By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
Potential conflict of interest: Nothing to report.
Capillarization precedes hepatic fibrosis. We hypothesize that capillarization of sinusoidal endothelial cells (SEC) is permissive for hepatic stellate cell (HSC) activation and therefore permissive for fibrosis. We examined whether freshly isolated SECs prevent activation of HSCs and promote reversion to quiescence, and whether this effect was lost in capillarization. HSCs were cultured alone or co-cultured with differentiated or capillarized SECs. Results: Co-culture with freshly isolated SECs markedly decreased HSC activation after 3 days in culture, but co-culture with capillarized SEC had no effect. Inhibition of nitric oxide (NO) synthesis abolished SEC suppression of HSC activation. Activated HSCs reverted to quiescence when co-cultured with SEC plus vascular endothelial growth factor (VEGF) (that is, with SECs that maintained differentiation), but co-culture with capillarized SECs did not. Reversion of activated HSCs to quiescence in the presence of SECs plus VEGF was abolished by inhibition of NO synthesis. To establish whether there was indeed reversion, activated and quiescent HSCs were counted before and 3 days after adding freshly isolated SECs plus VEGF to activated HSCs, and proliferation was quantified in quiescent HSCs; the stoichiometry demonstrated reversion. Conclusion: Differentiated SECs prevent HSC activation and promote reversion of activated HSCs to quiescence through VEGF-stimulated NO production. Capillarized SECs do not promote HSC quiescence, because of loss of VEGF-stimulated NO production. (HEPATOLOGY 2008.)
Activation of the quiescent hepatic stellate cell (HSC) to an activated, collagen-producing HSC is considered the pivotal event leading to fibrosis. The activated HSC is defined by a variety of phenotypical markers, such as expression of α-smooth muscle actin (ASMA), type I collagen, tissue inhibitor of matrix metalloproteinase-1 (TIMP-1), and increased F-actin stress fibers. In addition to HSC activation, another change that precedes fibrosis is capillarization, which is defined as the formation of an organized basement membrane in the space of Disse with characteristic changes in the sinusoidal endothelial cell (SEC) phenotype.1–3 Several phenotypical changes have been noted in the capillarized SEC, but the defining morphological change is the loss of the characteristic open fenestrae organized in sieve plates.1
Because capillarization precedes fibrosis,1, 2, 4, 5 this raises the question of whether capillarization itself might predispose to activation of HSCs. A causal link between capillarization and activation of HSC is biologically plausible. Studies over the last 15 years have shown in other vascular beds that endothelial cells and neighboring mural cells (pericytes, vascular smooth muscle cells) maintain each other's phenotype.6–8
In the liver, the SEC phenotype is maintained by paracrine production of vascular endothelial growth factor (VEGF) by hepatocytes and HSCs and autocrine production of VEGF-stimulated nitric oxide (NO).9 The current study examines whether there is reciprocity, that is, whether SECs might maintain and promote the quiescent HSC phenotype, whether this effect is lost when SECs undergo capillarization, and what the mechanism is by which SECs promote the quiescent phenotype of HSC.
Chemicals were obtained from the Sigma Chemical company (St. Louis, MO) unless stated otherwise.
SECs were isolated by collagenase perfusion, iodixanol density gradient centrifugation, and centrifugal elutriation as previously described.10 The average yield of SEC per rat was 88 × 106 with greater than 98% viability and 97% or greater purity. SEC purity was assessed by fluorescent acetylated low-density lipoprotein followed by a peroxidase stain to reveal any contaminating Kupffer cells.
SEC cultured for 3 days in the presence of VEGF or in the presence of freshly isolated, quiescent HSCs remain fenestrated for 3 days9 and are considered differentiated SECs, whereas SECs cultured alone for 3 days without VEGF lack fenestration,9 the definitive marker of a differentiated or noncapillarized SEC, and are considered capillarized. SECs isolated from thioacetamide-induced cirrhotic liver are also capillarized.
HSCs were isolated by collagenase/pronase digestion and Stractan density gradient centrifugation.11 The yield of HSC is typically between 4 × 106 and 10 × 106 with 95% or greater viability and 90% or greater purity; purity is assessed by vitamin A autofluorescence. Hepatocytes were isolated by collagenase perfusion, gravity sedimentation, and Percoll density gradient centrifugation. Viability and purity of hepatocytes is 80% or greater and 95% or greater, respectively; purity is assessed by morphology on light microscopy.
