Cytokine-induced arginase activity in pulmonary endothelial cells is dependent on Src family tyrosine kinase activity

Abstract

We hypothesized that the Src family tyrosine kinases (STKs) are involved in the upregulation of arginase and inducible nitric oxide synthase (iNOS) expression in response to inflammatory stimuli in pulmonary endothelial cells. Treatment of bovine pulmonary arterial endothelial cells (bPAEC) with lipopolysaccharide and tumor necrosis factor-α (L/T) resulted in increased urea and nitric oxide (NO) production, and this increase in urea and NO production was inhibited by the STK inhibitor PP1 (10 μM). The STK inhibitors PP2 (10 μM) and herbimycin A (10 μM) also prevented the L/T-induced expression of both arginase II and iNOS mRNA in bPAEC. Together, the data demonstrate a central role of STK in the upregulation of both arginase II and iNOS in bPAEC in response to L/T treatment. To identify the specific kinase(s) required for the induction of urea and NO production, we studied human pulmonary microvascular endothelial cells (hPMVEC) so that short interfering RNA (siRNA) techniques could be employed. We found that hPMVEC express Fyn, Yes, c-Src, Lyn, and Blk and that the protein expression of Fyn, Yes, c-Src, and Lyn could be inhibited with specific siRNA. The siRNA targeting Fyn prevented the cytokine-induced increase in urea and NO production, whereas siRNAs specifically targeting Yes, c-Src, and Lyn had no appreciable effect on cytokine-induced urea and NO production. These findings support our hypothesis that inflammatory stimuli lead to increased urea and NO production through a STK-mediated pathway. Furthermore, these results indicate that the STK Fyn plays a critical role in this process.

microbial infection activates neutrophils and macrophages through recognition of microbial components such as lipopolysaccharide (LPS) by pattern recognition receptors of these innate immune cells, resulting in the release of various cytokines, including the proinflammatory cytokines IL-1β, IL-6, and tumor necrosis factor-α (TNF-α). These proinflammatory cytokines can act locally or circulate in the bloodstream and induce the expression of a wide variety of proteins in many cell types by interacting with cell surface receptors. Endothelial cells are the first cells in the lung to encounter circulating LPS and/or cytokines, and the response of endothelial cells to these inflammatory stimuli includes the activation of a number of intracellular signaling pathways (25). Activation of these signaling pathways leads to the expression of a wide variety of stress response proteins either on the cell surface or within the cell. These stress response proteins include arginase and inducible nitric oxide synthase (iNOS).

Nitric oxide (NO) is a potent microbicidal substance and a powerful vasodilator. NO production by iNOS in various cell types, particularly macrophages and neutrophils, plays an important role in host defense. However, the excessive production of NO has also been implicated in the pathogenesis of many lung diseases, including acute respiratory distress syndrome, asthma, and bronchopulmonary dysplasia. Thus a great deal of attention has been paid to the signal transduction pathways involved in iNOS upregulation, particularly in macrophages (2, 13, 21, 34). Our group has previously shown in macrophage cell lines that activation of the Src family tyrosine kinases (STK) is necessary for LPS-induced iNOS expression (22, 34, 40). In endothelial cells, it has been demonstrated that the STK are expressed (26, 28) and may be involved in cytokine-induced iNOS expression (18, 42). The STK include at least nine members: c-Src, Fyn, Yes, Lyn, Lck, Hck, Blk, Fgr, and Yrk, although to date Yrk has been only found in chickens (33). Of the eight members so far described in mammalian cells, c-Src, Fyn, and Yes are thought to be widely distributed, whereas the expression of Lyn, Lck, Hck, Blk, and Fgr are thought to be essentially limited to hematopoietic cells (29, 33).

Arginase metabolizes l-arginine to l-ornithine and urea, and mammalian cells express two isoforms: arginase I and arginase II (27). The arginases are the first step in the generation of proline and polyamines, which are crucial for cell proliferation, and thus the induction of arginase is vital for tissue repair after injury (14, 45). Stimuli that lead to lung injury have been shown to upregulate arginase. For example, hyperoxia led to increased expression of arginase I and arginase II (36). Our group (9) has recently demonstrated that treatment of bovine pulmonary arterial endothelial cells (bPAEC) with LPS and TNF-α (L/T) results in increased protein expression of both arginase I and arginase II.

