Autotaxin activity increases locally following lung injury, but is not required for pulmonary lysophosphatidic acid production or fibrosis

Abstract

Lysophosphatidic acid (LPA) is an important mediator of pulmonary fibrosis. In blood and multiple tumor types, autotaxin produces LPA from lysophosphatidylcholine (LPC) via lysophospholipase D activity, but alternative enzymatic pathways also exist for LPA production. We examined the role of autotaxin (ATX) in pulmonary LPA production during fibrogenesis in a bleomycin mouse model. We found that bleomycin injury increases the bronchoalveolar lavage (BAL) fluid levels of ATX protein 17-fold. However, the LPA and LPC species that increase in BAL of bleomycin-injured mice were discordant, inconsistent with a substrate-product relationship between LPC and LPA in pulmonary fibrosis. LPA species with longer chain polyunsaturated acyl groups predominated in BAL fluid after bleomycin injury, with 22:5 and 22:6 species accounting for 55 and 16% of the total, whereas the predominant BAL LPC species contained shorter chain, saturated acyl groups, with 16:0 and 18:0 species accounting for 56 and 14% of the total. Further, administration of the potent ATX inhibitor PAT-048 to bleomycin-challenged mice markedly decreased ATX activity systemically and in the lung, without effect on pulmonary LPA or fibrosis. Therefore, alternative ATX-independent pathways are likely responsible for local generation of LPA in the injured lung. These pathways will require identification to therapeutically target LPA production in pulmonary fibrosis.—Black, K. E., Berdyshev, E., Bain, G., Castelino, F. V., Shea, B. S., Probst, C. K., Fontaine, B. A., Bronova, I., Goulet, L., Lagares, D., Ahluwalia, N., Knipe, R. S., Natarajan, V., Tager, A. M. Autotaxin activity increases locally following lung injury, but is not required for pulmonary lysophosphatidic acid production or fibrosis.

Idiopathic pulmonary fibrosis (IPF) and other fibrotic lung diseases cause significant morbidity and mortality, rivaling that of many cancers (1, 2). Despite advances in understanding IPF pathogenesis and recent progress made in developing therapies able to slow its progression (3, 4), there is an urgent, unmet need for curative treatment. IPF is believed to result from nonresolving alveolar epithelial damage coupled with aberrant or overexuberant fibroblast wound-healing responses, including excessive fibroblast accumulation, myofibroblast differentiation, and deposition of extracellular matrix (5, 6). Recent identifications of mediators of these aberrant repair processes, such as lysophosphatidic acid (LPA), have identified potential targets for effective new IPF therapies (7).

LPA species consist of glycerophosphate backbones esterified with a single fatty acid at either the sn-1 or -2 position. Individual LPA molecular species are distinguished by their fatty acid carbon chain length and degree of unsaturation. For example, 16:0 LPA describes a species with 16 carbons and no double bonds in its fatty acid side chain, while 22:5 LPA has 22 carbons and 5 double bonds in its fatty acid (Fig. 1). All LPA species act via specific G protein-coupled receptors. Six are currently recognized and designated LPA_1–6 (8). These receptors may have different affinities for different LPA species; LPA₃ and LPA₆, for example, appear to have increased affinities for species with polyunsaturated fatty acids at the sn-2 position (9). By signaling through different LPA receptors, various LPA molecular species may exert distinct biologic effects.

Figure 1.

Pathways for LPA generation. In the ATX-dependent pathways shown on the left, PC is hydrolyzed by PA, either by a PLA₁, which typically removes a saturated sn-1 fatty acid to generate an unsaturated LPC species (e.g., 22:5 LPC; top), or by a PLA₂, which typically removes an unsaturated sn-2 fatty acid to generate a saturated LPC (16:0 LPC, for example, bottom). The LPC headgroup (e.g., choline), is then removed by the lysoPLD activity of ATX, generating either unsaturated or saturated LPA (22:5 or 16:0 LPA, respectively; center). Alternatively, in the ATX independent pathways (right), PA is generated from PC through PLD. PA is then directly converted to LPA by PLA₁, again generating an unsaturated LPA, or by PLA₂, typically generating a saturated LPA.

LPA signaling through the LPA₁ receptor appears to direct many of the aberrant wound-healing processes that drive pulmonary fibrosis. In previous work, we have generated evidence that mice genetically deficient in LPA₁ (10) or LPA₂ (11) are significantly protected from pulmonary fibrosis in the bleomycin mouse model. LPA levels are increased in BAL fluid recovered from mice after bleomycin injury, and LPA-LPA₁ signaling is necessary for alveolar epithelial cell apoptosis (12), pulmonary vascular leak (10), and lung fibroblast recruitment in this model (10). We found evidence that LPA-LPA₂ signaling drives TGF-β expression and myofibroblast differentiation and also contributes to lung epithelial cell apoptosis and vascular leak in the bleomycin model (11). In humans, LPA levels are increased in the BAL fluid of patients with IPF, compared with control subjects (10), and LPA is responsible for most of the fibroblast chemoattractant activity that was observed in the BAL fluid recovered from these patients (10, 13).

Although LPA is generated through de novo biosynthesis as an intermediate in phospholipid synthesis, the LPA species that signal through LPA_1–6 appear to be formed through the degradation of more complex phospholipids (14). LPA can be formed from complex phospholipids by 2 major pathways (15–17). In one pathway (shown in Fig. 1, left), a fatty acid from a phospholipid, such as phosphatidylcholine (PC), is cleaved either by a phospholipase A₁ (PLA₁), which removes the (typically saturated) fatty acid from the sn-1 position (18), or by a PLA₂, which removes the (typically unsaturated) fatty acid from the sn-2 position. This cleavage step generates a lysophospholipid, usually lysophosphatidylcholine (LPC) (as shown in Fig. 1), lysophosphatidylethanolamine (LPE), or lysophosphatidylserine (LPS). Cleavage of the lysophospholipid head group by a lysophospholipase D (lysoPLD) then generates LPA (15). In 2002, the lysoPLD catalyzing LPA generation in this pathway was found to be autotaxin (ATX), also known as ecto-nucleotide pyrophosphatase/phosphodiesterase 2 (ENPP2), initially isolated as a promotility factor from melanoma cells (19–21).

The other major pathway of LPA formation from complex phospholipids is ATX-independent (Fig. 1, right). In this pathway, a phospholipase D (PLD1 or PLD2) first hydrolyzes the distal phosphodiester bond of a phospholipid such as PC, phosphatidylethanolamine, phosphatidylserine, or phosphatidylinositol, generating phosphatidic acid (PA). Then, the (typically saturated) fatty acid from the sn-1 position of PA is removed by a PLA₁ (18), or the (typically unsaturated) fatty acid from the sn-2 position is removed by a PLA₂, in either case producing LPA.

