1521-0103/359/1/207–214$25.00 http://dx.doi.org/10.1124/jpet.116.234013THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS J Pharmacol Exp Ther 359:207–214, October 2016Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics

Pharmacological Characterization of a Potent Inhibitor ofAutotaxin in Animal Models of Inflammatory Bowel Diseaseand Multiple Sclerosis

Kannan Thirunavukkarasu, Bailin Tan, Craig A. Swearingen, Guilherme Rocha, Hai H. Bui,Denis J. McCann, Spencer B. Jones, Bryan H. Norman, Lance A. Pfeifer, and Joy K. SahaLilly Research Laboratories, Eli Lilly and Company, Indianapolis, Indiana

Received April 29, 2016; accepted August 10, 2016

ABSTRACTAutotaxin is a secreted enzyme that catalyzes the conversion oflysophosphatidyl choline into the bioactive lipid mediator lyso-phosphatidic acid (LPA). It is the primary enzyme responsiblefor LPA production in plasma. It is upregulated in inflamma-tory conditions and inhibition of autotaxin may have anti-inflammatory activity in a variety of inflammatory diseases. Todetermine the role of autotaxin and LPA in the pathophysiology ofinflammatory disease states, we used a potent and orally bio-available inhibitor of autotaxin that we have recently identified, andcharacterized it in mouse models of inflammation, inflammatorybowel disease (IBD), multiple sclerosis (MS), and visceral pain.Compound-1, a potent inhibitor of autotaxinwith an IC50 of∼2 nM,

has good oral pharmacokinetic properties in mice and results in asubstantial inhibition of plasma LPA that correlates with drugexposure levels. Treatment with the inhibitor resulted in significantanti-inflammatory and analgesic effects in the carrageenan-induced paw inflammation and acetic acid-induced visceral paintests, respectively. Compound-1 also significantly inhibited dis-ease activity score in the dextran sodium sulfate–induced modelof IBD, and in the experimental autoimmune encephalomyelitismodel of MS. In conclusion, the present study demonstrates theanti-inflammatory and analgesic properties of a novel inhibitor ofautotaxin that may serve as a therapeutic option for IBD, MS, andpain associated with inflammatory states.

IntroductionAutotaxin is a secreted enzyme that produces the lipid

mediator lysophosphatidic acid (LPA) from extracellularlysophosphatidyl choline (LPC) (Umezu-Goto et al., 2002). Itis the primary enzyme responsible for LPA production inplasma (van Meeteren et al., 2006). LPA acts through Gprotein-coupled receptors on the cell surface and inducesdownstream effects, influencing a variety of physiologic andpathologic processes (van Meeteren et al., 2006; Chun et al.,2010; Moolenaar and Perrakis, 2011). Accumulating evidencehas shown that aberrant autotaxin expression and elevatedLPA production and signaling are involved in multiplepathologic conditions, including chronic inflammation, fibro-sis, colitis, and cancer (Lin et al., 2009; Gierse et al., 2010; Chuet al., 2015; Castagna et al., 2016).Inflammatory bowel disease (IBD) is a chronic and relapsing

intestinal disease that is accompanied by abdominal pain,diarrhea, rectal bleeding, and loss of the ability to digest food(Mozdiak et al., 2015). IBD includes ulcerative colitis andCrohn’s disease, both of which are chronic diseases of thegastrointestinal (GI) tract. The pathophysiologicalmechanisms

of IBD are not completely understood. Enhanced expression ofautotaxin mRNA and aberrant lymphocyte migration havebeen observed in the inflamed mucosa in patients with Crohn’sdisease or ulcerative colitis. Autotaxin expression is signifi-cantly higher in the actively inflamed mucosa compared withthe quiescentmucosa in the same patient (Hozumi et al., 2013).In addition, autotaxin in the vascular endothelium has beenshown to be involved in pathologic lymphocyte migration to theinflamed mucosa in two murine models of colitis, namely theCD41 CD25– T-cell transfer model and the dextran sodiumsulfate (DSS)–induced colitis model (Hozumi et al., 2013).Pharmacological intervention with bithionol, a weak inhibitorof autotaxin, has been shown to decrease lymphocyte in-filtration and ameliorate colitis in the T-cell transfer model.However, there is no data describing effects on functionalparameters of disease activity in the DSS model, and no otherstudy exists describing pharmacological activity of a potentand selective inhibitor of autotaxin in models of IBD.Multiple sclerosis is an autoimmune and chronic inflamma-

tory neurologic disorder in which the immune system destroysthe myelin and axons in the gray and white matter of thecentral nervous system (CNS) leading to progressive disability(Compston and Coles, 2002). Tissues of the CNS, such as thebrain and spinal cord, express abundant levels of autotaxindx.doi.org/10.1124/jpet.116.234013.

ABBREVIATIONS: CGN, carrageenan; CNS, central nervous system; DAI, disease activity index; DSS, dextran sodium sulfate; EAE, experimentalautoimmune encephalomyelitis; GI, gastrointestinal; HEC, 1% hydroxyethyl cellulose, 0.25% Tween-80, and 0.05% antifoam; IBD, inflammatorybowel disease; LPA, lysophosphatidic acid; LPC, lysophosphatidyl choline; MOG, myelin oligodendrocyte glycoprotein; MS, multiple sclerosis;TNF, tumor necrosis factor.

