Laboratory Investigations

Novel phenanthridinone inhibitors of poly(adenosine 5'-diphosphate-ribose) synthetase: Potent cytoprotective and antishock agents*Prakash Jagtap, PhD; Francisco Garcia Soriano, MD, PhD; László Virág, MD, PhD; Lucas Liaudet, MD;Jon Mabley, PhD; Éva Szabó, MD; György Haskó, MD, PhD; Anita Marton, MD; Clara Batista Lorigados, MD;Ferenc Gallyas Jr, PhD; Balázs Sümegi, MD, PhD; Dale G. Hoyt, PhD; Erkan Baloglu, PhD;John VanDuzer, PhD; Andrew L. Salzman, MD; Garry J. Southan, PhD; Csaba Szabó, MD, PhD

Poly(adenosine 5'-diphosphate[ADP]-ribose) synthetase (PARS),also known as poly(ADP-ribose)polymerase (PARP), is an abun-

dant nuclear enzyme of eukaryotic cells,which has been implicated in the cellularresponse to DNA injury (1, 2). PARS, when

activated in response to free radical andoxidant-induced DNA single-strand breaks,initiates an energy-consuming cycle bytransferring ADP ribose units to nuclearproteins. The result of this process is arapid depletion of the intracellular nicotin-amide adenine dinucleotide (oxidized)

(NAD�) and adenosine 5'-triphosphatepools, which slows the rate of glycolysis andmitochondrial respiration, culminating incell necrosis (3, 4).

Activation of PARS recently has beenshown to underlie certain shock andreperfusion states. Our group recently

*See also p. 1163.From Inotek Corporation (PJ, FGS, LV, LL, JM, ES,

GH, AM, EB, JV, ALS, GJS, CS), Beverly, MA; theDepartment of Surgery (FGS, LL, CS), New JerseyMedical School, UMDNJ, Newark, NJ; Hospital dasClinicas FMUSP (FGS, CBL), Sao Paulo, Brazil; theDepartment of Biochemistry (FG, BS), Pécs University

Medical School, Pécs, Hungary; and the College ofPharmacy (DGH), Ohio State University, Columbus, OH.

Supported, in part, by grants R44NS37635,R01GM57407, and R44GM58986 from the Nation-al Institutes of Health, Bethesda, MD; by a grant from theHungarian Ministry of Health; by a grant from the AD-UMED foundation (LL), Switzerland; by a fellowship from

FAPESP (Brazil) (FGS), and by grant ETT 32/2000 from theMinistry of Health (Hungary) (FG, BS).

Address requests for reprints to: Csaba Szabó, MD,PhD, Inotek Corporation, Suite 419E, 100 CummingsCenter, Beverly, MA 01915. E-mail: szabocsaba@aol.com

Copyright © 2002 by Lippincott Williams & Wilkins

Objective: To synthesize novel inhibitors of the nuclear enzymepoly(adenosine 5'-diphosphate [ADP]-ribose) synthetase (PARS),also known as poly(ADP-ribose) polymerase (PARP), and to testthem in in vitro models of oxidant-induced cytotoxicity and inendotoxin and splanchnic occlusion-reperfusion-induced shock.

Design: Randomized, prospective laboratory study.Setting: Research laboratory.Subjects: Murine macrophages, thymocytes, and endothelial

cells; Balb/c mice and Wistar rats.Interventions: Macrophages and endothelial cells were treated

with peroxynitrite and bleomycin to induce PARS activation, andthymocytes were treated with peroxynitrite to induce cell necrosis.Novel PARS inhibitors were synthesized and used to reduce PARSactivation and to reverse cytotoxicity. Balb/c mice were subjected tosplanchnic occlusion and reperfusion and were pretreated with var-ious doses (1–10 mg/kg intraperitoneally) of PJ34, a selected, potent,water-soluble PARS inhibitor. The passage of fluorescein isothiocya-nate-conjugated dextran (4 kDa) was analyzed in everted gut ilealsacs incubated ex vivo as an index of gut permeability. Wistar ratswere subjected to Escherichia coli bacterial lipopolysaccharide (40mg/kg intraperitoneally). PJ34 was also used at 10 mg/kg intraperi-toneally, 1 hr before lipopolysaccharide or at 25 mg/kg intraperito-neally 1 hr after lipopolysaccharide treatment. Serum concentrationsof indicators or multiple organ injury, concentrations of variousproinflammatory mediators, and tissue concentrations of myeloper-oxidase and malondialdehyde were measured. In addition, survivalrates and vascular contractile and relaxant responses were recorded.

Measurements and Main Results: Appropriate modifications ofthe phenanthridinone core structure yielded significant increases in

the potency of the compounds, both as PARS inhibitors and ascytoprotective agents. The compound N-(6-oxo-5,6-dihydro-phenan-thridin-2-yl) -N,N-dimethylacetamide (designated as PJ34) was oneof the potent PARS inhibitors of the series, and it dose-dependentlyprotected against thymocyte necrosis, with a half-maximal restora-tion of cell viability of 35 nM and complete protection at 200 nM.PARS activation also was visualized by immunohistochemistry andwas dose-dependently suppressed by PJ34. The effect of PJ34 wasdose-dependently reversed by excess nicotinamide adenine dinucle-otide (oxidized). The PARS inhibitors dose-dependently suppressedproinflammatory cytokine and chemokine production and restoredviability in immunostimulated macrophages. PJ34 was selected forthe subsequent in vivo studies. PJ34 significantly protected againstsplanchnic reperfusion-induced intestinal hyperpermeability in themouse. PJ34 reduced peak plasma concentrations of tumor necrosisfactor-�, interleukin-1�, and nitrite/nitrate in the plasma of lipopo-lysaccharide-treated rats. PJ34 ameliorated the lipopolysaccharide-induced increases in indexes of liver and kidney failure and concen-trations of myeloperoxidase and malondialdehyde in the lung andgut. Lipopolysaccharide elicited vascular dysfunction, which wasnormalized by PJ34. Lipopolysaccharide-induced mortality was re-duced by PJ34 (both pre- and posttreatment).

Conclusions: The novel series of phenanthridinone PARS in-hibitors have potent cytoprotective effects in vitro and significantprotective effects in shock and reperfusion injury in rodent mod-els in vivo. (Crit Care Med 2002; 30:1071–1082)

KEY WORDS: shock; endotoxin; gut; vascular; endothelial; lung;liver; myeloperoxidase; malondialdehyde

1071Crit Care Med 2002 Vol. 30, No. 5

demonstrated that pharmacologic inhi-bition, or the absence of PARS, preventsthe functional alterations associatedwith ischemia-reperfusion injury in theheart and gut (4 –7). In addition, wehave demonstrated that inhibition orinactivation of PARS provides potentprotective effects in various forms ofendotoxic and hemorrhagic shock (8 –10). Similar findings have been ob-tained in a number of other laborato-ries, by using various pharmacologicinhibitors of PARS or PARS-deficientanimals (11–15).

Most commercially available PARS in-hibitors (e.g., 3-aminobenzamide or nic-otinamide) have low potency and shortcellular residence time (16). Potent, wa-ter-soluble PARS inhibitors that can beused as tools to define the role of PARS invarious pathophysiological states arelacking. Phenanthridinone was describedas a fairly potent inhibitor of PARS (14).However, the compound is not water sol-uble, and it has not been tested in modelsof shock and reperfusion injury in vivo.In the present report, we describe the invitro characterization of a novel series ofpotent, water-soluble PARS inhibitors(designated as the PJ series of inhibitors),and we describe our rodent studies test-ing PJ34, a selected member of the PJseries in rodent models of shock andreperfusion injury.

MATERIALS AND METHODS

Reagents and Drugs. Lipopolysaccharide(LPS; from Escherichia coli, serotype 0127:B8) and all compounds not specified otherwisewere purchased from Sigma-Aldrich (St.Louis, MO). Peroxynitrite was a kind gift of Dr.H. Ischiropoulos (University of Pennsylvania,Philadelphia). 2-Substituted 6(5H)-phenan-thridinones were prepared according to previ-ously described methods and principles (17–21): 9-Fluorenone (1) was dissolved in sulfuricacid (98%) and treated with sodium azide at0°C over several hours. The resulting 6(5H)-phenanthridinone (PJ97A) was precipitated bypouring the mixture slowly on crushed ice. Asuspension of 2 in acetic acid was nitrated byusing nitric acid to give 2-nitro-6(5H)-phenanthridinone (PJ97B) as a major product.The nitro compound then was reduced by ironand NH4Cl and by using dimethylformamideas solvent to produce 2-amino-6(5H)-phenan-thridinone (PJ97AC). Chloroacylation of 4with chloroacetyl chloride in pyridine/DMFyielded compound PJ97D. Compound PJ97Dwas treated with the appropriate amine inMeOH or DMF to furnish the correspondingamino substituted compound, which was thentreated with HCl in ether in presence of meth-anol to give the PJ compounds as the water-soluble HCl salts. Product identity and puritywere assessed by using nuclear magnetic res-onance, mass spectroscopy, thin-layer chro-matography, and/or high-performance liquidchromatography. Compounds used in the cur-rent study were found �98% pure.

