiorucci, 2001, 2009; Wallace, 2008;Fiorucci et al., 1999; Santucci et al., 1996, 1994). Importantly, we havefound that ritonavir but not naproxen, per se, increased the activity ofcaspase-3 and -8, and that the activities of these pro-apoptoticsignals were further enhanced by TNFα or the combination ofTNFα plus naproxen with ritonavir (Fiorucci, 2009; Wallace, 2008;Fiorucci et al., 1999; Santucci et al., 1996, 1994). Noteworthy,Ritonavir also has been reported to inhibit the activation of NF-kB

0

25

50

75

100

naive

napro

x.alo

ne

napro

x.

misopro

stol

ritonavir

inte

stin

al d

amag

e(m

m2 )

0

50

100

150

MPO

(mU

nits

/mg

of p

rote

in)

0

25

50

75

*

Hae

mat

ocrit

(%)

naive

napro

x.alo

ne

napro

x.

misopro

stol

ritonavir

naive

napro

x.alo

ne

napro

x.

misopro

stol

ritonavir

0.0

2.5

5.0

7.5

CO

X-2

mR

NA

leve

ls

0

1

2

3

4

CO

X-1

mR

NA

leve

ls

01234567

TNF

mR

NA

leve

ls

0

10

20

MC

P-1

mR

NA

leve

ls

naive

napro

x.alo

ne

napro

x.

misopro

stol

ritonavir

naive

napro

x.alo

ne

napro

x.

misopro

stol

ritonavir

naive

napro

x.alo

ne

napro

x.

misopro

stol

ritonavir

naive

napro

x.alo

ne

napro

x.

misopro

stol

ritonavir

Fig. 4. A PGE1 analogue misoprostol prevents ritonavir induced intestinal damage. (A) Intestinal mucosal damage. (B) Haematocrit. (C) Mucosal myeloperoxidase activity(MPO). ((D)–(G)) Relative mRNA expression of COX-1, COX-2, MCP-1, and TNFα was assayed by quantitative Real-Time PCR. Values are normalized relatively to GAPDH mRNA.Data are expressed as means7S.E.M. of 6 mice/group. nPo0.05 vs not treated mice. #Po0.05 vs ritonavir administered mice.

B. Renga et al. / European Journal of Pharmacology 723 (2014) 194–201 199

induced by TNFα (Pati et al., 2002). Thus, it is possible that inhibitionof NF-kB activation by ritonavir is an additional pathway which couldexplain the increase in apoptosis induced by TNFα in HT29 cells.

Previous studies have shown that PI-induced barrier disruptionin intestinal epithelial cells is not due to tight junction alterations,but to induction of poly(ADP-ribose) polymerase (PARP)-depen-dent apoptosis (Zhang et al., 2013; Bode et al., 2005). Further on,exposure of intestinal epithelial cells to lopinavir and ritonavir, butnot amprenavir, is known to cause endoplasmic reticulum (ER)stress and increase in the unfolded protein response (Wu et al.,2010). This activated ER stress partially impairs the epithelialbarrier integrity by promoting intestinal epithelial cell apoptosis(Wu et al., 2010). Activation of the of the transcription factorCCAAT/enhancer binding protein homologous protein (CHOP) isrequired for ER damaged induced by HIV-PI, and CHOP silencingby specific small hairpin RNA prevents lopinavir- and ritonavir-induced barrier dysfunction in cultured intestinal epithelial cells,while mice deficient for CHOP exhibits decreased mucosal injuryafter exposure to lopinavir and ritonavir (Sung et al., 2012). It hasbeen shown that CHOP directly binds to death receptor 5 (DR5)promoter through ER-stress signaling and sensitizes cells to deathsignals activated by TNFα and TNF-related apoptosis-inducingligand (TRAIL) (Sung et al., 2012). Thus, the synergistic enhance-ment of apoptosis of HT29 cells by the combination of ritonavirwith naproxen and TNFα supports the notion that these factorsactivate intertwined pathways that converge on effector caspases,such as caspase 3 (Fiorucci, 2001).