Arginase and iNOS share a common substrate, l-arginine. It has been suggested that the coinduction of iNOS and arginase is a mechanism to limit NO production in macrophages, to avoid NO overproduction (7, 31, 37). Our group (9) has recently found that L/T triggered the upregulation of iNOS as well as arginase I and II proteins in bPAEC and that inhibiting iNOS increased urea production, whereas inhibiting arginase increased NO production. Thus understanding the mechanisms that lead to cytokine-induced arginase and iNOS expression is imperative to understanding the pathogenesis, as well as the development of novel therapeutic targets, for inflammatory lung diseases like acute respiratory distress syndrome, asthma, and bronchopulmonary dysplasia. We hypothesized that L/T treatment would lead to an increase in both iNOS and arginase expression through activation of the STK. We first tested this hypothesis in bPAEC by utilizing pharmacological inhibitors of the STK and found that pharmacological inhibition prevented L/T-induced arginase II and iNOS expression. To determine whether a specific STK member was responsible for cytokine-induced arginase and/or iNOS expression, we switched to human cells since the bovine genes for all of the STK have not been cloned. We tested the hypothesis that a specific STK member was required for cytokine-dependent induction of urea and NO production. Using human pulmonary microvascular endothelial cells (hPMVEC)-specific short interfering RNA (siRNA) for the various STK, we characterized the role of different STK family members in the induction of arginase and iNOS. Our results indicate that Fyn appears to be the only STK that is necessary for cytokine-induced urea and NO production in hPMVEC.

METHODS

bPAEC culture.

bPAEC were cultured as previously described (9, 32, 38). Briefly, bPAEC were obtained from Lonza (Allendale, NJ). After arrival, bPAEC were placed in T25 flasks with 5 ml of endothelial growth media (EGM; Lonza). When the bPAEC were 80–90% confluent, the bPAEC were passaged using trypsin-EDTA followed by trypsin neutralizing solution. The bPAEC were centrifuged at 1200 g for 5 min, and the bPAEC pellet was resuspended in EGM. Nine milliliters of EGM were placed in a T75 flask, and then 1 ml of the resuspended bPAEC pellet was added, after which the T75 flask was returned to the incubator at 37°C in 5% CO₂-balance air. bPAEC between passages 3 and 8 were used for these studies.

On the day of study, bPAEC were washed 3 times with 4 ml of HEPES balanced salt solution (HBSS; Lonza). Then, 5 ml of EGM were placed on the cells (control), and the bPAEC were returned to the incubator at 37°C in 5% CO₂-balance air for 24 h. In the L/T-treated bPAEC, 1.5 μg/ml LPS and 1.5 ng/ml TNF-α (both from Sigma, St Louis, MO) were included in the EGM as previously described (9, 32). After 24 h, the medium was removed and stored at −70°C. bPAEC were washed three times with 4 ml HBSS and lysed to either extract proteins or purify total RNA using Trizol (Life Technologies).

hPMVEC.

hPMVEC were purchased from Lonza and grown in six-well plates according to the manufacturer's recommendations using microvascular endothelial cell medium-2 (Lonza). On the day of study, the hPMVEC were washed three times with 2 ml of HBSS. Then, 1 ml of medium was placed on the hPMVEC (control), and cells were returned to the incubator at 37°C in 5% CO₂-balance air for 24 h. In the cytomix-treated hPMVEC, 1.5 μg/ml LPS, 3 ng/ml TNF-α, 2 ng/ml IL-1β, and 30 ng/ml IFN-γ (all from Sigma) were included in the medium. The components of the cytomix and their respective concentrations were selected from the literature (4, 6, 12, 15) and adjusted based on preliminary studies documenting augmented NO production without substantial cell loss. After 24 h, the medium was removed and frozen at −70°C. hPMVEC were washed three times with 4 ml of HBSS and treated with lysis buffer for protein extraction.

Protein isolation.

Protein was isolated from the bPAEC or hPMVEC as previously described (9, 32, 38). Briefly, cells were washed with HBSS and lysis buffer (0.2 M NaOH, 0.2% SDS) was added. Thirty minutes before use, the following protease inhibitors were added to each milliliter of lysis buffer: 0.2 μl aprotinin (10 mg/ml double distilled H₂O), 0.5 μl leupeptin (10 mg/ml double distilled H₂O), 0.14 μl pepstatin A (5 mg/ml methanol) and 5 μl of phenylmethylsulfonyl fluoride (34.8 mg/ml methanol). The cells were scraped and placed in sterile centrifuge tubes on ice. The supernatant was stored in 1-ml tubes at −70°C for Western blot analysis. Total protein concentration was determined by the Bradford method using a commercially available assay (BioRad, Hercules, CA).

RNA isolation.

RNA was isolated from bPAEC or hPMVEC as previously described (9, 31, 32). Briefly, Trizol (Life Technologies) was added to the cells and incubated for 5 min at room temperature. Chloroform (0.2 ml) was added, and the tubes were shaken for 15 s and then incubated at room temperature for 3 min. The mixture was centrifuged at 12,000 g for 15 min at 4°C. The supernatant was transferred to a fresh tube. Isopropyl alcohol (0.5 ml) was added, and the mixture was incubated at room temperature for 10 min and then centrifuged at 12,000 g for 15 min at 4°C. The supernatant was discarded, and the pellet was washed with 75% ethanol and centrifuged at 7,500 g for 5 min at 4°C. The supernatant was discarded, and the pellet was partially dried, dissolved in RNase-free water, and stored at −70°C.

Nitrite assay.

