Inducible NOS mediates CNP-induced relaxation of intestinal myofibroblasts

Abstract

Contraction of intestinal myofibroblasts (IMF) contributes to the development of strictures and fistulas seen in inflammatory bowel disease, but the mechanisms that regulate tension within these cells are poorly understood. In this study we investigated the role of nitric oxide (NO) signaling in C-type natriuretic peptide (CNP)-induced relaxation of IMF. We found that treatment with ODQ, a soluble guanylyl cyclase (sGC) inhibitor, or N^G-nitro-l-arginine (l-NNA) or N^G-monomethyl-l-arginine (l-NMMA), inhibitors of NO production, all impaired the relaxation of human and mouse IMF in response to CNP. ODQ, l-NNA, and l-NMMA also prevented CNP-induced elevations in cGMP concentrations, and l-NNA or l-NMMA blocked CNP-induced decreases in myosin light phosphorylation. IMF isolated from transgenic mice deficient in inducible nitric oxide synthase (iNOS) had reduced relaxation responses to CNP compared with IMF from control mice and were insensitive to the effects of ODQ, l-NNA, and l-NMMA on CNP treatment. Together these data indicate that stimulation of sGC though NO produced by iNOS activation is required for maximal CNP-induced relaxation in IMF.

proper maintenance of tension is critical for numerous cellular processes, including cytokinesis, migration, and contraction. Failure to control the tension of some intestinal cell types contributes to development of motility disorders and inflammatory bowel disease (IBD) (27, 39). Intestinal myofibroblasts (IMF) exist between the matrix and smooth muscle cells of the muscularis mucosae and are characterized by the expression of α-smooth muscle actin, desmin, and S100A4 (8, 12, 27). Several factors associated with the development of IBD (e.g., endothelin-1, transforming growth factor-β, and tumor necrosis factor-α) stimulate the contraction or migration of IMF, suggesting that changes in IMF cell tension contribute to IBD-associated pathology (4, 10, 15). IMF also control villous contractions that propel lymph toward lymph nodes and promote the resolution of intestinal injury through de novo production of extracellular matrix and tissue remodeling (7, 11, 28, 39). Therefore, understanding the pathways that regulate tension within IMF could lead to the development of new therapeutics for IBD.

Nitric oxide (NO) production in the intestine is implicated in the pathogenesis of IBD (1). Both inducible nitric oxide synthase (iNOS) and the NO reaction product nitrotyrosine are found at sites of intestinal inflammation in patients with ulcerative colitis or Crohn's disease but are undetected in nearby noninflamed tissue (35). NO has been proposed to mediate a wide range of effects on numerous cell types in the intestine, including the regulation of vasodilation, neurotransmission, and apoptosis (6). In other tissues such as the liver and lung, exposure of myofibroblasts to NO stimulates a relaxation response (i.e., a decrease in the amount of force placed upon a substrate) (14, 26). However, regulation of NO signaling in IMF, specifically in relation to control of cellular tension, is unclear.

We previously showed that C-type natriuretic peptide (CNP), a factor typically associated with relaxation in cardiac muscle cells and smooth muscle cells, promotes relaxation of IMF (5). CNP and other natriuretic peptides (A-, B-, and C-type) function by activating the guanylyl cyclase (GC) activity of their membrane-bound (i.e., particulate) receptors. Natriuretic peptide receptors (NPRs) convert GTP to cGMP, which promotes cellular relaxation through activation of protein kinase G (PKG) and/or alterations in calcium mobilization (33). NO also stimulates relaxation by increasing the intracellular concentration of cGMP, but unlike NPRs, NO is primarily thought to signal through intracellular soluble guanylyl cyclase (sGC) (29). Therefore, membrane-bound GC and sGC are regarded as two parallel pathways responsible for the production of cGMP and cell relaxation. In this study we investigated whether sGC and NO production mediate the effects of CNP on relaxation in IMF. We found that CNP-induced relaxation in IMF is controlled by a pathway involving NO production by iNOS and subsequent activation of sGC.

MATERIALS AND METHODS

Materials.