Of these 2 enzymatic pathways, the ATX-dependent pathway is thought to be responsible for most of the LPA present in the circulation (9, 16, 22–26). Several prior investigations have also suggested a role for ATX in fibrotic diseases. In patients with hepatic fibrosis secondary to hepatitis C, serum ATX levels correlate with the severity of fibrosis (27), and ATX has been shown to be up-regulated in synovial fibroblasts in patients with rheumatoid arthritis (28). We have recently found that ATX expression is increased in the fibrotic skin of patients with scleroderma and that pharmacologic inhibition of ATX with a small molecule inhibitor, PAT-048 [3-[6-chloro-7-fluoro-2-methyl-1-(1-propyl)-1H-pyrazol-4-yl-1H-indol-3-ylsulfanyl]-2-fluoro-benzoic sodium salt], markedly attenuates dermal fibrosis in the bleomycin mouse model of dermal fibrosis (unpublished data). In the current study, we investigated the role of ATX in the production of LPA, and of lung fibrosis, in the bleomycin mouse model. We found unexpectedly that this enzyme is not necessary for LPA production or the development of pulmonary fibrosis in this model, despite ATX protein and activity levels both being significantly increased after bleomycin lung injury.

MATERIALS AND METHODS

Animal experiments

All animal protocols were approved by the Massachusetts General Hospital Subcommittee on Research Animal Care, and all mice were maintained in a specific pathogen-free environment certified by the American Association for Accreditation of Laboratory Animal Care. Adult 7–12-wk-old male C57Bl/6 mice obtained from the National Cancer Institute were used in all experiments. Mice were administered a single intratracheal dose of bleomycin at 1.0 or 1.2 U/kg in a total volume of 50 μl; control mice were injected intratracheally with 50 μl sterile saline. For FTY720 experiments, mice were injected intraperitoneally with FTY720 (1 mg/kg; Cayman Chemical, Ann Arbor, MI, USA) or vehicle on d −3, 0, and 3 after bleomycin injection; these mice received only 0.1 U/kg bleomycin. For ATX-inhibitor experiments, PAT-048 (kind gift of PharmAkea, San Diego, CA, USA) was suspended in 0.5% methylcellulose (Sigma-Aldrich, St. Louis, MO, USA) and given by oral gavage to mice at a dose of 20 mg/kg once daily in a volume of 200 μl. Gavages were begun the day of bleomycin challenge and continued through the day before euthanasia. At death, mouse lungs were lavaged with six successive 0.5 ml aliquots of PBS. Recovered BAL fluid was centrifuged at 540 g at 4°C for 5 min, and supernatants were transferred into siliconized, low-binding microcentrifuge tubes (Fisher Scientific, Pittsburgh, PA, USA) and stored at −80°C for subsequent analysis. BAL cell pellets were resuspended in RLT buffer (Qiagen, Valencia, CA, USA) for RNA analysis, if warranted. Whole blood was collected via cardiac puncture with a heparinized syringe. Blood was placed on ice and spun at 1900 g for 15 min at 4°C. Plasma was collected and stored at −80°C until further use. Lungs were flushed with PBS via the right ventricle and harvested, flash frozen on dry ice, and stored at −80°C.

LPA and LPC mass spectrometry analyses of BAL and plasma

These analyses were performed by electrospray ionization-liquid chromatography-tandem mass spectrometry (ESI-LC/MS/MS) (29, 30). In brief, ESI-LC/MS/MS was performed with a 6500 QTRAP (AB Sciex, Framingham, MA, USA) hybrid triple quadrupole/ion trap mass spectrometer coupled with a 1290 liquid chromatography system (Agilent, Santa Clara, CA, USA). Lipids were separated on an Ascentis Express C8 (75 × 2.1 mm, 2.7 μm; Sigma-Aldrich) column with methanol:water:HCOOH, 60:40:0.5 v/v, with 5 mM NH₄COOH as solvent A and acetonitrile:chloroform:water:HCOOH, 80:20:0.5:0.5 v/v, with 5 mM NH₄COOH as solvent B. LPA molecular species were analyzed in negative ionization mode with declustering potential and collision energy optimized for each LPA molecular species. Individual saturated and unsaturated LPA molecular species (16:0, 17:0, 18:0, 18:1, 20:4, and 22:5 LPA, all obtained from Avanti Polar Lipids, Inc. (Alabaster, AL, USA) were used as reference compounds. 17:0 LPA was used as the internal standard, and LPA quantitation was performed by creating standard curves with variable amounts of each available LPA molecular species vs. a fixed amount of the internal standard. Total lipid extract from fetal bovine serum was used as a source of otherwise unavailable LPA molecular species to determine their chromatographic behavior and parameters of ionization and collision-induced decomposition, and the quantitation of these LPA molecular species was achieved via the use of the best possible approximation from the standard curves obtained with available individual LPA standards. The identification of LPA molecular species was achieved by monitoring for selected transitions from molecular to product (m/z 153) ions that are specific to each LPA molecular species and by the analyte retention time compared with the available LPA standards and with LPA extracted from bovine serum.

LPC levels were determined by ESI-LC/MS/MS in positive-ion mode. Lipids were separated on the same Ascentis Express C8 (75 × 2.1 mm, 2.7 μm) column using methanol:water, 60:40 v/v, as system A and methanol:water, 99.5:0.5 v/v, as system B. Individual saturated and unsaturated LPC molecular species (13:0, 16:0, 18:0, and 18:1-LPC, all obtained from Avanti Polar Lipids, Inc.) were used as reference compounds. Fetal bovine serum lipid extract was used as a reference source for the identification of polyunsaturated LPC molecular species.13:0 LPC was used as the internal standard, and LPC quantitation was performed by creating standard curves with variable amounts of each available LPA molecular species vs. a fixed amount of the internal standard. The identification of LPC molecular species was also achieved by monitoring for select transitions from molecular to product (m/z 184) ions that are specific for each LPC molecular species and by the analyte retention time compared with the available LPC standards and with LPC extracted from bovine serum.

Total protein and ATX protein measurements

Concentrations of total protein were determined by using a commercially available BCA Protein Assay Kit (Thermo Fisher Scientific Life Sciences, Rockford, IL, USA), according to the manufacturer’s instructions. ATX protein levels were determined with a commercially available ELISA kit (Echelon Biosciences, Salt Lake City, UT, USA) according to the manufacturer’s instructions.