HSC were plated in the wells of 24-well plates at a density of 130,000/cm2. SECs (400,000 cells/cm2) or hepatocytes (105,000 cells/cm2) were plated in collagen-coated Transwell inserts with 3-μm pore size (Costar, Fisher scientific, Pittsburgh, PA). This allows cells to maintain contact through shared culture medium without mixing of cells. Transwells were added to the culture plates after the SECs or hepatocytes had adhered.
Alpha-Smooth Muscle Actin Staining.
HSCs were cultured on glass coverslips and fixed with 10% formaldehyde (EM Science, Gibbstown, NJ) for 20 minutes at room temperature. Cells were incubated with a mouse monoclonal anti–alpha smooth muscle (ASMA) (1:100), followed by a rabbit anti-mouse immunoglobulin G tetramethylrhodamine isothiocyanate conjugate (1:50, Santa Cruz Scientific, Santa Cruz, CA). Controls were stained with a mouse isotype control antibody, a nonspecific immunoglobulin G2a kappa immunoglobulin (Sigma), as a primary antibody. Slides were examined using a Nikon PCM-2000 confocal microscope with a Nikon Eclipse TE300 microscope with a plan Apo 60×/1.4 aperture oil immersion objective, 543-nm laser, and Simple PCI software from the C-Imaging series from Nikon/Compix Inc (Cranberry Township, PA). Values were obtained by counting the number of ASMA-positive cells in 15 randomly selected fields.
F-Actin Stress Fibers.
HSC were grown on glass coverslips, fixed with 2% paraformaldehyde/0.1% glutaraldehyde (Electron Microscopy grade, Polysciences, Warrenton, PA) and incubated with 50 mM ammonium chloride for 5 minutes. Cells were permeabilized with 0.1% Triton X-100, blocked with 5% fetal bovine serum in phosphate-buffered saline, stained with Alexa Fluor-488 phalloidin (Molecular Probes, Eugene, OR), and counterstained with propidium iodide (Molecular Probes). Fluorescence was examined using a Nikon TE 300 Quantum inverted fluorescence microscope, with a plan Apo 60×/1.4 aperture oil immersion objective, a Hamamatsu Orca digital CCD camera (C4742-95-12), and Metamorph Meta imaging version 6.1 software (Universal Imaging Corporation, West Chester, PA). Fluorescence was quantified in 10 fields with exclusion of dead cell fluorescence and normalized to the number of HSCs counted per field.
HSC were cultured in a six-well plate for 3 or 6 days. For detection of ASMA and type I collagen, cells were harvested in lysis buffer (20 mM Tris-HCL, pH 7.6, 20 mM NaF, 20 mM β-glycerophosphate, 0.5 mM Na3VO4, 2.5 mM metabisulfite, 5 mM benzamidine, 1 mM ethylenediaminetetra-acetic acid, 0.5 mM ethylene glycol tetra-acetic acid, 300 mM NaCl, with 10% glycerol, protease inhibitor, and 1% triton X-100). An equal amount of the whole cell protein (100 μg/lane) was separated by 4% to 12% NuPage Bis-Tris Gel (Invitrogen, Carlsbad, CA) under reducing conditions and transferred to nitrocellulose. Protein was detected by incubating with 1:10,000 monoclonal anti-ASMA and 1:5000 rabbit polyclonal anti-type I collagen (Rockland, Gilbertsville, PA), followed by incubation with a horseradish peroxidase conjugated secondary antibody (Santa Cruz) at 1:10,000. For detection of TIMP-1, cell culture medium was centrifuged at 16,800g for 15 minutes at 4°C. Supernatants were mixed with loading buffer (Invitrogen) and transferred to nitrocellulose. Protein was detected by incubating 1 μg/mL mouse monoclonal anti-rat TIMP-1 (R & D System Inc, Minneapolis MN) followed by incubation with a horseradish peroxidase–conjugated secondary antibody (Santa Cruz) at 1:1000. The protein was visualized by a commercial chemiluminescent method using the Pierce ECL kit (Amersham Bioscience, Piscataway, NJ).