Testing of the Effects of the Compoundsfor PARS Activation and Cytotoxicity In Vitro.Testing the compounds for PARS inhibitorypotency and prevention of peroxynitrite-induced cytotoxicity was performed as previ-ously described (22, 23). Briefly, PARS activa-tion in cultured RAW murine macrophageswas measured as follows: RAW mouse macro-phages were cultured in Dulbecco’s modifiedeagle medium with high glucose and weresupplemented with 10% fetal bovine serum.Cells were used at 80% confluence in 12-well

Figure 1. Basic structure of the phenanthridi-none-based compounds used in the presentstudy. The various R groups are listed in Table 1,column 3.

Table 1. Inhibitory effect of various phenanthridinone derivative poly-(adenosine 5�-diphosphate-ribose) synthetase (PARS) inhibitor compounds onperoxynitrite-induced PARS activation in murine J774 macrophages and on peroxynitrite-induced cytotoxicity in freshly isolated murine thymocytes

CompoundDesignation Compound Chemical Name R Group

PARS Activationin Macrophages,

% inhibitionat 1 �M

EC50 forCytoprotectionin Thymocytes

Cytoprotectionin Thymocytes,% protection

at 200 nM

PJ97A 6(5H)-Phenanthridinone H 71 40 nM 73PJ97B 2-Nitro-6(5H)-phenanthridinone NO2 58 400 nM 45PJ97C 2-Amino-6(5H)-phenanthridinone NH2 70 36 nM 78PJ97D N-(6-oxo-5,6-dihydro-phenanthridin-2-yl)-

chloroacetamideNHCOCH2Cl 89 40 nM 91

PJ34 N-(6-oxo-5,6-dihydro-phenanthridin-2-yl)-N,N-dimethylacetamide

NHCOCH2NMe2 90 35 nM 94

PJ36 N-(6-oxo-5,6-dihydro-phenanthridin-2-yl)-1-piperidinylacetamide

NHCOCH2piperidine 86 30 nM 97

PJ38 N-(6-oxo-5,6-dihydro-phenanthridin-2-yl)-4-morpholinylacetamide

NHCOCH2-morpholine 83 35 nM 94

PJ44 N-(6-oxo-5,6-dihydro-phenanthridin-2-yl)-N,N-diethylacetamide

NHCOCH2NEt2 82 35 nM 94

PJ45 N-(6-oxo-5,6-dihydro-phenanthridin-2-yl)-N-methyl-N-benzylacetamide

NHCOCH2N(Me)CH2Ph 93 35 nM 78

PJ46 N-(6-oxo-5,6-dihydro-phenanthridin-2-yl)-1-(4-methyl-piperazinylacetamide)

NHCOCH2piperazine 48 170 nM 58

3-AB 3-aminobenzamide 0 (31% at 10 �M;63% at 100 �M)

15 �M 8

Values in the table are shown as EC50 (half maximal restoration of cell viability) or as percentage inhibition of PARS activation or percentagecytoprotection at a fixed inhibitor concentration. Data are shown as mean of triplicate determinations.

1072 Crit Care Med 2002 Vol. 30, No. 5

plates. Cells were pretreated with various con-centrations (1–100 �M) of the PARS inhibitorcompounds for 10 mins. Peroxynitrite, a pro-totypical oxidant, which induces DNA single-strand breakage and activates PARS (22), wasused to induce PARS activation in J774 mac-rophages (present section) and freshly isolatedthymocytes (see following section). Peroxyni-trite was diluted in phosphate-buffered saline(PBS; pH 11.0) and added to the cells in abolus of 50 �L. Cells were then incubated for20 mins. Decomposed peroxynitrite (incubat-ed for 30 mins at pH 7.0) served as controlsand failed to influence the variable studied.After 20 mins of incubation, cells were spun,medium was aspirated, and cells were resus-pended in 0.5 mL assay buffer (56 mM N-hydroxy ethyl piperazine-N-2-ethane sulfonicacid pH 7.5, 28 mM KCl, 28 mM NaCl, 2 mMMgCl2, 0.01% w/v digitonin, 0.125 �M NAD�,and 0.5 �Ci/mL 3H-NAD�). PARS activity wasthen measured (22). After incubation (10 minsat 37°C), 200 �L of ice-cold 50% w/v trichlo-roacetate was added and samples were incu-bated for 4 hrs at 4°C. Samples were then spun(10,000 � g, 10 mins) and pellets washedtwice with ice-cold 5% w/v TCA and solubi-lized overnight in 250 �L of 2% w/v sodiumdodecyl sulfate /0.1 N NaOH at 37°C. Contentsof the tubes were added to 6.5 mL of ScintiSafePlus scintillation liquid (Fisher Scientific, FairLawn, NJ), and radioactivity was determinedby using a liquid scintillation counter (Wallac,Gaithersburg, MD).

For the cytotoxicity studies in thymocytes,thymi from C57/BL6 mice were used. Animalsreceived food and water ad libitum, and light-ing was maintained on a 12-hr cycle. Thymifrom 3- to 4-wk-old male mice were removedaseptically and placed into ice-cold RPMI me-dia supplemented with 10% v/v fetal calf se-rum, 10 mM glutamine, 10 mM N-hydroxyethyl piperazine-N-2-ethane sulfonic acid, 100units/mL penicillin, and 100 �g/mL strepto-mycin. Single cell suspensions were preparedby sieving the organs through a stainless wiremesh. Cells isolated this way were routinely95% viable, as assessed by Trypan blue exclu-

sion assay. Thymocytes (106 cells in 0.5 mL ofmedium) were seeded in 24-well plates. Per-oxynitrite was diluted in PBS (pH 11.0) andadded to the cells in a bolus of 50 �L. Cellsthen were incubated for 4 hrs to measure cellnecrosis by the propidium iodide stainingmethod. Decomposed peroxynitrite (incubatedfor 30 mins at pH 7.0) served as control andfailed to influence the variable studied. Thy-mocytes were stained with 5 �g/mL pro-pidium iodide for 15 mins at 37°C, washedonce with PBS, and cell necrosis analyzed witha FacsCalibur flow cytometer (Beckton-Dickinson, San Jose, CA), as described (24).

PARS inhibition has been known to sup-press proinflammatory mediator production(25–27). We have tested a number of selectedPARS inhibitors for suppression of proinflam-matory mediator production in cultured J774macrophages. Macrophages were incubatedwith LPS (10 �g/mL). Microphage inflamma-tory protein (MIP) production in the macro-phages was measured at 3 hrs, and tumornecrosis factor (TNF)-� and nitrite/nitrateconcentrations and cell viability in the mac-rophages were measured at 24 hrs. Cell viabil-ity was measured by using the MTT method aspreviously described (22).

One of the most potent, water-solublePARS inhibitor compounds (PJ34) was se-lected for additional, detailed in vitro and invivo studies. First, the inhibitory potency ofthe compound on purified PARS enzyme wasdetermined and its potency was comparedwith that of 3-aminobenzamide, a prototypicalbenchmark PARS inhibitor on the purifiedPARS enzyme by using a commercially avail-able PARS inhibition assay kit (Trevigen,Gaithersburg, MD). The assay was carried outin 96-well enzyme-linked immunosorbent as-say plates following manufacturer’s instruc-tions. Briefly, wells were coated with 1 mg/mLhistone (50 �L/well) at 4°C overnight. Platesthen were washed four times with PBS andthen blocked by adding 50 �L of Strep-Diluent(supplied with the kit). After incubation (1 hr,room temperature) plates were washed fourtimes with PBS. Appropriate dilutions of PARS

inhibitors were combined with 2� PARS cock-tail (1.95 mM NAD�, 50 �M biotinylatedNAD� in 50 mM TRIS pH 8.0, 25 mM MgCl2)and high specific activity PARS enzyme (bothsupplied with the kit) in a volume of 50 �L.Reaction was allowed to proceed for 30 mins atroom temperature. After four washes in PBS,incorporated biotin was detected by peroxi-dase-conjugated streptavidine (1:500 dilution)and TACS Sapphire substrate.