We have then investigated whether intestinal injury caused bythe combination of HIV-PI with naproxen could be prevented byadministering mice with misoprostol. Data shown in Figs. 4 and 5

demonstrate that misoprostol effectively reduced the severity ofintestinal injury cauised by ritonavir and naproxen as measured byassessing the severity of injury score, MPO activity and hematocritlevels. Misoprostol administration was also effective in attenuatingbiomarkers of intestinal injury including COX-1 and COX-2, MCP-1and TNFα. Of relevance PGE1 and PGE2 effectively rescued HT-29cells from apoptosis induced by naproxen and ritonavir as well asby the combination of TNFα and ritonavir indicating that damagecaused by these two agents are PGE-dependent. Thus, thesefindings are consistent with human studies and current guidelinesthat support the use of prostaglandin derivative in the preventionof gastrointestinal injury caused by tNSAIDs (Lanza et al., 2009).

These observations might have a clinical readout. Indeed, combi-nation of COX inhibitors with ritonavir-boosted HAART might exposeHIV infected persons to the detrimental effects of the two agents inthe small intestine. Despite the present observation awaits confirma-tion from clinically relevant observations, our data suggest thatprostaglandin derivatives could be effective in protecting againstintestinal damage caused by the combination of these two agents.

Previous studies have shown that intestinal injury caused bytNASIDs could be prevented by shifting from tNSAIDs to a coxib(Goldstein et al., 2007; Laine et al., 2008) or novel NSAIDs thatassociate COX inhibition with the donation of hydrogen sulfide ornitric oxide (Fiorucci and Santucci, 2011; Wallace et al., 2010).Whether, these agents will spare the intestine in the present ofritonavir, however, is unknown.

Additionally, it has been shown that PPI, such as omeprazole,while protect the stomach against injury caused by ASA andtNSAIDs, exacerbates the intestinal injury caused by these agents(Wallace et al., 2011). In a recent study we have shown thatomeprazole exacerbates the intestinal injury caused by naproxen(Cipriani et al., 2013b) and that severity of intestinal injury causedby the combination these agents parallels that of ritonavir andnaproxen, making unlikely that PPI might be effective in reducingintestinal damage caused by ritonavir.

In conclusion, this animal study shed light on a novel source ofintestinal injury caused by the combination of HIV-PI withtNSAIDs and ASA. In animal models this HIV-PI/NSAIDs entero-pathy is efficiently prevented by the use of the prostaglandinderivatives. Further studies are needed to confirm the clinicalrelevance of this novel source of intestinal injury.

Appendix A. Supplementary material

Supplementary data associated with this article can be found inthe online version at http://dx.doi.org/10.1016/j.ejphar.2013.11.035.

References

Abraham, N.S., Hlatky, M.A., Antman, E.M., Bhatt, D.L., Bjorkman, D.J., Clark, C.B.,Furberg, C.D., Johnson, D.A., Kahi, C.J., Laine, L., Mahaffey, K.W., Quigley, E.M.,Scheiman, J., Sperling, L.S., Tomaselli, G.F., 2010. ACCF/ACG/AHA 2010 expertconsensus document on the concomitant use of proton pump inhibitors andthienopyridines: a focused update of the ACCF/ACG/AHA 2008 expert con-sensus document on reducing the gastrointestinal risks of antiplatelet therapyand NSAID use. Am. J. Gastroenterol. 105, 2533–2549.

Bode, H., Lenzner, L., Kraemer, O.H., Kroesen, A.J., Bendfeldt, K., Schulzke, J.D.,Fromm, M., Stoltenburg-Didinger, G., Zeitz, M., Ullrich, R., 2005. The HIVprotease inhibitors saquinavir, ritonavir, and nelfinavir induce apoptosis anddecrease barrier function in human intestinal epithelial cells. Antiviral Ther. 10,645–655.

Catella-Lawson, F., Reilly, M.P., Kapoor, S.C., Cucchiara, A.J., DeMarco, S., Tournier, B.,Vyas, S.N., FitzGerald, G.A., 2001. Cyclooxygenase inhibitors and the antiplateleteffects of aspirin. N. Engl. J. Med. 345, 1809–1817.

