Production of endothelium-derived factors from sodded expanded polytetrafluoroethylene grafts

Production of endothelium-derived factors from sodded expanded polytctrafluoroethylene grafts D. A. Lewis, P h D , R. C. Lowell, M D , R. A. C a m b r i a , M D , P. C. Roche, P h D , P. Gloviczki, M D , and V. M. Miller, P h D , Rochester, M i n n . Purpose: Experiments were designed to determine whether endothelium isolated from adipose tissue and sodded onto expanded polytetrafluoroethylene grafts release endothelium-derived vasoactive factors. Methods: Thin-walled expanded polytetrafluoroethylene grafts (6 m m internal diameter, 6 cm length, 30 ~m pore size), one sodded with autogenous endothelial cells, the other unsodded, were implanted bilaterally in carotid arteries of 30 male mongrel dogs. Dogs were treated with 325 mg aspirin daily. After 6 weeks grafts were excised and perfused in a bioassay system. Effluent from the grafts stimulated with either acetylcholine, thrombin, adenosine 5-diphosphate, or the calcium ionophore A23187 was superfused over rings of canine femoral arteries without endothelium contracted with phenylephrine. Effluent from the grafts was analyzed by radioimmunoassay for thromboxane B2, 6-ketoprostaglandin FI~ , endothelin-1, and C-type natriuretic peptide. Results: Ninety percent of the sodded grafts and 87% of the unsodded grafts were patent after 6 weeks. Bioassay rings superfused with effluent from sodded grafts stimulated with acetylcholine relaxed significantly more than rings superfused with effluent from similarly stimulated unsodded grafts. Biochemical analysis of the effluent showed an increase in 6-keto prostaglandin F h and C-type natriuretic peptide and a decrease in endothelin-1 and thromboxane B 2 release from the sodded compared with the unsodded grafts. Scanning electron microscopy showed a continuous layer of endothelial cells lining only the sodded grafts. Staining for ot-actin and heavy-chain myosin showed a differentiated layer of smooth muscle below the endothelial layer on the sodded grafts. Finally, there was positive staining for C-type natriuretic peptide and endothelin-1 in the endothelium of the sodded grafts. Conclusions: These results indicate that endothelial cells of sodded expanded polytetrafluoroethylene grafts produce endothelium-derived vasoactlve factors. In addition, receptor-coupled synthesis/release of these factors is retained in sodded endothelial cells. (J Vasc Surg 1997;25:187-97.)

When autogenous blood vessels are not available for arterial reconstruction in peripheral arterial occlusive disease, prosthetic grafts offer an alternative conduit. These grafts, however, have low long-term patency. 1,2 T o improve patency "seeding" or "sodding" with autogenous endothelial cells onto the From the Departments of Surgery and Physiology & Biophysics and Laboratory Medicine, Mayo Clinic and Foundation. Funded in part by NIH grant HL42614 and the Mayo Clinic and Foundation. Reprint requests: Virginia M Miller, PhD, Departments of Surgery, Physiologyand Biophysics,Mayo Clinic and Foundation, Medical Sciences Building 4-57, 200 First Street, SW, Rochester, MN 55905. Copyright © 1997 by The Societyfor Vascular Surgeryand International Society for Cardiovascular Surgery, North American Chapter. 0741-5214/97/$5.00 + 0 24/1/74860

prosthetic graft lumen has been suggested. Clinical results with small-diameter (6 m m or less) endothelial seeded grafts vary. One-year studies have shown a 82% patency rate for sodded grafts placed as femoropopliteal bypass versus 31% in unsodded expanded polytetrafluoroethylene (ePTFE) grafts? Another study showed no difference in patency rates. 4 Studies examining sodding o f prosthetic grafts have focused on the clinical impact o f the technique, patency, and defining anatomic characteristics and function o f these endothelial monolayers as it relates to antithrombogenicity. This study was designed (1) to identify factors released by the cells lining sodded grafts that might contribute to their antithrombogenicity, and (2) to identify stimuli that might p r o m o t e release o f antithrombogenic factors. It was hypothesized that endothelial cells on the sodded grafts 187

