Integrin Ligands Mobilize Ca2+ from Ryanodine Receptor-gated Stores and Lysosome-related Acidic Organelles in Pulmonary Arterial Smooth Muscle Cells*

Extracellular matrix (ECM) protein receptors, or integrins, participate in vascular remodeling and the systemic myogenic response. Synthetic ligands and ECM fragments regulate the vascular smooth muscle cell contractile state by altering intracellular Ca2+ levels ([Ca2+]i). Information on the Ca2+ effect of integrins in vascular smooth muscle cells is limited, but nonexistent in pulmonary arterial smooth muscle cells (PASMCs). We therefore characterized integrin expression in endothelium-denuded pulmonary arteries, and explored [Ca2+]i mobilization pathways induced by soluble ligands in rat PASMCs. Reverse transcriptase-PCR showed mRNA expression of integrins α1, α2, α3, α4, α5, α7, α8, αv, β1, β3, and β4, and immunoblots of α5, αv, β1, and β3 confirmed protein expression. Exposure of PASMCs to integrin-binding peptides (0.5 mm) containing the arginine-glycine-aspartate (RGD) motif elicited [Ca2+]i responses with an order of potency of GRGDNP > GRGDSP > GRGDTP = cyclo-RGD. Pharmacological analysis revealed that the GRGDSP-induced Ca2+ response was unrelated to Ca2+ influx and the inositol triphosphate receptor-gated Ca2+ store, but partially blocked by ryanodine or inhibition of lysosomerelated acidic organelles with bafilomycin A1. Simultaneous inhibition of both pathways was necessary to abolish the response. GRGDSP treatment increased cyclic ADP-ribose, the endogenous activator of ryanodine receptors, by 70%. GRGDSP also rapidly reduced Lysotracker Red accumulation, confirming direct modulation of acidic organelles. These data are the first demonstration of integrin-mediated Ca2+ regulation in PASMCs. The presence of an array of integrins, and activation of ryanodine-sensitive Ca2+ stores and lysosome-like organelles by GRGDSP suggest important roles for integrin-dependent Ca2+ signaling in regulating PASMC function.

transmembrane receptors that mediate cell-cell and cell-ECM interactions. Integrins physically bridge the ECM and cytoskeleton, and act as transducers of "outside-in" and "inside-out" signaling (1). Extracellular integrin ligation changes [Ca 2ϩ ] i in a variety of cell types, including platelets, neutrophils, monocytes, lymphocytes, fibroblasts, endothelial cells, osteoclasts, neurons, glomerulosa cells, epithelial and smooth muscle cells (2)(3)(4)(5). Integrin ligation also activates other intracellular signaling molecules, including H ϩ and a plethora of protein kinases such as the focal adhesion kinase, Rac, and extracellular signal regulated kinases (1,6). As a result, a myriad of cellular functions such as differentiation, proliferation, migration, apoptosis, and mechanosensing for shear and tension are affected (6,7).
Apart from regulating tone, integrins play an important role in vascular pathogenesis. This is evidenced by the remodeling and deposition of ECM components seen in various vascular diseases such as atherosclerosis, hypertension, and restenosis (5). In the pulmonary vasculature, remodeling occurs in chronic obstructive pulmonary disease, asthma, scleroderma, and pulmonary hypertension (14). Interestingly, pulmonary vascular remodeling has been linked to integrin activation in rat models recapitulating pulmonary hypertension. Previous studies in monocrotaline-induced pulmonary hypertension show that degraded ECM components activate vascular smooth muscle cell integrin ␤ 3 , causing focal adhesion formation and initiation of the extracellular signal-regulated kinase signal transduction cascade (15). Incidentally, the increased PASMC proliferation and medial hypertrophy related to pulmonary hypertension also involves altered [Ca 2ϩ ] i homeostasis in PASMCs (16).
The influence of integrins on [Ca 2ϩ ] i homeostasis in governing systemic vascular tone, and the participation of integrins in pulmonary vascular remodeling suggest that integrin ligation would affect [Ca 2ϩ ] i homeostasis in PASMCs. However, information to this end is lacking. In this article, we therefore test this hypothesis by monitoring [Ca 2ϩ ] i levels in rat PASMCs upon exposure to soluble integrin ligands. We subsequently characterize the intracellular Ca 2ϩ signaling cascade initiated by GRGDSP, and identify the ryanodine-sensitive Ca 2ϩ stores and lysosome-related acidic organelles as the key mediators of this response.