Testing of PJ34 on poly(ADP-ribose) syn-thesis in situ (PARSIS) in murine lung micro-vascular endothelial cells, isolated as previ-ously described (22, 28), was also performed(29). Bleomycin, a prototypical cytotoxic agentthat induces DNA single-strand breakage (30),was used to induce PARS activation in theendothelial cells. Cells were treated for 45mins with 0.5 mg/mL bleomycin in the pres-ence of various concentrations of PJ34 (0–10�M), followed by incubation for 0.5–2 hrs inbleomycin-free medium, rinsed with PBS andfixed in acetone at 4°C. After washing oncewith PARSIS buffer (100 mM Tris, pH 8, 10mM MgCl2, 1 mM dithiothreitol), they wereincubated for 0 or 10 mins in PARSIS buffercontaining various concentrations of NAD�

(0–200 �M) in the presence of various con-centrations of PJ34 (0–10 �M) and then werewashed with PBS. Poly(ADP-ribose) was mea-sured by fluorescence immunostaining (31).Cells were incubated with 10% goat serum/PBS for 30 mins and then incubated withrabbit anti-poly(ADP-ribose) at 4° for 18 hrsand then washed. Conjugated goat anti-rabbitimmunoglobin G (Fab2') in 10% goat serum/PBS was then applied for 1 hr at 37°. Thepolymerization of poly(ADP-ribose) detectedin this manner is linear up to 10 mins and 200�M NAD�.

Animals. Male Wistar rats (280 –320 g)were obtained from Charles River Laboratories(Wilmington, MA). Male BALB/c mice (8 wks)were purchased from Taconic Laboratories(Georgetown, NY). In vivo studies were per-formed in accordance with National Institutes

Figure 2. Dose-dependent cytoprotection with the phenanthridinone-based poly(adenosine 5'-diphosphate-ribose) synthetase (PARS) inhibitor compounds and a reference PARS inhibitor com-pound (3-aminobenzamide, 3AB) in thymocytes exposed to peroxynitrite. Results (shown as means oftriplicate determinations) represent n � 3 independent determinations.

Figure 3. Dose-dependent inhibition of isolatedphenanthridinone-based poly(adenosine 5'-diphosphate-ribose) synthetase (PARS) enzymeactivity with the phenanthridinone-based PARSinhibitor compound PJ34 and a reference PARSinhibitor compound (3-aminobenzamide, 3AB).Results are shown as mean � SEM of triplicatedeterminations.

1073Crit Care Med 2002 Vol. 30, No. 5

of Health Guidelines and with the approval ofthe local Institutional Animal Care and Usecommittee.

Animal Preparation. For the splanchnicocclusion shock studies in mice, anesthesiawas induced with 2% isoflurane and main-tained with a mixture of 1.5% isoflurane and50% oxygen by using the anesthesia port of aVentimeter ventilator (Air Shields, Hatboro,PA). Animals were maintained at 37°C on aheating pad. After a median laparotomy, theintestine was exteriorized on a plastic sheetand a mechanical vascular occlusion was pro-duced by placing an aneurysm miniclip on themesenteric artery near its aortic origin. Forty-five minutes after the initiation of ischemia,the peritoneal cavity was reopened and thebowel examined for manifestations of isch-emia. Then, the miniclip was removed, 2 mLof warmed saline was poured into the abdom-inal cavity, and the laparotomy was closed bysuture. After 180 mins of reperfusion, intesti-nal samples were removed for analysis ofchanges in intestinal permeability (7). ForPJ34 treatment, mice were injected intraperi-toneally with saline or various doses of PJ34(1–10 mg/kg intraperitoneally) in a volume of0.1 mL/10 g body weight 60 mins before oc-clusion.

For the endotoxic shock studies, rats wereinjected intraperitoneally with LPS or saline.Rats were killed and tissues obtained and har-vested after 8 hrs of 40 mg/kg LPS. For PJ34pretreatment, rats were injected intraperito-neally with saline or PJ34 (10 mg/kg intraperi-toneally) in a volume of 0.1 mL/100 g bodyweight 60 mins before LPS. In the survivalstudies, in an additional group of rats, PJ34was administered 1 hr after LPS, at a dose of25 mg/kg intraperitoneally (posttreatmentgroup). Animals were monitored for 48 hrs,after which time they were monitored daily for2 wks and then killed. No late deaths (betweenthe period of 48 hrs to 2 wks post-LPS) wereobserved in any of the experimental groups.

In an additional series of experiments,mice were pretreated with PJ34 (10 mg/kg, asdescribed previously), challenged with LPS(40 mg/kg, as described previously), and killedat 8 and 24 hrs. Plasma was taken and ana-lyzed for albumin and total protein concentra-tions as well as for indexes of multiple organfailure.

In a set of experiments specifically de-signed to detect PARS activation in vivo andthe inhibitory effect of PJ34 on this response,in separate subsets (n � 5 per group) animalswere pretreated with PJ34 and then treatedwith LPS (as described previously) and killedat 5 or 24 hrs. Lungs were taken and processedfor the immunohistochemical detection ofpoly(ADP-ribose) polymers, as previously de-scribed (32, 33). Briefly, paraffin sections (3�m) were deparaffinized in xylene and rehy-drated in decreasing concentrations (100%,95%, and 70%) of ethanol followed by a 10-min incubation in PBS (pH7.4). Sections weretreated with 0.3% hydrogen peroxide for 15

Figure 4. a, dose-dependent inhibition of phenanthridinone-based poly(adenosine 5'-diphosphate-ribose) synthetase (PARS) activation with the phenanthridinone-based PARS inhibitor compoundPJ34; partial reversal of the effect of PJ34 with excess nicotinamide adenine dinucleotide (NAD). PARSactivation was detected by quantifying poly(adenosine 5'-diphosphate-ribose) (PAR)-positive nuclei.The number of PAR-positive nuclei per tissue culture well is presented. Results are shown as mean �SEM of triplicate determinations. b, representative PAR staining in endothelial cells without PARSinhibition (left panel) and in the presence of 1 �M PJ34 (right panel).

Table 2. Inhibitory effect of selected phenanthridinone derivative poly-(adenosine 5�-diphospate-ribose) synthetase (PARS) inhibitor compounds on lipopolysaccharide (LPS)-induced tumor necrosisfactor (TNF)-�, microphage inflammatory protein (MIP)-1�, MIP-2, and nitrite (the breakdownproduct of nitric oxide [NO]) production in immunostimulated J774 macrophages

Inhibition ofTNF-�,

EC50, �M

Inhibition ofMIP-1�,

EC50, �M

Inhibition ofMIP-2, EC50,

�M

Inhibition ofNO EC50,

�M

Restorationof cell

viability,EC50, �M

PJ 34 5.4 � 0.6 10.2 � 1.4 7.2 � 0.88 15.2 � 2.8 12.5 � 1.4PJ 36 11.2 � 1.1 15.2 � 2.1 21.2 � 2.7 25.4 � 2.9 20.2 � 2.3PJ 38 6.2 � 0.42 11.3 � 1.4 14.9 � 1.3 20.2 � 1.5 22.1 � 2.7PJ 44 11.1 � 0.67 10.2 � 0.95 22.4 � 1.2 32.8 � 2.1 16.2 � 1.9PJ 46 9.2 � 0.88 37.1 � 2.1 39.5 � 1.6 42.3 � 0.2 39.1 � 1.0

Macrophages were treated with the various compounds for 30 mins at different concentrations(0.1–40 �M) for 30 mins, followed by exposure of the cells to immunostimulation (with LPS, 10�g/mL). MIP-1� and MIP-2 production in the macrophages was measured at 3 hrs., TNF-� and NOproduction and cell viability in the macrophages were measured at 24 hrs. Values in the table areshown as EC50 (half-maximal inhibition of mediator production or half-maximal restoration of cellviability; in �M). Mean � SEM or n � 4–6 determinations.