LysoPLD activity assay

ATX lysoPLD activity was measured by the release of choline from the substrate, 14:0 LPC, modified from published methods (31). In brief, 25 µl of plasma was diluted 1:10 in PBS, and the BAL fluid was used neat. Sample (50 μl) was incubated with 50 μl lysoPLD buffer [100 mM Tris (pH 9), 500 mM NaCl, 5 mM MgCl₂, 5 mM CoCl₂, and 0.05% Triton X-100] (21, 31) containing 2 mM 1-myristoyl (14:0)–LPC (Avanti Polar Lipids, Inc.), for 4–6 h at 37°C (total volume of each reaction was 100 µl). A control sample was prepared to account for baseline choline concentration, using 50 μl BAL fluid or diluted plasma and lysoPLD buffer alone. The liberated choline was then detected by the addition of an equal volume of 4.5 mM TOOS detection reagent containing 4-aminoantioyrine, 2.7 mM N-ethyl-N-(2-hydroxy-3-sulfopropyl)-m-toluidine), 20 U ml⁻¹ horseradish peroxidase, and 3 U ml⁻¹ choline oxidase in 50 mM Tris (pH 8), and 4.5 mM MgCl₂, incubated for 10 min at room temperature. Color development was detected with a SpectraMax Plus (Thermo Fisher Scientific Life Sciences) by measuring the absorbance at a wavelength of 555 nm. The absorbance was then converted to nanomoles of choline by comparison to a standard curve generated using choline chloride. Baseline choline concentration was subtracted by using measurements from control samples, and ATX activity reported as nanomoles/milliliters choline per minute. In separate experiments to test the linearity of this assay, we found that the concentrations of lysoPLD-produced choline in these assays increased linearly over time up to 6 h after the start of the assays, covering our 4–6 h assay duration.

RNA isolation and quantitative PCR analyses

Mouse solid organs (lung, liver, kidney, small intestine, colon, and spleen) were harvested and stored in RNALater (Qiagen); RLT buffer was added to harvested BAL cells, bone marrow, and whole blood. Solid organs were then homogenized in Trizol (Thermo Fisher Scientific Life Sciences, Grand Island, NY, USA), and RNA was isolated by using chloroform and isopropanol according to the manufacturer’s instructions. BAL, blood, and bone marrow RNA was extracted using RNeasy Mini Kits (Qiagen) according to the manufacturer’s instructions. Quantitative real-time PCR analyses of RNA were performed using a Mastercycler realplex2 (Eppendorf, Hauppauge, NY, USA), with the following primers: mouse ENPP2 (ATX)-forward TCTAGCATCCCAGAGCACCT; mouse ENPP2-reverse CGTTTGAAGGCAGGGTACAT; mouse GAPDH-forward AACTTTGGCATTGTGGAAGG; mouse GAPDH-reverse GGATGCAGGGATGATGTTCT.

Assay of PAT-048 inhibition of ATX

Inhibition of plasma ATX activity by the selective ATX inhibitor PAT-048 (patent publication number WO 2012024620 A2; Amira Pharmaceuticals, San Diego, CA, USA) was evaluated by measuring the concentration of 20:4 LPA produced from the endogenous LPC in mouse plasma after incubating for 4 h at 37°C (concentrations of 18:1 and 18:0 LPA were also measured, with similar results). In brief, BALB/c blood was collected into heparinized Vacutainer tubes (BD Biosciences, San Diego, CA, USA) by cardiac puncture and centrifuged at 800 g for 10 min, to prepare the plasma. An aliquot of plasma was mixed with vehicle and precipitated immediately with ice-cold methanol to determine baseline concentrations of 20:4 LPA. For inhibition studies, 40 µl plasma was mixed with 1 µl PAT-048 or vehicle control (DMSO) and incubated at 37°C for 4 h before precipitating with 5 volumes of ice-cold methanol containing 17:0 LPA as the internal standard. The samples were incubated for 10 min on ice before centrifugation at 4000 g for 10 min at 4°C. The supernatant (150 µl) was mixed with 100 µl 90:10:0.1 water:acetonitrile:ammonium hydroxide then centrifuged at 4000 g for 10 min at 4°C. The concentration of 20:4 LPA was measured by LC-MS/MS by injecting 20 µl onto an Xbridge 2.1 × 50 mm, 5 μm, C8 column (Waters, Milford, MA, USA). Mobile phase A was 90:10:0.1 water:acetonitrile:ammonium hydroxide. Mobile phase B was 10:90:0.1 water:acetonitrile:ammonium hydroxide. Flow rate was 800 μl/min with initial conditions of 90% mobile phase A and 10% mobile phase B. Initial conditions were held for 0.5 min then mobile phase B concentration was increased linearly to 90% over 1 min to elute 20:4 LPA and its internal standard, 17:0 LPA. This condition (10% A, 90% B) was held for 0.5 min. Initial conditions were restored and held for 0.9 min to equilibrate the column. An API-4000 Q-trap mass spectrometer (AB Sciex) was operated in negative-ion multiple-reaction monitoring mode to detect 20:4 LPA and 17:0 LPA. The transitions monitored were 457.2 to 153.0 and 423.2 to 153.0 for 20:4 LPA and 17:0 LPA, respectively. Source conditions were optimized by infusing 20:4 LPA. Declustering potential was −60 and collision energy was −30.

Determination of PAT-048 pharmacokinetics and pharmacodynamics

BALB/c mice (n = 3 per time point) were administered PAT-048 by oral gavage (10 mg/kg in 0.5% methylcellulose), and blood was collected by cardiac puncture under anesthesia in EDTA Vacutainer tubes at 0.5, 1, 2, 4, 8, 16, and 24 h after dosing. Plasma samples were prepared and stored at −80°C before analysis of PAT-048 concentrations by LC-MS/MS and ATX activity by the TOOS choline release assay. Known amounts of PAT-048 were added to mouse plasma to yield a concentration range from 0.8 to 4000 ng/ml. Plasma samples were precipitated using 70:30 acetonitrile:methanol containing another small molecule ATX inhibitor with a distinct charge mass ratio as the internal standard. The supernatant (150 µl) was mixed with 100 µl water, then centrifuged at 4000 g for 10 min at 4°C. The analyte mixture (10 µl) was injected with a PAL autosampler (Leap Technologies, Carrboro, NC, USA). Calibration curves were constructed by plotting the peak–area ratio of analyzed peaks against known concentrations. The lower limit of quantitation was 20 ng/ml. The data were subjected to linear regression analysis with 1/x² weighting. ATX activity in each plasma sample was analyzed with the lysoPLD activity assay described above. Plasma from nondosed animals was used to determine maximum ATX activity.

Statistics

All statistical analyses were performed with Prism (GraphPad Software, La Jolla, CA, USA). Differences among more than 2 groups were evaluated for significance with ANOVA, followed by post hoc Bonferroni tests or by multiple Student’s t tests when specified. Differences between treatment and control groups were evaluated for significance using 2-tailed Student’s t test, with P ≤ 0.05 considered significant. Correlations between measured variables were calculated, and Spearman’s r was reported.