Nitric Oxide Assay.
NO was determined as the sum of nitrite plus nitrates. Escherichia coli nitrate reductase and nicotinamide adenine dinucleotide phosphate, reduced form were used to convert nitrate to nitrite in culture medium, and nitrite was quantified according to the Griess reaction.12 Fifty microliters culture supernatant is mixed with equal volumes of 0.1% N-1-naphthylenediamine hydrochloride and 1% sulfanilamide in 5% H3PO4. After 5 minutes at room temperature, absorbance is measured at 540 nm on a Biorad 3550 microplate reader (Biorad, Hercules, CA).
Rats (240-260 g males, Sprague Dawley, Harlan, Indianapolis, IN) were treated with thioacetamide 200 mg/kg intraperitoneally three times weekly for 9 weeks.
HSCs were plated on glass coverslips in 24-well plates. On the day of isolation (day 0), HSCs were incubated with 5 μM Vybrant carboxyfluorescin diacetate, succinimidyl ester (CFDA SE) cell tracer (Molecular Probes) for 15 minutes at 4°C. This dye binds irreversibly to intracellular amines, it cannot be transmitted to surrounding cells, and dye fluorescence per cell is reduced by half with each cell division. From day 0 to day 3, HSCs were cultured alone, and from day 3 to day 6 activated HSC were co-cultured with freshly isolated SECs in a Transwell insert with 40 ng/mL VEGF added to the medium.
The number of quiescent and activated HSC were counted in 15 high-power fields on day 3 and day 6. Compact cells with cytoplasmic lipid droplets were considered quiescent HSCs. At the conclusion of the experiment on day 6, cells were fixed with 2% paraformaldehyde. Vybrant CFDA fluorescence was examined using a Nikon TE 300 Quantum inverted fluorescence microscope with a plan Apo 60×/1.4 aperture oil immersion objective, a Hamamatsu Orca digital CCD camera (C4742-95-12), and Metamorph Meta imaging software version 6.1 software (Universal Imaging Corporation, West Chester, PA). Fluorescence was quantified in 10 fields with exclusion of dead cell fluorescence and normalized to the number of HSCs counted in each field.
All data, expressed as mean ± standard error of the mean, were from at least three separate experiments. Groups were compared by analysis of variance (ANOVA) with a posteriori contrast by least significant difference; or by Student t test using the Microsoft Excel Analysis ToolPak (Microsoft, Redmond, WA). P < 0.05 was considered significant.
Maintenance of Stellate Cell Phenotype.
A much larger proportion of HSCs maintained quiescence when co-cultured with freshly isolated, differentiated SECs compared with either HSCs cultured alone, co-cultured with capillarized SECs, or co-cultured with hepatocytes (Figs. 1, 2, 6). When HSCs were cultured on plastic in homotypic culture, 70.5% ± 5.8% of cells expressed ASMA after 3 days as assessed by confocal microscopy. In contrast, only 29.6% ± 1.8% of HSCs co-cultured together with freshly isolated SEC expressed ASMA after 3 days (HSC cultured alone versus co-culture with freshly isolated SECs, P < 0.01). Previous studies have demonstrated that SECs cultured with quiescent HSCs for 3 days under these conditions maintain fenestrae in sieve plates.9 Two models were used to obtain capillarized SECs. As previously reported,9 SECs cultured alone without exogenous VEGF for 3 days lack fenestrate (Fig. 3). SECs isolated from thioacetamide-treated cirrhotic liver (n = 5) were examined for fenestrae: there was a complete lack of fenestrae and sieve plates in 15 electron microscopy (EM) pictures per liver (Fig. 3). When HSCs were co-cultured with capillarized SECs from in vitro capillarization or from cirrhotic rats (Fig. 3), 79.9% ± 4.0% and 81.2% ± 2.1%, respectively, of HSCs were ASMA positive on confocal microscopy after 3 days (Fig. 1). Similarly, 75.6% ± 5.3% of HSCs co-cultured with hepatocytes for 3 days were ASMA positive on confocal microscopy (Fig. 6).