Cipriani, S., Francisci, D., Mencarelli, A., Renga, B., Schiaroli, E., D’Amore, C., Baldelli, F.,Fiorucci, S., 2013a. Efficacy of the CCR5 antagonist maraviroc in reducing early,ritonavir-induced atherogenesis and advanced plaque progression in mice.Circulation 127, 2114–2124.

0

10

20

30

40

50

ritonavir

% o

f cel

ls (s

ub-G

1pha

se)

0

25

50

75

TNF

ritonavirTNF

% o

f cel

ls(s

ub-G

1pha

se)

naive alone alone PGE-2 PGE-1

naive naprox. alone naprox. PGE-2 PGE-1

Fig. 5. Prostaglandins PGE1 and PGE2 prevent apoptosis mediated by ritonavir.(A) HT29 cells were left untreated (ctrl) or incubated 36 h with ritonavir, naproxen,ritonavir plus naproxen, ritonavir plus PGE1 and ritonavir plus PGE2 as described inmaterials and methods. (B) HT29 cells were left untreated (ctrl) or incubated 36 hwith ritonavir, TNFα, ritonavir plus TNFα, ritonavir plus TNFα plus PGE1 andritonavir plus TNFα plus PGE2 as described in materials and methods. Apoptosiswas determined by staining the cells with propidium iodide and assaying thenumber of apoptotic/necrotic cells at flow cytometry. Data are means7S.E.M. of4 experiments. nPo0.01 vs control cells. #Po0.05 vs cells stimulated withritonavir. nPo0.05 vs cells stimulated with ritonavir plus TNFα.

B. Renga et al. / European Journal of Pharmacology 723 (2014) 194–201200

Cipriani, S., Renga, B, D’Amore, C., Distrutti, E., Santucci, L., Fiorucci, S., 2013b.Instruction of intestinal microbiota by vsl#3 protects against nsaids-ppienteropathy. Abstr. Gastroenterol. 144 (5), S-345. (Suppl. 1).

Clevenbergh, P., Garraffo, R., Dellamonica, P., 2003. Impact of various antiretroviraldrugs and their plasma concentrations on plasma lipids in heavily pretreatedHIV-infected patients. HIV Clin. Trials 4, 330–336.

Cohen, C.J., 2005. Ritonavir-boosted protease inhibitors, Part 2: Cardiac implicationsof lipid alterations. AIDS Read. 15, 528–532.

Dubé, M.P., 2000. Disorders of glucose metabolism in patients infected with humanimmunodeficiency virus. Clin. Infect. Dis. 31, 1467–1475.

Fiorucci, S., Antonelli, E., Santucci, L., Morelli, O., Miglietti, M., Federici, B., Mannucci, R.,Del Soldato, P., Morelli, A., 1999. Gastrointestinal safety of nitric oxide-derivedaspirin is related to inhibition of ICE-like cysteine proteases in rats. Gastroenter-ology 116, 1089–1106.

Fiorucci, S., 2001. NO-releasing NSAIDs are caspase inhibitors. Trends Immunol. 22,232–235.

Fiorucci, S., Distrutti, E., de Lima, O.M., Romano, M., Mencarelli, A., Barbanti, M.,Palazzini, E., Morelli, A., Wallace, J.L., 2003. Relative contribution of acetylatedcyclo-oxygenase (COX)-2 and 5-lipooxygenase (LOX) in regulating gastricmucosal integrity and adaptation to aspirin. FASEB J 17, 1171–1173.

Fiorucci, S., Antonelli, E., Distrutti, E., Rizzo, G., Mencarelli, A., Orlandi, S., Zanardo, R.,Renga, B., Di Sante, M., Morelli, A., Cirino, G., Wallace, J.L., 2005. Inhibition ofhydrogen sulfide generation contributes to gastric injury caused by anti-inflammatory nonsteroidal drugs. Gastroenterology 129, 1210–1224.

Fiorucci, S., Santucci, L., Distrutti, E., 2007. NSAIDs, coxibs, CINOD and H2S-releasingNSAIDs: what lies beyond the horizon. Dig. Liver Dis. 39, 1043–1051.