1074 Crit Care Med 2002 Vol. 30, No. 5

mins to block endogenous peroxidase activityand then were rinsed briefly in 10 mM PBS.Nonspecific binding was blocked by incubat-ing the slides for 1 hr in PBS containing 2%horse serum. Mouse monoclonal anti-poly-(ADP-ribose) antibody (Alexis, San Diego, CA)and isotype-matched control antibody wereapplied in a dilution of 1:100 for 2 hrs at roomtemperature. After extensive washing (5 � 5mins) with PBS, immunoreactivity was de-tected with a biotinylated horse antimousesecondary antibody and the avidin-biotin-peroxidase complex, both supplied in the Vec-tor Elite kit (Vector Laboratories, Burlingame,CA). Color was developed by using Ni-DABsubstrate (95 mg of diaminobenzidine, 1.6 gNaCl, 2 g of nickel sulfate in 200 mL of 0.1 Macetate buffer). Sections then were counter-stained with nuclear fast red, dehydrated, andmounted in Permount. Photomicrographswere taken with a Zeiss Axiolab microscope(Carl Zeiss North America, New York, NY)equipped with a Fuji HC-300C digital camera.Staining intensity was quantified and repre-sentative sections were selected by a personblinded to the treatment protocol.

Gut Permeability Measurements. In themurine splanchnic occlusion-reperfusionstudies, intestinal mucosal barrier functionwas assessed by the mucosal-to-serosa clear-ance of fluorescein isothiocyanate-conjugateddextran (4 kDa, FD4) in everted gut ileal sacsincubated ex vivo (7). The everted gut sacswere prepared in ice-cold modified Krebs-Henseleit bicarbonate buffer (KHBB containedin mM: N-hydroxy ethyl piperazine-N-2-ethane sulfonic acid 10, NaCl 137, KCl 5.5,NaHCO3 4.2, Na2HPO4 0.3, KHPO4 0.4, MgSO4

0.4, MgCl2 0.5, CaCl2 1.3, glucose 19.5). Theileal segment was first lavaged with 3 mL ofisotonic saline to remove fecal material andthen was closed at one end with a 4–0 silkligature. The gut sac was everted onto a thinplastic rod and then connected to a 3-mLsyringe containing 0.4 mL of the KHBB solu-tion, via a male luer fitting for 1/16 inches(approximately 1.5 mm) internal diametertubing (World Precision Instrument, Sarasota,FL) and secured with a 4–0 silk ligature. Theeverted gut sac was then gently distended with0.4 mL of KHBB and suspended in a 50-mLbeaker containing FD4 (20 �g/mL) in KHBB,continuously gassed with 95% oxygen and 5%CO2, and maintained at 37°C in a water bath.

At the beginning of the incubation, a 1-mLsample was withdrawn from the beaker to deter-mine the initial external (i.e., mucosal) FD4 con-centration ([FD4muc]). After a 30-min incuba-tion period, the gut sac was removed from thebeaker, its diameter and length were measured,and the KHBB solution (0.4 mL) was withdrawnfrom within the sac to determine the internal(i.e., serosal) FD4 concentration ([FD4ser]). Aftercentrifugation (1000 � g, 10 mins), 300 �L ofthe clear supernatants were diluted with 3 mL ofPBS (10 mM, pH7.4) and fluorescence was mea-sured (ex � 492 nm, slit width � 1.5 nm; em� 515 nm, slit width � 10 nm) in a spectrofluo-

rimeter (RF 5301, Shimadzu, Columbia, MD).The mucosal-to-serosal clearance of FD4 wascalculated using the following equations:

Mucosal surface area (A) � πLD [1]

Mass of FD4 in the gut sac after 30 min

incubation (M) � �FD4ser� � 0.4 [2]

Mucosal-to-serosal permeation rate of

FD4 (PR, ng/min) � M/30 mins [3]

Mucosal-to-serosal clearance of FD4

(C, nL�min�1�cm�2) � PR/[FD4muc])/A [4]

Vascular Reactivity in Isolated AorticRings. In the rat studies of endotoxic shock,the thoracic aortae were harvested, clearedfrom periadvential fat, and mounted in organbaths filled with warmed (37°C) and gas equil-

ibrated (95% oxygen, 5% CO2) Krebs’ solution(mM): CaCl2 1.6, MgSO4 1.17, EDTA 0.026,NaCl 130, NaHCO3 14.9, KCl 4.7, KH2PO4

1.18, glucose 11. Isometric tension was mea-sured with isometric transducers (Kent Scien-tific Corporation, Litchfield, CT), digitized byusing a MacLab A/D converter and store anddisplayed on a Macintosh computer (8). Thepreload was 1 g. The rings were equilibratedfor 60 mins, and the solution was changedevery 15 mins. Dose-response curves to phen-ylephrine (10�10 to 3�10�5 M) and acetylcho-line (10�9 to 3�10�4 M) were obtained.

Myeloperoxidase (MPO) Assay. Samplesfrom gut, liver, and lung were harvested, snapfrozen in liquid nitrogen, and stored at �70°Cbefore determination of MPO activity as de-scribed (10). Tissues were homogenized (50mg/mL) in 0.5% hexadecyltrimethylammo-nium bromide in 10 mM 3-[N-Morpholino]

Figure 5. Gut mucosal permeability to isothiocyanate fluorescein dextran (FD4) in everted gut sacsincubated ex vivo. A segment of ileum was harvested after 45 mins of splanchnic artery occlusion(SAO) and 3 hrs of reperfusion in mice, followed by measurement of the mucosal-to-serosal passageof FD4. Mean � SEM. *p � .05 control vs. SAO: significant alteration in response to SAO, whencompared with vehicle-treated control animals. #p � .05 SAO vs. SAO�PJ34: significant improvementby PJ34 in the presence of SAO, vs. SAO alone. n � 4–6 determinations per group.

1075Crit Care Med 2002 Vol. 30, No. 5

propanesulfonic acid and centrifuged at15,000 � g for 40 mins. The suspension thenwas sonicated three times for 30 secs. An ali-quot of supernatant was mixed with a solution

of 1.6 mM tetra-methyl-benzidine and 1 mMhydrogen peroxide. Activity was measuredspectrophotometrically as the change in ab-sorbance at 650 nm at 37°C, by using a Spec-

tramax microplate reader (Molecular Devices,Sunnyvale, CA). Results are expressed as mil-liunits of MPO activity per milligram of pro-tein, which were determined with the Bio-Radassay (Bio-Rad, Hercules, CA).

Malondialdehyde (MDA) Assay. MDA for-mation was used to quantify the lipid peroxi-dation in tissues and measured as described(10). Tissues were homogenized (100 mg/mL)in 1.15% KCl buffer. Then, 200 �L of thehomogenates were added to a reaction mix-ture consisting of 1.5 mL of 0.8% thiobarbi-turic acid, 200 �L of 8.1% sodium dodecylsulfate, 1.5 mL of 20% acetic acid (pH 3.5),and 600 �L of distilled water. The mixturethen was heated at 90°C for 45 mins. Aftercooling to room temperature, the sampleswere cleared by centrifugation (10,000 � g for10 mins) and their absorbance was measuredat 532 nm, by using 1,1,3,3-tetramethoxypro-pane as an external standard. The concentra-tion of lipid peroxides was expressed as nmolMDA/mg of tissue.

Measurements of Proinflammatory Media-tor Production and Nitrite/Nitrate Concentra-tions. In tissue culture medium supernatantsand in plasma samples, murine TNF-�, inter-leukin (IL)-1�, MIP-1�, and MIP-2 concentra-tions were measured by specific enzyme-linked immunosorbent assay kits according tothe manufacturers’ instructions (25). Nitriteand nitrate was measured by using the modi-fied Griess reaction as previously described(25).

Measurements of Plasma Albumin and To-tal Protein Concentrations and Indexes ofLiver and Kidney Failure. In plasma samples,concentrations of albumin, alanine amino-transferase, amylase, blood urea nitrogen, andcreatinine were measured by using a veteri-nary blood analyzer system (VetScan systemand matching cartridges, Abaxis, Sunnydale,CA). Control experiments (plasma samplesspiked with 10 �M PJ34 in vitro) confirmedthat the PARS inhibitor compound did notinterfere with the assays used.

Statistical Evaluation. Values in the fig-ures, tables, and text are expressed as mean �SEM of n observations. Statistical analysis wasperformed by analysis of variance followed byTukey’s test. Differences in survival rate were

Figure 6. Vascular rings experiments. Thoracic aortae were harvested after 8 hrs of lipopolysaccharide(LPS) injection to measure their contraction to phenylephrine (a) or relaxation to acetylcholine (b).Mean � SEM. *p � .05 LPS vs. LPS�PJ34: significant improvement by PJ34 in the presence of LPS,vs. LPS alone. PJ34 alone did not affect contractions in relaxations in vehicle-treated control animals.n � 4–6 determinations per group.