The effect of co-culture with freshly isolated SECs on HSC activation was also examined using immunoblot for ASMA and TIMP-1 (Fig. 4) or by quantitation of fluorescence of F-actin stress fibers (Fig. 5). HSC expression of ASMA and of TIMP-1 was markedly reduced by co-culture with freshly isolated SECs compared with HSCs cultured alone, but co-culture with SECs obtained from thioacetamide-treated cirrhotic liver did not reduce HSC expression of ASMA or TIMP-1 (Fig. 4). Quantitation of F-actin stress fiber fluorescence demonstrated a 5-log difference when comparing HSCs cultured alone versus co-culture with freshly isolated SECs for 3 days (Fig. 5, P < 0.05). There was no difference in HSC stress fiber formation between HSCs cultured alone compared with HSCs co-cultured with SECs that had been allowed to capillarize in vitro.
Paracrine Regulation of Stellate Cell Phenotype by Nitric Oxide.
To determine whether nitric oxide (NO) contributes to maintenance of HSC phenotype, nitric oxide synthase was inhibited with 3 mM NG-nitro-L-arginine methyl ester (L-NAME). The number of ASMA-positive HSCs in co-culture with SECs in the presence and absence of L-NAME was 64.2% ± 2.1% versus 23.0% ± 2.3% (P < 0.0001). Thus, the addition of L-NAME blocked SEC suppression of HSC activation (Fig. 6), indicating that NO is an essential mediator of the SEC effect. Freshly isolated HSC produce very low amounts of NO, whereas SECs are the major hepatic source of NO13; thus L-NAME is acting on NO produced by SECs.
Hepatocytes did not prevent HSC activation (Fig. 6). However, when V-PYRRO/NO is added to hepatocytes, NO is liberated by P450 cleavage. When HSCs were co-cultured with hepatocytes in the presence of 100 μM V-PYRRO/NO, that is, when hepatocytes produced NO, HSC activation was prevented (Fig. 6). V-PYRRO/NO in the absence of hepatocytes had no effect on HSC. This further supports the concept that NO plays a major role in suppressing HSC activation.
As described above, differentiated SEC promote quiescence in HSC, but capillarized SECs do not. SECs maintain a differentiated phenotype by day 3 when co-cultured with quiescent HSCs, whereas SECs cultured alone are capillarized by day 3.9 After 3 days in culture, NO in the medium of SECs co-cultured with quiescent HSCs is 40.1 ± 0.3 nmole/million SECs (9.4 ± 0.7 μM) versus 30.1 ± 2.1 nmole/million cells (7.5 ± 0.5 μM) from SECs cultured alone that have capillarized by day 3 (n = 3; P < 0.01). Thus, if the ability of differentiated, but not of capillarized, SECs to prevent HSC activation is due to NO, it is due to a difference in NO production of approximately 30%.
SECs Induce Reversal of Stellate Cell Phenotype.
HSCs in homotypic culture were allowed to activate on plastic over 3 days. From day 3 to day 6, HSCs were cultured under four different conditions: cultured alone, cultured in the presence of VEGF, in co-culture with SECs freshly isolated on day 3 but without exogenous VEGF, or in co-culture with SECs freshly isolated on day 3 plus exogenous VEGF (Figs. 7–9). When HSCs were cultured alone for 3 days followed by 3 days of co-culture with freshly isolated SECs, the SECs were capillarized after 3 days as ascertained by surface expression of CD31 (data not shown); the use of CD31 as a marker of capillarization has been previously validated.9 In contrast, when VEGF is added, SECs do not capillarize.9 By day 6, HSCs co-cultured from day 3 to day 6 with freshly isolated SECs plus VEGF had significantly fewer ASMA-positive cells compared with HSCs cultured alone (Fig. 8; 38.3% ± 5.2% versus 88.7% ± 5.2% ASMA-positive cells; n = 4, P < 0.0001). In contrast, there was no significant decrease in the number of ASMA-positive HSCs on day 6 after culture from day 3 to day 6 with either SECs without VEGF or VEGF alone, when compared with HSCs cultured alone. These findings suggest that SECs that remain differentiated in the presence of VEGF induce reversal of the stellate cell phenotype. Of note, the addition of hepatocytes on day 3 did not reduce the number of ASMA-positive HSCs by day 6 (n = 3, data not shown). Morphology of HSC was also significantly different: HSCs cultured with SECs plus VEGF from day 3 to day 6 regained the appearance of quiescent HSC with compact cytoplasm and fat droplets (Fig. 8), whereas HSC cultured alone, co-cultured with SECs without VEGF, or with VEGF alone maintained F-actin stress fibers (Fig. 8).