JOURNAL OF VASCULAR SURGERY

188

Lewiset al.

would produce endothelium-derived relaxing factors in response to pharmacologic-receptor stimulation. METHODS

All animals were cared for following The Principles of Laboratory Animal Care and The Guide For the Care and Use of Laboratory Animals. Surgery. Dogs (n = 30, mean weight 23.2 _+ 0.7 kg) were anesthetized with methohexital sodium (10 m g / k g ) and intubated. Surgical levels of anesthesia were maintained with halothane (1% to 1.5%) for the remainder of the surgical procedure. Bicillin (1.2 million U) was given prophylactically by intramuscular injection to each animal before surgery. A catheter was inserted into the femoral artery to monitor arterial pressure and obtain blood samples. A catheter was also placed in the cephalic vein so that a continuous infusion of saline solution (7 c c / k g / h r ) could be maintained. A midline epigastric incision was made, and a minimum of 20 gm of omental or falciform fat was removed for harvesting of endothelial cells. Bilateral incisions were made in the neck to expose the carotid arteries. Blood flow was measured through the common carotid artery with an ultrasonic flow probe (Transonic Systems Inc., Ithaca, N. Y.). Heparin (100 U / k g intravenously) was administered and permitted to circulate for 5 minutes, after which the carotid artery was clamped and bisected. Thin-walled ePTFE (Impra Inc., Tempe, Ariz.) interposition grafts (6 cm in length, 6 mm internal diameter, and 30 ~m pore size) were placed in the carotid arteries with continuous nmning 6.0 prolene sutures in an end-to-end fashion. One graft was implanted unsodded, whereas the contralateral side was sodded with autogenous endothelial cells and then implanted. After both grafts were implanted, measurements of blood flow were repeated distal to the grafts. Dogs recovered from surgery and were given aspirin (325 rag/day) for the remainder of the study. This treatment mimics treatments used clinically. After 6 weeks animals were anesthetized with pentobarbital (30 m g / k g ) , and the grafts were removed either for histology or bioassay experiments. Endothelial sodding. Isolation of endothelial cells and sodding onto ePTFE grafts was performed according to the method of Williams. s In brief, 100 ml of arterial blood was removed and centrifuged to obtain serum. Fat from the omentum or falciform pad was placed in a sterile beaker with "harvest media" (20% fetal bovine serum, GIBCOBRL, Life Technologies, Grand Island, N. Y. in Media 199; GIBCOBRL), chopped into fine pieces, and washed