CNP, 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ), N^G-nitro-l-arginine (l-NNA), N^G-monomethyl-l-arginine (l-NMMA), and monoacetate salt were obtained from Calbiochem (La Jolla, CA). Lipopolysaccharide (LPS), interferon-γ (IFN-γ), 4′,6′-diphenyleneiodonium chloride (DPI), and cytochalasin D were from Sigma Chemical (St. Louis, MO). 2′,5′-ADP-Sepharose 4B, Biotrak cGMP competitive enzyme immunoassay (EIA) system, enhanced chemiluminescence (ECL Plus) reagent, and anti-rabbit secondary antibodies were from GE Healthcare (Piscataway, NJ). iNOS, neuronal nitric oxide synthase (nNOS), and endothelial nitric oxide synthase (eNOS) primers used for RT-PCR and iNOS primers used for genotyping iNOS^-/- mice and detecting iNOS expression were from Integrated DNA Technologies (Coralville, IA). For Western blot analysis, an antibody directed against iNOS was obtained from BD Biosciences, and the β-actin (clone sc-1616) antibody was from Santa Cruz Biotechnology (Santa Cruz, CA). Type I collagen was from Millipore Upstate (Temecula, CA).

Cell culture.

Co18 cells were obtained at passage five from the American Type Culture Collection (ATCC, Rockville, MD). These cells maintain most known characteristics of intestinal subepithelial myofibroblasts in situ after several passages (34, 41). Cells were grown as previously described and were discarded by passage 15, when they have been reported to undergo significant phenotypic changes (15, 34, 41). RAW 264.7 cells were obtained from ATCC and cultured in DMEM containing 10% of FBS.

Primary IMF isolation.

Mouse primary IMF were prepared with modification of a previously described method (38). Mice were housed within an environmentally controlled specific pathogen-free animal facility at the University of California, San Francisco, and all experiments were approved by the Institutional Animal Care and Use Committee of the University of California, San Francisco. Briefly, the small intestines of 25- to 35-day-old C57BL/6 mice or C57BL/6 iNOS^-/- mice (20) were removed and flushed with washing medium (Dulbecco's modified Eagle's medium, DMEM H-21, University of California, San Francisco, Cell Culture Facility) plus penicillin/streptomycin. Intestines were cut into ∼1 × 1 mm² pieces and washed two times with washing medium. Tissue pieces were then digested in 10∼15 ml of washing medium containing 4 U/ml Dispase and 300 U/ml collagenase I (Worthington Biochemicals, Lakewood, NJ) at 25°C for 25 min. Digestion medium with single cells released from the tissue was collected into a 15-ml conical tube, and two volumes of washing medium were added to the remaining digestion mixture. Tissue pieces were then broken into a single cell suspension by vigorous pipetting. Cells and tissue debris were spun down and washed two times with washing medium plus 5% FBS and one time with growth medium, which consisted of DMEM plus 10% of FBS and penicillin/streptomycin. Cells were plated in two T-75 tissue culture flasks or 12 to 14 P-35 tissue culture plates in growth medium. Growth medium was changed after 48 h to remove nonadherent and dead cells. Adherent cells were maintained in growth medium, and media was replaced two times a week. Subconfluent cells were for experiments at passages 3–5.

RT-PCR measurement of NOS isoforms.

RNA was isolated with an RNeasy MiniKit (Qiagen, Valencia, CA) according to the manufacturer's instructions. RT-PCR reactions were conducted on a GeneAmp PCR System 9700 (Applied Biosystems, Carlsbad, CA) with a OneStep RT-PCR Kit (Qiagen) and the following primers: mouse nNOS, forward, 5′-ACC AGC TCT TCC CTC TAG CC-3′ and reverse, 5′-ATG GGT ACT TCC AGC ACC AG-3′; mouse iNOS, forward, 5′-CCT TGT TCA GCT ACG CCT TC-3′ and reverse, 5′-GGC TGG ACT TTT CAC TCT GC-3′; mouse eNOS, forward, 5-TCT TCG TTC AGC CAT CAC AG-3′ and reverse, 5′-CAC AGG GAT GAG GTT GTC CT-3′.

Western blot.

