INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

HYPERHOMOCYSTEINEMIA IS an independent and graded risk factor for development of cardiovascular disease such as stroke, myocardial infarction, peripheral vascular disease, and atherosclerosis (5, 23, 42). Homocysteine (Hcy) increases the production of reactive oxygen species in cultured human umbilical vein endothelial cells (HUVEC) and induces endothelial dysfunction (54). However, the underlying mechanisms responsible for Hcy-induced endothelial dysfunction are not fully understood.

DNA microarray technology provides a high-throughput tool for an unbiased screening of differentially displayed genes (11, 51). This approach, as well as other techniques, has been successfully utilized in large-scale gene expression analyses of various pathological conditions, including effects of some proatherogenic agents on endothelial cells (16, 23, 30, 44). Accordingly, we examined HUVEC treated with Hcy with the complementary DNA microarrays (Clontech) and, among other findings, detected overexpression of mRNA encoding connexin43 (Cx43) (see RESULTS).

Gap junctional communication is essential for the embryonic development, cell proliferation, differentiation, and coordination of individual cell behavior within a cell population. Several members of the connexin family of proteins, including Cx43, participate in the formation of gap junctions (37). Cx43 is widely expressed in endothelial cell in vitro and in vivo (6, 20, 35). Like other Cx proteins, the intercellular junctional hemichannels formed by Cx43 allow ions and small molecules of <1 kDa to pass through (3, 21, 31, 39, 41, 47, 55). Gap junctional communication was found to be disrupted by phosphorylation of Cx43 caused by the oxidant stress (48). Cx43 gene upregulation has been previously found in several proatherosclerotic pathophysiological conditions (1, 13, 14, 33, 40, 52). Most recently, Cx43 has been linked to myoendothelial junctions, where it regulates the action of endothelium-derived hyperpolarizing factor (EDHF) (9, 46). Using cardiovascular cDNA array, we detected upregulation of Cx43 in the Hcy-treated endothelial cells. In this report, we demonstrate that Hcy-induced Cx43 overexpression is not associated with an increase in gap junctional communication. Overexpressed Cx43 redistribute to the mitochondria. Furthermore, defective acetylcholine-evoked nitric oxide synthase (NOS)- and cyclooxygenase (COX)-independent vasorelaxation suggest the loss of myoendothelial gap junctional communication in Hcy-treated resistance arteries.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell culture. HUVEC were cultured in EBM-2 media (Clonetics) and used between passages 3 and 8.

Cardiovascular microarray analysis. The cell monolayers were washed twice with PBS and then scraped. Atlas cardiovascular microarray (Clontech) was carried out according to the manufacturer's instructions. Differentially displayed genes were studied in Hcy-treated HUVEC and control companion cells at the third passage. The design of the cDNA array and the complete transcription list of genes are available at www.clontech.com.

Semiquantitative RT-PCR. Total RNA was isolated from HUVEC cells using Trizol (GIBCO). RT-PCR was performed using Cx43 primers 5'-CATGGGTGACTGGAG-3' and 5'-AGGACCCAGAAGCGCA-3', generating a product of ~238 bp (27). Primers to human glyceraldehyde-3-phosphate dehydrogenase (GAPDH) 5'-CAACTACATGGTTTACATGTTC-3'and 5'-GCCAGTGGACTCCACGAC-3', generating a product of 180 bp, were used as an internal control. Each PCR cycle consisted of denaturing at 94°C for 30 s, annealing at 50°C for 30 s, and elongation at 68°C for 30s. The linear exponential phrase was 30 cycles for Cx43 and 22 cycles for GAPDH. Equal amounts of corresponding Cx43 and GAPDH reverse PCR products were loaded on 1.8% agarose gels. Optical densities of ethidium bromide-stained DNA bands were quantitated using the National Institutes of Health IMAGE program.