207

at ASPE

T Journals on O

ctober 24, 2020jpet.aspetjournals.org

Dow

nloaded from

transcripts, more thanmost other tissues (Mills andMoolenaar,2003; Kanda et al., 2008). PatientswithMShave been shown tohave increased levels of LPA in serum (Balood et al., 2014) andincreased expression and activity of autotaxin in the cerebro-spinal fluid (Hammack et al., 2004; Zahednasab et al., 2014). Inaddition, in the mouse experimental autoimmune encephalo-myelitis (EAE) model of MS, LPA1-receptor knockout micewere less affected by disease (Garcia-Diaz et al., 2010). LPCandLPA have also been shown to cause demyelination in vivo andin ex vivo cultures of dorsal root fibers (Inoue et al., 2004; Fujitaet al., 2007; Nagai et al., 2010). Dorsal root demyelinationcaused by nerve injury or intrathecal injection of LPC ismarkedly attenuated in autotaxin heterozygote mice and inLPA1-receptor knockout mice (Nagai et al., 2010). These dataimplicate the autotaxin/LPA pathway in the pathogenesis ofMS. It is currently not known whether pharmacological in-hibition of autotaxin (and the resultant decrease in LPA levels)would ameliorate disease in an animal model of MS.In the present study, we have demonstrated the anti-

inflammatory and analgesic properties of an autotaxin in-hibitor (compound-1) in animal models. To our knowledge,this is the first report demonstrating the pharmacologicalefficacy of a potent, selective and orally bioavailable inhibitorof autotaxin in models of IBD, MS, and visceral pain.

Materials and MethodsGeneral Procedure and Animal Maintenance

Experiments were carried out in C57/Black6 female mice that were9–10 weeks old (Charles River Breeding Laboratories, Portage, MI)unless otherwise mentioned elsewhere. Following shipment, the micewere housed for at least 1 week to allow acclimatization prior toexperimental studies. The animals were maintained on a regular12-hour light/dark cyclewith food andwater ad libitum, unless otherwisenoted. Protocols were approved by the Animal Care and Use Committee(IACUC) of Eli Lilly and Company (Indianapolis, IN) and were inagreement with the guidelines for the proper use and care of animals inbiomedical research. Blood samples were collected by cardiac puncturefor measurement of analytes under isoflurane anesthesia. Studies wereperformed in a blinded fashion, with the group and compound identitykept unknown to the individual collecting data from the studies.Compound and group identities were uncovered after analysis of theresults. At the end of the studies, all animals were euthanized by CO2,followed by cervical dislocation, per IACUC guidelines. Statisticalanalyses were performed using Student’s t test, Dunnett’s test, andMann-Whitney’s U test in the GraphPad Prism program. The analyzeddataare presentedasmean6 standard errorwith thenumber of animals(n) in parenthesis, and values considered significant when P , 0.05.

Reagents

The autotaxin inhibitor (compound-1) (Fig. 1A) was discovered andcharacterized at Lilly Research Laboratories (Beauchamp et al., 2015).Vehicle HEC (1% hydroxyethyl cellulose, 0.25% Tween-80, and 0.05%antifoam) was prepared internally at Lilly Research Laboratories.Compound-1 was dissolved in HEC and vortexed at room temperatureto prepare a fine suspension, and was dosed by oral gavage using a #22gavage needle. Compound suspension was prepared fresh every dayprior to dosing to avoid any contamination and precipitation. Carra-geenan (CGN) was purchased from Sigma-Aldrich (cat. no. C1867; St.Louis, MO), glacial acetic acid from Mallinckrodt Chemicals (cat. no.V183-14z; St. Louis, MO), dextran sulfate sodium (DSS; molecularweight 30–40K) fromMPBiomedicals (cat. no. 160110; Santa Ana, CA),MOG35–55 peptide (MEVGWYRSPFSRVVHLYRNGK) from PolyPep-tide Group (cat. no. SC1272; Torrance CA), complete Freund’s adjuvant(cat. no. 231131) and heat-killed Mycobacterium tuberculosis (H37RA,cat. no. 231141) from Difco Laboratories/Becton Dickinson and Com-pany (Franklin Lakes, NJ), and pertussis toxin from Life Technolo-gies/Thermo Fisher Scientific, Waltham, MA (cat. no. PHZ1174).

Pharmacokinetic/Pharmacodynamic Studies in Mice

Pharmacokinetic studies were performed in 7–8 week old maleC57/BL6mice fromHarlan/Envigo (Indianapolis, IN). Compound-1wasdosed orally at the indicated doses and bloodwas collected at 0, 0.5, 1, 2,3, 6, and 24 hours post-dose (n 5 3 per time point). Plasma levels ofcompound-1 were determined bymass spectrometry, and LPA levels inplasma were measured as described previously (Chen et al., 2008).

Carrageenan-Induced Paw Inflammation

Body weights were taken and themice were randomized into variousstudy groups on the basis of their body weight. Carrageenan wasprepared on the day of the study prior to injection. Eachmouse receivedCGN (0.5%, 40 ml/paw) in the right hind paw (intraplantar injection)immediately after the initial paw thicknesswasmeasured (time0). Testcompoundswere freshly prepared on the day of the study andmiceweredosed with the compounds on day 0 and 16 hours later on day 1. Onehour later, CGN was injected. Paw inflammation (paw thickness) wasmeasured by a minicaliper with digital display at different time points,for a period of six hours post-CGN. The changes in paw inflammationare expressed in millimeters as changes in paw thickness (differencebetween the time 0 and various time points post-CGN).