Fiorucci, S., 2009. Prevention of nonsteroidal anti-inflammatory drug-inducedulcer: looking to the future. Gastroenterol. Clin. North Am. 38, 315–332.

Fiorucci, S., Santucci, L., 2011. Hydrogen sulfide-based therapies: focus on H2S-releasing NSAIDs. Inflamm. Allergy Drug Targets 10, 133–140.

Goldstein, JL, Eisen, GM, Lewis, B, et al., 2007. Small bowel mucosal injury isreduced in healthy subjects treated with celecoxib compared with ibuprofenplus omeprazole, as assessed by video capsule endoscopy. Aliment. Pharmacol.Ther 25, 1211–1222.

Laine, L., Curtis, S.P., Langman, M., Jensen, D.M., Cryer, B., Kaur, A., Cannon, C.P.,2008. Lower gastrointestinal events in a double-blind trial of the cyclo-oxygenase-2 selective inhibitor etoricoxib and the traditional nonsteroidalanti-inflammatory drug diclofenac. Gastroenterology 135, 1517–1525.

Langenbach, R., Morham, S.G., Tiano, H.F., Loftin, C.D., Ghanayem, B.I., Chulada, P.C.,Mahler, J.F., Lee, C.A., Goulding, E.H., Kluckman, K.D., Kim, H.S., Smithies, O.,1995. Prostaglandin synthase 1 gene disruption in mice reduces arachidonicacid-induced inflammation and indomethacin-induced gastric ulceration. Cell83, 483–492.

Lanza, F.L., Chan, F.K., Quigley, E.M., 2009. Practice parameters committee of theAmerican College of Gastroenterology. Guidelines for prevention of NSAID-related ulcer complications. Am. J. Gastroenterol 104, 728–738.

Lipshultz, S.E., Mas, C.M., Henkel, J.M., Franco, V.I., Fisher, S.D., Miller, T.L., 2012.HAART to heart: highly active antiretroviral therapy and the risk of cardiovas-cular disease in HIV-infected or exposed children and adults. Expert Rev. AntiInfect. Ther. 10, 661–674.

Mencarelli, A., Cipriani, S., Renga, B., Francisci, D., Palladino, G., Distrutti, E., Baldelli, F.,Fiorucci, S., 2010. The bile acid sensor FXR protects against dyslipidemia and aorticplaques development induced by the HIV protease inhibitor ritonavir in mice. PLoSOne 5, e13238.

Mencarelli, A., Francisci, D., Renga, B., D’Amore, C., Cipriani, S., Basile, F., Schiaroli, E.,Baldelli, F., Fiorucci, S., 2012. Ritonavir-induced lipoatrophy and dyslipidaemiais reversed by the anti-inflammatory drug leflunomide in a PPAR-γ-dependentmanner. Antiviral Ther. 17, 669–678.

Moyle, G., Carr, A., 2002. HIV-associated lipodystrophy, metabolic complications,and antiretroviral toxicities. HIV Clin. Trials 3, 89–98.

Mugavero, M.J., Napravnik, S., Cole, S.R., Eron, J.J., Lau, B., Crane, H.M., Kitahata, M.M.,Willig, J.H., Moore, R.D., Deeks, S.G., Saag, M.S., 2011. Centers for AIDS researchnetwork of integrated clinical systems (CNICS) cohort study. Viremia copy-yearspredicts mortality among treatment-naive HIV-infected patients initiating anti-retroviral therapy. Clin. Infect. Dis. 53, 927–935.

Pati, S., Pelser, C.B., Dufraine, J., Bryant, J.L., Reitz, J.M.S., Weichold, F.F., 2002.Antitumorigenic effects of HIV protease inhibitor ritonavir: inhibition of Kaposisarcoma. Blood 99, 3771–3779.

Riddle, T.M., Kuhel, D.G., Woollett, L.A., Fichtenbaum, C.J., Hui, D.Y., 2001. HIVprotease inhibitor induces fatty acid and sterol biosynthesis in liver and adiposetissues due to the accumulation of activated sterol regulatory element-bindingproteins in the nucleus. J. Biol. Chem 276, 37514–37519.