Preparation of subcellular fractions. HUVEC cells were washed twice in ice-cold PBS and incubated in 2 ml of hypotonic buffer containing 5 mM Tris (pH 7.4), 5 mM KCl, 1.5 mM MgCl₂, 0.1 mM EGTA (pH 8.0), 1 mM dithiothreitol, and proteinase inhibitor cocktail (Roche), on ice for 30 min. After homogenization with 30 strokes in a Dounce homogenizer, samples were centrifuged at 640 g for 10 min at 4°C. The supernatant was centrifuged at 10,000 g for 30 min at 4°C to obtain the heavy membrane (HM) fraction (pellet), and the supernatant was then centrifuged for 1.5 h at 150,000 g to obtain the light membrane (LM) and cytosolic (the pellet and supernatant, respectively) fractions. The HM and LM material were resuspended in 0.1 ml of Triton X-100 lysis buffer (50).

Mitochondrial isolation. After being washed twice with ice-cold PBS, the cells were scraped into buffer containing 0.2% BSA, 200 mM mannitol, 70 mM sucrose, 10 mM HEPES-KOH, and 1 mM EGTA and were incubated on ice. The lysates were homogenized with 20-30 strokes in a Dounce homogenizer, followed by centrifugation at 1,800 rpm for 10 min using Sorvall SS34 rotor. The supernatant was centrifuged at 7,000 rpm for 10 min. The pellet was washed in buffer without BSA. Mitochondria were collected in the lysis buffer, and protein concentration was measured together with the enrichment of mitochondrial markers, according to previously described protocols (22).

Western blotting. Cells were lysed in the following ice-cold lysis buffer: 20 mM Tris, pH 7.8, 140 mM NaCl, 1 mM EDTA, complete miniprotease inhibitor cocktail, 1% Triton X-100, 0.1% SDS, 1% sodium deoxycholate, 1 mM NaF, and 1 mM orthovanadate. The protein concentration of the lysates was determined with Pierce bicinchoninic protein assay against BSA standards. The samples were diluted with SDS sample buffer and stored at 20°C. Twenty micrograms of total cellular protein were separated in a 4-20% Tris-glycine gel (Invitrogen) and electroblotted to Immobilon-P membranes (Millipore). The membranes were blocked with 1% casein in PBS for 1 h, incubated with the primary antibodies for 1 h [rabbit polyclonal Cx43 antibody (Zymed, dilution 1:1,000; of note, the reaction of this antibody to Cx43 is independent of phosphorylation status), rabbit polyclonal tubulin antibody (Sigma, dilution 1:2,000), and mouse monoclonal cytochrome oxidase subunit IV (Molecular Probes, diluted 1:1,000)], and incubated with horseradish peroxidase-conjugated goat anti-rabbit antibody (Amersham) for 30 min. The membranes were then washed three times with 0.1% Tween 20 in PBS, pH 7.4, for 5 min each, and protein-antibody conjugates were detected by chemiluminescence (Super Signal CL-HRP, Pierce Chemical).

Immunostaining. HUVEC were seeded on glass coverslips and treated with 50 µM Hcy. HUVEC were incubated with 25 nM MitoTracker (Molecular Probes) and 5 nM ER-Tracker-DiO₆C (Molecular Probes) for 15 min. Cells were then washed with PBS, fixed with 3.7% paraformaldehyde for 15 min, washed three times in PBS, permeablized by incubation with 1% Triton-X in PBS for 20 min, and blocked with 2% FBS in PBS for 30 min at room temperature. Cells were stained with polyclonal anti-Cx43 antibody (Zymed, 1:200) for 1 h and detected with a Cy-5-conjugated secondary antibody (Amersham). After washing with PBS, the coverslips were mounted using Vectashield mounting medium (Vector Labs) for laser confocal microscopy (Odyssey). Images were obtained at different focal plains shifting by 0.2 µm, and stacks of images were analyzed for the number of pixels displaying an overlap in fluorescence of MitoTracker, ER-Tracker, and Cy-5 using Laser-Pix software from Bio-Rad.

Green fluorescent protein-Cx43 construct. Cx43 cDNA was PCR amplified using oligonucleotides GAATTCCACGCCACCATGGGTGACTGGAGTGCCTTG to create a EcoR I site and ACGGATCCACAATCTCCCAGGTCATCAGG to create a BamH I site at the 5' and 3' ends of Cx43, respectively. PCR products and the vector pEGFP-N1 (Clontech) were digested with EcoR I and BamH I. After ligation, competent Escherichia coli (Invitrogen) was transformed with the plasmid, and selected positive colonies were amplified. Finally, the cDNA encoding the chimeric protein was verified by the Applied Biosystems (Foster City, CA) dye terminator cycle sequencing method.