Measurement of LPS-Induced Tumor Necrosis Factor aProduction

Studies were conducted to determine the plasma levels of theinflammatory cytokine tumor necrosis factor (TNF)a induced by LPStreatment in mice. After 7 days of acclimatization, body weights weretaken and the mice were randomized into two study groups (n 54–5/group) on the basis of body weight. Control group received vehicle

Fig. 1. (A) Chemical structure ofcompound-1. (B) Pharmacokinetics andpharmacodynamics of compound-1(30 mg/kg) dosed orally in mice.Plasma samples were collected atvarious time points post-dosing, andcompound levels and LPA levels weremeasured as described. The horizon-tal dotted line indicates 50% inhibi-tion of plasma LPA. Data show aninverse relationship between com-pound exposure levels and the corre-sponding LPA levels in plasma.

208 Thirunavukkarasu et al.

at ASPE

T Journals on O

ctober 24, 2020jpet.aspetjournals.org

Dow

nloaded from

HEC and the treatment group received compound-1. Immediatelyafter compound administration, mice in both groups received anintraperitoneal injection of LPS (10 mg/kg). Two hours following theinjection of LPS, animals were euthanized with CO2, and blood wascollected by cardiac puncture in heparin-coated tubes (Becton Dick-inson). Plasma was prepared and stored frozen at –20°C, untilneeded for cytokine analysis. TNFa levels in the plasma weremeasured using the mouse TNFa Quantikine ELISA Kit (R&DSystems, Minneapolis, MN).

DSS-Induced Inflammatory Bowel Disease Model

TheDSS-IBDprotocol wasmodified fromMurthy et al. (1993). Bodyweights were taken a day before the study start, and mice wererandomized into five study groups: sham (none), vehicle HEC, andthree groups receiving different doses of compound-1. DSS solution(5%) was prepared in drinking water on the day of the study and every2–3 days thereafter. IBD was induced by allowing the mice to drinkDSS-water ad libitum given in their drinking water bottle. Dosingwith the autotaxin inhibitor was started immediately after the micewere allowed to drink DSS-containing water. The compounds wereadministered orally by a gavage needle twice a day, in a dose volume of200 ml/mouse, at final doses of 3, 10, or 30 mg/kg. Body weights,stool consistency, occult blood (Hemoccult II SENSA rapid test kit;Beckman Coulter Inc., Brea, CA), or gross rectal bleeding were de-termined daily in a blinded fashion. The study was terminated on day6 post-DSS, and the colon tissuewas collected. Parametersmeasured atthe end of the study include body weight, stool consistency, stools forhemoccults (if no rectal bleeding), and colon length. Colon tissue fromthe ileocecal junction to the anal verge was quickly excised and thelength was measured using a ruler.

Disease Severity Measurement. Disease severity is expressedas disease activity index (DAI) on the basis of the changes in threeimportant parameters of IBD, namely: 1) body weight, 2) stoolconsistency, and 3) rectal bleeding. DAI5 (A1 B1 C)/3. DAI scoringscheme was adapted from the literature (Eijkelkamp et al., 2007), andis shown in Table 1.

Acetic Acid Writhing Model

Mice were randomized into four study groups (n 5 10 mice/group)on the basis of body weight. Dosing was staggered to facilitate thebehavioral observations. Control group received vehicle HEC and theother three study groups received three different doses (3, 10,30 mg/kg) of compound-1. The next morning (after 16–18 hours), themice were given another dose of the vehicle and test compound. Twohours later, themicewere given 0.9% glacial acetic acid (100ml/mouse,i.p.) that was freshly prepared and dosed at 5 ml/kg by i.p. injection toinduce writhing. The mice were placed immediately in plasticobservation cages. A timer was turned on after the first mousereceived acetic acid, and the writhing response (stretching/twistingof the hind leg and contraction of abdomen) was cumulatively countedfor a period 10 minutes after 5 minutes of acetic acid injection. Wehave previously validated the model with an NSAID (diclofenac,30 mg/kg, PO) as a positive control that results in a significantinhibition of the writhing response (data not shown).

Experimental Autoimmune Encephalomyelitis

The myelin oligodendrocyte glycoprotein (MOG)–induced experi-mental autoimmune encephalomyelitis (EAE)model is widely used forthe evaluation of different drugs that show efficacy in human MS(Floris et al., 2002). The mouse EAE study was performed at CovanceLaboratories (Greenfield, IN), following IACUC animal study guide-lines as approved by Eli Lilly & Company. Studies were conducted in7–8 week old female C57BL/6mice that were group housed (3–5/cage).Following shipment, the mice were acclimatized at Covance for3–7 days prior to inclusion in experimental studies. EAE was inducedin mice by injecting subcutaneously into the flanks an emulsioncontaining MOG35–55 peptide (MEVGWYRSPFSRVVHLYRNGK)(cat. no. SC1272; PolyPeptide Group). Briefly, the MOG peptide wasdiluted to 2 mg/ml in saline. Equal volumes of complete Freund’sadjuvant and MOG peptides were mixed and homogenized. Two-hundred microliters of the MOG homogenate was dosed subcutane-ously (100 ml each into the right and left flanks) on day 1. Pertussistoxin was administered intraperitoneally on days 1 and 3. Tenanimals were used for each of the treatment groups and three naiveanimals (without any MOG peptide injection) were used as “nodisease” comparators.

Prednisolone (10 mg/kg) was used as a positive control in the studyand was dosed orally, once daily from days 1 to 30. The autotaxininhibitor compound-1 was resuspended in HEC and was dosed orally,twice daily at 30 mg/kg on days 1–30. Body weights were measuredonce on day 1 and then daily starting on day 8. Clinical signs weremonitored twice daily to check for dead and/or moribund animals.Clinical scoring for disease was done daily starting from day 8 usingthe scale in Table 2.