Western blot was performed as previously described with modification (36). Detection of iNOS in IMF was performed with anti-mouse iNOS (clone 2/iNOS; BD Pharmingen, San Jose, CA) after lysates from cells were affinity precipitated with 2′5′-ADP-Sepharose 4B per the manufacturer's instructions. NPR Western blots were performed by AllCells (Emeryville, CA) with the following antibodies: anti-mouse NPR-A (clone ab116680; Abcam, Cambridge, MA), anti-mouse NPR-B (clone sc-16870; Santa Cruz Biotechnology), and anti-mouse NPR-C (clone sc-16871; Santa Cruz Biotechnology). An antibody directed against actin (clone I-19; Santa Cruz Biotechnology) was used as a control to measure total protein loaded in each lane.

Relaxation measurements.

Relaxation of IMF was measured in a direct and quantitative fashion as previously described (5, 15). Briefly, myofibroblasts (1 × 10⁶ Co18 cells and 0.4 × 10⁶ primary mouse IMF) were cultured for 3 days in an elastic gel made of type I collagen. Collagen gels with IMF were attached to an isometric force transducer, stretched to their original length, and submerged in an organ bath containing 35 ml of physiological salt solution buffered with HEPES. Organ baths were maintained at 37°C, and gels were allowed 1 h to achieve a stable baseline tension. CNP or inhibitors were added directly to the bath as indicated. Contractile tension was recorded digitally using an analog-to-digital converter (DAQ-500; National Instruments, Austin, TX) and Virtual Bench Logger software (National Instruments). Relaxation was defined as a sustained decrease (i.e., negative change) in contractile tension.

cGMP measurement.

IMF were cultured overnight in 96-well Microtest III tissue culture plates (BD Falcon; Becton-Dickson, Franklin Lakes, NJ). Cells were then washed two times and incubated in serum-free media for 1 h at 37°C before cGMP measurement. Intracellular cGMP levels were assayed with a commercial cGMP EIA system and an AutoReader II (Ortho-Clinical Diagnostic, Raritan, NJ) at an optical density of 450 nm as previously reported (5). Measurements were begun 15 min after CNP or carrier was added to each well. A standard curve for cGMP was generated by plotting the optical density values relative to log cGMP concentrations (fmol/well) of the standards. The cGMP concentrations of the samples were then calculated from the standard curve.

Myosin light chain phosphorylation measurements.

Myosin phosphorylation was determined as previously described (5, 15). Briefly, IMF cultures were washed two times and incubated in serum-free media for 1 h before each experiment. Lysates were separated by urea-glycerol-acrylamide gel electrophoresis. After immunoblot with an antibody directed against the myosin regulatory light chain (MLC), unphosphorylated and phosphorylated forms of the MLC were detected by ECL. Bands were quantified by densitometry using image analysis software (Scion, Frederick, MD). Myosin phosphorylation is presented as the sum of phosphorylated MLCs as a percentage of the total MLC signal.

Statistical analysis.

Data are presented as means ± SD. Statistical comparisons were made by Student's t-test. Statistical significance was defined as P < 0.05.

RESULTS

CNP induces relaxation of human IMF through sGC and NO production.

To determine if CNP induces relaxation of IMF through sGC, we treated human Co18 IMF grown in three-dimensional collagen gels with ODQ, a specific heme-site inhibitor of sGC (32). Treatment with ODQ had no effect on the resting tension of IMF (Fig. 1A, inset), but incubation with ODQ for 15 min followed by the addition of CNP significantly inhibited the CNP-induced relaxation response (Fig. 1A). The addition of the actin microfilament destabilizer cytochalasin D further reduced cellular tension by 182 ± 18 dynes, indicating that CNP-induced relaxation comprised about 22% of the total cellular tension of IMF within the collagen gels. Because activation of sGC by NO drives the conversion of GTP to cGMP, we next tested whether NO mediated relaxation through production of cGMP. We pretreated IMF in collagen gels for 15 min with l-NNA or l-NMMA, l-arginine analogs that inhibit NO production (30), and then measured the change in cellular tension in response to CNP. Both l-NNA and l-NMMA had no effect on the baseline tension of IMF but dose-dependently impaired CNP-induced relaxation (Fig. 1B). These findings suggest that CNP induces the relaxation of IMF through a pathway involving activation of sGC and production of NO.

Fig. 1.