Table 3. Effect of PJ34 on changes in plasma concentrations of albumin, alanine aminotransferase (ALT), amylase, blood urea nitrogen (BUN), creatinine,and total protein in lipopolysaccharide (LPS)-treated mice at 8, 16, and 24 hrs

Control PJ Control

8 Hrs 16 Hrs 24 Hrs

LPS LPS � PJ LPS LPS � PJ LPS LPS � PJ

Albumin, g/dL 5.4 � 0.15 5.03 � 0.17 4.12 � 0.05a 4.32 � 0.14 3.94 � 0.1a 4.3 � 0.1b 2.95 � 0.09a 4.45 � 0.28b

ALT, units/L 37.67 � 5.24 49 � 11.15 114.4 � 9.47a 60.8 � 5.44b 1432.4 � 320.3a 498.4 � 2.0b 1003.5 � 3.16a 133.75 � 9.73b

Amylase units/L 1066.7 � 61.5 1020.3 � 21.8 823.2 � 62.9 1278 � 544.3 954.8 � 85.7 1212.4 � 81.2 1577.5 � 142.4a 1154.7 � 124.8BUN, mg/dL 19 � 0.58 18 � 1.53 54.2 � 3.4a 46.8 � 3.27 81.8 � 2.74a 62.6 � 2.1b 123.5 � 1.58a 109.25 � 2.29b

Creatinine, mg/dL 0.23 � 0.03 0.23 � 0.03 0.32 � 0.04 0.34 � 0.04 0.72 � 0.06a 0.38 � 0.04b 0.75 � 0.03a 0.45 � 0.09Total protein, g/dL 5.73 � 0.09 5.43 � 0.09 4.42 � 0.08a 4.84 � 0.07 4.38 � 0.07a 4.76 � 0.08b 5.05 � 0.03a 5.15 � 0.21b

ap � .05 control vs. LPS: significant alteration in response to LPS, when compared with vehicle-treated control animals in the absence of LPS; bp � .05 LPSvs. PJ34 � LPS: significant improvement by PJ34 in the presence of LPS, vs. LPS alone, at a given time point. Mean � SEM or n � 5–6 determinations.

1076 Crit Care Med 2002 Vol. 30, No. 5

compared by the chi-square test. We consid-ered p � .05 to be statistically significant.

RESULTS

Testing of Novel Phenanthridinone-Based PARS Inhibitors In Vitro. Thestructures and the chemical names of thenovel inhibitors synthesized are shown inTable 1. All compounds exhibited a dose-dependent inhibitory effect on PARS ac-tivity in the macrophages stimulated withperoxynitrite. Many of the compounds ex-hibited potency that was higher than that

of the base phenanthridinone ring (Table1; Fig. 1). The PARS inhibitors induced adose-dependent protection against theperoxynitrite-induced cell necrosis in iso-lated thymocytes (Table 1, Fig. 2); therewas a good correlation between the po-tency of the compounds as PARS inhibi-tors and the potency of the compounds ascytoprotective agents (Table 1). BecausePJ34 emerged as one of the most potentderivatives, we tested PJ34 in the isolatedPARS enzyme in vitro and compared itspotency with that of the prototypical

PARS inhibitor compound 3-aminoben-zamide (Fig. 3). PJ34 dose-dependentlyinhibited PARS activity, with a half-maximal restoration of cell viability(EC50) of 20 nM. The prototypical PARSinhibitor 3-aminobenzamide was 10,000weaker than PJ34 and inhibited PARS ac-tivity with an EC50 of 200 �M. A compa-rable potency difference also was noted inthe thymocyte assay (Fig. 2): both PJ34and 3-aminobenzamide dose-dependentlyinhibited peroxynitrite-induced cell ne-crosis, with respective EC50 values of 20nM and 30 �M.

In addition to inhibiting oxidant-induced cell necrosis, PARS inhibitors orPARS-deficient cells previously have beenshown to produce significantly lowerconcentrations of various proinflamma-tory mediators (20 –22). In immuno-stimulated macrophages, a number of thePARS inhibitors were tested for produc-tion of proinflammatory cytokines (TNF-�), chemokines (MIP-1� and MIP-2), andfree radicals (nitric oxide, or NO, pro-duced by inducible NO synthase, as mea-sured by the detection of nitrite and ni-trate concentrations, stable breakdownproducts of NO) in the cell culture super-natants. The PARS inhibitors dose-dependently suppressed the production ofthese inflammatory mediators and alsodose-dependently restored the LPS-induced suppression of cell viability (Ta-ble 2). The potency differences seen inPARS inhibition also were reflected intheir potency as inhibitors of inflamma-tory mediator production. For example,PJ44 and especially PJ46, which had rel-atively lower potency as PARS inhibitorin the macrophages and as inhibitors ofthymocyte necrosis (Table 1), also tendedto be less potent on inflammatory medi-ator production (Table 2).

We next visualized the inhibition pro-vided by PJ34 by using the in situ PARSassay and determined the reversibility ofthe inhibitory effect of PJ34 by using thismethod in endothelial cells stimulatedwith bleomycin (Fig. 4). The potent PARSinhibitory effect of PJ34 was confirmed inthis assay. Significant PARS inhibitionwas observed with PJ34 concentrations aslow as 1 nM (Fig. 4a); EC50 values of thecompound amounted to 2 nM and 100nM at 20 �M and 200 �M NAD� concen-trations, respectively (Fig. 4b).

From the various series of in vitroassays, we concluded that PJ34 is a potentPARS inhibitor, and we have designedfollow-up in vivo studies to test its anti-shock efficacy in in vivo experiments.

Figure 7. Concentrations of malondialdehyde (MDA, left panels) and myeloperoxidase (MPO) activity(right panels) in the lung (top), gut (middle), and liver (bottom) obtained from vehicle-treated rats andrats at 8 hrs after lipopolysaccharide (LPS) injection, with and without PJ34 pretreatment. Mean �SEM. *p � .05 control vs. LPS: significant alteration in response to LPS when compared withvehicle-treated control animals in the absence of LPS. #p � .05 LPS vs. PJ34�LPS: significantimprovement by PJ34 in the presence of LPS, vs. LPS alone.

1077Crit Care Med 2002 Vol. 30, No. 5

Testing of PJ34 in the Splanchnic Oc-clusion-Reperfusion Model In Vivo.Splanchnic ischemia for 45 mins followedby 3 hrs of reperfusion markedly in-creased gut permeability. Previous stud-ies demonstrated that this effect is depen-dent on PARS activation in the intestinalepithelial cells: PARS-deficient mice werefound to be resistant against this type ofinjury (7). PJ34, when administered 10mins before reperfusion (intravenously),

provided a dose-dependent, potent pro-tection against the splanchnic ischemiaand reperfusion induced intestinal hyper-permeability (Fig. 5a). PJ34 fully pro-tected against the splanchnic occlusionand reperfusion induced mortality, evenat the lowest dose used (1 mg/kg; Fig.5b).

Effect of PJ34 on Vascular Reactivityin Endotoxic Shock. We next tested theeffect of PJ34 on various variables associ-

ated with endotoxic shock induced by E.coli LPS in the rat. The aortic rings takenfrom LPS animals showed a significantlyreduced peak response to phenylephrine,compared with controls. Animals pre-treated with PJ34 showed improved con-tractility (Fig. 6a).

The endothelial function was analyzedby relaxant responsiveness of precon-tracted vascular rings to the endotheli-um-dependent vasodilator, nitric oxideliberating hormone acetylcholine. LPS-treated animals developed an endothelialdysfunction with a reduced relaxant re-sponse compared with controls, and PJ34treatment improved the endothelial func-tion (Fig. 6b).

PJ34 Reduces the Severity of LPS-Induced Intraorgan Inflammatory Pro-cesses. Plasma chemistry values in theLPS-challenged animals with and with-out PJ34 treatment are shown in Table 3.PJ34 provided significant, partial protec-tion against the LPS-induced suppressionof albumin and total protein content andagainst the LPS-induced elevations inalanine aminotransferase, amylase, bloodurea nitrogen, and creatinine plasmaconcentrations. The activity of MPO inselected tissues is demonstrated in Figure7. In the lung there is an increase in MPOconcentrations in LPS-treated animals,and treatment with PJ34 protectedagainst this alteration. Also in the gut,MPO activity was reduced in the presenceof PJ34. In the liver there was no increasein MPO concentrations in response toLPS, and there were no significant differ-ences among the groups studied (Fig. 7).In gut and lung, there were significantincreases in the concentrations of MDA inresponse to LPS. PJ34 treatment reducedtissue MDA concentrations. In the liver,LPS failed to significantly increase MDAgeneration, and PJ34 did not exert anysignificant effects (Fig. 7).