During the study, if the animals showed severe signs of disease andcouldn’t move to food and water, they were sacrificed and were gradedas 6.0 for the remainder of the study. All the animalswere sacrificed onday 31 at the end of the study. Eight of 10 (80%) of the vehicle-treatedmice demonstrated a daily clinical score of $1, indicating diseaseincidence, and demonstrated a mean peak score of 3.2 on day 15.

ResultsPharmacokinetics and Pharmacodynamics of

Compound-1 in Mice. We recently described the identifica-tion of compound-1 (Fig. 1A) as an orally available smallmolecule that inhibits autotaxin with an IC50 of ∼2 nM in anex vivo human whole blood assay and does not inhibit relatedproteins ENPP1 and ENPP7 (Beauchamp et al., 2015; Joneset al., 2016). Administration of this compound in rats results in adose- and time-dependent inhibition of plasma LPA levels. Toconfirm that the compound has a good oral exposure profile inmice, we performed a pharmacokinetic/pharmacodynamic study

TABLE 1DAI scoring schemeNormal stool: firm and well-formed pellets. Loose stool: pasty/semi-formed that do notstick to the anus. Diarrhea: liquid stool that sticks to the anus.

Score Weight Loss Stool Consistency Fecal Blood

0 None Normal Normal1 1–5%2 5–10% Loose stool Hemoccult +3 10–15%4 15–20% Diarrhea Gross bleeding

TABLE 2Clinical scoring guide

Score Characteristics

0 No weakness0.5 Distal limp or spastic, curling tail1.0 Complete tail limpness/paralysis1.5 Easily rights and complete tail paralysis2.0 Easily flips onto back but no hind limb weakness2.5 Easily flips onto back and stumbles slightly when walking3.0 Easily flips onto back with moderate hind limb weakness3.5 Severe hind limb weakness4.0 Brings feet fully forward and feet will drag4.5 Partial hind limb paralysis5.0 Complete hind limb paralysis5.5 Complete hind limb paralysis and mild front limb weakness6.0 Complete both hind and front limb paralysis and/or dead

Autotaxin Inhibitor for IBD and MS 209

at ASPE

T Journals on O

ctober 24, 2020jpet.aspetjournals.org

Dow

nloaded from

inmice.Oral dosing of compound-1 (30mg/kg) inmice resulted ingood plasma exposurewith aCmax of∼24mMat 1 hour post-dose(Fig. 1). To assess the functional activity of the compound, wemeasured LPA levels in plasma collected at the various timepoints, and observed an inverse relationship between drugexposure levels and LPA levels in plasma (Fig. 1). LPA levelsrebounded as the compound got cleared from the system. The30-mg/kg dose was chosen for testing in efficacy studies as thisdose resulted in a $50% inhibition of plasma LPA levels over a12-hour period andwould facilitate a twice daily dosing regimen.Anti-Inflammatory Activity of Compound-1. Effect of

compound-1 was investigated in the CGN-induced paw in-flammation model to determine its anti-inflammatory activ-ity. Intraplantar injection of CGN in the right hind paw ofmiceinduced a progressive increase in paw inflammation (mea-sured as paw thickness) in a time-dependent manner (up to6 hours post-CGN) in the vehicle-treated group. Treatmentwith compound-1 (30 mg/kg) produced a significant inhibitionof the inflammatory response compared with vehicle-treatedanimals. As shown in Fig. 2, the inhibitory response persistedfor the entire duration of the study, depicting an anti-inflammatory effect of the compound at all time points post-treatment.

Further studies were conducted to gain a mechanisticunderstating of the anti-inflammatory activity of compound-1by investigating its effect on the production of the inflammatorycytokine TNFa in mice treated with LPS. Administration of asingle dose of compound-1 (30 mg/kg) resulted in a significantinhibition of LPS-induced TNFa release in plasma (Fig. 3A).Concurrent measurement of plasma LPA from the samesamples showed a reduction in LPA (Fig. 3B), suggesting acorrelation between inhibition of TNFa release and inhibitionof plasma LPA.Efficacy of Compound-1 in the DSS-Induced IBD

Model. The DSS model is one of the most commonly usedmodels of IBD, wherein administration of DSS in the drinkingwater leads to lethal inflammatory bowel disease resemblinghuman IBD. To test the effect of autotaxin inhibition ondisease incidence and severity in the DSS model, compound-1was dosed orally (30 mg/kg, twice daily for 6 days) startingfrom the initiation of DSS administration. There was a time-dependent increase in diarrhea and rectal bleeding thatcontributed to an increase in disease activity score invehicle-treated animals (Fig. 4, A–C). Colon-length shorteningwas also observed at the end of the study (on day 6) in thevehicle-treated animals (Fig. 4D). However, treatment withcompound-1 resulted in a significant decrease in all measuredparameters of disease activity, and also led to a partialreversal of Colon-length shortening (Fig. 4, A–D).We then tested whether the inhibitory effects of the

autotaxin inhibitor on IBD progression and severity is dose-dependent. So, we tested the inhibitor at 3-, 10-, and 30-mg/kgdoses and monitored the effects. Treatment with compound-1resulted in a dose-dependent inhibition of the disease activityscores (Fig. 5A). The 10- and 30-mg/kg doses of the compoundresulted in significant inhibition compared with the vehicle-treated group, and the highest inhibition was observed at the30 mg/kg dose. Colon-length shortening was also partiallyreversed in the compound-treated groups in a dose-dependentfashion (Fig. 5B).Analgesic Activity of Compound-1 in Acetic

Acid–Induced Visceral Pain. Since the autotaxin/LPApathway has been shown to play a role in mediating in-flammatory and neuropathic pain (Inoue et al., 2004, 2008b;Ueda, 2011, 2013), we tested the autotaxin inhibitor in theacetic acid–induced writhing model of acute visceral painin mice. Intraperitoneal injection of acetic acid producedwrithing responses in the vehicle-treated group, whereas

Fig. 2. Inhibition of carrageenan-induced paw inflammation in mice byautotaxin inhibitor compound-1. Compound-1 was dosed orally at30 mg/kg, as indicated in Materials and Methods, followed by injectionof carrageenan into the paw. Paw thickness was measured at baseline andat 2, 4, and 6 hours post–carrageenan injection. Statistical comparisonswere made to the vehicle group using an unpaired t test (*P, 0.05; **P,0.01; ***P , 0.001 versus vehicle; n = 10 animals/group).