Green fluorescent protein-Cx43 transfection and analysis. HUVEC were seeded in the glass-bottomed petri dishes. The green fluorescent protein (GFP)-Cx43 fusion construct (1 µg) was used with the FuGENE 6 transfection reagent (Boehringer Mannheim), according to manufacturer's instructions. The transfection was carried out in EBM-2 medium plus 10 mM HEPES buffer. Hcy (50 µM) was added to the cells 8-24 h after transfection. The GFP-Cx43 fusion construct was described previously (8).

Imaging of Cx43-GFP in living cells. HUVEC transiently expressing Cx43-GFP were grown on a tissue-culture dish with a glass bottom, which contained 2 ml EBM-2 medium supplemented with 10 mM HEPES, pH 7.2. Fluorescence microscopy of HUVEC was performed using a Nikon epifluorescence inverted microscope (Diaphot) equipped with a silicon-intensified target camera (Hamamatsu) and temperature-controlled incubator. Cells were illuminated at an excitation wavelength of 485 nm in 10-min intervals for 30 ms using an automatic shatter (Lambda 10-2, Sutter Instruments) interfaced to Image-1-Fluor software (Universal Imaging). Images were collected at the wavelength of 530 nm using an appropriate dichroic mirror, stored, and analyzed using Image-1 (Universal Imaging) software.

Lucifer yellow transfer and electrical measurements. HUVEC cultured in the dishes with glass bottoms and treated with Hcy for 24-48 h were scrape-loaded with Lucifer yellow. Briefly, a 18-gauge needle was used to scrape confluent monolayers 4 times/dish in the presence of Lucifer yellow (1%), followed by an additional 5-min incubation. Cells were washed with PBS and fixed with 4% paraformaldehyde for 10 min (29).

Assessment of nitric oxide- and prostanoid-independent vasorelaxing responses to acetylcholine. Experiments were conducted on microdissected interlobar arteries (259 ± 2 µm ID) obtained from kidneys of rats anesthetized with pentobarbital sodium (60 mg/kg ip). Freshly dissected vessels were cut into rings 2 mm in length and placed in culture dishes (35 mm) containing DMEM (with 10% Nu-serum, 100 µg/ml streptomycin, and 100 µg/ml penicillin) without rest drugs, with Hcy (100 µM), or with 18-glycyrrhetinic acid (18-GA, 100 µM), an inhibitor of gap junction. After incubation for 6 h at 37°C in an atmosphere of 95% air-5% CO₂, the vessels were mounted on 25-µm stainless steel wires in the chambers of a multivessel myograph (JP Trading, Aarhus, Denmark) filled with Krebs buffer (composition, in mM: 118.5 NaCl, 4.7 KCl, 2.5 CaCl₂, 1.2 KH₂PO₄, 1.2 MgSO₄, 25.0 NaHCO₃, and 11.1 dextrose) containing N^G-nitro-L-arginine methyl ester (1 mM) and indomethacin (10 µM) to inhibit NOS and COX, respectively, and gassed with 95% O₂-5% CO₂. After equilibration (30 min), the vascular rings were stretched radially so that the internal circumference was 90% of that the vessels would have when relaxed under a transmural pressure of 80 mmHg (40). Isometric tension was monitored continuously. Experiments were initiated by exposing the vessels to Krebs buffer modified by increasing the concentration of KCl to 60 mM (by equimolar exchange with NaCl) to ascertain reproducibility of contractile responses. Subsequently, the vessels were washed with regular Krebs buffer and induced to contract by the addition of phenylephrine (0.5 µM) to the buffer. Once the phenylephrine-induced isometric tension development had reached a stable plateau, acetylcholine (10⁹ to 10⁵ M) or sodium nitroprusside (10¹⁰ to 10⁵ M) was added cumulatively. Agonist-induced vasorelaxation was expressed as percent reduction of phenylephrine-induced tone (38).