PJ34 Reduces the LPS-Evoked Sys-temic Inflammatory Response. Peak se-rum concentrations of TNF-�, IL-1�, andnitrite/nitrate (breakdown products of ni-tric oxide) were all reduced by PJ34 treat-ment (Fig. 8).

PJ34 Improves Survival Rate in Endo-toxic Shock. Pretreatment of the animalswith PJ34 provided a significant survivaladvantage to the animals treated withLPS (Fig. 9). Because many potential an-tishock treatments lose their efficacy inthe posttreatment regimen, in an addi-tional set of experiments, we also testedwhether the effect of PJ34 was main-tained in the posttreatment regimen.

Figure 8. Plasma concentrations of (a) tumor necrosis factor (TNF)-�; (b) interleukin (IL)-1�, and (c)nitrite/nitrate (NOx; circulating breakdown products of nitric oxide) in rats at various times afterlipopolysaccharide (LPS) injection, in the presence or absence of PJ34 treatment. In control animals(without LPS injection), cytokine concentrations were below the detection limit, in the absence orpresence of PJ34. Mean � SEM. *p � .05 LPS vs. PJ34-treated LPS animals. n � 4–6 determinationsper group.

1078 Crit Care Med 2002 Vol. 30, No. 5

When PJ34 was injected to the animals 1hr after the administration of LPS, a clearand significant survival benefit was main-tained (Fig. 9). No additional changes inmortality rate occurred after 48 hrs, andall animals were killed after a 2-wk addi-tional observation period.

PJ34 Blocks PARS Activation in Endo-toxic Shock. There was a marked degreeof PARS activation, visualized as nuclearpoly(ADP-ribose) immunostaining, in theanimals injected with LPS at 24 hrs (butnot at 5 hrs; Fig. 10). Treatment withPJ34 provided complete blockade of poly-(ADP-ribose) accumulation, indicative of

the compound’s in vivo PARS inhibitoryactivity (Fig. 10).

DISCUSSION

PARS Inhibitors Exert Cytoprotectiveand Anti-Inflammatory Effects In Vitro.PARS recently has emerged as a cytopro-tective and antishock drug target. Wesynthesized a range of novel PARS inhib-itors based on the phenanthridinonestructure and tested its effects in a varietyof in vitro systems of oxidant-induced cellnecrosis and endotoxin-induced proin-flammatory cytokine production. In addi-

tion, one selected compound, PJ34, wasfurther characterized in vitro and wastested in shock models in vivo. The re-sults demonstrated that the PJ series arepotent, water-soluble inhibitors of PARS,with potent cytoprotective and anti-inflammatory effects in vitro and withpotent antishock effects in vivo.

The cytoprotective potential of PARSinhibitors emerged from early studies byBerger (3) and Cochrane (34). Thesestudies implicated the PARS pathway inthe process of cell death induced by hy-drogen peroxide and related oxidants.The mode of this PARS-dependent, oxi-dant-induced cell death has been clarifiedonly recently: It has been demonstratedthat PARS inhibitors protect against oxi-dant-induced cell necrosis (rather thaninhibiting apoptosis) (23, 35–38). Per-oxynitrite (produced when nitric oxidereacts with superoxide) is a key, patho-physiologically relevant trigger of PARSactivation (22, 25). Most previous in vitrostudies used millimolar concentrations ofthe weak PARS inhibitors 3-aminobenz-amide or nicotinamide (18, 31, 32). In anumber of prior studies, we used a ben-zopyrone derivative compound, whichwas cytoprotective and mimicked thePARS negative phenotype, in concentra-tions of approximately 100 �M (25). Thecurrent studies, demonstrating potent (inthe low nM range) cytoprotective actions

Figure 9. Survival rate over 48 hrs in lipopolysaccharide (LPS)-treated rats in the presence or absenceof PJ34 treatment. a, PJ34 (10 mg/kg intraperitoneally) was injected 60 mins before LPS (pretreatmentwith the PARS inhibitor). b, PJ34 (25 mg/kg intraperitoneally) was injected 60 mins after LPS(posttreatment with the PARS inhibitor). *p � .05 significant difference in the survival rate of LPStreated vs. PJ34 treated LPS animals. n � 18 animals per group.

Figure 10. Poly(adenosine 5'-diphosphate-ribose)(PAR) formation, an indicator of poly(adenosine5'-diphosphate-ribose) synthetase (PARS) activa-tion, as determined in lung sections from controlanimals and animals injected with lipopolysac-charide (LPS) with and without PJ34 treatmentfor 5 or 24 hrs (as in Fig. 9). There was a markedactivation of PARS, evidenced by increased PARstaining, at 24 but not at 5 hrs after LPS injec-tion. Treatment with 3-AB abolished PARS acti-vation. Immunohistochemical pictures representrepresentative sections from n � 4 determina-tions per group.

1079Crit Care Med 2002 Vol. 30, No. 5

of the PJ series, can be considered animportant further step toward the devel-opment of potent, water-soluble, antine-crotic, cytoprotective PARS inhibitorpharmacologic agents, which may beused in the future as research tools orpossibly as drug candidates.

When we consider the relative poten-cies of PJ34 with those of the prototypicalPARS inhibitor 3-aminobenzamide (3-AB), it may be surprising that althoughthe EC50 of PJ34 on both the isolatedenzyme assay and the cell-based cytopro-tection assay is similar, 3-AB appears tobe more potent in cells than in enzymes.The possibility that 3-AB exerts additionaleffects (e.g., as an antioxidant) may beresponsible for this difference. In fact, wehave compared the potential antioxidanteffect of various PARS inhibitors, and wefound that 3-AB indeed acts as an antiox-idant in vitro, whereas PJ34 does not (21,25).

Initial studies on the role of PARS inexperimental models of disease focusedon its effects on intracellular energeticsand resultant cellular dysfunction. It isalso noteworthy, however, that the ab-sence of PARS or its pharmacologic inhi-bition has been shown to suppress theactivation of MAP kinase (26), AP-1 com-plex (39), and nuclear factor-�B (27).Consequently, PARS inhibition interfereswith the expression of proinflammatorygenes, such as inducible NO synthase (20,22) and intercellular adhesion mole-cule-1 (34), that are dependent on thesesignaling pathways. The regulation byPARS of gene expression may involve thepoly-ADP ribosylation of transcriptionfactors or the repair of DNA strand breaksthat interfere with transcription. PARSinhibition and genetic PARS deficiencyhave been shown, for example, to sup-press endotoxin-induced expression ofTNF-�, IL-6, inducible NO synthase, andcyclooxygenase-2 (26, 27). Our currentstudies in which we used immunostimu-

lated murine macrophages and rodentstreated with PJ34 confirm and extend thisconcept: PJ34 suppressed the productionof TNF-�, NO, IL-1�, and the chemokinesMIP-1� and MIP-2. Because these medi-ators are important in the pathogenesisof shock and various forms of inflamma-tion (40–42), we hypothesize that thesuppression by PARS inhibitors of che-mokine production may play a role insuppressing neutrophil recruitment byPARS inhibitors into the inflamed or-gans. It is important to note that theinhibition of mediator production by thePJ series of PARS inhibitors was seen atdoses somewhat higher than the dosesseen for cytoprotection. This observationmay be related to the fact that a near-complete degree of the inhibition of thecatalytic activity of PARS may be neededto affect signal transduction pathways(e.g., because of the high sensitivity andredundancy of these pathways). Alterna-tively, pharmacologic actions of the com-pounds on cellular targets other thanPARS may be involved.