Fig. 3. Effect of compound-1 in inhibiting LPS-induced TNFa release (A) and LPA levels in plasma(B). Immediately after oral dosing of vehicle (n = 4) orcompound-1 (30 mg/kg, n = 5), LPS was injectedintraperitoneally (10 mg/kg). Two hours later, plasmawas collected and TNFa levels were measured byELISA, and LPA levels were measured by massspectrometry.

210 Thirunavukkarasu et al.

at ASPE

T Journals on O

ctober 24, 2020jpet.aspetjournals.org

Dow

nloaded from

pretreatment of mice with compound-1 resulted in a dose-dependent reduction in the number of writhing responses at 3,10, or 30 mg/kg (Fig. 6). The inhibitory effects observed weresignificantly different from the vehicle group at the 10- and30-mg/kg doses of compound-1.Efficacy of Compound-1 in the MOG-EAE Model.

EAE induced by myelin oligomeric glycoprotein (MOG35–55

peptide) is an animal model for the human inflamma-tory demyelinating disease, multiple sclerosis. EAE is acomplex condition in which both immunopathological and

neuropathological mechanisms lead to inflammation, demye-lination, axonal loss, and gliosis as the key pathologic featuresof MS. To query the potential role of autotaxin in thepathogenesis of MS, we tested compound-1 in the MOG-induced EAE model in mice. EAE was induced via theinjection of MOG peptide and pertussis toxin, as described inMaterials andMethods, and the animalswere dosed daily withvehicle, prednisolone, or compound-1. Prednisolone, a gluco-corticoid, was used as a positive control in this study. Clinicalscoring was done daily starting from day 8. As shown in Fig. 7,

Fig. 4. Effect of autotaxin inhibitor compound-1 (30 mg/kg, dosed orally, twice daily for 6 days) in the DSS-induced IBD model. Effects on diarrhea (A),rectal bleeding (B), DAI score (C), and colon length (D) are shown. Data from individual animals, as well as the group mean values (horizontal bars), areshown. Statistical comparisons were made to the vehicle group using an unpaired t test (*P, 0.05; **P, 0.01; ***P, 0.001 versus vehicle; n = 8/group).Five animals that were given regular water (without DSS) were used as sham controls.

Fig. 5. Dose-dependent effect of autotaxininhibitor compound-1 (3, 10, and 30 mg/kg,dosed orally, twice daily) in the DSS-induced IBDmodel. Effects on the DAI score(A) and colon length (B) are shown. Datafrom individual animals, as well as thegroup mean values (horizontal bars), areshown. Statistical comparisons were madeto the vehicle group using an unpaired t test(*P , 0.05; **P , 0.01; ***P , 0.001 versusvehicle; n = 10/group). Four animals givenregular water (without DSS) were used assham controls.

Autotaxin Inhibitor for IBD and MS 211

at ASPE

T Journals on O

ctober 24, 2020jpet.aspetjournals.org

Dow

nloaded from

therewas a time-dependent increase in the disease activity scorestarting from day 9 in the vehicle-treated animals. However,treatment with compound-1 resulted in a slight rightward shiftin the disease progression curve with a significant decrease indisease score observed starting from day 20 onwards andmaintained until the end of the study. Prednisolone treatment,as reported before (Tsutsui et al., 2008), led to a completeinhibition of disease development in the animals.

DiscussionWe have characterized the pharmacokinetics and pharma-

codynamics of a potent inhibitor of autotaxin (compound-1) inmice and have demonstrated its anti-inflammatory activity inresponse to carrageenan and LPS administration. Consistentwith the expression of autotaxin in the inflamed intestinalmucosa of IBD patients and the involvement of the autotaxin-LPA pathway in pathologic lymphocyte migration in animalmodels of colitis (Hozumi et al., 2013), we have demonstrated asignificant dose-dependent inhibition of the disease activityscore by compound-1in the DSS-induced model of IBD. Theinhibitor also resulted in a partial reversal of colon-lengthshortening observed in the model. The fact that the inhibitordid not lead to complete amelioration of the disease isconsistent with the involvement of additional mechanismsthat play a role in disease pathogenesis. In this short-termmodel of colitis, the inhibitor was dosed in the preventionmode (with dosing initiated when DSS administration indrinking water was started). In addition, although the DSSmodel displays some similarities to human IBD, it is primarilya model of acute inflammation and epithelial damage. Hence,it would be interesting to test the autotaxin inhibitor in a long-term model of colitis. In such a model, initiation of autotaxininhibitor treatment either before or after establishment of thedisease would help distinguish the role of autotaxin/LPA inthe initiation versus progression of the disease.Autotaxin is a secreted protein and is the primary enzyme

responsible for the generation of LPA in plasma (vanMeeteren et al., 2006). It is not known whether circulatingLPA levels are increased in humans with IBD. It is probablethat the increased levels of autotaxin in the inflamed mucosain IBD patients would cause an increase in local production of