The samples of medium were assayed in duplicate for nitrite (NO₂⁻) using a chemiluminescence NO analyzer (model 280i; Sievers Instruments, Boulder, CO) as previously described (31, 38). Briefly, 100 μl of sample were placed in a reaction chamber containing a mixture of NaI in glacial acetic acid to reduce NO₂⁻ to NO. The NO gas was carried into the NO analyzer by a constant flow of He gas. The analyzer was calibrated with the use of a NaNO₂ standard curve.

Urea assay.

The samples of medium were assayed in duplicate for urea colorimetrically as previously described (9, 31, 32, 38). Briefly, 100 μl of sample were added to 3 ml of chromogenic reagent [5 mg thiosemicarbazide, 250 mg diacetyl monoxime, 37.5 mg FeCl₃ in 150 ml 25% (vol/vol) H₂SO₄, 20% (vol/vol) H₃PO₄] or the same reagents with 0.5 U urease added. After 1 h at 37°C, the mixtures were vortexed and then boiled at 100°C for 5 min. The mixtures were cooled to room temperature, and the difference in absorbance (530 nm) with and without urease was determined and compared with a urea standard curve.

Western blotting.

Cell lysates were assayed for Fyn, c-Src, Yes, Lyn, Lck, Hck, Blk, Fgr, or total phosphorylated STK protein using Western blot analysis as previously described (9, 31, 38). Aliquots of cell lysate were diluted 1:1 with SDS sample buffer, heated to 80°C for 15 min, and then centrifuged at 10,000 g at room temperature for 2 min. Aliquots of the supernatant were used for SDS-PAGE. The proteins were transferred to PVDF membranes and blocked overnight in PBS with 0.1% Tween 20 containing 5% non-fat dried milk and 3% albumin. The membranes were then incubated with the primary antibody against total Fyn, c-Src, Yes, Lyn, Lck, Hck, Blk, or Fgr (1:1,000; Transduction Laboratories, Lexington, KY). Active/tyrosine-phosphorylated STK was detected by Western blot analysis using an antibody (1:1,000) that recognizes phosphotyrosine at the activation site of all of the STKs (Cell Signaling). This site corresponds to tyrosine 416 on c-Src and is highly conserved throughout the STK family. The blots were then washed with PBS-0.1% Tween 20 with 1% non-fat dried milk. The membranes were then incubated with the biotinylated IgG secondary antibody (1:5,000; Vector Laboratories, Burlingame, CA) for 1 h and washed and then incubated with streptaviden-horseradish peroxidase conjugate (1:1,500; Bio-Rad) for 30 min. The bands of interest were visualized with enhanced chemiluminescence (Amersham, Piscataway, NJ) and quantified by densitometry (Sigma Gel, Jandel Scientific, San Rafael, CA). To control for protein loading, the blots were then stripped using a stripping buffer (each 100 ml contained 6.25 ml 1 M Tris·HCl pH 6.8, 20 ml 10% SDS, 0.7 ml 2-β-mercaptoethanol, and 73 ml double distilled H₂O). The blots were reprobed for β-actin (1:10,000; Abcam, Cambridge, MA) as described above.

RT-PCR.

RT-PCR was performed as previously described (9, 32). Briefly, 2 μg of total RNA were reverse transcribed in 2.5 μM dT₁₆ (Applied Biosystems, Foster City, CA), 20 U AMV-RT, 1 mM dNTP, 1× buffer (Promega, Madison, WI), and balance RNase-free water (total volume 40 μl). The samples were incubated in a PCR-iCycler (Bio-Rad) at 42°C for 60 min and 95°C for 5 min and stored at −20°C. Multiplex PCR was used to assess the expression of the arginase II and iNOS genes. PCR reactions (total volume of 50 μl) contained 5 μl of RT product, 1 mM MgCl₂, 1.25 U AmpliTaqGold (Applied Biosystems), 0.2 mM dNTP (Promega), and 15 μM forward (5′-TTGTGTTGATCTGGGTTGATGC-3′) and reverse (5′-TGCCTTCTCGATAGGTCAGTCC-3′) primers for arginase II or 0.3 μM of forward (5′-TGGACTTGGCTACGGAACTGG-3′) and reverse (5′-TTCTGGTGAAGCGTGTCTTGG-3′) primers for iNOS. The mixed samples were heated to 94°C for 4 min and then cycled as follows: 94°C for 1 min, 53°C for 1 min, and 72°C for 2 min for 35 cycles. The PCR products were sized by electrophoresis using 2.0% agarose gel electrophoresis and poststained with Syber gold (Molecular Probes, Eugene, OR) for 30 min. The gels were scanned and densitized using a MultiGenius Bio imaging system (Syngene, Frederick, MD), and band density analysis was performed on a personal computer with SigmaGel (Jandel Scientific) software. Preliminary PCR reactions run at various total cycle numbers between 20 and 45 demonstrated that the total of 35 cycles was well within the linear range for both arginase II and iNOS.

Statistical analysis.

Values are expressed as means ± SE. One-way ANOVA was used to compare the densitometry data between control and treated cells and to compare the effect of the additives on either nitrites or urea production. Significant differences were identified by a Neuman-Keuls post hoc test. Differences were considered significant when P < 0.05.