The effect of SECs plus VEGF on HSC reversal was also examined by immunoblot for ASMA and type I collagen expression (Fig. 7) and by detection of F-actin stress fibers in HSCs (Fig. 9). Co-culture with capillarized SECs or the addition of VEGF to HSCs did not reduce the expression of ASMA or type I collagen (Fig. 7A, lanes 2 and 3, and Fig. 7B, C) compared with HSCs cultured alone for 6 days (Fig. 7A lane 1, and Fig. 7B, C). In contrast, there was a significant reduction in ASMA and type I collagen in HSCs co-cultured with SEC plus VEGF from day 3 to day 6 (Fig. 7A, lane 4, and Fig. 7B, C) There was a 7-log drop in F-actin stress fiber fluorescence in HSCs cultured from day 3 to day 6 with SECs plus VEGF compared with HSCs cultured alone (Fig. 9). There was no effect of co-culture with capillarized SECs or the addition of VEGF on HSC F-actin stress fiber expression (Fig. 9).
Role of Nitric Oxide in Reversal of the Activated HSC Phenotype.
To determine whether NO plays a role in reversal of activated HSC to a quiescent phenotype, L-NAME was added to the co-culture system (Fig. 10). HSCs were cultured alone from day 0 to day 3 and then cultured with freshly isolated SECs and VEGF with or without L-NAME. L-NAME abolished the reduction in ASMA-positive HSCs seen in the co-culture of SECs with VEGF. The stability of L-NAME over 3 days was confirmed by dissolving L-NAME in culture medium, leaving it in an incubator for 3 days, and then comparing the ability to suppress NO production with that of freshly dissolved L-NAME; L-NAME activity was stable for 3 days (data not shown).
As described, freshly isolated SECs added to HSCs on day 3 were capillarized by day 6 unless VEGF was added. NO concentration was 32.1 ± 0.6 nmol/million cell (7.1 ± 0.9 μM) on day 6 in the medium after co-culture of HSCs with SECs without exogenous VEGF from day 3 to day 6 and was 48.8 ± 3.2 nmol/million cells (10.0 ± 2.1 μM) on day 6 in the medium after co-culture of HSC with SEC plus VEGF from days 3 to 6 (n = 3; P < 0.01).
Reversal of Stellate Cell Phenotype.
Studies were done to determine whether the decrease in activated HSCs on day 6 after co-culture from day 3 to day 6 with SECs plus VEGF was indeed attributable reversion to a quiescent phenotype (Table 1). There were 7 times more quiescent HSCs per high-power field on day 6 than on day 3 and less than half the number of activated HSC. There are two possible explanations for the increase in quiescent HSCs and the decrease in activated HSCs. One possibility is that activated HSCs reverted to a quiescent phenotype. The second possibility is that half the activated HSCs underwent apoptosis (half the number of activated HSCs were present on day 6) and that the quiescent HSCs proliferated. The average number of quiescent HSCs per high-power field increased from 2.5 on day 3 to 17.5 on day 6. According to the formula, doublings = log(final number/starting number)/log 2 = log (17.5/2.5)/log 2 = log7/log 2 = 2.8. Thus, the quiescent HSCs would have to undergo an average of 2.8 doublings to account for the number of quiescent HSC found on day 6.
Table 1. Quantitation of Quiescent and Activated HSCs
Average Cell No/HPF
HSCs were cultured alone on plastic from day 0 to day 3. From day 3 to day 6, HSCs were co-cultured with SECs isolated on day 3 plus VEGF. HSCs were counted on day 3 and day 6 in 15 randomly selected high-power fields (HPF). Quiescent HSCs were considered compact cells with cytoplasmic lipid droplets (n = 3).