Statistical analysis. All the data are presented as means ± SE of the number (n) of replicative samples. When appropriate, the differences between experimental groups were analyzed using two-way ANOVA, followed by a t-test. P < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Induction of Cx43 gene and protein expression by Hcy in HUVEC. We have previously demonstrated that 24-h exposure to 50 µM Hcy resulted in endothelial dysfunction characterized by the reduced ability to generate nitric oxide in response to bradykinin or a calcium ionophore (54). Therefore, HUVEC were cultured in the presence of 50 µM Hcy for 24 h before RNA isolation. mRNA obtained from control and Hcy-treated cells was reverse transcribed, and a hybridization reaction was performed on an array containing "cardiovascular"-relevant genes (600 genes). Cx43 message level was moderate in untreated cells but increased 1.8-fold in Hcy-treated cells (relative intensity 60 and 105, respectively), well above the confidence level of the technique (1.5-fold change). The semiquantitative analysis of Cx43 expression was done by densitometric scanning of the Cx43 bands and GAPDH bands, and the density ratio of the Cx43 band to the GAPDH band was used as a measure for Cx43 mRNA expression. Cx43 mRNA level increased by 70 and 120% after 12 and 24 h, respectively.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Gene microarray and Western blot analyses of connexin43 (Cx43) in human umbilical vein endothelial cells (HUVEC) treated with homocysteine (Hcy). A: Cx43 mRNA abundance in control and Hcy-treated HUVEC. Hcy was applied at a concentration of 50 µM for 24 h. Total RNA was extracted and subjected to hybridization on "cardiovascular" microarrays. Hcy treatment resulted in a 1.8-fold increase in Cx43 message. B: representative semiquantitative RT-PCR of Cx43 mRNA levels in the control and Hcy-treated HUVEC, which were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA level (representative of 3 independent experiments). * P < 0.05 vs. control. C: representative Western blot analysis of Cx43 expression in control and Hcy-treated HUVEC. Cell proteins were separated by 4-20% Tris-glycine gel electrophoresis, and Cx43 was immunodetected with the polyclonal antibody (representative of 3 independent experiments). 24M: Hcy-treated HUVEC cells coincubated with 50 µM Mn(111)tetrakis(4-benzoic acid) porphyrin chloride for 24 h. * P < 0.05 vs. control. ** P < 0.001 vs. control. ##P < 0.05 vs. Hcy 24 h.

Upregulation of Cx43 does not increase gap junctional cell-cell communication. Cx43 has been shown to participate in formation of gap junctions between adjacent endothelial cells both in vivo and in vitro (6, 20, 35). To examine the functional consequences of the above overexpression of Cx43, gap junctional communication was evaluated by two independent techniques. With the use of a scrape-loading technique to study the extent of Lucifer yellow transfer, no detectable differences were observed between control and Hcy-treated HUVEC in their cell-to-cell communication (Fig. 2A). Furthermore, electrophysiological studies of gap junctional communication using whole cell patch clamp revealed no significant differences between control and Hcy-treated HUVEC (Fig. 2B). In an additional series of experiments, both groups of cells were treated with 0.1 mM 8-bromo-cAMP to stimulate gap junctional communication, with no difference recorded between the groups (data not shown). This apparent lack of functional correlate of overexpressed Cx43 in Hcy-treated cells could indicate either 1) its inability to form hemichannels, 2) a relatively insignificant role of Cx43 for endothelial cell-cell communication, or 3) its redistribution away from the plasma membrane. This prompted us to examine the intracellular distribution of Cx43 in HUVEC.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Gap junctional communication capacity of Hcy-treated HUVEC determined by Lucifer yellow transfer and patch-clamp measurements in cell pairs. A: Lucifer yellow transfer experiments showed no significant differences between control and Hcy-treated HUVEC. The extent of Lucifer yellow transfer was determined by the number of fluorescently labeled neighboring cells after scrape-loading 1% Lucifer yellow to HUVEC monolayers. Values are averages from 12 scrape-loading experiments. B: patch-clamp currents between HUVEC pairs were not affected by pretreatment with Hcy. Values represented 6 pairs in each group.