PARS Inhibitors Exert Cytoprotectiveand Anti-Inflammatory Effects In Vivo.Circulatory shock, a systemic inflamma-tory condition, is associated with a re-duced responsiveness of arteries andveins to exogenous or endogenous vaso-constrictor agents (vascular hyporeactiv-ity), myocardial dysfunction, and dis-rupted intracellular energetic processes,culminating in multiple organ failure anddeath. Some of these alterations previ-ously have been suggested to be related toNO or peroxynitrite overproduction, at-tributable to the activation of the endo-thelial isoform of NO in the early stageand expression of a distinct inducible iso-form of NO synthase in the later stage orthe disease (43). Shock is associated withproinflammatory cytokine productionand subsequent stimulation of oxygen-centered free radical and peroxynitriteproduction. The vascular contractile fail-ure associated with circulatory shock iswell known to be closely related to over-production of NO within the blood ves-sels. The evidence that PARS is involvedin the peroxynitrite-induced vascular hy-poreactivity in shock is multiple. First, instudies in anesthetized rats, inhibition ofPARS with 3-aminobenzamide and nico-tinamide reduced the suppression of thevascular contractility of the thoracicaorta in ex vivo experiments (8). Thesefindings are similar to the in vitro resultswith authentic peroxynitrite, which alsocauses a vascular hyporeactivity in tho-

racic aortic rings and which can be re-duced by pharmacologic inhibition ofPARS (9). Second, the vascular hyporeac-tivity associated with hemorrhagic shockis attenuated in animals deficient in func-tional PARS (10). Peroxynitrite produc-tion and PARS activation also have beensuggested to contribute to endothelial in-jury in various pathophysiological statesincluding shock (44–49). Finally, the in-testinal leakage, which is considered bymany investigators as an important trig-ger of multiple organ failure, is markedlyreduced in PARS-deficient animals sub-jected to shock (7, 10). The current stud-ies with PJ34 in splanchnic occlusionshock and endotoxin-treated rats confirmand extend these observations: PARS in-hibition significantly and markedly im-proved vascular reactivity, protectedagainst intestinal hyperpermeability, andreduced proinflammatory mediator pro-duction, systemic organ failure, infiltra-tion of neutrophils into the tissues, andintraorgan oxidant generation. Finally,PARS inhibition provided significant sur-vival benefit, an effect that was main-tained in the posttreatment therapeuticregimen. This last point is important, be-cause many investigators believe that onereason for the recent clinical failures ofantishock “magic bullets” was the factthat the therapies tested were targetingearly pathways of systemic inflammation,and they lost their efficacy when admin-istered at later stages of the disease.

Although the current models of shockare useful and satisfactory for preliminarystudies and screening purposes, rats areremarkably resistant to endotoxin, andchallenge with high doses of one selectedbacterial proinflammatory component(e.g., endotoxin) does not fully mimic theseptic condition. Thus, large animal stud-ies, testing PJ34 in models of bacteria-induced sepsis, are required to furthervalidate the concept. In this respect it isimportant to note that in a model ofsepsis induced by live E. coli sponge im-plantation in pigs, pharmacologic inhibi-tion of PARS with PJ34 provides markedhemodynamic improvements and mas-sive survival benefit (50). It is likely thatthe improved hemodynamic status attrib-utable to improved vascular function, andpossibly the improved cellular energeticstatus in some organs, results in an over-all survival benefit in these various con-ditions of LPS-induced shock and bacte-rial sepsis.

When comparing the effects of PJ34with those of the prototypical PARS in-

P oly(ADP-ribose)

synthetase emerges

as a drug target

for the clinical treatment

of shock and systemic

inflammation.

1080 Crit Care Med 2002 Vol. 30, No. 5

hibitor 3-AB in vivo, we find that theeffective doses of 3-aminobenzamide arehigher than the effective doses of PJ34,but—unlike the in vitro situation—notby several orders of magnitude. For in-stance, in a previous survival study inLPS shock that used 3-AB, a total dose of30 mg/kg was administered over a 24-hrperiod (8). In contrast, here we used adose of 10 mg/kg, administered onlyonce. Yet, the degree of protection seenwith PJ34 was much more robust thanthe protection seen with 3-AB previously.The differences were even more pro-nounced when the effects of 3-AB andPJ34 on reperfusion injury were com-pared: Whereas PJ34 exerts protective ef-fects in 1–3 mg/kg doses, the effects of3-AB can be seen at 20–100 mg/kg doses(32). Furthermore, whereas 3-aminoben-zamide, even at 100 mg/kg dose, does notexert a full inhibitory effect on PARS ac-tivation—as assessed by immunohisto-chemistry—(32), PJ34 exerts a completeinhibitory effect on PARS activation atthe 10 mg/kg dose (Fig. 10). The fact that3-aminobenzamide also exerts additionalantioxidant effects that may contribute tosome of its protective effects was dis-cussed earlier.

Summary and Implications: Role ofPARS in Shock and Systemic Inflamma-tion. The following working hypothesissummarizes our current understandingon the role of PARS activation in thedevelopment of shock and systemic in-flammatory injury. Proinflammatory cy-tokines stimulate free radical and oxidantformation. Oxidant stress triggers proin-flammatory gene expression and alsogenerates DNA single-strand breaks. DNAstrand breaks then activate PARS, withsubsequent cell dysfunction and ulti-mately necrosis. PARS also potentiatesnuclear factor-�B activation and AP-1 ex-pression, resulting in greater expressionof the AP-1 and nuclear factor-�B depen-dent genes, such as inducible NO syn-thase, intercellular adhesion molecule-1,MIP-2, MIP-1�, and TNF-�. The chemo-kines, in combination with increased en-dothelial expression of intercellular adhe-sion molecule-1, recruit more activatedleukocytes, producing greater oxidantstress. The cycle is thus renewed as theincrease in oxidant stress triggers moreDNA strand breakage. Thus, PARS occu-pies a critical position in the various pos-itive-feedback loops of shock and sys-temic inflammatory injury.

In this report, we described the invitro characterization of a novel series of

potent, water-soluble PARS inhibitorsand presented rodent studies demonstrat-ing the protective effects of the com-pounds in endotoxic shock and splanch-nic reperfusion injury. PARS emerges asa drug target for the clinical treatment ofshock and systemic inflammation.

REFERENCES

1. Lautier D, Lageux J, Thibodeau J, et al: Mo-lecular and biochemical features of poly(ADP-ribose) metabolism. Mol Cell Biochem1993; 122:171–193

2. Ueda K, Hayaishi O: ADP-ribosylation. AnnuRev Biochem 1985; 54:73–100

3. Berger NA: Oxidant-induced cytotoxicity: Achallenge for metabolic modulation. Am JRespir Cell Mol Biol 1991; 4:1–3

4. Szabo C (Ed): Cell Death: The Role of PARP.Boca Raton, FL, CRC Press, 2000

5. Zingarelli B, Cuzzocrea S, Zsengeller Z, et al:Beneficial effect of inhibition of poly-ADPribose synthetase activity in myocardial isch-emia-reperfusion injury. Cardiovasc Res1997; 6:205–215

6. Zingarelli B, Salzman AL, Szabo C: Geneticdisruption of poly (ADP-ribose) synthetaseinhibits the expression of P-selectin and in-tercellular adhesion molecule-1 in myocar-dial ischemia/reperfusion injury. Circ Res1998; 83:85–94

7. Liaudet L, Szabo A, Soriano FG, et al: Poly(ADP-ribose) synthetase mediates intestinalmucosal barrier dysfunction after mesentericischemia. Shock 1999; 14:134–141

8. Szabó C, Zingarelli B, Salzman AL: Role ofpoly-ADP ribosyltransferase activation in thevascular contractile and energetic failureelicited by exogenous and endogenous nitricoxide and peroxynitrite. Circ Res 1996; 78:1051–1063

9. Szabó C, Cuzzocrea S, Zingarelli B, et al:Endothelial dysfunction in endotoxic shock:Importance of the activation of poly (ADPribose) synthetase (PARS) by peroxynitrite.J Clin Invest 1997; 100:723–735

10. Liaudet L, Soriano FG, Szabo E, et al: Pro-tection against hemorrhagic shock in micegenetically deficient in poly(ADP-ribose)polymerase. Proc Natl Acad Sci U S A 2000;97:10203–10208

11. Thiemermann C, Bowes J, Myint FP, et al:Inhibition of the activity of poly(ADP ribose)synthetase reduces ischemia-reperfusion in-jury in the heart and skeletal muscle. ProcNatl Acad Sci U S A 1997; 94:679–683

12. Pieper AA, Walles T, Wei G, et al: Myocardialpostischemic injury is reduced by polyADPri-bose polymerase-1 gene disruption. Mol Med2000; 6:271–282

13. Pulido EJ, Shames BD, Selzman CH, et al:Inhibition of poly (ADP-ribose) synthetaseattenuates endotoxin-induced dysfunction ofpulmonary vasorelaxation. Am J Physiol1999; 277:L769–L776

14. Watts JA, Grattan RM II, Whitlow BS, et al:

Activation of poly(ADP-ribose) polymerase insevere hemorrhagic shock and resuscitation.Am J Physiol 2001; 281:G498–G506