RESULTS

Bleomycin injury increased BAL but not plasma LPA levels

Having previously shown that total LPA concentrations increase in BAL fluid from mice after bleomycin-induced lung injury (10), we compared distributions of individual LPA species between bleomycin-injured and control mice in both BAL fluid and peripheral blood. As we saw previously, total LPA in BAL fluid was significantly higher in bleomycin-challenged mice, increasing ∼4-fold to 40.8 ± 9.9 pmol/ml in d 7 bleomycin-challenged mice from 10.3 ± 1.6 pmol/ml at d 7 after the PBS challenge (Fig. 2 and Table 1). The predominant species of LPA in BAL after either challenge was 22:5 LPA, which accounted for 55 and 66% of total BAL LPA in bleomycin- vs. PBS-challenged mice, respectively (Fig. 2A). This species of LPA was also ∼4-fold higher in bleomycin-challenged compared to PBS-challenged mice (23.7 ± 7 pmol/ml vs. 6.8 ± 1.2 mol/ml). BAL 16:0 LPA concentration was higher in bleomycin- vs. PBS-challenged mice (4.8 ± 0.8 vs. 0.7 ± 0.3 pmol/ml), although this LPA species represented only 12 and 7% of total BAL LPA in bleomycin- and PBS-challenged mice, respectively. In contrast, plasma LPA was mostly 16:0, 18:0, and 18:2 LPA species. 16:0 LPA was the predominant species in plasma after either bleomycin or PBS challenge, accounting for 55 and 59% of total LPA after these 2 challenges, respectively. Total plasma LPA concentrations, and plasma concentrations of individual LPA species, did not change after bleomycin injury (Fig. 2B and Table 1).

Figure 2.

LPA levels were elevated in the lung after intratracheal bleomycin instillation. Mice were challenged intratracheally with PBS or 1.0 U/kg bleomycin. BAL fluid and blood were collected 14 d after challenge, and plasma was prepared from the blood. A) BAL fluid LPA. Left: concentrations for each LPA species measured in the BAL fluid; right: total BAL LPA concentrations. The predominant LPA species in BAL fluid from both PBS- and bleomycin-challenged mice was 22:5 LPA. B) Plasma LPA. Left: concentrations for each LPA species measured in plasma; right: total plasma LPA concentrations. No significant differences were seen in plasma LPA levels between PBS- and bleomycin-challenged mice. C) Relative distribution of BAL fluid and plasma LPA species. Plasma and BAL fluid LPA species had markedly different distributions of their fatty acid side chains, with most of the plasma LPA being the unsaturated species 16:0 LPA, and most of BAL fluid LPA being the polyunsaturated species 22:5 and 22:6 LPA. For all panels, data are presented as mean ± sem concentration (n = 6 mice in the bleomycin-challenged group; n = 5 mice in the PBS-challenged group).

TABLE 1.

BAL and plasma concentration of individual LPA species after PBS or bleomycin challenge

LPA species	BAL: post-PBS (n = 5)				BAL: post-bleomycin (n = 6)				Plasma: post-PBS (n = 5)				Plasma: post-bleomycin (n = 6)
	Mean	sem	Min	Max	Mean	sem	Min	Max	Mean	sem	Min	Max	Mean	sem	Min	Max
14:0	0.01	0.01	0.00	0.04	0.14	0.04	0.04	0.26	1.77	0.51	0.00	3.27	0.14	0.04	0.04	0.26
16:0	0.74	0.28	0.27	1.81	4.79	0.78	2.60	7.20	215.30	8.58	190.22	242.60	4.79	0.78	2.60	7.20
16:1	0.01	0.00	0.00	0.01	0.67	0.17	0.19	1.22	1.32	0.35	0.45	2.65	0.67	0.17	0.19	1.22
18:0	1.07	0.17	0.71	1.54	2.56	0.34	1.59	3.71	61.16	2.28	55.07	69.55	2.56	0.34	1.59	3.71
18:1	0.03	0.01	0.02	0.06	0.24	0.07	0.03	0.40	9.62	0.23	8.65	10.04	0.24	0.07	0.03	0.40
18:2	0.02	0.01	0.00	0.04	0.53	0.12	0.11	0.89	42.57	3.89	33.19	56.79	0.53	0.12	0.11	0.89
20:2	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
20:3	0.01	0.00	0.00	0.03	0.02	0.01	0.00	0.04	2.32	0.22	1.79	2.97	0.02	0.01	0.00	0.04
20:4	0.27	0.16	0.00	0.78	0.31	0.05	0.17	0.49	9.87	0.88	7.76	13.32	0.31	0.05	0.17	0.49
20:5	0.06	0.01	0.02	0.09	0.62	0.13	0.28	1.05	1.95	0.20	1.18	2.48	0.62	0.13	0.28	1.05
22:4	0.21	0.03	0.15	0.34	0.70	0.22	0.00	1.57	0.29	0.14	−0.05	0.79	0.70	0.22	0.00	1.57
22:5	6.81	1.23	4.52	11.30	23.74	7.02	6.85	54.66	2.79	0.44	1.62	4.32	23.74	7.02	6.85	54.66
22:6	1.11	0.10	0.88	1.43	6.47	1.33	2.98	11.55	16.28	0.88	13.54	18.01	6.47	1.33	2.98	11.55
Total	10.35	1.63	7.52	16.54	40.78	9.88	16.13	82.28	10.35	1.63	7.52	16.54	40.78	9.88	16.13	82.28

Data are expressed in pmol/ml.

To highlight the differences between BAL fluid and plasma distributions of LPA species, Fig. 2C shows the percentages that each measured species of LPA contributed to BAL fluid or plasma total LPA levels in bleomycin-challenged mice. The plasma-predominant species 16:0, 18:0, and 18:2 LPA represented 55, 18, and 12% of total plasma LPA, but only 13, 7, and 1.4% of total BAL LPA, respectively. In contrast, the BAL-predominant species of 22:5 and 22:6 LPA represented 55 and 16.5% of total BAL LPA in bleomycin-injured mice, but only 1.6 and 5% of total plasma LPA, respectively. This discordance between the distributions of LPA species in the BAL and plasma is most consistent with increases in airspace LPA levels induced by lung injury caused by increased local LPA production in the lung, rather than increased entry of LPA from the circulation into the lung caused by injury-induced increases in pulmonary vascular permeability.