Santucci, L., Fiorucci, S., Giansanti, M., Brunori, P.M., Di Matteo, F.M., Morelli, A.,1994. Pentoxifylline prevents indomethacin induced acute gastric mucosaldamage in rats: role of tumour necrosis factor alpha. Gut 35, 909–915.

Santucci, L., Fiorucci, S., Patoia, L., Di Matteo, F.M., Brunori, P.M., Morelli, A., 1995.Severe gastric mucosal damage induced by NSAIDs in healthy subjects isassociated with Helicobacter pylori infection and high levels of serum pepsino-gens. Dig. Dis. Sci 40, 2074–2080.

Santucci, L., Fiorucci, S., Chiorean, M., Brunori, P.M., Di Matteo, F.M., Sidoni, A.,Migliorati, G., Morelli, A., 1996. Interleukin 10 reduces lethality and hepaticinjury induced by lipopolysaccharide in galactosamine-sensitized mice. Gastro-enterology 111, 736–744.

Spector, A.A., 2003. HIV protease inhibitors and hyperlipidemia: a fatty acidconnection. Arterioscler. Thromb. Vasc. Biol. 26, 7–9.

Sung, B., Prasad, S., Ravindran, J., Yadav, V.R., Aggarwal, B.B., 2012. Capsazepine,a TRPV1 antagonist, sensitizes colorectal cancer cells to apoptosis by TRAILthrough ROS-JNK-CHOP-mediated upregulation of death receptors. Free RadicalBiol. Med. 53, 1977–1987.

Wallace, M.R., Brann, O.S., 2000. Gastrointestinal manifestations of HIV infection.Curr. Gastroenterol. Rep 2, 283–293.

Wallace, J.L., 2008. Prostaglandins, NSAIDs, and gastric mucosal protection: whydoesn't the stomach digest itself? Physiol. Rev 88, 1547–1565.

Wallace, J.L., Caliendo, G., Santagada, V., Cirino, G., 2010. Markedly reduced toxicityof a hydrogen sulphide-releasing derivative of naproxen (ATB-346). Br. J.Pharmacol. 159, 1236–1246.

Wallace, J.L., Syer, S., Denou, E., de Palma, G., Vong, L., McKnight, W., Jury, J., Bolla, M.,Bercik, P., Collins, S.M., Verdu, E., Ongini, E., 2011. Proton pump inhibitorsexacerbate NSAID-induced small intestinal injury by inducing dysbiosis. Gastro-enterology 141, 1314–1322. (1322.e1-5).

Wolfe, M.M., Lichtenstein, D.R., Singh, G., 1999. Gastrointestinal toxicity of non-steroidal antiinflammatory drugs. N. Engl. J. Med. 340, 1888–1899.

Wu, X., Sun, L., Zha, W., Gurley, E., Chen, L., Wang, X., Hylemon, P.B., Pandak Jr., W.M.,Sanyal, A.J., Zhang, L., Wang, G., Chen, J., Wang, J.Y., Zhou, H., 2010. HIV proteaseinhibitors induce endoplasmic reticulum stress and disrupt barrier integrity inintestinal epithelial cells. Gastroenterology 138, 197–209.

Yiannoutsos, C.T., Johnson, L.F., Boulle, A., Musick, B.S., Gsponer, T., Balestre, E., Law, M.,Shepherd, B.E., Egger, M., 2012. International epidemiologic databases to evaluateAIDS (IeDEA) collaboration. Estimated mortality of adult HIV-infected patientsstarting treatment with combination antiretroviral therapy. Sex Transm. Infect. 88(Suppl. 2), i33–i43.

Zhang, S., Chao, G.Q., Lu, B., 2013. Proton pump inhibitors are not the key fortherapying non-steroidal anti-inflammatory drugs-induced small intestinalinjury. Rheumatol. Int 33, 2513–2521.

B. Renga et al. / European Journal of Pharmacology 723 (2014) 194–201 201