RNA Isolation and RT-PCR-Intralobar pulmonary arteries and aorta were removed, cleaned of connective tissue, denuded of endothelium as above, frozen in liquid nitrogen, and kept at Ϫ80°C until use. Tissues were homogenized in TRIzol reagent (Invitrogen) using a PowerGen homogenizer (Fisher Scientific), and total RNA was isolated according to the protocol supplied with the TRIzol reagent. Ethanol-precipitated RNA samples were dissolved in diethyl pyrocarbonate-treated water and quantified using a BioPhotometer spectrophotometer (Eppendorf, Hamburg, Germany).
Total RNA samples were subsequently used for cDNA synthesis with SuperScript II (Invitrogen). Primers for PCR (Table  1) were carefully designed to regions specific for each integrin subtype, based on sequence alignments generated by a Clustal W algorithm. At least one of each primer pair was designed to exon-exon junctions, to minimize the possibility of amplifying genomic DNA. Where it was not feasible to design primers to exon-exon junctions, primers were designed to span intronic regions, to differentiate between genomic DNA and cDNA. PCR was carried out for 30 cycles using PlatinumTaq DNA polymerase (Invitrogen), which involved denaturation at 94°C for 1 min, annealing at 55°C for 1 min, and extension at 72°C for 2 min. The resulting RT-PCR products were analyzed by 1.5% agarose gel electrophoresis.
Preparation of Protein Samples and Western Analysis-Total protein was isolated from endothelium-denuded pulmonary arteries and aorta. Briefly, upon dissection, tissues were frozen in liquid nitrogen and pulverized. Samples were subsequently homogenized with a Dounce homogenizer (30 strokes) in icecold Tris-HCl buffer (50 mM, pH 7.4) containing phenylmethylsulfonyl fluoride (1 mM) and protease mixture inhibitor (Roche Applied Sciences). Homogenized tissues were centrifuged (3000 ϫ g, 4°C, 10 min), upon which the protein concentrations of the post-nuclear supernatant were measured with the BCA Protein Assay Kit (Pierce). Total protein from cultured PASMCs were isolated by scraping cells with a rubber policeman in the ice-cold Tris-HCl buffer and processed as described for the smooth muscle tissue.
Laser scanning confocal microscopy was used to monitor the fluorescence changes upon treatment of PASMCs with the respective agents. Images were acquired with a Zeiss LSM-510 inverted confocal microscope (Carl Zeiss Inc., Germany) using a Zeiss Plan-Neofluor ϫ40 oil immersion objective (NA ϭ 1.3) and an excitation wavelength of 488 nm. Fluorescence, measured at wavelengths Ͼ505 nm, was acquired in a framescan mode at either 5-(Figs. 3-6) or 10-s (Fig. 7) intervals. Photobleaching and laser damage to cells were minimized by attenuating the laser to ϳ1% of its maximum power (25 milliwatts). In all cases during fluorescence recordings, reagents were applied manually but with care, so as not to disturb the cells or cause an artificial change in fluorescence.
In some experiments, Fluo-3 fluorescence was detected using a Nikon Diaphot microscope equipped with epifluorescence attachments and a microfluorometer (Biomedical Instrument Group, University of Pennsylvania). Protocols were executed and data collected on-line with a Digidata analog-to-digital interface and the pClamp software package (Axon Instruments, Inc., Foster City, CA).
Changes in fluorescence intensity were used to calculate the cellular concentrations of Ca 2ϩ , using the following equation:  (18). Concentrations of Ca 2ϩ were similarly calculated for data obtained with the epifluorescence microscope, upon obtaining a background fluorescence value in an area devoid of cells by using the Ca 2ϩ ionophone 4-Br-A23187 (EMD Biosciences) followed by Mn 2ϩ quenching.
Imaging Lysosomes-Acidic organelles were labeled with the acidotropic dye Lysotracker Red DND-99 (Molecular Probes) diluted in Tyrode's solution (50 nM) at room temperature for 30 -45 min. Labeled PASMCs were imaged with a Zeiss LSM-510 inverted microscope as described above, but with excitation and emission wavelengths of 543 and Ͼ560 nm, respectively. Images were acquired in framescan mode at 2-min intervals, and cellular changes in fluorescence intensity were represented as F/F 0 , with F 0 being the fluorescence level immediately preceding treatment with the respective agent.