LPA that would play a role in disease pathogenesis. PlasmaLPA levels could, however, serve as an easily accessiblebiomarker of drug activity and may reflect tissue autotaxininhibition.Pain is a common symptom that significantly decreases the

quality of life in about 50–70% of patients experiencing theinitial onset or exacerbations of IBD. Pain is usually caused byabdominal cramps, owing to irritation of the nerves byinflammatory cytokines and mediators sensitizing the pri-mary afferents andmuscles controlling intestinal contractions(Bielefeldt et al., 2009; Docherty et al., 2011; Glocker andGrimbacher, 2012). Pain may be the only symptom of diseaseactivity in some patients, and about one-sixth of IBD patientsare chronically treated with opioids (Edwards et al., 2001;Cross et al., 2005; Lix et al., 2008; Bielefeldt et al., 2009).Despite advances in medical management and surgical treat-ment of IBD, pain management is still a challenge (Lix et al.,2008). Traditional analgesic agents such asNSAIDs or COX-2-selective inhibitors have been linked to a higher likelihood ofGI bleeding and disease exacerbation in IBD patients (Matuket al., 2004; Guslandi, 2006; Takeuchi et al., 2006; Bielefeldtet al., 2009), and systemic treatment with steroids areassociated with severe side effects such as changes in metab-olism, blood pressure, and even mood. Consistent with theinvolvement of the autotaxin/LPA pathway in mediatinginflammatory/neuropathic pain (Inoue et al., 2004, 2008a,b;Gierse et al., 2010; Thirunavukkarasu et al., 2016, underrevision, Osteoarthritis Cartilage), we have shown that auto-taxin inhibitor compound-1 can dose dependently inhibitacute visceral pain in the acetic acid–induced writhing model.Our data suggest that the autotaxin inhibitor may provide asafer alternative option to treat pain associated with IBD, inaddition to its potential disease-modifying benefits.The presence of LPA and autotaxin in the nervous system

and CSF, the effects of LPA in inducing inflammation anddemyelination, and the observed increase in autotaxin levelsin the CSF of MS patients prompted us to test the effect of theautotaxin inhibitor in the MOG-induced EAE model of MS inmice. Using a highly potent and orally bioavailable inhibitor of

Fig. 6. Effect of autotaxin inhibitor compound-1, dosed orally at 3, 10, or30 mg/kg, in the acetic acid–induced writhing model. Data from individualanimals, as well as the group mean values (horizontal bars), are shown.Statistical comparisons were made to the vehicle group using an unpairedt test (*P , 0.05; **P , 0.01 versus vehicle; n = 10/group).

Fig. 7. Effect of compound-1 in inhibiting disease activity in the mouseMOG-EAE model. Vehicle, compound-1, or prednisolone was dosed orallystarting from day 1 and the clinical disease scores were assessed dailystarting from day 8. Treatment with compound-1 resulted in a significantinhibition of the clinical scores compared with vehicle on days 20–31 (n =10 animals/group).

212 Thirunavukkarasu et al.

at ASPE

T Journals on O

ctober 24, 2020jpet.aspetjournals.org

Dow

nloaded from

autotaxin, we have demonstrated an inhibition of the EAEdisease activity score that takes into account behavioralobservations on the mobility of the animals. We also observeda trend toward a decrease in inflammation scores in the spinalcord sections and a possible decrease in demyelination in thelumbar region of the spinal cord in animals treated withcompound-1 (data not shown). It is probable that autotaxinand LPA are potentially involved in the initiation of thedisease process, at least in part, via the effects of LPA ininducing inflammation and/or demyelination. Diagnosis ofMS typically occurs between the ages of 20 and 40 with afemale/male ratio of 2:1 (Compston and Coles, 2002). At thetime of diagnosis, 85% of the patients have the relapsing-remitting form of MS (RRMS) that is characterized byrepeating cycles of acute exacerbations (relapses) of neuro-logic dysfunction that is followed by recovery. It is conceiv-able that the autotaxin inhibitors may be efficacious in therelapsing-remitting form of MS if dosing is initiated prior tothe onset of disease relapse. To confirm this possibility,testing of the compound in an animal model that displays therelapsing-remitting form of disease, such as the RR-EAEmodel, would be required.There is a significant inflammatory component in the EAE

model of MS that eventually leads to demyelination. Sinceautotaxin and LPA are known to be involved in the inflam-matory cell infiltration and demyelination, it is probable thatboth mechanisms are at play in the efficacy seen with theautotaxin inhibitors. Considering the role of the autotax-in/LPA pathway in mediating inflammatory and neuropathicpain, it is tempting to speculate that inhibition of autotaxinmay also have an analgesic benefit in MS, in addition to thepotential disease-modifying effects.Mice that lack both copies of the autotaxin gene are not

viable and have defects in neural tube closure and neuriteoutgrowth, probably attributable to a need for autotaxinfunction during development. However, the heterozygote miceare viable and show a 50% decrease in autotaxin activity andplasma LPA levels (van Meeteren et al., 2006). In addition,animals have recently been generated wherein autotaxin canbe knocked out after birth, and these animals also have anormal phenotype without any obvious abnormalities (Katsifaet al., 2015). Subjecting the autotaxin heterozygotes and theadult-specific autotaxin knockout mice to models of IBD andEAE would help further confirm the role of the autotaxin/LPApathway in the pathogenesis of these diseases and determinewhether a decrease or lack of autotaxin activity would resultin a delay in disease onset or progression and/or a decrease inseverity of the disease.In summary, autotaxin inhibition (either alone or in

combination with drugs targeting other mechanisms) mayprovide an oral therapeutic option for IBD and MS that wouldprobably lead to a potential analgesic benefit as well.