Inhibition of soluble guanylyl cyclase (sGC) or nitric oxide (NO) production impairs intestinal myofibroblast (IMF) relaxation in response to C-type natriuretic peptide (CNP). Contractile force measurements of human Co18 IMF grown in 3-dimensional (3D) collagen gels after the following treatments are shown. A: treatment with 500 nM CNP followed by the addition of 1 μM cytochalasin D or pretreatment for 15 min with 10 μM 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ) followed by treatment with 500 nM CNP. Representative contractile force tracings for each of these conditions are shown below the bar graph. The scale bar indicates the duration of force measurement and change in tension (measured in dyne units) for each tracing. B: pretreatment for 15 min with the indicated doses of N^G-nitro-l-arginine (l-NNA) or N^G-monomethyl-l-arginine (l-NMMA) followed by the addition of 500 nM CNP. Representative contractile force tracings for IMF pretreated with 5 μM l-NNA or l-NMMA followed by the addition of 500 nM CNP are shown below the bar graph. Data are presented as means ± SD.

To determine if decreased CNP-induced relaxation after treatment with l-NNA and l-NMMA was the result of diminished activity of sGC and regulation of the contractile apparatus, we measured the effects of these compounds on cGMP production and myosin light chain phosphorylation. CNP-induced production of cGMP was dose-dependently blunted by both l-NNA and l-NMMA (Fig. 2A). Elevated cGMP signals to the contractile apparatus of cells by activating PKG, which then regulates intracellular calcium stores and myosin light chain phosphatase (25). Pretreatment of IMF with l-NNA or l-NMMA significantly abrogated CNP-induced decreases in myosin light phosphorylation to levels that were ∼16% below basal untreated cells (Fig. 2B). Together these data indicate that CNP triggers human IMF relaxation in part through a pathway involving stimulation of sGC and reduced myosin activity.

Fig. 2.

Inhibitors of NO production block synthesis of cGMP and prevent decreased myosin light chain (MLC) phosphorylation in response to CNP. A: measurements of cGMP in human Co18 IMF after pretreatment for 15 min with the indicated doses of l-NNA or l-NMMA followed by the addition of 500 nM CNP. B: quantification of MLC phosphorylation in human Co18 IMF after pretreatment for 15 min with 2 μM of l-NNA or l-NMMA followed by the addition of 500 nM CNP. *P ≤ 0.05 compared with MLC phosphorylation levels after CNP treatment. Data are presented as means ± SD.

sGC and NO production are required for CNP-induced relaxation in mouse IMF.

To determine if this pathway for CNP-induced relaxation also operates in primary cells, we developed a procedure to isolate primary mouse IMF. IMF freshly isolated from mouse intestine began to express the myofibroblast markers α-smooth muscle actin and S100A4 after 5 days in culture (Fig. 3A). By 12 days of culture, nearly every cell stained positive for these markers. These cells also displayed characteristic spindle-shaped myofibroblast morphology by day 8 (Fig. 3A). To determine if primary mouse IMF express NPR-A, -B, or -C, we performed Western blots for these proteins using IMF cell lysates. Because NPR-A, -B, and -C are expressed in the brain, we also tested mouse brain tissue as a positive control for the Western blots (42). Mouse IMF expressed relatively low, but detectable, levels of NPR-A and NPR-B compared with brain tissue (Fig. 3B). In contrast, IMF expressed high levels of NPR-C compared with NPR-A or NPR-B (Fig. 3B). These data indicate that primary mouse IMF express receptors for CNP.

Fig. 3.

Primary mouse IMF express myofibroblast markers and natriuretic peptide receptor (NPR)-A, NPR-B, and NPR-C. A: fluorescence images of primary mouse IMF isolated as described in materials and methods and stained with antibodies to α-smooth muscle actin and S100A4 at the indicated days in culture. Phase contrast (PC) images are shown below the fluorescence images and correspond to the same fields. B: Western blots of primary mouse IMF and brain lysates with antibodies to NPR-A, NPR-B, and NPR-C. The three IMF lanes correspond to IMFs prepared from three different mice and are shown next to whole brain lysates from two different mice for comparison of expression levels. Below the NPR-A, NPR-B, and NPR-C blots are images of the same blots, stripped of antibody and relabeled with an antibody to actin to show the amount of total protein present in each lane.