Detection of Cyclic ADP-ribose (cADPR)-cADPR levels were measured in cultured PASMCs at passage 3-4. Cells were isolated and cultured in Ham's F-12 media overnight as described above, and maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum thereafter. Prior to treatment with GRGDSP (1.25 mM), cells were washed three times with PBS, and equilibrated in Tyrode's solution for at least 30 min. Upon GRGDSP treatment, nucleotides were extracted from cells with 10% (w/v) trichloroacetic acid at 4°C, followed by removal with watersaturated diethyl ether. The aqueous layer containing cADPR was adjusted to pH 8 with 1 M Tris and contaminating nucleotides other than cADPR were removed with a mixture containing hydrolytic enzymes with the following final concentrations: 0.44 unit/ml nucleotide pyrophosphatase, 12.5 units/ml alkaline phosphatase, 0.0625 unit/ml NADase, 2.5 mM MgCl 2 . Incubation proceeded at 37°C for 2 h. Detection of cADPR was performed with some modifications to the cycling method described recently (19). Briefly, 0.1 ml of cADPR standard or nucleotides extracted from cell samples were incubated with 100 l of cycling reagent containing (final concentrations) 0.3 g/ml ADP-ribosyl cyclase, 4 mM  Table 1 gives the specifics of the primers used, as well as the expected sizes, which correspond with each of the products obtained in this figure. All primers were tested in other tissue types to verify their capability in efficiently and robustly amplifying the product of the expected size. This figure is a representative image of an RT-PCR, which was repeated a total of 3 times.

Integrin Expression in PASMCs-
To determine the repertoire of integrins expressed in endothelium-denuded rat pulmonary arteries, RT-PCR was performed using primers designed to integrins Table 1). Fig. 1 shows a representative image of an RT-PCR, in which positive mRNA expression was observed for all integrin subtypes examined (Fig. 1a). We further compared the integrin expression profile to that in endothelium-denuded aorta (Fig. 1b). As illustrated in Fig. 1, the pattern of integrin expression was similar in both tissue types, except that integrin ␣ 4 consistently could not be detected in the aortic smooth muscle. Sizes of the PCR products obtained in all cases corresponded to that expected (see Table 1). All PCR primers were tested prior to use in other tissues (e.g. heart, small intestine, and brain) and shown to efficiently amplify products of the expected size (data not shown). Negative controls, in which the reverse transcriptase was omitted, yielded no PCR products (data not shown).
We focused our attention on a few integrin subtypes that have been characterized most extensively in other smooth muscles and examined the protein expression in endothelium-denuded pulmonary arteries and aorta. Western analysis of integrins ␣ 5 , ␣ v , ␤ 1 , and ␤ 3 confirmed the presence of these proteins in both tissue subtypes (Fig. 2). Immunoblots were performed using both reduced and nonreduced protein samples, as certain antibodies were generated against reduced or non-reduced antigens. Some integrin ␣ subunits including ␣ 5 and ␣ v contain a disulfide-linked cleavage site at the C terminus of the protein, yielding 2 polypeptides in the presence of reducing agents (20). ␤ Subunits have internally disulfide-bonded cysteine-rich domains in the C termini that influence mobility on SDS-PAGE gels depending on the presence of reducing agents (21). A number of antibodies from various companies were tested for each of the integrin subtypes, and those giving the most specific products of the expected size were chosen. The anti-␣ 5 antibody recognized a broad band of ϳ150 kDa under non-reducing conditions in both PA and aorta, but was not reactive with the reduced protein as indicated by the manufacturer. Two bands of about 120 and 150 kDa were immunoreactive to the anti-␣ v antibody, presumably representing the reduced and non-reduced forms, respectively, because the 150-kDa band was more prominent under nonreducing conditions. A single band of about 120 kDa was detected by anti-␤ 1 in both tissues regardless of the presence of  reducing agents, and a single 100-kDa band was recognized by the anti-␤ 3 antibody only under reducing conditions.