Authorship Contributions

Participated in research design: Thirunavukkarasu, Swearingen,McCann, Norman, Saha.

Conducted experiments: Tan, Swearingen, Bui, Saha.Contributed new reagents or analytic tools: Jones, Pfeifer, Norman.Performed data analysis: Thirunavukkarasu, Tan, Swearingen,

Rocha, McCann, Saha.Wrote or contributed to the writing of themanuscript: Thirunavukkarasu,

Rocha, Jones, Saha.

References

Balood M, Zahednasab H, Siroos B, Mesbah-Namin SA, Torbati S, and HarirchianMH (2014) Elevated serum levels of lysophosphatidic acid in patients with multiplesclerosis. Hum Immunol 75:411–413.

Beauchamp TJ, Dao Y, Jones SB, Norman BH, and Pfeifer LA (2015) inventors, EliLilly and Company, assignee. Bicyclic pyrimidine compounds. U.S. patent8,969,555. 2015 Mar 3.

Bielefeldt K, Davis B, and Binion DG (2009) Pain and inflammatory bowel disease.Inflamm Bowel Dis 15:778–788.

Castagna D, Budd DC, Macdonald SJ, Jamieson C, and Watson AJ (2016) Develop-ment of Autotaxin Inhibitors: An Overview of the Patent and Primary Literature. JMed Chem 59:5604–5621.

Chen YQ, Kuo MS, Li S, Bui HH, Peake DA, Sanders PE, Thibodeaux SJ, Chu S,Qian YW, Zhao Y, et al. (2008) AGPAT6 is a novel microsomal glycerol-3-phosphateacyltransferase. J Biol Chem 283:10048–10057.

Chu X, Wei X, Lu S, and He P (2015) Autotaxin-LPA receptor axis in the patho-genesis of lung diseases. Int J Clin Exp Med 8:17117–17122.

Chun J, Hla T, Lynch KR, Spiegel S, and Moolenaar WH (2010) International Unionof Basic and Clinical Pharmacology. LXXVIII. Lysophospholipid receptor nomen-clature. Pharmacol Rev 62:579–587.

Compston A and Coles A (2002) Multiple sclerosis. Lancet 359:1221–1231.Cross RK, Wilson KT, and Binion DG (2005) Narcotic use in patients with Crohn’sdisease. Am J Gastroenterol 100:2225–2229.

Docherty MJ, Jones RC, 3rd, and Wallace MS (2011) Managing pain in inflammatorybowel disease. Gastroenterol Hepatol (N Y) 7:592–601.

Edwards JT, Radford-Smith GL, and Florin TH (2001) Chronic narcotic use in in-flammatory bowel disease patients: prevalence and clinical characteristics. JGastroenterol Hepatol 16:1235–1238.

Eijkelkamp N, Kavelaars A, Elsenbruch S, Schedlowski M, Holtmann G, and HeijnenCJ (2007) Increased visceral sensitivity to capsaicin after DSS-induced colitis inmice: spinal cord c-Fos expression and behavior. Am J Physiol Gastrointest LiverPhysiol 293:G749–G757.

Floris S, Ruuls SR, Wierinckx A, van der Pol SM, Döpp E, van der Meide PH, DijkstraCD, and De Vries HE (2002) Interferon-beta directly influences monocyte in-filtration into the central nervous system. J Neuroimmunol 127:69–79.

Fujita R, Kiguchi N, and Ueda H (2007) LPA-mediated demyelination in ex vivoculture of dorsal root. Neurochem Int 50:351–355.

Garcia-Diaz B, Hurtado-Guerrero I, Gómez-Conde AI, Castilla-Ortega E, LeyvaL, Fernández O, Chun C, Rodríguez De Fonseca F, and Estivill-Torrús G (2010)Effects of lysophosphatidic acid LPA1 receptor absence in the onset and clin-ical course of experimental autoimmune encephalomyelitis. FENS Abstr 5:050.048.

Gierse J, Thorarensen A, Beltey K, Bradshaw-Pierce E, Cortes-Burgos L, Hall T,Johnston A, Murphy M, Nemirovskiy O, Ogawa S, et al. (2010) A novel autotaxininhibitor reduces lysophosphatidic acid levels in plasma and the site of in-flammation. J Pharmacol Exp Ther 334:310–317.

Glocker E and Grimbacher B (2012) Inflammatory bowel disease: is it a primaryimmunodeficiency? Cell Mol Life Sci 69:41–48.

Guslandi M (2006) Exacerbation of inflammatory bowel disease by nonsteroidal anti-inflammatory drugs and cyclooxygenase-2 inhibitors: fact or fiction? World JGastroenterol 12:1509–1510.

Hammack BN, Fung KY, Hunsucker SW, Duncan MW, Burgoon MP, Owens GP,and Gilden DH (2004) Proteomic analysis of multiple sclerosis cerebrospinal fluid.Mult Scler 10:245–260.

Hozumi H, Hokari R, Kurihara C, Narimatsu K, Sato H, Sato S, Ueda T,Higashiyama M, Okada Y, Watanabe C, et al. (2013) Involvement of autotaxin/lysophospholipase D expression in intestinal vessels in aggravation of intestinaldamage through lymphocyte migration. Lab Invest 93:508–519.

Inoue M, Ma L, Aoki J, Chun J, and Ueda H (2008a) Autotaxin, a synthetic enzyme oflysophosphatidic acid (LPA), mediates the induction of nerve-injured neuropathicpain. Mol Pain 4:6.

Inoue M, Rashid MH, Fujita R, Contos JJ, Chun J, and Ueda H (2004) Initiation ofneuropathic pain requires lysophosphatidic acid receptor signaling. Nat Med 10:712–718.