Effect of Hcy on Cx43 distribution in endothelial cells. Tricolor fluorescence confocal microscopy was employed to evaluate the distribution of Cx43 in HUVEC. Mitochondria and endoplasmic reticulum (ER) were identified with MitoTracker (Fig. 3B) and ER-DiO₆C Tracker (Fig. 3C), respectively, while Cx43 was detected using polyclonal antibody counterstained with Cy-5-conjugated IgG, as detailed in MATERIALS AND METHODS. In control HUVEC, Cx43 was predominantly localized to the plasma membrane, where it formed typical gap junctional plaques between the adjacent cells. (Fig. 3A). The frequency of colocalization with the mitochondrial marker was 3.6% pixels, and the ER marker averaged 33.6% pixels (Fig. 3D). Preincubation with Hcy decreased the typical punctate Cx43 localization to the plasma membrane and increased Cx43 colocalization with the mitochondrial marker (50.7% pixels), with only a minor increase in ER localization (45.9% pixels) (Fig. 3, E-H).

View larger version (90K): [in this window] [in a new window]   Fig. 3.   a: immunolocalization of Cx43 in control (A-D) and Hcy-treated (E-H) HUVEC. Anti-Cx43 antibody (Cy5-linked secondary antibody, blue; A and E), MitoTracker (red; B and F) , ER-Tracker (DiO6C, green; C and G), and a composite overlay image of all 3 staining patterns (D and H). Scale bar, 50 nm throughout. b: a diagrammatic representation of Cx43 colocalization in control (left) and Hcy-treated (right) cells. The radius of each pie diagram corresponds to the net fluorescence intensity of Cx43. Note that the proportion of Cx43 localized to the plasma membrane is decreased in Hcy-treated HUVEC, and its colocalization with the mitochondrial marker increased. The proportion of Cx43 colocalized with the endoplasmic reticulum marker did not change.

Intravital analysis of Cx43 distribution. To further examine the effect of Hcy on the Cx43 localization in the living cells, we transiently transfected HUVEC with the cDNA encoding Cx43-GFP construct. Transfected HUVEC showed Cx43-GFP assembled into typical gap junctional plaques after overnight incubation (Fig. 4A). Addition of 50 µM Hcy resulted in a punctate appearance of Cx43-GFP, which became scattered within the cells (Fig. 4, B and C). Time-lapse fluorescence microscopy of transfected cells further confirmed these observations and refined the dynamics of Cx43-GFP redistribution (Fig. 5). Fluorescently tagged Cx43 underwent intracellular redistribution from the plasma membrane to the cell interior within 5-6 h.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Distribution of green fluorescent protein-Cx43 in the endothelial cells treated with homocysteine. Hcy-treated cells expressing GFP-Cx43 fusion protein exhibited a dramatic shift of Cx43 from the intercellular plaques to the intracellular compartment(s). As detailed in MATERIALS AND METHODS, HUVEC were transfected with GFP-Cx43 construct and incubated with 50 µM Hcy for 24 and 48 h. In control cells, GFP-Cx43 did not undergo this transition (not shown).

View larger version (88K): [in this window] [in a new window]   Fig. 5.   Time-lapse video frames of GFP-Cx43 trafficking in the Hcy-treated endothelial cells. Cell transfection was performed as in Fig. 4. Fluorescence microscopy of transfected cells was combined with videomicroscopy. A few exhibited representative frames indicate that the translocation of Cx43 from the plasma membrane to intracellular compartment(s) is almost complete within 4 h after application of Hcy. Time unit is minute 0.

Cx43 is found in the mitochondria of Hcy-treated HUVEC. Immunocytochemical analysis of Hcy-pretreated HUVEC revealed Cx43 partially redistributed to mitochondria. When analyzed by cell fractionation and immunoblotting, the mitochondrial fraction (20 µg) isolated from these cells revealed the presence of Cx43 as well as cytochrome-c oxidase subunit IV (Fig. 6). Targeting of Cx43 to mitochondria was further evaluated by subcellular fractionation (Fig. 7). Cytosolic, light membrane (LM), and mitochondria-enriched heavy membrane (HM) fractions were prepared and analyzed by immunoblotting. Cx43 was present in both LM and HM fractions. Hcy stimulation increased twofold Cx43 abundance in the HM, but not in the LM. Furthermore, mitochondria isolated from HUVEC pretreated with Hcy exhibited a high level of Cx43 expression, in sharp contrast to control mitochondria, where it was expressed in trace amounts. Taken together, these data indicate that overexpressed Cx43 in response to Hcy is redistributed to mitochondria.