Effect of Integrin Ligands on [Ca 2ϩ ] i -One of the initial events in the signal transduction cascade following integrin activation is a change in [Ca 2ϩ ] i levels (6,22). For many integrin ligands, the amino acid sequence immediately following the active RGD motif determines receptor selectivity (23). To see if integrin activation was capable of eliciting changes in [Ca 2ϩ ] i in PASMCs, 4 RGD peptides (GRGDSP, GRGDNP, GRGDTP, and cyclo-RGD) and 2 control peptides (GRGESP, GRADSP) were exogenously applied at a concentration of 0.5 mM and [Ca 2ϩ ] i levels were monitored (Fig. 3)   20 min), a non-selective Ca 2ϩ channel blocker inhibiting both L-type Ca 2ϩ channels and capacitative Ca 2ϩ entry, also had no significant effect on the GRGDSP-induced [Ca 2ϩ ] i increase (Fig. 5B). GRGDSP caused [Ca 2ϩ ] i elevations of 142.7 Ϯ 16.6 nM in SKF 96365-treated cells (n ϭ 34 cells), compared with 136.7 Ϯ 14.9 nM (n ϭ 46 cells; p Ͼ 0.05) in time-matched controls. To further confirm that GRGDSP was not causing Ca 2ϩ entry, PASMCs were exposed to the peptide under nominally Ca 2ϩ -free conditions (Fig. 5C) (Fig. 6B).
It has recently been shown that in addition to ryanodine and IP 3 receptor-gated intracellular Ca 2ϩ stores, that PASMCs contain an active intracellular Ca 2ϩ store of lysosomal origin that is sensitive to nicotinic acid adenine dinucleotide phosphate (NAADP), but insensitive to thapsigargin treatment (24,25). To test the possibility that GRGDSP was promoting Ca 2ϩ release through a lysosome-related acidic organelle, we treated the PASMCs with bafilomycin A1 (Fig. 6C). Bafilomycin A1 is a macrolide antibiotic that selectively inhibits the vacuolar H ϩ -ATPase (V-ATPase) at nanomolar to low micromolar concentrations, and the P-type ATPases at high micromolar concentrations (26,27). It potently inhibits both ATPase and H ϩ pumping activities of the V-ATPase, causing dissipation of the proton gradient that energizes Ca 2ϩ uptake into these compartments, thereby inhibiting the NAADP-sensitive Ca 2ϩ activity (26 -29). As reported by Kinnear et al. (24), bafilomycin A1 treatment alone elevated [Ca 2ϩ ] i in PASMCs, presumably arising from the lysosome-related acidic compartments as a result of inhibiting the lysosomal proton motive force. In addition, the GRGDSP-induced response in the bafilomycin A1-treated cells was significantly reduced to 50.4% of control values (⌬[Ca 2ϩ ] i (control) ϭ 79.0 Ϯ 15.9 nM, n ϭ 11 dishes versus ⌬[Ca 2ϩ ] i (bafilomycin) ϭ 39.9 Ϯ 11.6 nM, n ϭ 11 dishes, p Ͻ 0.1). The GRGDSP-induced [Ca 2ϩ ] i response was taken to be the peak [Ca 2ϩ ] i elicited by GRGDSP relative to the [Ca 2ϩ ] i just prior to GRGDSP application. The inset to Fig. 6C shows the same dishes treated with the Ca 2ϩ ionophore 4Br-A23187 upon completion of the experiment. Treatment with 4Br-A23187 caused [Ca 2ϩ ] i changes in the micromolar range in both the presence and absence of bafilomycin A1, demonstrating that Fluo-3 fluorescence intensity had not been saturated by GRGDSP treatment.