January1997

through a sieve. The resulting slurry was placed in "digesting media" (Dulbecco's phosphate-buffered saline solution, GIBCOBRL; collagenase CLS1; 8 m g / m l , Worthington Biochemical Corp., Freehold, N. J.; human serum albumin; 8 m g / m l , Sigma Chemical Corp., St. Louis, Mo.) and digested by gently shaking the slurry for 30 minutes at 37 ° C. The digested slurry was centrifuged at 1800 RPM for 4 minutes, and the resulting pellet was resuspended in Dulbecco's solution. The resuspended pellet/ Dulbecco's was again centrifuged at 1800 RPM for 4 minutes. This final pellet was resuspended in sodding media (autogenous dog serum and Eagle's media 199, GIBCOBRL; 1:6). Cells were then infused into the ePTFE graft followed by four graft volumes of sodding media infused (1 m l / m i n ) at a pressure of - 5 psi to force the endothelial cells into the graft wall. Bioassay e x p e r i m e n t s . Grafts were excised, gently irrigated with a chilled modified Kreb's Ringer bicarbonate solution in txmol/L: 118.3 NaC1, 4.7 KCI, 2.5 CaC12, 1.2 MgSO4, 1.2 KH2PO4, 25.0 NaHCOs, 0.026 edetate calcium disodium, and 11.1 glucose (control solution). Sodded and unsodded ePTFE grafts from the same animal were cut 0.5 cm from each anastomosis and cannulated onto stainless steel tubes. The cannulated grafts then were placed in a bioassay apparatus. 6 Grafts were maintained at 37 ° C in the control solution, which was gassed with 95% 02 and 5% CO 2. The bioassay apparatus also contained two stainless steel tubes through which control solution was perfused directly. The tubes and grafts were perfused with the control solution at a constant rate of 3 m l / m i n with a roller pump (Gilson Miniplus 2, Middleton, Wis.). Care was taken to ensure flow through the graft was in the same direction in vitro as in vivo. Two rings of femoral artery (bioassay rings) were removed from an unoperated, unmedicated dog anesthetized with pentobarbital and then exsanguinated on the same day as the experiment. The endothelium was removed, and both were suspended for the measurement of isometric force directly below the perfused stainless steel. The ring/stirrup/force transducer assembly could move freely, permitting the bioassay rings to be superfused with either effluent from a graft (graft superfusion) or from a stainless steel tube (direct superfusion). Drugs were added to the perfusing fluid of the stainless steel tubing to determine their direct effect on the vascular smooth muscle of the bioassay ring or above the graft, exposing the graft lining to the agonist to determine whether the cells on the graft lining produce endothelium-de-

JOURNAL OF VASCULARSURGERY Volume 25, Number 1

Sodded

Lewis et al.

189

Unsodded

Fig. 1. Scanning electron micrographs (×500) of sodded (left panel) and unsodded (right panel) ePTFE grafts. Sodded graft is lined by cells forming typical cobblestone appearance of endothelial cells. No consistent lining was observed on unsodded grafts. rived factors that affect the smooth muscle of the bioassay ring. The bioassay ring was stretched to 10 gm of resting tension in 2-gin increments while being superfused with effluent from the direct line. 7 Bioassay rings and not the grafts were contracted with phenylephrine (1 × 10 -6 mol/L). The absence ofendothelium on the bioassay rings was determined by lack of relaxation to acetylcholine (1 × 10 -6 mol/L) during direct superfusion. Bioassay rings were contracted with phenylephfine under their respective direct lines. Once the contraction was stable, an agonist was injected into the perfusate above the direct lines, and the response was recorded. Rings were then placed under their respective grafts, and contracted with phenylephrine, and the same agonist was injected into the perfusate above the grafts. The effluent from the sodded and unsodded grafts also was collected for subsequent biochemical analysis. These steps were repeated for each agonist. Agonists were administered in the same order for all sodded and unsodded grafts. All sodded and unsodded grafts were stimulated with acetylcholine (1 × 10 -6 mol/L) and the calcium ionophore A23187 (i × 10 -6 mol/L). Half of the grafts were also stimulated with adenosine 5-diphosphate (1 × 10 -4 mol/L) and the other half with thrombin (0.1 U/ml). Thus each graft was stimulated with only three agonists. No experiment lasted more than 5 hours. A minimum of 20 minutes was permitted to elapse between agonists. Drugs and chemicals. The following drugs were used: acetylcholine, adenosine 5-diphosphate,