We then tested whether CNP induces relaxation of primary mouse IMF. Addition of CNP to primary mouse IMF in collagen gels triggered a rapid reduction in cellular tension of 15 ± 3 dynes with a time course that was nearly identical to human IMF (Fig. 4). We next pretreated cells with ODQ or l-NMMA and measured the response to CNP. Both ODQ and l-NMMA impaired CNP-induced relaxation, but only l-NMMA had a statistically significant effect that resulted in a near 60% inhibition of CNP-induced relaxation (Fig. 4). These data suggest that CNP induces relaxation through a similar pathway in human and mouse IMF.

Fig. 4.

NO production and sGC activation modulate mouse IMF relaxation in response to CNP. Contractile force measurements of primary mouse IMF grown in 3D collagen gels after treatment with 500 nM CNP alone or pretreatment for 15 min with 10 μM ODQ or 2 μM l-NMMA followed by treatment with 500 nM CNP. *P ≤ 0.05 compared with CNP alone. Representative contractile force tracings of IMF treated with CNP alone or pretreated for 15 min with 10 μM ODQ or 2 μM l-NMMA followed by the addition of 500 nM CNP are shown below the bar graph. Data are presented as means ± SD.

Mouse IMF express iNOS, but not eNOS or nNOS.

To investigate which NOS isoform(s) was responsible for mediating CNP-induced relaxation, we performed RT-PCR for eNOS, nNOS, and iNOS in primary mouse IMF. Because expression of all isoforms occurs in the brain, mouse brain tissue was used as a positive control (3). We observed mRNA expression of eNOS, nNOS, and iNOS in brain samples, but only the iNOS isoform was found in primary mouse IMF (Fig. 5A). To test if iNOS mediates CNP-induced relaxation, we obtained transgenic iNOS-deficient mice, which carry a neomycin cassette in place of exons 12 and 13 of the iNOS gene (20). Transcripts for iNOS were undetectable by RT-PCR in IMF isolated from iNOS-deficient mice both under basal conditions and after stimulation with CNP or LPS and IFN-γ (Fig. 5B). However, iNOS transcripts were highly expressed in RAW 264.7 cells, a macrophage cell line, and wild-type IMF (Fig. 5B). RAW 264.7 cells and wild-type IMF also expressed iNOS protein that was found in pellet lysate fractions, but iNOS protein was absent in both supernatant and pellet lysate fractions of IMF from iNOS-deficient mice (Fig. 5C). These results suggest that iNOS is the solitary NOS isoform expressed in primary mouse IMF under basal and cytokine-stimulated conditions.

Fig. 5.

Primary mouse IMF express inducible nitric oxide synthase (iNOS), but not endothelial nitric oxide synthase (eNOS) or neuronal nitric oxide synthase (nNOS). A: images of gels stained with ethidium bromide after electrophoresis of RT-PCR products generated with RNA from primary mouse IMF and primers specific to eNOS, nNOS, or iNOS. RNA isolated from murine brain and reactions with no RNA input (H₂O) were included as controls. The approximate size (no. of base pairs) of each band is indicated next to the images. Images for each set of primers are taken from a different part of the same gel, as indicated by a space between the IMF and brain samples. B: RT-PCR samples with RNA isolated from control (+/+) or iNOS-deficient (−/−) primary mouse IMF cells after treatment with combinations of 100 ng/ml interferon-γ (IFN-γ) and 100 ng/ml lipopolysaccharide (LPS), or 2 μM CNP for 24 h. RNA isolated from RAW 264.7 cells before and after treatment with 100 ng/ml IFN-γ and 100 ng/ml LPS was included as a control. C: Western blots with antibodies specific to iNOS or β-actin using 2′5′-ADP Sepharose 4B pull-down (PD) or supernatant (SN) fractions from control (+/+) or iNOS-deficient (−/−) IMF. RAW 264.7 cell lysate was included as a positive control for iNOS expression.

Impaired CNP-induced relaxation responses in iNOS-deficient IMF.