The fact that none of the above pharmacological treatments completely eliminated the response to GRGDSP suggested that multiple Ca 2ϩ signaling pathways were involved. This was verified by depletion of both intracellular and extracellular Ca 2ϩ sources (Fig. 7A). Upon washing the Fluo-3-loaded cells with 2 mM calcium-Tyrode, thapsigargin (20 M) was applied under nominally Ca 2ϩ -free conditions in the presence or absence of bafilomycin A1. After an additional 30 min, caffeine (5 mM) was applied to ensure complete depletion of the ryanodine-sensitive Ca 2ϩ store, upon which PASMCs were ultimately exposed to GRGDSP (1.25 mM). The initial peak ⌬[Ca 2ϩ ] i upon application of the 0 Ca/EGTA/thapsigargin/bafilomycin A1 mixture was higher in the presence of bafilomycin A1 (⌬[Ca 2ϩ ] i ϭ 251.3 Ϯ 8.5 nM, n ϭ 202 cells) compared with that in its absence (⌬[Ca 2ϩ ] i ϭ 180.6 Ϯ 7.0 nM, n ϭ 160 cells; p Ͻ 0.05), presumably due to the additional release of Ca 2ϩ by bafilomycin A1 from the acidic organelles. We proceeded to test if inhibition of the ryanodine receptor (RyR) and acidic organelle-mediated intracellular [Ca 2ϩ ] i stores was sufficient to abolish the GRGDSP-induced Ca 2ϩ transients. In the presence of 2 mM extracellular Ca 2ϩ (Fig. 7B), treatment of PASMCs with both ryanodine (50 M) and bafilomycin A1 (3 M) completely abrogated the GRGDSP-elicited response (⌬[Ca 2ϩ ] i ϭ Ϫ4.4 Ϯ 1.7 nM, n ϭ 9 dishes) in comparison to control cells (⌬[Ca 2ϩ ] i ϭ 53.9 Ϯ 11.2 nM, n ϭ 8 dishes, p Ͻ 0.05). As in Fig. 6C, treatment of cells with ryanodine and bafilomycin A1 caused elevations in [Ca 2ϩ ] i , whose peak ([Ca 2ϩ ] i ϭ 48.8 Ϯ 5.5 nM) was close to that elicited by GRGDSP in control cells ([Ca 2ϩ ] i ϭ 54.8 Ϯ 9.9 nM), suggesting that the levels of intracellular Ca 2ϩ released by GRGDSP was equivalent to that released by bafilomycin A1 and ryanodine. The results from Fig. 7 collectively show that simultaneous disruption of both ryanodine/cADPR-and bafilomycin A1-sensitive intracellular Ca 2ϩ stores was necessary and sufficient to abolish the GRGDSP-induced Ca 2ϩ response in rat PASMCs.
Effect of GRGDSP on cADPR Levels-The pyridine nucleotide, cADPR, is thought to be the endogenous activator of ryanodine receptors (30). To test whether GRGDSP affected cADPR levels and thereby verify the pharmacological results of Figs. 6A and 7, we measured cADPR directly in cultured PASMCs upon exposure to GRGDSP. In unstimulated PASMCs, the concentrations of cADPR ranged from 0.5 to 2 pmol/mg of protein, with an average of 1.2 Ϯ 0.3 pmol/mg of total protein. Treatment of cultured PASMCs indeed increased cADPR levels as early as 30 s after treatment with GRGDSP (Fig. 8B). Maximum levels were reached by 1 min, and were sustained through 5 min, but were decreased by 45 min post-application. GRGDSP treatment elevated cADPR levels by ϳ70% compared with untreated controls. CD38 is a transmembrane glycoprotein residing on the plasma membrane that possesses ADP ribosyl cyclase activity (30). Cultured PASMCs were shown to express CD38 via Western analysis (Fig. 8A).  (Fig. 9A, lower panel) significantly reduced the fluorescence intensity over 30 min as compared with vehicle (Me 2 SO, Fig. 9A, upper panel). A similar initial drop in Lysotracker Red fluorescence was seen in both vehicle and bafilomycin A1-treated cells. However, fluorescence levels were sustained thereafter in the vehicle-treated cells, whereas levels continued to decrease significantly in bafilomycin A1-treated cells (Fig. 9B). Bafilomycin A1 reduced the fluorescence by ϳ80% (n ϭ 56 cells) over time compared with vehicle (n ϭ 45), which caused an approximate 30% decrease.
Most intriguingly, treatment of PASMCs with GRGDSP ( Fig. 10) affected Lysotracker Red fluorescence intensities in a concentration-dependent manner. The decrease in Lysotracker Red fluorescence elicited by GRGDSP was transient in nature as it showed a partial recovery with time. Effects of photobleaching were evident from control experiments in which vehicle (Tyrode's solution, Fig.  10B) was applied, but significantly differed from the peptide-induced effects (Fig. 10C). The degree to which GRGDSP affected the Lysotracker Red fluorescence, as well as its transient nature differed from the effect elicited by bafilomycin A1 (Fig. 9), which dissipates the lysosomal pH gradient through inhibition of the V-ATPase. These data directly demonstrate the modulation of the properties of lysosomerelated acidic organelles by GRGDSP.