HSC day 3
2.5 ± 0.1
10.2 ± 0.2
12.8 ± 0.3
HSC day 6/SEC day 3/VEGF
17.5 ± 1.4
4.4 ± 0.3
21.7 ± 1.6
To test these two possibilities, cell division of quiescent HSCs was tracked with a fluorescent dye. The dye fluorescence on day 3 in quiescent HSCs was the same as on day 0, indicating that none of the quiescent HSCs had divided by day 3. On day 6, half of the quiescent HSCs had undergone two doublings, and half of the quiescent HSCs present had never divided, for an average of 1 doubling, that is, less than the average 2.8 doublings needed if the explanation was apoptosis and proliferation of the quiescent HSCs present on day 3. Thus, reversion of activated HSCs best accounts for the findings.
The data presented here demonstrate that SECs prevent HSC activation and also promote reversion of activated HSCs to a quiescent phenotype, whereas capillarized SECs lose this effect. Coupled with the in vivo observation in humans and in animal models that capillarization precedes fibrosis,1, 2, 4, 5 the findings in this study suggest that capillarization of sinusoidal endothelial cells may be permissive for HSC activation and fibrosis. The current study examined whether the decrease in the number of activated HSCs and the increase in quiescent HSCs in the presence of SECs with VEGF was indeed attributable to reversion of activated HSCs to the quiescent phenotype. Enumeration of quiescent and activated HSCs and quantitation of proliferation of quiescent HSCs showed that only reversion of activated HSCs could account for the stoichiometry.
NO production by SECs was necessary to maintain the quiescent HSC phenotype. A study by Langer et al.14 examined whether NO generated by SECs might induce apoptosis of activated HSCs.14 HSC apoptosis was not significantly increased over background at 5 μM NO delivered by an exogenous NO donor, but was induced in a dose-dependent fashion with NO concentrations ranging from 50 to 500 μM.14 In the current study, NO derived from SECs co-cultured for 3 days with quiescent HSCs or cultured with VEGF was 9.5 and 10 μM, respectively. These NO concentrations are within the same order of magnitude as those found in vivo in the rat hepatic vein13 and 5-fold to 50-fold lower than the concentrations that induced HSC apoptosis.14
In normal liver, paracrine production of VEGF by hepatocytes and HSCs stimulates NO by SECs, and VEGF-stimulated NO production by SECs is essential for maintenance of differentiated SECs.9 In capillarization before cirrhosis, paracrine production of VEGF is markedly decreased, VEGF stimulation of NO is lost, and only basal NO production persists.15 Capillarized SEC did not maintain a quiescent HSC phenotype and did not promote reversion of activated HSC to a quiescent phenotype in co-culture. NO concentration in the medium of capillarized SECs was 7 to 7.5 μM, that is, 25% to 30% lower than NO production by differentiated SECs either cultured with VEGF or co-cultured with a cellular source of VEGF, and this reflects the decrease in NO production when VEGF stimulation is lost.9 Thus, the current study suggests that VEGF-stimulated NO production is also necessary for preservation of the quiescent HSC phenotype. Another change in capillarized SECs that promotes HSC activation is expression of fibronectin containing an extra type III domain (EIIIA fibronectin).16
Interdependence in maintenance of phenotype between a differentiated SEC and a quiescent HSC is consistent with studies from Dr. Eli Keshet's laboratory. In these studies, a transgenic system is used for conditional switching of a soluble VEGF receptor in the liver. Secretion of the soluble VEGF receptor blocks VEGF, which leads to loss of SEC fenestration and a marked increase in ASMA-positive stellate cells; these changes are fully reversible with lifting of the VEGF blockade (E. Keshet, personal communication).
A variant of capillarization called pseudocapillarization is seen with aging.17, 18 Future studies will need to examine whether aging-related capillarization causes similar changes to disease-related capillarization. If so, this may explain why age is a risk factor for fibrosis in hepatitis C19 and in nonalcoholic fatty liver disease.20–22
In summary, the data presented here demonstrate that differentiated SECs prevent HSC activation and promote reversion of activated HSC to quiescence. The paracrine effect of differentiated SECs on HSC phenotype requires NO production by the SEC. Capillarized SECs lose the ability to maintain quiescence of HSC and to promote reversion and this may be attributable to the decline in NO production that occurs with loss of responsiveness to VEGF. One implication of these findings is that approaches to reverse capillarization might promote resolution of fibrosis.
The authors thank Robert S. McCuskey for his invaluable assistance with scanning electron microscopy.