15. Soriano FG, Liaudet L, Szabó É, et al: Resis-tance against acute septic peritonitis in poly-(ADP-ribose) polymerase deficient mice.Shock, In Press

16. Banasik M, Komura H, Shimoyama M, et al:Specific inhibitors of poly(ADP-ribose) syn-thetase and mono(ADP-ribosyl)transferase.J Biol Chem 1992; 267:1569–1575

17. Andrievski AM, Poplavski AN, Dyumaev KM,et al: Synthesis and structure of nitro-substituted 6(5H)-phenanthridinones. ChemHeterocycl Compd 1985; 8:924–931

18. Migachev GI, Terenev AM: Studies onphenanthridinone and dioxotetrahydrodiaza-pyrene. 4. Synthesis of amino-substituted de-rivatives of phenanthridinone and dioxotet-rahydrodiazapyrene. Chem HeterocyclCompd 1981; 17:295–298

19. Harfenist M, Joyner CT, Mize PD, et al: Se-lective inhibitors of monoamine oxidase. 2.Arylamide SAR. J Med Chem 1994 37:2085–2089

20. Li J, Serdyuk L, Ferraris DV, et al: Synthesisof substituted 55(H)phenanthridin-6-ones.Bioorg Med Chem Lett 2001; 11:1687–1690

21. Soriano FG, Virág L, Jagtap P, et al: Diabeticendothelial dysfunction: The role of poly(ADP-ribose) polymerase activation. NatureMed, 2001; 7:108–113

22. Szabó C, Zingarelli B, O’Connor M, et al:DNA strand breakage, activation of poly-ADPribosyl synthetase, and cellular energy deple-tion are involved in the cytotoxicity in mac-rophages and smooth muscle cells exposed toperoxynitrite. Proc Natl Acad Sci U S A 1996;93:1753–1758

23. Virag L, Scott GS, Cuzzocrea S, et al: Per-oxynitrite-induced thymocyte apoptosis: Therole of caspases and poly (ADP-ribose) syn-thetase (PARS) activation. Immunology1998; 94:345–355

24. Zhang J: Use of biotinylated NAD to label andpurify ADP-ribosylated proteins. MethodsEnzymol 1997; 280:255–265

25. Szabó C, Virág L, Cuzzocrea S, et al: Protec-tion against peroxynitrite-induced fibroblastinjury and arthritis development by inhibi-tion of poly (ADP-ribose) synthetase. ProcNatl Acad Sci U S A 1998; 95:3867–3872

26. Szabó C, Wong H, Bauer PI, et al: Regulationof components of the inflammatory responseby 5-iodo-6-amino-1,2-benzopyrone, an in-hibitor of poly (ADP-ribose) synthetase andpleiotropic modifier of cellular signal path-ways. Int J Oncol 1997; 10:1093–1104

27. Oliver FJ, Menissier-de Murcia J, Nacci C, etal: Resistance to endotoxic shock as a conse-quence of defective NF-kappaB activation inpoly (ADP-ribose) polymerase-1 deficientmice. EMBO J 1999; 18:4446–4454

28. Gerritsen ME, Shen C-P, McHugh MC, et al:Activation-dependent isolation and culture ofmurine pulmonary microvascular endothe-lium. Microcirculation 1995; 2:151–163

29. Hoyt DG, Lazo JS: Murine strain differences

1081Crit Care Med 2002 Vol. 30, No. 5

in acute lung injury and activation of poly-(ADP-ribose) polymerase by in vitro exposureof lung slices to bleomycin. Am J Respir CellMol Biol 1992;7: 645–651

30. Hoyt DG, Lazo JS: NAD depletion after invitro exposure of murine lung slices to bleo-mycin. Biochem Pharmacol 1993; 46:1819–1824

31. Kupper JH, van Gool L, Muller M, et al:Detection of poly(ADP-ribose) polymeraseand its reaction product poly(ADP-ribose) byimmunocytochemistry. Histochem J 1996;28:391–395

32. Liaudet L, Szabo E, Timashpolsky L, et al:Suppression of poly (ADP-ribose) polymeraseactivation by 3-aminobenzamide in a ratmodel of myocardial infarction: Long-termmorphological and functional consequences.Br J Pharmacol 2001; 133:1424–1430

33. Liaudet L, Yang Z, Al-Affar EB, et al: Myocar-dial ischemic preconditioning in rodents isdependent on poly (ADP-ribose) synthetase.Mol Med 2001; 7:406–417

34. Cochrane CG: Cellular injury by oxidants.Am J Med 1991; 91:23S–30S

35. Virag L, Salzman AL, Szabo C: Poly(ADP-ribose) synthetase activation mediates mito-chondrial injury during oxidant-induced celldeath. J Immunol 1998; 161:3753–3759

36. Ha HC, Snyder SH: Poly(ADP-ribose) poly-merase is a mediator of necrotic cell death byATP depletion. Proc Natl Acad Sci U S A1999; 36:13978–13982

37. Radons J, Heller B, Burkle A, et al: Nitricoxide toxicity in islet cells involves poly(ADP-ribose) polymerase activation and con-

comitant NAD depletion. Biochem BiophysRes Comm 1994; 199:1270–1277

38. Zhang J, Dawson VL, Dawson TM, et al: Ni-tric oxide activation of poly (ADP-ribose) syn-thetase in neurotoxicity. Science 1994; 263:687–689

39. Roebuck KA, Rahman A, LakshminarayananV, et al: H2O2 and tumor necrosis factor-alpha activate ICAM-1 gene transcriptionthrough distinct cis-regulatory elementswithin the ICAM-1 promoter. J Biol Chem1995; 270:18996–18974

40. Zingarelli B, Szabo C, Salzman AL: Blockadeof poly (ADP-ribose) synthetase inhibits neu-trophil recruitment, oxidant generation, andmucosal injury in murine colitis. Gastroen-terology 1999; 116:335–345

41. Szabó C, Lim LH, Cuzzocrea S, et al: Inhibi-tion of poly (ADP-ribose) synthetase attenu-ates neutrophil recruitment and exerts anti-inflammatory effects. J Exp Med 1997; 186:1041–1049

42. Cuzzocrea S, Zingarelli B, Gilad E, et al:Protective effects of 3-aminobenzamide, aninhibitor of poly (ADP-ribose) synthase in acarrageenan-induced model of local inflam-mation. Eur J Pharm 1998; 342:67–76

43. Kilbourn RG, Szabó C, Traber D: Beneficialversus detrimental effects of nitric oxide syn-thase inhibitors in circulatory shock: Les-sons learned from experimental and clinicalstudies. Shock 1997; 7:235–246

44. Zingarelli B, Day BJ, Crapo JD, et al: Thepotential role of peroxynitrite in the vascularcontractile and cellular energetic failure in

endotoxic shock. Br J Pharmacol 1997; 120:259–267

45. Villa LM, Salas E, Darley-Usmar M, et al:Peroxynitrite induces both vasodilatationand impaired vascular relaxation in the iso-lated perfused rat heart. Proc Natl AcadSci U S A 1994; 91:12383–12387

46. Chabot F, Mitchell JA, Quinlan GJ, et al:Characterization of the vasodilator proper-ties of peroxynitrite on rat pulmonary ar-tery: Role of poly (adenosine 5'-diphospho-ribose) synthase. Br J Pharmacol 1997;121:485– 490

47. Spragg RG: DNA strand break formation fol-lowing exposure of bovine pulmonary arteryand aortic endothelial cells to reactive oxy-gen products. Am J Respir Cell Mol Biol1991; 4:4–10

48. Thies RL, Autor AP: Reactive oxygen injuryto cultured pulmonary artery endothelialcells: Mediation by polyADP-ribose polymer-ase activation causing NAD depletion andaltered energy balance. Arch BiochemBiophys 1991; 286:353–363

49. Zingarelli B, O’Connor M, Wong H, et al:Peroxynitrite-mediated DNA strand breakageactivates poly-ADP ribosyl synthetase andcauses cellular energy depletion in macro-phages stimulated with bacterial lipopolysac-charide. J Immunol 1996; 156:350–358

50. Goldfarb RD, Marton A, Szabó E, et al: Pro-tective effect of a novel, potent inhibitor ofpoly(adenosine 5�-diphosphate-ribose) syn-thetase in a porcine model of severe bacterialsepsis. Crit Care Med 2002; 30:974–980

1082 Crit Care Med 2002 Vol. 30, No. 5