Bleomycin lung injury increases local ATX protein and activity

Given our evidence that LPA is produced locally in the lung after bleomycin injury, we compared ATX protein and activity levels in BAL fluid, lung homogenates, and plasma of unchallenged or PBS-challenged to those of bleomycin-challenged mice. Unchallenged and PBS-challenged mice were statistically indistinguishable in these analyses and were therefore analyzed together as control mice for these experiments. Compared to the controls, ATX protein levels increased in BAL fluid after bleomycin challenge at d 3, 7, and 14 (Fig. 3A) and in lung homogenates at d 7 and 14 (Fig. 3B). In contrast, plasma ATX concentrations did not increase between control and bleomycin-challenged mice at any time point assessed (Fig. 3C). Consistent with these results for ATX protein levels, ATX activity was increased in the BAL fluid after bleomycin challenge at d 3, 7, and 14 (Fig. 3D), whereas plasma ATX activity was similar in control and bleomycin-challenged mice at all time points; with a small decrease in activity levels measured at d 0 vs. d 3 (Fig. 3E).

Figure 3.

ATX protein and activity increased in the lung after bleomycin injury. Mice in these experiments were either unchallenged (intact) or challenged intratracheally with PBS or 1.0 U/ kg bleomycin. BAL fluid, plasma, and lung homogenates were collected at the indicated time points. A) BAL ATX protein concentrations increased after bleomycin injury (n = 10 mice for the intact/PBS-challenged group and the groups euthanized at d 3 or 7 after bleomycin challenge; n = 6 for group euthanized at d 14 after bleomycin challenge). B) ATX protein levels increased in lung homogenates after bleomycin injury (n = 12 mice for the intact/PBS-challenged group; n = 9, 10, and 6 for the groups of mice euthanized at 3, 7, or 14 d after bleomycin challenge, respectively. C) Plasma ATX protein concentrations did not increase in following bleomycin injury (n = 15 mice for the intact/PBS-challenged group; n = 10, 15 and 6 for the groups of mice euthanized at d 3, 7, or 14 after bleomycin challenge, respectively). D) BAL fluid ATX activity was elevated 3 d after bleomycin injury, was increased further at d 7, and remained elevated at d 14 (n = 10 mice for the intact/PBS-challenged group, and the groups of mice euthanized at d 3 or 7 after bleomycin challenge; n = 6 for the group of mice euthanized at d 14 after bleomycin challenge). E) Plasma ATX activity did not increase after bleomycin injury to the mice. There was actually a small decrease in ATX activity seen at d 3 after bleomycin challenge (n = 15 mice for the intact/PBS-challenged group; n = 10, 15 and 6 for the groups of mice euthanized at d 3, 7, and 14 after bleomycin challenge, respectively). Data were pooled from multiple experiments, with each point showing a value from an individual mouse; lines and error bars show means ± sem. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001, by 1-way ANOVA with Bonferroni correction for multiple comparisons.

Bleomycin lung injury did not increase ATX mRNA expression

To determine whether the localized increases in lung ATX protein and activity levels that we observed after bleomycin injury were related to local de novo ATX synthesis, we compared lung and BAL cell ATX mRNA expression in mice 7 d after bleomycin or PBS challenge, when lung and BAL ATX protein and activity levels were significantly elevated in bleomycin-injured mice. Lung ATX mRNA was significantly reduced in bleomycin- compared to PBS-challenged mice at this time point (Fig. 4A), similar to recently reported findings that increased LPA production leads to decreased ATX mRNA in cultured carcinoma cells (32). Far lower levels of ATX mRNA were present in BAL cells than in lung and were similar in bleomycin- and PBS-challenged mice (Fig. 4B). We surveyed ATX mRNA levels in multiple other organs, including kidney, liver, small intestine, colon, and spleen, as well as in bone marrow and whole blood cells. No organ showed a significant increase in ATX mRNA after bleomycin treatment when corrected for multiple comparisons (Fig. 4A, B). These data suggest that the increases in local ATX protein and activity levels induced by bleomycin lung injury are not a result of increased ATX transcription in any of the organs assessed.

Figure 4.

ATX mRNA expression decreased in the lung after bleomycin injury and was unchanged in other organs. Mice were challenged intratracheally with PBS or 1.2 U/kg bleomycin and harvested at d 7, the time of peak ATX protein, and activity in the BAL fluid. A) ENPP2 mRNA expression in solid organs. RNA was isolated from lung, kidney, small intestine, and spleen. There were no significant differences between bleomycin- and PBS-challenged mice for any organ except the lung (n = 10 mice per group for lung and kidney, n = 5 mice per group for all other organs). B) ENPP2 mRNA expression in leukocytes. RNA was isolated from BAL cells, peripheral blood cells, and cells flushed from the bone marrow. ENPP2 expression was minimal in bleomycin- and PBS-challenged mice and was not different between groups. In both panels, data presented are ENPP2 mRNA levels normalized to levels of GAPDH mRNA; points are values from individual mice, with lines and error bars representing means ± sem. Significance of comparisons was determined by multiple Student’s t tests with Sidak-Bonferroni correction for multiple tests.

Bleomycin-induced increases in BAL ATX correlated with lung vascular leak

Bleomycin lung injury induces an increase in pulmonary vascular permeability, which is reflected by an increase in BAL fluid total protein after challenge (33). We found close correlations between both BAL fluid ATX activity and protein levels and total protein concentrations (Fig. 5A, B). Taken together, these correlations and the absence of increased ATX transcription induced by bleomycin suggest that the increases in BAL and lung ATX protein and activity we observed after bleomycin challenge are predominantly attributable to increased amounts of ATX entering the lung interstitium and air spaces from the plasma as a result of bleomycin-induced vascular leak. To further support this hypothesis, we examined BAL total protein and ATX activity levels in mice receiving sustained treatment with the S1P receptor functional antagonist FTY720, which, as we have demonstrated, augments the vascular leak induced by bleomycin injury (33). As described, treatment with FTY720 significantly augmented the postbleomycin increase in BAL total protein (Fig. 5C). FTY720 similarly augmented bleomycin-induced increases in ATX activity in BAL fluid (Fig. 5D), with ATX activity and total protein levels again correlating closely (r = 0.96, data not shown). These results further support the hypothesis that BAL fluid ATX protein increases after bleomycin injury, primarily because of increased lung vascular permeability, which allows ATX to leak from the plasma into the alveolar space.

Figure 5.

BAL fluid levels of ATX protein and activity strongly correlated with BAL fluid total protein. A, B) BAL fluid samples from 29 PBS- and bleomycin-challenged mice (n = 8 mice for the PBS-challenged group; n = 5, 8, and 8 for the groups of mice euthanized at d 3, 7, or 14 postbleomycin challenge, respectively) were analyzed for ATX protein and activity levels (A) and total protein levels (B). Correlation between the measured values was calculated; shown are the Spearman’s r between BAL total protein and ATX activity (A) and BAL fluid total protein and ATX concentration (B). C, D) BAL fluid samples from mice 7 d after intratracheal challenge with PBS or low-dose bleomycin (0.1 U/kg), and concurrent treatment intraperitoneally with FTY720 (1 mg/kg) or vehicle, were analyzed for total protein (C) and ATX activity (D) (n = 3 mice per group). **P < 0.01, ***P < 0.001, and ****P < 0.0001 by Student's t test.