RESULTS

Effect of PP1 and dexamethasone on urea and NO production.

We have previously shown that L/T treatment of bPAEC increased both NO and urea production (9, 32). Therefore, in the first set of experiments, we determined the effect of the STK inhibitor PP1 and a well-described inhibitor of L/T-induced iNOS expression, dexamethasone, on the production of urea and NO during a 24-h incubation period. The samples were assayed for NO₂⁻ and urea production, which were normalized to total protein content. Treatment of bPAEC with L/T resulted in an approximately fivefold increase in urea production (Fig. 1A). The L/T-induced increase in urea production was not significantly affected by dexamethasone. However, PP1 completely inhibited the L/T-induced increase in urea production in bPAEC (Fig. 1A). The changes in urea production correlated with L/T-induced changes in arginase II mRNA levels, L/T treatment increased arginase II mRNA levels, dexamethasone had no apparent effect on the L/T-induced increase in arginase mRNA levels, and PP1 prevented the L/T-induced increase in arginase II mRNA levels (Fig. 1B). Treatment of the bPAEC with L/T resulted in an approximately fivefold increase in NO production (Fig. 1B). Treatment of bPAEC with either dexamethasone or PP1 significantly attenuated the L/T-induced increase in NO production (Fig. 1C).

Fig. 1.

A: Src family tyrosine kinase (STK) inhibition but not dexamethasone (dex) prevented LPS and tumor necrosis factor-α (L/T)-induced urea production. Urea production is shown in control, L/T-treated, L/T + 3 μM dex-treated, and L/T + 10 μM PP1-treated (STK inhibitor) bovine pulmonary arterial endothelial cells (bPAEC) after a 24-h incubation (n = 6 in each group). *Different from control, P < 0.01; #L/T+PP1 different from L/T, P < 0.01. B: PP1 but not dex prevented L/T-induced arginase II mRNA expression. Shown are a representative RT-PCR blot and a bar graph with mean densities for all 4 experiments. *Different from control, P < 0.05; #different from L/T, P < 0.05. C: PP1 and dex prevented L/T-induced NO production. Nitrite production is shown in control, L/T-treated, L/T + 3 μM dex-treated, and L/T + 10 μM PP1-treated bPAEC after a 24-h incubation (n = 4 in each group). *Different from control, P < 0.005; #different from L/T, P < 0.05.

The effect of PP2 and herbimycin A on arginase II and iNOS mRNA levels.

To determine the effect of STK inhibitors on arginase II and iNOS expression in bPAEC, the following study was done using PP2 and herbimycin A. bPAEC were incubated either in EGM, EGM with L/T, L/T with 10 μM PP2, or L/T with 10 μM herbimycin A. After 24 h, the cells were harvested for RNA and RT-PCR analysis for arginase II, and iNOS mRNA levels were carried out as described above. PP2 or herbimycin A treatment prevented the L/T-induced increases in arginase II mRNA (Fig. 2, A and B) and iNOS mRNA (Fig. 2, C and D) in bPAEC. These results demonstrate the central role of the STKs in cytokine-induced arginase and iNOS expression.

Fig. 2.

STK inhibition prevented L/T-induced arginase II and iNOS mRNA expression in bPAEC. A: representative RT-PCR for arginase II. Lane 1, control bPAEC; lane 2, L/T-treated bPAEC; lane 3, L/T + 10 μM PP2-treated bPAEC; lane 4, L/T + 30 μM herbimycin A (HA)-treated bPAEC. B: normalized densitometry data (arginase II density/18S rRNA density) for 6 separate experiments. *Different from control, P < 0.05. C: representative RT-PCR for iNOS. Lane 1, control bPAEC; lane 2, L/T-treated bPAEC; lane 3, L/T + 10 μM PP2-treated bPAEC; lane 4, L/T + 30 μM HA-treated bPAEC. D: normalized densitometry data (iNOS density/18S rRNA density) for 6 separate experiments. *Different from control, P < 0.001.

The use of specific STK siRNA in endothelial cells.

To utilize siRNA approaches to knock down specific STK family members in pulmonary endothelial cells, we needed to change to a species in which the genes encoding the specific STK family members had been cloned. We chose to study hPMVEC. To ascertain the validity of this model, we verified the induction of arginase and iNOS in response to inflammatory stimuli in hPMVEC. Interestingly, when hPMVEC were treated with L/T (as done in bPAEC), there was an increase in urea production (Fig. 3A), but no increase in NO production was observed (Fig. 3B). However, when the hPMVEC were treated with cytomix (LPS, TNF-α, IL-1β, and IFN-γ), there was an increase in both urea (Fig. 3C) and NO production (Fig. 3D). The relative hyporesponsiveness of the human iNOS gene to inflammatory stimuli is consistent with previous studies (6). Thus, for the remainder of the hPMVEC studies, cytomix was used to stimulate the cells.

Fig. 3.