View larger version (49K): [in this window] [in a new window]   Fig. 6.   Identification of mitochondria as the site of Cx43 translocation in Hcy-treated HUVEC. A and B: HUVEC treated with 50 µM Hcy for 24 h. Mitochondrial fractions were isolated and subjected to Western blot analysis (representative of 3 independent experiments). The membrane was stained with Coomassie blue to demonstrate the protein loading in each lane. * P < 0.05 vs. control.

View larger version (44K): [in this window] [in a new window]   Fig. 7.   Analysis of Cx43 expression using cell fractionation. A and B: cytosolic (Cyto), mitochondria-enriched heavy membrane (HM), and light membrane fractions (LM) were isolated by differential centrifugation. Cx43 was distributed in both HM and LM. A 24-h incubation with 50 µM Hcy increased the abundance of Cx43 in the HM and had no effect on the LM. Protein loading was shown by staining the membrane with Coomassie blue. * P < 0.05 vs. control.

Effect of Hcy on nitric oxide- and prostanoid-independent vasorelaxing response to acetylcholine. One of the typical sites of Cx43 expression in the vasculature is the myoendothelial junction, where it has been shown to participate in the nitric oxide- and prostanoid-independent vasorelaxation mediated by an EDHF (12, 42, 51). To elucidate the impact of the observed Cx43 redistribution on EDHF-mediated vasorelaxation, rat renal interlobar arteries maintained for 6 h before experimentation in culture media with and without Hcy (100 µM) were compared in terms of relaxing responses to acetylcholine or sodium nitroprusside under conditions in which both NOS and COX activities were inhibited. As shown in Fig. 8, vasorelaxing responses to acetylcholine were greatly attenuated in vessels treated with Hcy. In contrast, vasorelaxing responses to sodium nitroprusside were nearly identical in vessels treated and not treated with Hcy. Relaxing responses to acetylcholine were similarly reduced in vessels treated with 18-GA (100 µM), whereas relaxing responses to sodium nitroprusside were not.

View larger version (17K): [in this window] [in a new window]   Fig. 8.   Effect of Hcy or 18-glycyrrhetinic acid (18-GA) pretreatment on the nitric oxide synthase (NOS)- and cyclooxygenase (COX)-independent relaxation of rat interlobar arteries evoked by acetylcholine. A: acetylcholine-evoked relaxation of interlobar arteries pretreated with NG-nitro-L-arginine methyl ester and indomethacin is significantly diminished in vessels pretreated with Hcy or 18-GA. * P < 0.05 vs. control. B: sodium nitroprusside-evoked relaxation of interlobar arteries pretreated with NG-monomethyl-L-arginine and indo is unaffected by Hcy or 18-GA treatment.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Screening with complementary DNA microarray for the gene expression modified by Hcy in HUVEC revealed that a gap junctional protein, Cx43, was upregulated. This unbiased selection from an array of 600 cardiovascular-relevant genes was confirmed by Western analysis of Cx43 expression. Because Cx43 has been recently found to exert diverse cellular functions, i.e., participation in EDHF-induced vasorelaxation, transport of NAD, and tumor suppression (7, 17, 28, 36, 56), we felt that this unexpected finding is worth pursuing. Neither Lucifer yellow transfer nor whole cell patch-clamp currents, however, were enhanced in Hcy-treated cells compared with intact HUVEC. These apparently contradictory findings were reconciled by the observation of redistribution of Cx43 in Hcy-treated HUVEC. Both Cx43-GFP and immunocytochemical data showed its internalization to an intracellular compartment, which was identified as mitochondria by means of 1) colocalization with the MitoTracker, 2) shift of immunodetectable Cx43 to the mitochondria-enriched HM fraction, and 3) appearance of Cx43 in the isolated mitochondria. Ex vivo studies of renal interlobar arteries pretreated with Hcy showed reduced NOS- and COX-independent vasorelaxation to acetylcholine, thus providing a potential link between cultured cells and resistance arteries. A hypothetical mechanism of Hcy-induced defective EDHF signal transduction, possibly caused by the redistribution of Cx43 from the myoendothelial junctions to mitochondia, is schematically illustrated (Fig. 9).