To determine whether iNOS mediates CNP-induced relaxation in primary mouse IMF, we measured cellular tension in wild-type and iNOS-deficient IMF grown in collagen gels following CNP treatment. CNP decreased cellular tension by 11 ± 1 dynes in wild-type IMF, and this response was attenuated in a dose-dependent fashion with l-NMMA (Fig. 6A). However, iNOS-deficient IMF had a significantly reduced response to CNP (7 ± 1 dynes) that was not further inhibited by l-NMMA (Fig. 6A). We next examined if the decreased response to CNP in iNOS-deficient IMF was the result of a lack of cGMP production. As we observed in human IMF, primary mouse IMF dose-dependently synthesized cGMP following the addition of CNP (Fig. 6B). Conversely, iNOS-deficient IMF produced significantly lower concentrations of cGMP under basal conditions, and these amounts were only marginally enhanced by the addition of CNP (Fig. 6B). We next determined if CNP-induced cGMP generation in wild-type and iNOS-deficient IMF was modulated by ODQ, l-NMMA, and DPI, an iNOS inhibitor (23). Treatment with ODQ and DPI blocked CNP-induced cGMP generation to concentrations similar to those observed in iNOS-deficient IMF (Fig. 6C). l-NMMA was less potent, but did diminish cGMP production in wild-type IMF (Fig. 6C). Collectively, these data indicate that primary mouse IMF relax in response to CNP through a pathway involving activation of iNOS and sGC (Fig. 7).

Fig. 6.

IMF deficient in iNOS have an attenuated CNP relaxation response that is not further reduced by inhibition of sGC or NO production. A: contractile force measurements of wild-type (WT) and iNOS-deficient primary mouse IMF grown in 3D collagen gels after treatment with 500 nM CNP or pretreatment for 15 min with the indicated doses of l-NMMA followed by the addition of 500 nM CNP. The no. of replicates from at least 3 independent experiments is shown next to each data point. B: measurements of cGMP in wild-type and iNOS-deficient primary mouse IMF after treatment with the indicated doses of CNP. C: measurements of cGMP in wild-type and iNOS-deficient primary mouse IMF after pretreatment for 15 min with the indicated doses of ODQ, l-NMMA, or 4′,6′-diphenyleneiodonium chloride (DPI), and followed by addition of 500 nM CNP. *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.005 for iNOS-deficient primary mouse IMF compared with wild-type IMF. Data are presented as means ± SD.

Fig. 7.

Model of CNP-induced relaxation in IMF. Activation of NPR-B or NPR-C by CNP triggers a pathway that promotes relaxation of IMF. NPR-B, but not NPR-C, contains a transmembrane receptor with an intracellular GC subunit (i.e., particulate GC) that synthesizes cGMP from GTP. Through an as yet identified mechanism involving NPR-B or NPR-C, iNOS is activated and produces NO from l-arginine. NO stimulates sGC to produce cGMP that in turn activates protein kinase G (PKG). PKG phosphorylates myosin light chain phosphatase (MLCP), which dephosphorylates myosin light chain and results in cellular relaxation.

DISCUSSION

In this study we confirm a previous finding that human IMF Co18 cells relax in response to CNP and show for the first time that primary mouse IMF behave in a similar fashion. Our data demonstrate that CNP-induced relaxation of both cell types was reduced by multiple compounds that either inhibited NO synthesis or blocked sGC. Moreover, we found that primary IMF from iNOS-deficient mice had impaired responses to CNP and were also insensitive to compounds that target NOS. Taken together, these data suggest a model whereby IMF relax in response to CNP in part through the rapid activation of sGC by iNOS-generated NO.

A key aspect to this model is that iNOS is activated following ligation of a CNP receptor. Of the three members in the NPR family, only two receptors are known to bind CNP, NPR-B and NPR-C (16, 21). NPR-B contains an intracellular GC domain that synthesizes cGMP from GTP and is believed to be responsible for most of the vasoactive effects of CNP (16). In contrast, NPR-C lacks an intracellular GC domain and is primarily thought to function as a scavenger receptor, although it can couple to inhibitory G protein receptors and regulate adenylyl cyclase activity (31). Our finding that NPR-C is highly expressed compared with NPR-A or NPR-B is in contrast to another myofibroblast cell type, the hepatic stellate cell, where NPR-B is expressed while NPR-C is not (37). Future studies should focus on elucidating the relative contribution of NPR-B and NPR-C to CNP-induced IMF relaxation and how these receptors function to activate iNOS.