L/T treatment increases urea production but not NO production in human pulmonary microvascular endothelial cells (hPMVEC), whereas cytomix (LPS, TNF-α, IL-1β, and IFN-γ) treatment increases both urea and NO production in hPMVEC. A: urea production by hPMVEC (n = 3 or 4) after a 24-h incubation with either vehicle (control) or L/T. B: nitrite production by hPMVEC (n = 3 or 4) after a 24-h incubation with either vehicle (control) or L/T. C: urea production by hPMVEC (n = 4–6) after a 24-h incubation with either vehicle or cytomix. *Cytomix different from control, P < 0.05. D: nitrite production by hPMVEC (n = 4–6) after a 24-h incubation with either vehicle or cytomix. *Cytomix different from control, P < 0.05.

As the second step in validating this model for delineating the role of specific STK members, we verified that the STK were activated by cytomix and that specific STK members were expressed in hPMVEC. hPMVEC were incubated with cytomix for 0, 0.5, 1, 2, 4, 6, and 24 h, and protein was harvested for Western blotting. Using an antibody for activated phosphorylated STK, which does not differentiate the various family members, we found that maximal STK activation occurred at 2 h after cytomix treatment (Fig. 4A). STK activation decreased at 4 and 6 h and was minimal by 24 h (Fig. 4A). Protein from hPMVEC was then harvested and probed for the various STK members using antibodies specific for various STK members, including Yes, c-Src, Fyn, Lyn, Lck, Hck, Fgr, and Blk. As expected, hPMVEC expressed easily detectable levels of Yes, c-Src, and Fyn protein and did not express detectable levels of Lck, Hck or Fgr protein (Fig. 4B). To our surprise, hPMVEC also expressed detectable levels of Lyn and Blk protein (Fig. 4B). To further corroborate this finding, we performed RT-PCR using primers specific for either Lyn or Blk and found that both Lyn and Blk mRNA was expressed in levels detectable in hPMVEC (Fig. 4C).

Fig. 4.

hPMVEC express at least 5 STK members, and individual family members can be specifically knocked down using short interfering RNA (siRNA). A: activation of the STK peaks at 2 h after treatment with cytomix in hPMVEC. Top blot was probed with an antibody against phosphorylated tyrosine 416 on Src or the equivalent tyrosine on the other STK (Cell Signaling Technology, Danvers, MA). Blot was then stripped and reprobed with an antibody against total c-Src (middle blot) and then stripped and reprobed with an antibody against β-actin (bottom blot) to confirm equal protein loading. B: hPMVEC express at least 4 STK members (Yes, c-Src, Fyn, Lyn, and Blk) but did not express detectable protein levels of Lck, Hck, or Fgr. Cell lysates from hPMVEC were probed with antibodies specific for Yes, c-Src, Fyn, Lyn, Blk, Lck, Hck, and Fgr. C: we did not expect to find Lyn or Blk protein expression in hPMVEC; therefore, we extracted RNA from cell lysates and performed RT-PCR using primers specific for Lyn or Blk. D: siRNA for the 4 STK demonstrated specificity for the particular STK. hPMVEC cells were either untreated (control) or treated with a scramble siRNA, Fyn siRNA, Yes siRNA, Lyn siRNA, or c-Src siRNA, and then Western blotting was performed for Fyn, Yes, Lyn, c-Src or β-actin. The scramble siRNA had no effect on any STK member, and the siRNA against specific STK members only knocked down the targeted STK. We were unable to purchase or develop an siRNA that reliably knocked down only Blk protein in hPMVEC.

Finally, we examined the specificity of our siRNA for selective knock down of the targeted STK family member. hPMVEC were incubated with the siRNA of interest (purchased from Dharmacon, Lafayette, CO); 48 h later, protein was harvested for Western blotting. We found that each siRNA was specific for the appropriate STK (Fig. 4D); i.e., the Fyn siRNA only knocked down the Fyn protein without effecting the protein levels of Yes, Lyn, or c-Src (Fig. 4D). Similar results were obtained with the siRNA for Yes, Lyn or c-Src (Fig. 4D).

Effect of specific STK siRNAs on urea and NO production in hPMVEC.

To examine the effect of knocking down a specific STK family member in cytomix-induced urea and NO production, hPMVEC were transfected with the siRNA of interest (Dharmacon) or a scramble siRNA as a control, using a transfection reagent provided by the manufacturer (Dharmacon). After 48 h, hPMVEC were washed three times and medium was added that contained either cytomix or the vehicle for cytomix. After a further 24-h incubation, the medium was harvested for urea and NO₂⁻ analysis, and the cells were lysed for determination of total protein.

Utilizing an siRNA against Fyn resulted in substantial knockdown of Fyn protein expression compared with untreated or scramble-treated cells, and cytomix treatment had no appreciable effect on the siRNA-mediated knockdown of Fyn protein (Fig. 5A). Addition of cytomix to nontransfected or scramble siRNA-transfected hPMVEC resulted in a significant increase in both urea (Fig. 5B) and NO (Fig. 5C) production. However, addition of cytomix to the hPMVEC that were transfected with the Fyn siRNA resulted in no appreciable change in either urea (Fig. 5B) or NO (Fig. 5C) production. These results clearly demonstrate a central role for Fyn in cytomix-induced urea and NO production in hPMVEC.