the calcium ionophore A23187, dimethyl sulfoxide (DMSO), phenylephrine, and thrombin (all from Sigma Chemical). Drugs were prepared fresh daily and kept on ice. The calcium ionophore 3,23187 was dissolved in DMSO and diluted in distilled water (final concentration of DMSO: 8.2 × 10 -3 mol/L). The final concentration of DMSO has been shown not to affect smooth muscle. 8 Concentrations of the drug are reported as the final molar concentration in the superfusing solution. Biochemical assays. Assays of effluent from the bioassay system were obtained for thromboxane B2 (the stable metabolite of thromboxane A2) , 6-keto prostaglandin Fz~ (the stable metabolite ofprostacyclin), endothelin-1, and C-type natriuretic peptide by radioimmunoassay (DuPont NEN, E.I DuPont deNemours & Co., Boston, Mass.; Amersham International, Amersham, U. K . ) . 9 The lower limits of the assays arc 10 p g / m l for thromboxane B2, 20 p g / m l for 6-keto prostaglandin FI~ , 0.5 p g / m l for endothelin-1, and 2 p g / m l for C-type natriuretic peptide. Histology. Sodded (n = 3) and unsodded grafts (n = 3) not used in bioassay experiments were cut into three 2-centimeter sections (proximal, mid, and distal), fixed in formalin, and prepared for scanning electron microscopy. Immunohistochemistry. All immunohistochemistry was performed in the Immunostaining Laboratory at the Mayo Clinic, Rochester, Minnesota. Midportions of sodded and unsodded grafts used in bioassay experiments were cut into two pieces at the conclusion of the experiment. One piece was snap-

190

Lewis et al.

JOURNAL OF VASCULARSURGERY January 1997

JOURNAL OF VASCULAR SURGERY Volume 25, Number 1

frozen, and the other was fixed in 10% formalin for 24 hours and subsequently embedded in paraffin. Frozen pieces of grafts were stained for c~-actin and heavy-chain myosin (the primary antibodies for these were gifts from Dr. Gary K. Owens, University of Virginia). Sections (5 Ixm) were placed on a silanized slide and stored in a dehumidified chamber overnight. Sections were fixed in acetone (4 ° C) for 5 minutes, allowed to air-dry, and then incubated in phosphate-buffered saline solution (PBS)/ethylenediamine tetraacetic acid for 11/2 minutes and washed with PBS. Sections were then incubated in 0.1% azide/3% H202 for 10 minutes and rinsed in PBS incubated with 5% goat serum for 15 minutes. The serum was drained, and the primary antibody (1:200 dilution) was applied and incubated for 60 minutes at room temperature. The sections were washed three times for 10 minutes each in PBS, and the secondary antibody (biotinylated F(ab)2 fragment) in a 1:200 dilution was added and incubated for 30 minutes at room temperature. The sections were washed three times for 10 minutes each in PBS. Streptavidin was added (1:250 dilution) and incubated for 15 minutes at room temperature. The sections were again washed three rimes for 10 minutes each in PBS. Finally, all slides were stained for 15 minutes in 3-amino-9 ethylcarbazole (AEC) solution and counterstained for 1 minute in mercury-ftee hematoxylin-eosin. The sections were washed three times for 10 minutes each in distilled water. Slides were mounted in aqueous glycerol gelatin media. Segments of grafts placed in formalin were examined for endothelin-1, endothelial nitric oxide synthase (ecNOS), and C-type natriuretic peptide-22 (CNP). All segments of grafts were cut into 5 txm sections, deparatfinized in xylene, and then rehydrated through graded alcohols. Endogenous peroxidase activity was blocked by incubating the sections in 50% M e O H / H 2 0 2 for 10 minutes and then rinsing with tap water. Slides that were to be stained for

Lewis et al.