Inoue M, Xie W, Matsushita Y, Chun J, Aoki J, and Ueda H (2008b) Lysophospha-tidylcholine induces neuropathic pain through an action of autotaxin to generatelysophosphatidic acid. Neuroscience 152:296–298.

Jones SB, Pfeifer LA, Bleisch TJ, Beauchamp TJ, Durbin JD, Klimkowski VJ,Hughes NE, Rito CJ, Dao Y, Gruber JM, et al. (2016) Novel autotaxin inhibitorsfor the treatment of osteoarthritis pain: Lead optimization via structure-baseddrug design. ACS Med Chem Lett DOI: 10.1021/acsmedchemlett.6b00207(epub).

Kanda H, Newton R, Klein R, Morita Y, Gunn MD, and Rosen SD (2008) Autotaxin,an ectoenzyme that produces lysophosphatidic acid, promotes the entryof lymphocytes into secondary lymphoid organs. Nat Immunol 9:415–423.

Katsifa A, Kaffe E, Nikolaidou-Katsaridou N, Economides AN, Newbigging S,McKerlie C, and Aidinis V (2015) The bulk of autotaxin activity is dispensable foradult mouse life. PLoS One 10:e0143083.

Lin S, Wang D, Iyer S, Ghaleb AM, Shim H, Yang VW, Chun J, and Yun CC (2009)The absence of LPA2 attenuates tumor formation in an experimental model ofcolitis-associated cancer. Gastroenterology 136:1711–1720.

Lix LM, Graff LA, Walker JR, Clara I, Rawsthorne P, Rogala L, Miller N,Ediger J, Pretorius T, and Bernstein CN (2008) Longitudinal study of qualityof life and psychological functioning for active, fluctuating, and inactivedisease patterns in inflammatory bowel disease. Inflamm Bowel Dis 14:1575–1584.

Matuk R, Crawford J, Abreu MT, Targan SR, Vasiliauskas EA, and Papadakis KA(2004) The spectrum of gastrointestinal toxicity and effect on disease activity ofselective cyclooxygenase-2 inhibitors in patients with inflammatory bowel disease.Inflamm Bowel Dis 10:352–356.

Autotaxin Inhibitor for IBD and MS 213

at ASPE

T Journals on O

ctober 24, 2020jpet.aspetjournals.org

Dow

nloaded from

Mills GB and Moolenaar WH (2003) The emerging role of lysophosphatidic acid incancer. Nat Rev Cancer 3:582–591.

Moolenaar WH and Perrakis A (2011) Insights into autotaxin: how to produce andpresent a lipid mediator. Nat Rev Mol Cell Biol 12:674–679.

Mozdiak E, O’Malley J, and Arasaradnam R (2015) Inflammatory bowel disease.BMJ 351:h4416.

Murthy SN, Cooper HS, Shim H, Shah RS, Ibrahim SA, and Sedergran DJ (1993)Treatment of dextran sulfate sodium-induced murine colitis by intracolonic cyclo-sporin. Dig Dis Sci 38:1722–1734.

Nagai J, Uchida H, Matsushita Y, Yano R, Ueda M, Niwa M, Aoki J, Chun J,and Ueda H (2010) Autotaxin and lysophosphatidic acid1 receptor-mediated de-myelination of dorsal root fibers by sciatic nerve injury and intrathecal lysophos-phatidylcholine. Mol Pain 6:78.

Takeuchi K, Smale S, Premchand P, Maiden L, Sherwood R, Thjodleifsson B,Bjornsson E, and Bjarnason I (2006) Prevalence and mechanism of nonsteroidalanti-inflammatory drug-induced clinical relapse in patients with inflammatorybowel disease. Clin Gastroenterol Hepatol 4:196–202.

Tsutsui S, Vergote D, Shariat N, Warren K, Ferguson SS, and Power C (2008) Glu-cocorticoids regulate innate immunity in a model of multiple sclerosis: reciprocalinteractions between the A1 adenosine receptor and beta-arrestin-1 in monocytoidcells. FASEB J 22:786–796.

Ueda H (2011) Lysophosphatidic acid as the initiator of neuropathic pain. Biol PharmBull 34:1154–1158.

Ueda H, Matsunaga H, Olaposi OI, and Nagai J (2013) Lysophosphatidic acid:chemical signature of neuropathic pain. Biochim Biophys Acta 1831:61–73.

Umezu-Goto M, Kishi Y, Taira A, Hama K, Dohmae N, Takio K, Yamori T, Mills GB,Inoue K, Aoki J, et al. (2002) Autotaxin has lysophospholipase D activity leading totumor cell growth and motility by lysophosphatidic acid production. J Cell Biol158:227–233.

van Meeteren LA, Ruurs P, Stortelers C, Bouwman P, van Rooijen MA, Pradère JP,Pettit TR, Wakelam MJ, Saulnier-Blache JS, Mummery CL, et al. (2006) Auto-taxin, a secreted lysophospholipase D, is essential for blood vessel formation duringdevelopment. Mol Cell Biol 26:5015–5022.

Zahednasab H, Balood M, Harirchian MH, Mesbah-Namin SA, Rahimian N,and Siroos B (2014) Increased autotaxin activity in multiple sclerosis. J Neuro-immunol 273:120–123.

Address correspondence to: Dr. Kannan Thirunavukkarasu, Drop Code0403, Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, IN46285. E-mail: kannan@lilly.com

214 Thirunavukkarasu et al.

at ASPE

T Journals on O

ctober 24, 2020jpet.aspetjournals.org

Dow

nloaded from