Fig. 5.

Knockdown of Fyn using siRNA prevented cytomix-induced urea and NO production in hPMVEC. A: Western blot demonstrating knockdown of Fyn using the Fyn siRNA (top blot); blot was then stripped and reprobed for β-actin (bottom blot). B: Fyn siRNA prevented cytomix-induced urea production in hPMVEC. Urea production is shown in hPMVEC treated with vehicle (control), a scramble siRNA, or a Fyn siRNA for 48 h and then either untreated (black bars) or treated with cytomix (gray bars) for 24 h; n = 5 in each group. *Cytomix different from untreated same condition, P < 0.05. +Cytomix-treated Fyn siRNA cells different from cytomix-treated scramble siRNA cells, P < 0.05. C: Fyn siRNA prevented cytomix-induced NO production in hPMVEC. Nitrite production is shown in hPMVEC treated with vehicle (control), a scramble siRNA, or a Fyn siRNA for 48 h and then either untreated (black bars) or treated with cytomix (gray bars) for 24 h; n = 5 in each group. *Cytomix different from untreated same condition, P < 0.05. +Cytomix-treated Fyn siRNA cells different from cytomix-treated scramble siRNA cells, P < 0.05.

Utilizing an siRNA against Yes resulted in substantial knockdown of Yes protein levels compared with nontransfected or scramble siRNA-transfected cells, and cytomix-treatment had no effect on the knockdown of Yes (Fig. 6A). Contrary to what we found in the Fyn siRNA experiments, cytomix triggered increases in urea (Fig. 6B) and NO (Fig. 6C) production in all three groups: nontransfected, scramble-transfected, and Yes siRNA-transfected hPMVEC. Thus knockdown of Yes had no apparent effect on cytomix-induced increases in urea or NO production in hPMVEC.

Fig. 6.

Knockdown of Yes using siRNA had no effect on cytomix-induced urea or NO production in hPMVEC. A: Western blot demonstrating knockdown of Yes using the Yes siRNA (top blot); blot was then stripped and reprobed for β-actin (bottom blot). B: Yes siRNA had no effect on cytomix-induced urea production in hPMVEC. Urea production is shown in hPMVEC treated with vehicle (control), a scramble siRNA, or a Yes siRNA for 48 h and then either untreated (black bars) or treated with cytomix (gray bars) for 24 h; n = 4 in each group. *Cytomix different from untreated same condition, P < 0.05. C: Yes siRNA had no effect on cytomix-induced NO production in hPMVEC. Nitrite production is shown in hPMVEC treated with vehicle (control), a scramble siRNA, or a Yes siRNA for 48 h and then either untreated (black bars) or treated with cytomix (gray bars) for 24 h; n = 5 in each group. *Cytomix different from untreated same condition, P < 0.05.

The expression of c-Src could be substantially knocked down with the specific c-Src siRNA (Fig. 7A). However, knockdown of c-Src had no significant effect on cytomix-induced urea (Fig. 7B) or NO (Fig. 7C) production in hPMVEC. Similarly, knockdown of Lyn (Fig. 8A) had no appreciable effect on cytomix-induced urea (Fig. 8B) or NO (Fig. 8C) production in hPMVEC. We were unable to either purchase or develop an siRNA that specifically and reliably knocked down Blk protein expression in hPMVEC. These results demonstrate that, of the STK examined in this study (Fyn, Yes, c-Src, and Lyn), only Fyn knockdown prevented cytomix-induced urea and NO production in hPMVEC.

Fig. 7.

Knockdown of c-Src using siRNA had no effect on cytomix-induced urea or NO production in hPMVEC. A: Western blot demonstrating knockdown of c-Src using the c-Src siRNA (top blot); blot was then stripped and reprobed for β-actin (bottom blot). B: c-Src siRNA had no effect on cytomix-induced urea production in hPMVEC. Urea production is shown in hPMVEC treated with vehicle (control), a scramble siRNA, or a c-Src siRNA for 48 h and then either untreated (black bars) or treated with cytomix (gray bars) for 24 h; n = 4 in each group. *Cytomix different from untreated same condition, P < 0.05. C: c-Src siRNA had no effect on cytomix-induced NO production in hPMVEC. Nitrite production is shown in hPMVEC treated with vehicle (control), a scramble siRNA, or a c-Src siRNA for 48 h and then either untreated (black bars) or treated with cytomix (gray bars) for 24 h; n = 5 in each group. *Cytomix different from untreated same condition, P < 0.05.

Fig. 8.