Regulation of iNOS differs from eNOS and nNOS, which are constitutively expressed and quickly activated by increases in calcium concentration (3). The iNOS isoform binds tightly to calcium/calmodulin and is generally thought to be constitutively active once expressed (2). However, recent reports indicate that iNOS activity can be modulated by phosphorylation at Ser⁷⁴⁵ (19, 43) and Tyr¹⁰⁵⁵ (40), or binding of activated calcium/calmodulin kinase II (9). Rapid degradation of iNOS following ubiquitination is also known to regulate NO production (18). In the present study, relaxation began within 1 min of CNP treatment and in most cases was complete by 10 min. Therefore, it is unlikely that regulation of protein transcription mediated this response. While not directly examined in this study, it would seem that the most plausible explanation for our observations is that activation of NPR-B or NPR-C by CNP results in a posttranslational modification of iNOS. It is also important to note that we observed constitutive expression of iNOS mRNA and protein in nonstimulated IMF. This finding is in line with earlier work that showed iNOS mRNA and protein can be detected in healthy mouse ileum (13). LPS is a potent inducer of iNOS in multiple cell types, and we speculate that intestinal normal flora-derived LPS plays a role in the regulation of iNOS expression in IMF.

The IMF relaxation response attributable to cGMP is consistent with other models of cellular relaxation triggered by elevation of intracellular cGMP concentrations. We observed that CNP triggered a substantial reduction in MLC phosphorylation that was partially reversed by treating cells with inhibitors of NO production. cGMP signaling in the gut is known to promote relaxation through at least two distinct mechanisms (24). One mechanism is the cGMP-induced activation of PKG, which phosphorylates and activates MLC phosphatase. Enhanced activity of MLC phosphatase reduces the amount of phosphorylated MLC and results in relaxation. Another mechanism is cGMP-induced activation of voltage-gated potassium channels. Opening of these channels causes an influx of potassium and efflux of cytosolic calcium. Reduced calcium levels correspond to lower activity of MLC kinase and less phosphorylated MLC. The mechanism by which cGMP regulates myosin light chain phosphorylation after CNP signaling in IMF awaits future investigation, although reports in other nonmuscle cell types suggest that regulation of tension is independent of changes in intracellular calcium concentration (17, 22).

In this study we show that CNP-induced relaxation of IMF is partially mediated by activation of iNOS. If a similar pathway operates in vivo in humans, we believe these findings are relevant to the creation of new therapeutics for intestinal disease. Because iNOS is suggested to be a critical factor in the development of IBD, a therapeutic designed to inhibit its function may also impair the actions of CNP. Conversely, a therapeutic aimed at modulating the CNP signaling pathway may have the potential to promote the relaxation and breakdown of strictures. Continued investigation of the mechanisms controlling IMF tension should aid in the development of new intestinal disease therapies that target wound healing and villous motility.

GRANTS

H. F. Yee, Jr. was supported by National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Grant R01-DK-61532 and the William and Mary Ann Rice Memorial Distinguished Professorship. This work was also supported by the Hefni Technical Training Foundation and the Cell and Tissue Biology Core Facility of the University of California San Francisco Liver Center (NIDDK R01-P30-DK-26743).

DISCLOSURES

The authors declare that no conflict of interest exists.

AUTHOR CONTRIBUTIONS

Author contributions: Y.C., T.C., R.K.S.J., and H.F.Y.J. conception and design of research; Y.C., T.C., J.W., and R.K.S.J. performed experiments; Y.C., T.C., J.W., R.K.S.J., and H.F.Y.J. analyzed data; Y.C., T.C., J.W., R.K.S.J., A.C.M., and H.F.Y.J. interpreted results of experiments; Y.C. and A.C.M. prepared figures; Y.C. and A.C.M. drafted manuscript; Y.C., A.C.M., and H.F.Y.J. edited and revised manuscript; Y.C., T.C., J.W., R.K.S.J., A.C.M., and H.F.Y.J. approved final version of manuscript.