BAL LPC and LPA species after bleomycin injury were discordant

To investigate further whether increased ATX in the alveolar spaces was responsible for generating the increased 22:5 and 22:6 LPA found in BAL, we examined the BAL species distribution of LPC, ATX’s preferred substrate. Because ATX’s lysoPLD activity generates LPA from LPC by removing the choline headgroup, the resulting LPA molecule generated will have the same fatty acid as the parent LPC. For example, ATX would generate 22:5 LPA from 22:5 LPC (Fig. 1). Total BAL LPC levels were significantly higher in bleomycin- vs. PBS-challenged mice, as were all BAL levels of individual LPC species when adjusted for multiple comparisons, except 22:2 LPC (Fig. 6A). However, the distribution of LPC species in the alveolar space following bleomycin injury was quite different from that of LPA (Fig. 6B), which gives the percentages of total BAL LPA or LPC represented by each individual species of these 2 lysophospholipids. In bleomycin-challenged mice, the predominant LPC species in BAL fluid was 16:0, representing 55% of total BAL LPC, whereas 22:5 and 22:6 LPC species represented only 0.4 and 2.3% of total BAL LPC after bleomycin challenge, respectively. As described above, however, the corresponding long-chain polyunsaturated LPA species, 22:5 and 22:6 LPA, were the predominant LPA species in BAL, representing 55 and 16.5% of total BAL fluid LPA after bleomycin challenge (Figs. 2A and 6B). This discordance was not seen when comparing the LPA and LPC species that were present in the plasma (Fig. 6C). The discordance we observed between the fatty acids of the LPC and LPA species that are elevated in the alveolar space and detected in the BAL fluid after bleomycin challenge suggested to us that ATX-mediated hydrolysis of LPC is not the major mechanism through which LPA is produced in the air spaces in this model; however, we consider alternative explanations in our discussion below.

Figure 6.

BAL fluid LPC concentrations increase after bleomycin challenge, but elevated LPC species differ from elevated LPA species. Mice were challenged intratracheally with PBS or 1.0 U/kg intratracheal bleomycin; 14 d later, BAL fluid and blood were collected, and plasma was prepared from blood. A) BAL LPC concentrations. Bleomycin injury increased BAL concentrations of all individual LPC species measured except for 22:2 LPC (left graph), and increased BAL total LPC concentration (right graph). After either bleomycin or PBS challenge, the predominant BAL fluid LPC species was 16:0 LPC, which was ∼11-fold higher in bleomycin- than in PBS-challenged mice. B) Distribution of LPC and LPA species in the BAL fluid of bleomycin-challenged mice. Shown are the concentrations of each individual species as a percentage of the total concentration of that lysophospholipid. After bleomycin injury, BAL fluid LPC was primarily composed of 16:0 (55% of total BAL fluid LPC), 18:0 (14%), 18:2 (11.6%), and 18:1 LPC species (8.5%). In contrast, LPA in BAL fluid from bleomycin-injured mice was primarily composed of 22:5 (55% of total BAL fluid LPA) and 22:6 LPA species (16.5%); 16:0 LPA represented only 13% of total BAL fluid LPA. C) Distribution of LPC and LPA in plasma of bleomycin-challenged mice. Shown are the concentrations of each species as a percentage of the total concentration of that lysophospholipid. Plasma LPC in bleomycin-injured mice was primarily composed of 16:0 (33.8%), 18:2 (25%), 18:1 (14.8%), and 18:0 LPC (14.1%); similarly, plasma LPA in these mice was primarily composed of 16:0 (54.9%), 18:0 (18.1%), and 18.2 LPA (11.6%). For all panels, n = 6 mice in the bleomycin-challenged group; n = 5 in the PBS-challenged group; data presented are mean concentrations ± sem. *P ≤ 0.001.

PAT-048 inhibited bleomycin-induced BAL and plasma ATX activity

To test whether ATX activity is necessary to produce LPA in the bleomycin model of pulmonary fibrosis, we used the potent small molecule ATX inhibitor PAT-048 (Fig. 7A; Amira Pharmaceuticals) to inhibit ATX activity. PAT-048 has inhibitory concentration (IC)₅₀ and IC₉₀ values for ATX inhibition of 20 and 200 nM in mouse plasma, respectively (Fig. 7B). After a single 10 mg/kg oral dose, mouse plasma concentrations of PAT-048 peaked with a C_max of 16 µM at 30 min and decreased to a trough concentration of ∼100 nM (Fig. 7C). At this trough concentration, inhibition of plasma ATX activity remained >70% (Fig. 7D). However, to exceed the IC₉₀ at trough, we administered PAT-048 orally at 20 mg/kg once daily in the bleomycin lung fibrosis model. We have recently found that administration of PAT-048 in this dosing regimen was able to significantly decrease dermal fibrosis in a model of systemic sclerosis produced by daily subcutaneous injections of bleomycin (unpublished data). Other ATX inhibitors have recently been demonstrated to have decreased efficacy in whole blood compared to plasma, potentially because of decreased affinity of those molecules for cell-bound ATX (34). PAT-048 showed equivalent inhibition of ATX in whole blood and plasma (Fig. 7E), suggesting that ATX leukocyte binding does not interfere with PAT-048 activity.

Figure 7.

Structure, potency, and pharmacokinetics/pharmacodynamics of PAT-048. A) Chemical structure of the ATX inhibitor, PAT-048. B) Concentration–response curves of PAT-048 for inhibition of ATX activity in mouse plasma. C) Average PAT-048 plasma concentrations over time in BALB/c mice after a single oral dose of 10 mg/kg; n = 3 mice. D) Corresponding PAT-048 inhibition of plasma ATX activity from the same plasma samples shown in (C). E) PAT-048 inhibition of ATX in human whole blood compared to its inhibition of ATX in plasma.