Knockdown of Lyn using siRNA had no effect on cytomix-induced urea or NO production in hPMVEC. A: Western blot demonstrating knockdown of Lyn using the Src siRNA (top blot); blot was then stripped and reprobed for β-actin (bottom blot). B: Lyn siRNA had no effect on cytomix-induced urea production in hPMVEC. Urea production is shown in hPMVEC treated with vehicle (control), a scramble siRNA, or a Lyn siRNA for 48 h and then either untreated (black bars) or treated with cytomix (gray bars) for 24 h; n = 4 in each group. *Cytomix different from untreated same condition, P < 0.05. C: Lyn siRNA had no effect on cytomix-induced NO production in hPMVEC. Nitrite production is shown in hPMVEC treated with vehicle (control), a scramble siRNA, or a Lyn siRNA for 48 h and then either untreated (black bars) or treated with cytomix (gray bars) for 24 h; n = 5 in each group. *Cytomix different from untreated same condition, P < 0.05.

DISCUSSION

The main findings of this study were that 1) pharmacological inhibition of the STK family blocked L/T-induced iNOS and arginase II mRNA expression and prevented L/T-induced urea and NO production in bPAEC; 2) hPMVEC express five of the STK family members, including Fyn, Yes, c-Src, Lyn, and Blk; 3) knockdown of Fyn using specific siRNA prevents cytomix-induced urea and NO production in hPMVEC; and 4) knockdown of Yes, c-Src, or Lyn had no appreciable effect on cytomix-induced urea and NO production in hPMVEC. These findings support our hypothesis that inflammatory stimuli lead to increased urea and NO production through STK activation. Furthermore, activation of the specific STK family member Fyn appears to be responsible for the cytomix-induced urea and NO production in hPMVEC. Our data demonstrate that activation of specific STK family members in pulmonary endothelial cells can lead to unique and specific cellular responses.

We found that inhibiting STK activation using either PP1, PP2, or herbimycin A resulted in decreased cytokine-induced urea and NO production as well as decreased cytokine-induced arginase II and iNOS mRNA expression in bPAEC. To the best of our knowledge, we are the first to demonstrate that the pharmacological inhibition of the STK prevents cytokine-induced urea production and arginase expression in pulmonary endothelial cells. Consistent with our findings, it has recently been found that PP1 prevented LPS-induced iNOS expression in isolated rat tail arteries (23). The fact that nonspecific pharmacological inhibitors of the STK prevent both cytokine-induced arginase and iNOS expression in pulmonary arterial endothelial cells clearly indicates a central role for STK in the signaling mechanisms mediating arginase and iNOS expression. The L/T-induced, STK-dependent signaling pathway that leads to iNOS expression may include NF-κB activation, since dexamethasone treatment attenuated L/T-induced NO production. A role for NF-κB activation in cytokine-induced iNOS expression in endothelial cells has been previously reported (24). Although somewhat speculative, it may be that the NF-κB pathway is less important in cytokine-induced arginase expression, since dexamethasone treatment had less of an effect on L/T-induced urea production and arginase II mRNA expression. This observation would suggest that arginase and iNOS induction are regulated differently downstream of the STKs. Together, these data suggest a critical role for the STKs in an important endothelial cell physiological process: arginase and iNOS induction in response to LPS and/or cytokines. Interestingly, it has been reported that, in human umbilical vein endothelial cells (HUVEC), hepatocyte growth factor resulted in increased NO production and iNOS expression that was unaffected by PP1 treatment (30, 35). In light of the L/T data, the hepatocyte growth factor data suggest that the role of STK in iNOS induction in endothelial cells may depend on which cellular receptors are activated by the particular stimulus.

The STKs have been extensively studied in hematopoietic and tumor cells and cell lines (29). The STKs are involved in a variety of cell functions, including responses to inflammatory stimuli (29, 33). Phosphorylation of a tyrosine residue in the activation loop (Tyr⁴¹⁸ in human c-Src) leads to activation, whereas phosphorylation of a tyrosine residue in the COOH-terminal tail (Tyr⁵²⁹ in human c-Src) results in autoinhibition (33). Using an antibody specifically recognizing the phosphorylated tyrosine at the activation loop, we found that the hPMVEC STK are activated by cytomix with maximal activation occurring at ∼2 h. This finding is consistent with two recent studies examining STK activation using TNF-α stimulation: one in human pulmonary arterial endothelial cells (3) and another in hPMVEC (43). As expected, we found expression of Fyn, Yes, and c-Src in hPMVEC; these STKs have previously been detected in endothelial cells (3, 5, 8, 44). Interestingly, although endothelial cells may express Fyn, Yes, and c-Src, the relative subcellular localization of these proteins may vary. For example, Chen et al. (8) found that Fyn and Yes colocalized with flotillin-1 and caveolin-1, whereas c-Src did not, suggesting that of the three STKs studied, only Fyn and Yes were found in lipid rafts in human retinal endothelial cells. The subcellular distributions of STK members may have implications in relation to the unique cellular functions of the respective STK members. Surprisingly, we also found that hPMVEC express Lyn and Blk. To the best of our knowledge, this is the first description of the presence of Lyn and Blk in pulmonary endothelial cells. The finding of Lyn in endothelial cells is consistent with a previous study that found that fetal bovine brain endothelial cells expressed Lyn but that adult bovine brain endothelial cells did not express Lyn (1). We could find no studies reporting Blk expression in endothelial cells. It is unclear at present what physiological functions Lyn and Blk play in endothelial cells. Further studies are needed to address their roles in endothelial cell physiology.