DISCUSSION
Here we characterize intracellular Ca 2ϩ signaling by integrins in rat PASMCs. We show the presence of multiple integrins in PASMCs, and the initiation of intracellular Ca 2ϩ signaling by soluble ligands. The Ca 2ϩ response is independent of Ca 2ϩ entry and IP 3 receptorgated intracellular Ca 2ϩ stores, but dependent on both RyRgated Ca 2ϩ stores and the lysosome-related acidic organelles. We further demonstrate the ability of soluble integrin ligands to directly affect RyR-gated Ca 2ϩ stores by increasing levels of cADPR, as well as alter the properties of lysosome-related acidic compartments through Lysotracker Red imaging. These data illustrate a novel mode of Ca 2ϩ regulation by integrins, and underscore both the RyR-gated Ca 2ϩ store and lysosome-related acidic organelle as important players in influencing intracellular Ca 2ϩ homeostasis in PASMCs.

Integrin-mediated Ca 2؉ Mobilization in PASMCs
Four peptides were used to pinpoint the integrins involved, because the amino acid immediately succeeding the RGD motif, as well as peptide conformation (cyclic versus linear) determines receptor specificity. For instance, cyclo-RGD preferentially interacts with ␣ v ␤ 3 , whereas GRGDSP and GRGDNP target both ␣ v ␤ 3 and ␣ 5 ␤ 1 , with GRGDNP showing higher specificity for the latter (23). The four peptides elicit differential [Ca 2ϩ ] i changes in rat PASMCs, in which the order of potency (i.e. GRGDNP Ͼ GRGDSP Ͼ cyclo-RGD ϭ GRGDTP) suggests a vital role for ␣ 5 ␤ 1 rather than ␣ v ␤ 3 in PASMC function. Indeed, soluble fibronectin elicits concentrationdependent [Ca 2ϩ ] i changes, albeit of lower magnitude (data not shown). Despite the fact that GRGDNP elicited the greatest changes in [Ca 2ϩ ] i , we focused this study on the effects of GRGDSP, the most thoroughly characterized and widely occurring integrin-binding peptide sequence in endogenous ECM proteins. Micromolar concentrations of RGD peptides have been shown to inhibit, whereas millimolar concentrations have been demonstrated as activating integrins (41).
Integrin ligation causes [Ca 2ϩ ] i changes in many cell types, but the source of Ca 2ϩ mobilized is both integrin and cell-type specific. For example, endothelial integrin ligation causes Ca 2ϩ influx and Ca 2ϩ release from IP 3 -sensitive intracellular stores (42,43). In Jurkat T cells, integrin ligation activates IP 3 and cADPR-sensitive stores as well as capacitative Ca 2ϩ entry (44).
Smooth muscle Ca 2ϩ signaling by integrins has as yet only been studied in rat cremaster arterioles and renal SMCs. In rat cremaster arterioles, integrins differentially regulate the L-type Ca 2ϩ channel, where ␣ 4 ␤ 1 and ␣ 5 ␤ 1 activate and ␣ v ␤ 3 inhibits the L-type Ca 2ϩ channels, causing vasoconstriction and vasodilation, respectively (8,12,40). Stimulation of rat renal vascular smooth muscle cells by soluble GRGDSP causes Ca 2ϩ release entirely from RyR-sensitive stores (10,11).