Administration of PAT-048 in the bleomycin lung fibrosis model completely eliminated bleomycin-induced increases in BAL ATX activity levels, whereas BAL fluid ATX activity levels were so low in PBS-challenged mice, it was difficult to appreciate the effect of PAT-048 treatment (Fig. 8A). BAL fluid ATX activity levels in bleomycin-challenged mice treated with PAT-048 were reduced to the point where they did not significantly differ from the very low ATX activity levels of PBS-challenged mice. PAT-048 treatment also markedly decreased trough plasma ATX activity in PBS- and bleomycin-challenged mice, by 76 and 87%, respectively (Fig. 8B). As opposed to ATX activity levels, however, the concentrations of ATX protein in the BAL fluid of bleomycin-challenged mice were not significantly changed by PAT-048 treatment (Fig. 8C), suggesting that this inhibitor did not affect ATX production or entry in the air spaces in this model. In contrast, administration of different ATX inhibitor, ONO-8430506, to unchallenged mice was recently shown to increase ATX mRNA in adipose tissue, a major site of ATX production, and consequently to increase the amount of (inhibitor-bound) ATX protein in the plasma (32). The feedback inhibition of ATX expression by elevated LPA levels demonstrated by these experiments, however, was found to be able to be overcome by the inflammatory cytokines TNF-α or IL-1β (32). The preservation of ATX protein levels in the face of PAT-048 inhibition that we observed in the bleomycin model consequently could be attributable to the inflammatory cytokines induced by bleomycin lung injury, which include TNF-α and IL-1β.

Figure 8.

PAT-048 inhibits ATX activity in BAL fluid and plasma. Mice were treated with PAT-048 (20 mg/kg) or vehicle by oral gavage and then challenged with bleomycin (1.0 U/kg) or PBS intratracheally ∼5 h later. Oral PAT-048 or vehicle administration was continued once daily for 13 d, and mice were euthanized on d 14 (∼20 h after the last dose of PAT-048). ATX activity was determined in BAL, plasma, and lung homogenates. A) BAL ATX activity was significantly increased in bleomycin-challenged mice, and this increase was abrogated by PAT-048. B) Plasma ATX activity was not affected by bleomycin challenge, but was significantly decreased by PAT-048. C) BAL fluid ATX protein was increased in bleomycin-challenged mice, and this increase was not affected by PAT-048. Lines and error bars represent means ± sem of data pooled from 2 experiments for bleomycin-challenged, PAT-048–treated (n = 10 mice); bleomycin-challenged, vehicle-treated mice (n = 11 mice); and from a single experiment for PBS-challenged, PAT-048–treated and PBS-challenged, and vehicle-treated mice (n = 5 mice per group). ****P < 0.0001, ANOVA followed by the post hoc Bonferroni test.

PAT-048 did not inhibit bleomycin-induced LPA production or lung fibrosis

Inhibition of ATX with PAT-048 did not affect postbleomycin levels of total BAL fluid LPA or of any individual LPA species (Fig. 9A). ATX inhibition with PAT-048 also did not affect bleomycin-induced increases in BAL fluid total protein, a measure of the lung vascular leak produced by bleomycin injury (Fig. 9B), or lung collagen, as assessed by hydroxyproline content, a measure of the pulmonary fibrosis produced by bleomycin injury (Fig. 9C). Taken together, our data suggest that the increased LPA present in the lung after injury, which we have shown contributes to the development of pulmonary fibrosis (10), is produced independent of ATX.

Figure 9.

ATX inhibition with PAT-048 did not attenuate bleomycin-induced pulmonary fibrosis. Mice were treated PAT-048 (20 mg/kg) or vehicle by oral gavage and then challenged intratracheally with bleomycin (1.0 U/kg) or PBS 5 h later. Oral PAT-048 or vehicle administration was continued once daily for 13 d, and mice were euthanized on d 14. A) Analysis of individual LPA species (left graph), and total LPA levels (right graph), in BAL fluid from bleomycin-challenged mice treated with PAT-048 or vehicle. No differences between individual LPA species or total LPA concentrations were significantly different between the PAT-048– and vehicle-treated groups, as evaluated for significance with 2-way Student’s t test with Bonferroni correction for multiple comparisons. B) BAL from bleomycin-challenged, PAT-048– or vehicle-treated mice was analyzed for total protein, and no significant difference (ns) was seen. C) Lung homogenates were analyzed for total lung hydroxyproline as a measure of collagen content. Bleomycin-challenged mice had significantly higher lung hydroxyproline than PBS-challenged mice, but there was no significant difference (ns) in hydroxyproline between PAT-048 and vehicle treatment. For all panels, data shown are means ± sem pooled from 2 experiments for bleomycin-challenged, PAT-048–treated (n = 10 mice) and bleomycin-challenged, vehicle-treated mice (n = 11), and from a single experiment for PBS-challenged, PAT-048–treated and PBS-challenged, vehicle-treated mice (n = 5 mice per group). **P < 0.01; ***P < 0.001, by Student’s t test.

DISCUSSION

In this study, we provide evidence that ATX is not necessary to increase LPA levels in the bleomycin mouse model of pulmonary fibrosis, despite finding that BAL fluid and lung ATX activity and protein levels are significantly increased by bleomycin injury. We first confirmed our prior observation that bleomycin lung injury produces increased LPA concentrations in the lung lining fluid that is sampled by BAL (10). By determining the concentrations of multiple individual LPA species, we then demonstrated that the majority of the LPA produced in the lung after this injury consists of LPA species with the long-chain polyunsaturated acyl groups 22:5 and 22:6. Despite finding higher levels of ATX activity and protein in the BAL and lung, but not the plasma, of mice following bleomycin injury, we found 2 lines of evidence indicating that ATX is not responsible for increasing LPA levels in this model. First, we found a marked discordance between the individual species of LPC (shorter chain, saturated or monounsaturated species) and the individual species of LPA (longer chain, polyunsaturated species) that predominated in the BAL fluid after bleomycin challenge. This discordance is not supportive of a substrate–product relationship between LPC and LPA. Second, we found no reduction of bleomycin-induced increases in BAL fluid LPA with administration of PAT-048, a small molecule that very effectively inhibited both BAL and plasma ATX activity. Consistent with this latter finding, PAT-048 had no effect on lung fibrosis in the bleomycin model.

Longer chain polyunsaturated LPA species, similar or identical to those we found to predominate in the BAL fluid of bleomycin-injured mice, have recently been described in samples from human lungs in IPF and asthma. We characterized the LPA species present in the exhaled breath condensate (EBC), as well as the plasma, of patients with IPF (29). Of the 9 different LPA species that were detectable in EBC, only 22:4 LPA, another longer chain polyunsaturated species, was significantly elevated in the EBC of IPF subjects compared to controls. This species was particularly elevated in the EBC of an IPF patient assessed during an acute exacerbation. Of the 13 LPA species that were detectable in the plasma of these subjects, none were significantly different between IPF patients and controls; consistent with our finding, there no differences in plasma LPA concentrations of bleomycin- and PBS-challenged mice. The predominant LPA species we observed in the BAL fluid of bleomycin-challenged mice, 22:5 and 22:6 LPA, were the same LPA species that predominated in the dramatically increased LPA levels observed in BAL fluid recovered from mildly asthmatic human subjects after subsegmental allergen challenge (35). Also similar to our findings, there was a discordance in this study of human asthmatic subjects between these longer chain polyunsaturated LPA species that predominated in the BAL and the shorter chain saturated LPA species that predominated in the plasma, suggesting that elevations of 22:5 and 22:6 LPA species in the lung following allergen challenge were also due to local LPA production rather than to LPA extravasation from plasma into the lung.