Our findings clearly implicate Fyn as the predominant STK responsible for arginase and iNOS induction following cytokine stimulation in endothelial cells. It has been previously suggested that the presence in a cell of many STK members allows for redundancy in STK signaling; i.e., in this scenario the loss of one STK member will not result in the loss of STK signaling (29, 44). However, in agreement with our findings, recent data suggest that the individual STK members may have unique signaling functions in endothelial cells. In a brain microvascular endothelial cell line, expression of a kinase-dead Fyn mutant, but not of a kinase-dead Src mutant, blocked FGF-2-induced capillary morphogenesis (16). In human aortic endothelial cells, it has been found that a green tea polyphenol (epigallocathechin gallate) causes increased eNOS phosphorylation and NO production; with the use of siRNAs against either c-Src or Fyn, it was shown that Fyn but not c-Src was necessary for this increase in eNOS phosphorylation and NO production (18). With the use of knock-out mice (fyn^−/−, c-src^−/−, or yes^−/−), it has been found that the vascular permeability induced by VEGF in the skin required Src and Yes but not Fyn (10). Werdich and Penn (44) using siRNA techniques found that Fyn, c-Src, and Yes were involved in various aspects of VEGF affects on human retinal microvascular endothelial cells; for example, all three were required for VEGF-induced cell proliferation, whereas c-Src had no effect on VEGF-induced cell migration, Yes was necessary for migration, and Fyn was inhibitory for migration. Angelli et al. (3) found in human lung microvascular endothelial cells using Fyn, Yes, and c-Src siRNAs that only the Fyn siRNA attenuated TNF-α-induced loss of barrier function. Kinney et al. (20) found that, in HUVEC, VEGF-induced mitogen-activated protein kinase phosphatase-1 (MKP-1) expression was dependent on c-Src but not on Fyn, whereas thrombin-induced MKP-1 expression depends on Fyn but not on c-Src. Thus it appears that the various STK members may have important and unique functions within endothelial cells. Our results clearly indicate that, in hPMVEC, only Fyn activity is necessary for cytokine-induced urea and NO production.

The endothelial cells from the main pulmonary artery responded to L/T treatment with an increase in both urea and NO production, whereas in the pulmonary microvascular endothelial cells L/T treatment did not result in an increase in NO production. Indeed, to get an increase in NO production in the microvascular endothelial cells required the addition of IL-1β and IFN-γ. This difference in iNOS response to L/T between arterial and microvascular endothelial cells may represent species differences between bovine and human cells. We have previously shown robust iNOS induction with L/T in bPAEC (9), whereas Chan et al. (6) found that HUVEC were relatively resistant to cytokine upregulation of iNOS expression. Alternatively, this may represent segmental differences in pulmonary endothelial cell phenotypes; i.e., endothelial cells from the main pulmonary artery differ from endothelial cells from the lung microvasculature (11). Indeed, King et al. (19) demonstrated differences in structure and function between rat main pulmonary artery and rat pulmonary microvascular endothelial cells. Similarly, Kelly et al. (17) and Stevens et al. (39) have demonstrated very different barrier responses to either elevated Ca²⁺ or cAMP between endothelial cells from the main pulmonary artery vs. endothelial cells from the lung microvasculature. Thus it may be that segmental differences in cytokine-induced iNOS expression exist between endothelial cells of conduit vessels (such as the pulmonary artery) and endothelial cells from small alveolar vessels (PMVEC). This concept is consistent with a report by Geiger et al. (12) that demonstrated greater iNOS mRNA expression and NO production after LPS + IFN-γ treatment in rat pulmonary arterial cells vs. rat pulmonary microvascular endothelial cells.

In conclusion, our results demonstrate that the STK Fyn appears to play a critical role in cytokine-induced urea and NO production in hPMVEC. We speculate that various cell surface receptors (for example, toll-like receptors, TNF receptors) interact with docking proteins, such as MyD88 and TRAF6, to activate Fyn, which then leads to the downstream signal cascade(s) resulting in increased arginase and iNOS expression and activity. However, further studies are needed to elucidate the exact mechanisms involved in cytokine-induced, Fyn-dependent arginase and iNOS expression. Although speculative, given that Fyn gene polymorphisms have recently been described (41), alterations in Fyn activity may underlie differing human susceptibilities to inflammatory lung diseases. Thus Fyn may represent a unique and specific therapeutic target for inhibiting cytokine-induced arginase and/or iNOS expression in inflammatory lung diseases.

GRANTS

This study was supported by an Advancing Newborn Medicine Fellowship Grant (R. Chang), National Heart, Lung, and Blood Institute Grant HL-075261 (L. D. Nelin), and National Institute of Allergy and Infectious Diseases Grant AI-057798-01 (Y. Liu).