Our investigations into the source of Ca 2ϩ mobilized by GRGDSP offer unique information on signaling by integrins. Specifically, Ca 2ϩ entry and IP 3 -sensitive Ca 2ϩ stores are uninvolved, whereas inhibition of the RyR-gated Ca 2ϩ store and the lysosome-related acidic organelle each reduce the GRGDSPmediated Ca 2ϩ response by ϳ50%. Furthermore, simultaneous inhibition of the RyR-gated Ca 2ϩ store and the lysosome-related acidic organelle is essential to completely abolish the GRGDSP-induced response. Ryanodine receptor-gated intracellular Ca 2ϩ stores have been implicated pharmacologically as contributing to integrinmediated [Ca 2ϩ ] i increases, as in the case of rat renal vascular smooth muscle cells, cardiomyocytes, and T cells (10,11,44,45). However, our study is the first to directly demonstrate that integrin binding peptides increase the levels of cADPR, which is the endogenous activator of RyRs. It has been well established that cADPR contributes to the regulation of arterial smooth muscle tone by altering [Ca 2ϩ ] i homeostasis. In the pulmonary vasculature, cADPR plays important roles in the development of hypoxic pulmonary vasoconstriction (46). The resting cADPR levels obtained in this study from cultured PASMCs (ϳ1.2 pmol/mg of protein) is in agreement with that reported in the PASM tissue (ϳ1.5 pmol/mg of protein) (47). These concentrations are also similar to those described in airway smooth muscle (ϳ0.8 pmol/mg of protein) as well as uterine smooth muscle (ϳ1 pmol/mg of protein) (48 -50). GRGDSP causes increases in cADPR levels as early as 30 s post-application. This time course is consistent with that observed for [Ca 2ϩ ] i measurements, where [Ca 2ϩ ] i levels were on the rise by 30 s, and increased further at 1 and 5 min. The [Ca 2ϩ ] i at 5 min post-GRGDSP treatment is generally double of that observed at 1-min post-application. This differs from that seen for [cADPR], whose normalized values vary only by 5% between 1 and 5 min post-treatment, and can be due to a contribution of the cADPR-independent, bafilomycin-sensitive intracellular Ca 2ϩ store.
Thapsigargin-insensitive intracellular Ca 2ϩ stores have increasingly received attention over the last decade, and are responsive to NAADP, which is structurally related to the coenzyme NADP (51). This NAADP-sensitive intracellular Ca 2ϩ store has been described in various systems including PASMCs, and has been ascribed to a lysosome-related acidic compartment in sea urchin eggs, pancreatic acinar cells, MIN6 cells, and PASMCs (25,29,52). However, research in this field is still in its infancy and therefore controversial. Some studies show NAADP targeting a unique receptor in the lysosome, whereas others argue that it directly activates either the ryanodine or IP 3 receptors, as in skeletal muscle and T-lymphocytes (53,54). Furthermore, some assert that NAADP receptors are functionally coupled to IP 3 and/or ryanodine receptors, such that Ca 2ϩ mobilization by NAADP causes Ca 2ϩ -induced Ca 2ϩ -release, generating global [Ca 2ϩ ] i transients mediated by the IP 3 and/or ryanodine receptors (55)(56)(57).
The sensitivity of the GRGDSP-elicited [Ca 2ϩ ] i response to ryanodine and bafilomycin A1 is consistent with that observed by Evans and colleagues (25). They have proposed that NAADP evokes spatially localized "calcium bursts" that trigger RyR-mediated Ca 2ϩ -induced Ca 2ϩ -release through "lysosome-SR" junctions, because RyR inhibition eliminates the global Ca 2ϩ  transients without affecting Ca 2ϩ bursts (24,25). However, our results cannot be explained entirely by this model, because thapsigargin treatment or inhibition of RyRs alone did not eliminate the global [Ca 2ϩ ] i increase. Rather, in our preparation the RyR-gated Ca 2ϩ store and lysosome-related acidic compartments each contributed to half of the GRGDSP-induced Ca 2ϩ response, and had to be inhibited together to abrogate the response. Whereas it is conceivable that these two sources of intracellular Ca 2ϩ may be functionally coupled, GRGDSP does not seem to evoke Ca 2ϩ -induced Ca 2ϩ release from either store. Instead our results suggest that it activates Ca 2ϩ release from these two stores via independent pathways.