There are 2 alternative explanations for the discordance that we observed between the predominant LPA species and the predominant LPC species in the BAL fluid after bleomycin challenge that would not necessarily be inconsistent with ATX being responsible for LPA generation in this model. The first alternative explanation would be for ATX to have a much higher affinity for longer chain polyunsaturated LPC species than for shorter chain saturated species. If this were the case, ATX could preferentially act on the less abundant longer chain polyunsaturated species to produce mostly longer chain polyunsaturated LPA species. However, analyses of the structure of ATX’s LPC-binding pocket predict that saturated LPC species will have optimal ATX binding (36), and preferred LPC substrates for ATX have empirically been found to be 14:0 > 12:0 > 18:3 > 16:0 > 18:2,18:1 > 18:0 (37), making this alternative explanation for the BAL fluid levels of LPC-LPA discordance unlikely. A second alternative explanation for this discordance would be for the LPC species that predominate in the BAL fluid to be those that were preferentially discriminated against by ATX, rather than those that were preferentially used as substrates to generate the predominant BAL fluid LPA species. This would require ATX to prefer longer chain polyunsaturated LPC species, which as noted above, has neither been predicted nor seen. Further, the predominant species of LPC and LPA we observed were very similar to the plasma, where ATX-mediated hydrolysis of LPC has been established to be the major source of LPA, making this alternative explanation for the discordance in BAL fluid levels of LPC-LPA unlikely as well. We therefore conclude that the discordance between the fatty acids of the LPC and LPA species that are elevated in the alveolar space after bleomycin challenge most likely reflects ATX-mediated hydrolysis of LPC not being the major mechanism through which LPA is produced in this model.

In our studies, lung ATX mRNA levels were decreased by bleomycin injury, and BAL fluid ATX protein and activity levels correlated closely with total protein levels. These findings suggest that increases in BAL fluid and lung ATX induced by bleomycin lung injury are predominantly attributable to plasma ATX entering the alveolar space via vascular leak, in contrast to LPA, which, as discussed above, appears to be produced locally in the lungs after bleomycin injury. Prior evidence has indicated that levels of some of the other proteins that are present at sites of injury or inflammation are also predominantly increased by vascular leak rather than by local production. For example, in a mouse model of skin inflammation induced by topical application of the phorbol ester 12-O-tetradecanoylphorbol 13-acetate, skin levels of 55 of 976 proteins analyzed were reduced more than 2-fold in mice with decreased vascular permeability caused by reduced production of bradykinin (38).

There are well-described ATX-independent pathways that generate LPA. The major ATX-independent pathway of LPA production begins with the formation of PA by hydrolysis of the phosphodiester bond of a phospholipid by PLD (Fig. 1). The sn-1 fatty acid of PA is then removed by aPLA₁, or the sn-2 fatty acid is removed by a PLA₂, producing LPA. An important role for this ATX-independent pathway, specifically involving the PLA₁ family member membrane-associated PA-selective PLA₁α (mPA-PLA₁α), also known as lipase H (LIPH), has been demonstrated for hair growth in humans. Mutations in LIPH, which have been shown to cause hereditary hypotrichosis syndromes, manifest as diffuse hair loss in early childhood (39–41). Although ATX is highly expressed in the dermal papilla precursor cells that produce hair follicles, follicle development and hair growth proceed normally in mice with deletion of ATX specifically in these cells (42), suggesting that LPA generation can proceed through ATX-independent pathways in the hair follicle, as we have found for the lung. Moreover, the expression of LIPH was found to be up-regulated in the dermal papilla precursor cells of ATX-deficient mice compared to wild-type cells, suggesting that increases in LIPH expression may compensate for the loss of ATX-dependent LPA production in the hair follicle. As a PLA₁, LIPH preferentially removes the sn-1 saturated fatty acids of phospholipids, generating polyunsaturated LPA species, such as those that we observed to predominate in the lungs of bleomycin-challenged mice. In contrast to ATX-deficient mice, LIPH-deficient mice have abnormalities of hair follicle development and hair growth. LIPH-deficient mice have specific reductions in several unsaturated LPA species, including 22:5 LPA, in their hair follicles (39), further suggesting that LIPH preferentially generates polyunsaturated LPA species. LIPH is robustly expressed in the lungs at baseline (15, 43). These observations, taken together with our data, suggest that an ATX-independent, PLD-PLA₁-dependent pathway is responsible for the increased BAL fluid LPA observed after bleomycin injury. In human IPF, lung LIPH expression has been found to be 11.5-fold greater in rapidly compared to slowly progressing patients. In this study, rapidly progressing patients were identified as those who presented for medical attention within 6 mo of the onset of symptoms and slowly progressing patients as those who presented greater than 24 mo after symptom onset (44).

In summary, we found that the LPA required for the development of pulmonary fibrosis can be generated independent of ATX, in contrast to the requirement of ATX for the production of LPA present in the circulation (45, 46) and to our own recent finding that ATX activity is required for the development of dermal fibrosis in a bleomycin scleroderma model (unpublished data). Taken together, our data and these observations suggest that the pathways through which LPA is generated are context specific; they differ across tissues and disease states. Therapeutic approaches to treat diseases such as pulmonary or dermal fibrosis by inhibiting LPA generation will consequently require the identification of the enzymes responsible for producing LPA on a tissue-by-tissue and disease-by-disease basis. Investigation of the specific roles of the PLD enzymes, and of LIPH and other potentially relevant PLA₁ and PLA₂ family members, in LPA generation induced by lung injury may reveal new therapeutic targets for pulmonary fibrosis.

Glossary

ATX

autotaxin (ENPP2)

BAL

bronchoalveolar lavage

EBC

exhaled breath concentrate

ENPP2

ecto-nucleotide pyrophosphatase/phosphodiesterase 2 (autotaxin)

ESI-LC/MS/MS

electrospray ionization-liquid chromatography-tandem mass spectrometry

IC

inhibitory concentration

LPA

lysophosphatidic acid

LPC

lysophosphatidylcholine

LIPH

lipase H (mPA-PLA₁α)

lysoPLD

lysophospholipase D

mPA-PLA₁α

membrane-associated PA-selective PLA₁ (lipase H)

PA

phosphatidic acid

PAT-048

3-[6-chloro-7-fluoro-2-methyl-1-(1-propyl)-1H-pyrazol-4-yl-1H-indol-3-ylsulfanyl]-2-fluoro-benzoic sodium salt

PC

phosphatidylcholine

PLA

phospholipase A₁

PLD

phospholipase D