View larger version (22K): [in this window] [in a new window]   Fig. 9.   Hypothetical summary of Hcy-induced redistribution of Cx43 to the mitochondria, resulting in the defective propagation of endothelium-derived hyperpolarizing factor (EDHF) signaling via myoendothelial junction.

The fact that cell-to-cell communication was not affected despite the changes in expression of Cx43 is not unique. Several precedents exist when overexpression of Cx43 did not result in the increased gap junctional communication (28). Our findings may indicate, therefore, that Cx43 is not exclusive for cell-cell communication in HUVEC in vitro. Alternatively, the data may be suggestive of conformational changes, i.e., due to changes in phosphorylation (40) and/or ability to form hemichannels.

The finding of Cx43 internalization has previously been reported (2, 19), but the observed translocation from the plasma membrane to the mitochondrial compartment has not been appreciated. The functional consequences of such a translocation are unknown, have not been pursued in the present work, and require a separate study. The only relevant point to be made, however, is that this translocation to mitochondria did not result in a dramatic increase in the number of apoptotic cells (2.13 ± 0.5% in the 24-h Hcy-treated cells vs. 1.25 ± 0.85% in the control; unpublished observations), suggesting that within the reported time-frame there was no Cx43-dependent permeability transition and commitment of endothelial cells to apoptosis.

Endothelial cells within the vessels are richly interconnected via gap junctions (25) and express myoendothelial gap junctions (18, 45). Cx43 participates in formation of myoendothelial junctions (9), which appear to be indispensable for the relaxing action of a putative EDHF on vascular smooth muscle cells. Indeed, application of 18-GA resulted in both the inhibition of EDHF-induced relaxation (46) and the internalization of gap junction plaques (24). More recently, Chaytor et al. (10) have demonstrated that the combination of peptides targeting Cx43, Cx40, and Cx37 abolished NOS- and COX-independent relaxations evoked by acetylcholine in the rat hepatic artery. These data suggest the role of connexins in myoendothelial intercellular communication as a prerequisite for EDHF vasorelaxation. Our findings of suppressed COX- and NOS-independent vasorelaxation in response to acetylcholine of renal resistance arteries pretreated with Hcy are consistent with such a scenario. However, the elusive nature of EDHF and the use of blocking peptides (15) or chemical inhibitors of gap junctional communication alone, without solid support from Cx43 knockout mice (34), render certain uncertainty to the topic of EDHF signal transduction and thus hamper further interpretation of our findings. The possibility that the hyperpolarizing effect of nitric oxide per se (4) is reduced in Hcy-treated vessels, which are characterized by the "uncoupled" endothelial NOS (54), cannot be excluded.

We have previously demonstrated that Hcy reduces NO bioavailability through generation of superoxide (54), thus suggesting that in vivo NO-dependent vasorelaxation may be compromised in hyperhomocysteinemia. Indeed, vasorelaxant responses to acetylcholine were found to be reduced by Hcy in aortic rings (32). Our present finding of the Hcy-induced suppression of the NOS- and COX-independent relaxation to acetylcholine in renal interlobal arteries incriminates hyperhomocysteinemia in the defective EDHF-induced vasrelaxation. Collectively, these data are in accord with the emerging view on the vascular topography of mediators of relaxation, indicating that nitric oxide-dependent vasorelaxation predominates in conduit arteries, whereas EDHF-dependent relaxation may play a more prominent role in resistance arteries, where it involves myoendothelial gap junctions.

In summary, Hcy induces upregulation of Cx43 transcript and protein expression in the face of unchanged intercellular communication in vitro. In ex vivo renal interlobar arteries, however, Hcy results in the loss of EDHF-induced vasorelaxations evoked by acetylcholine. This paradoxical overexpression of a gap junction protein accompanied by the defective communication of EDHF signals, presumably, via myoendothelial junctions is reconciled by the observed redistribution of Cx43 into mitochondria of endothelial cells.