Based on the fact that: 1) bafilomycin A1 reduces the intracellular Ca 2ϩ response to GRGDSP, and 2) both bafilomycin A1 and GRGDSP directly affect the properties of acidic organelles as visualized by Lysotracker Red imaging, we can conclude that GRGDSP stimulates lysosome-related organelles. The seemingly transient nature of the GRGDSP-induced decrease in Lysotracker Red fluorescence (Fig. 10B) suggests that it causes a loss of organelle acidification, rather than disrupting the integrity of the lysosome as in the case of compounds such as glycyl-L-phenylalanine 2-napthylamide (58,59). A correlation between lysosomal pH and lysosomal [Ca 2ϩ ] ([Ca 2ϩ ] lys ) has been shown in macrophages, where alkalinization of lysosomes resulted in proportional decreases in [Ca 2ϩ ] lys (28). The mechanism of [Ca 2ϩ ] lys regulation in macrophages was likened to that in yeast vacuoles, in which the V-ATPase maintains a proton gradient that in turn drives [Ca 2ϩ ] lys uptake through a Ca 2ϩ /H ϩ exchanger in the lysosomal membrane (28). The fact that GRGDSP reduces the Lysotracker Red fluorescence therefore suggests a proportional decrease in [Ca 2ϩ ] lys , which is accompanied by a transient alkalinization of the lysosome. Because bafilomycin A1 also promotes lysosomal alkalinization, it is possible that pretreatment of PASMCs with bafilomycin A1 (Figs. 6C and 7B) causes concomitant [Ca 2ϩ ] lys depletion, such that this Ca 2ϩ source is no longer available for stimulation by GRGDSP. Given that NAADP evokes Ca 2ϩ release from lysosome-related acidic organelles in a bafilomycin A1-sensitive manner, we hypothesize that GRGDSP mobilizes lysosomal Ca 2ϩ in part via NAADP (24,25,29,52). However, we have yet to establish a direct link between GRGDSP treatment and induction of the NAADP pathway. Future studies on the change in cellular NAADP concentrations upon GRGDSP treatment should address this pressing issue.
This study raises interesting questions regarding the mechanism of integrin-mediated Ca 2ϩ signaling. In particular, how does integrin activation couple to Ca 2ϩ mobilization from both RyR-gated stores and acidic organelles? One plausible explanation is that GRGDSP activates ADP-ribosyl cyclase. ADP-ribosyl cyclase synthesizes both cADPR and NAADP from ␤-NAD and ␤-NADP, respectively, and its activation by GRGDSP may thus explain the near equivalent contribution of these two Ca 2ϩ stores to the peptide-induced response (60). High levels of ADP-ribosyl cyclase activity have in fact been reported in the PASM tissue (47). We have also evidenced robust levels of ADP-ribosyl cyclase activity in cultured PASMCs, using NGD ϩ as an alternative substrate to measure cyclic GDP ribose forma-tion (data not shown), and have, furthermore, shown the expression of CD38 in these cells (this study). The lung is among a group of tissues containing a novel enzyme with ADPribosyl cyclase activity in addition to CD38 (61)(62)(63). This notion is supported by data from CD38 Ϫ/Ϫ mice containing an appreciable amount of cADPR in the lung albeit at lower levels than in the wild type, as well as the fact that cADPR synthesis has been measured in microsomal fraction of PASM homogenates (47,63,64).
Stimulation by GRGDSP may result in Ca 2ϩ mobilization from the RyR-gated stores and acidic organelles, alternatively, by causing integrin clustering and subsequent protein recruitment to form focal adhesion complexes, which tether integrins to the cytoskeleton. Many components of the focal adhesion complex are either substrates of tyrosine kinases or tyrosine kinase themselves, and are activated upon integrin ligation in various cell types, including endothelial and vascular smooth muscle cells (65,66). Moreover, in rat cremaster arterial cells, [Ca 2ϩ ] i modulation through the activation of tyrosine kinase pathways has been shown for ␣ 4 ␤ 1 and ␣ 5 ␤ 1 stimulation (12,67).
In summary, we have demonstrated the modulation of [Ca 2ϩ ] i in PASMCs by integrin agonists. The information obtained adds a level of complexity in PASMC regulation, uncovering a unique mode of Ca 2ϩ signaling by integrins in which RyR-gated Ca 2ϩ stores and acidic organelles are recruited. Modulation of PASMCs by integrin ligands has important implications in pulmonary diseases including chronic obstructive pulmonary disease, pulmonary fibrosis, and pulmonary hypertension, where ECM remodeling as well as induction of matrix metalloproteinases that degrade ECM molecules is observed (68,69). ECM remodeling and degradation expose cryptic integrin binding sites including RGD motifs (69). Therefore our work on [Ca 2ϩ ] i regulation by soluble RGD peptides highlights the importance of the ECM in influencing pulmonary function and may provide valuable insights in the progression of pulmonary disease.