191

endothelin-1 and CNP were pretreated by microwaving in 10 txmol/L citrate (pH = 6.0) and were then washed in tap water for 1 to 2 minutes. Nonspecific protein binding was blocked on all slides with 5% normal goat serum (NGS)/PBS/0.05% Tween 20 (pH = 7.4) for 10 minutes at room temperature and then blotted dry. Slides that were stained for endothelin-1 were incubated in anti-endothelin-1 polyclonal antibody (Phoenix Pharmaceuticals, Inc., Mountain View, Calif.), and those stained for CNP were incubated in anti-C type natriuretic peptide-22 polyclonal antibody (Phoenix Pharmaceuticals, Inc., Mountain View, Calif.), both in a 1:2000 dilution overnight at room temperature and then drained, rinsed, and incubated in biotinylated goat anti-rabbit immunoglobulin G (1:200 dilution) for 30 minutes. Slides stained for eNOS were incubated in antieNOS monoclonal antibody (Transduction Laboratories, Lexington, Ky.) 5 txg/ml dilution for 60 minutes, then drained, rinsed, and incubated in goat anti-mouse immunoglobulin G (1:200 dilution) for 30 minutes. All slides were then rinsed well in tap water, incubated for 30 minutes in peroxidase-labeled streptavidin (1:500 dilution), stained for 15 minutes in AEC solution, and counterstained for 1 minute in mercury-free hematoxylin-eosin. After rinsing for 5 minutes in tap water, slides were mounted in aqueous glycerol gelatin media and silanized slides. For all immunostaining sections with a single antibody were batched to control for variability in antibody preparation. To control for specificity of the antibody stain adjacent sections of tissue were incubated in paral,lel in 1% goat serum/PBS/Tween in the absence of primary antibody or normal rabbit serum diluted 1:2000. Calculations and statistical analysis. Only data from paired grafts (sodded and unsodded from the same animal) were included in the analysis. For example, during data collection some bioassay rings

Big. 2. Micrographs (Xl00) of sodded (left panel) and unsodded (right panel) ePTFE grafts stained with antibody for heavy-chain myosin (a, b), C-type natriuretic peptide (CNP: c, d), endothelin-1 (e, f), and endothelial nitric oxide synthase (ecNOS: g, h). Bar denotes extent of ePTFE graft; arrow indicates infiltration of white cells. Sodded grafts exhibited positive, specific staining for heavy-chain myosin, indicating differentiated smooth-muscle cells. Sodded grafts also exhibited positive staining for CNP throughout section; only cells of adventia had positive staining CNP in unsodded grafts. Positive staining for endothelin-1 was present in cells lining lumen and smooth muscle of sodded graft. There was nonspecific staining along luminal border in unsodded grafts. Positive staining for ecNOS was present in cells lining lumen and smoothmuscle/ePTFE border in sodded graft. Unsodded grafts showed nonspecific staining throughout section.

192

JOURNAL OF VASCULAR SURGERY January 1997

Lewis et al.

Graft Superfusion:

Direct

Direct

ACh, 10 -6 M

\

lO-6M ,

I J/ __1,o 5 rain

T

Phenylephrine, 10 "5 M Fig. 3. Sample trace of response of bioassay ring to effluent from direct line (superfusion through metal tube) and effluent from sodded graft stimulated only by flow of perfusing solution and following stimulation with acetylcholine (1 × 10 -6 tool/L).

developed spontaneous activity; these data were not interpretable. Therefore data from the point where spontaneous activity occurred were not used from either graft for the bioassay experiments. Likewise, if one of the grafts was not patent, neither side was included in any of the analyses. All data are expressed as mean + SEM; n equals the number of dogs from which the grafts or samples were taken. Responses of the bioassay rings are expressed as a percent change in tension from the contraction to phenylephrine. Paired t tests were performed on the maximal relaxations and on all the effluent assay experiments. Two-way analysis of variance was used to compare blood flow and mean arterial pressure. The a level for all tests was set at O.O5. RESULTS Patency. Twenty-six (90%) of the 29 sodded grafts (one graft was not put into the analysis because of structural failure) were patent, and 26 (87%) of the 30 unsodded grafts were patent. Blood flow and blood pressure. Mean arterial pressure was not different at the time of surgery compared with 6 weeks after surgery (84.7 and 79.7 mm Hg, respectively). Blood flow did increase through the unsodded grafts from immediately after implantation (183.2 + 16.6 m l / m i n ; n = 19) to

death 249.7 + 29.4 ml/min; n = 19) but did not change in the sodded grafts (after implantation: 204.9 + 16.4 m l / m i n ; n = 19 and death: 227.7 + 24.8 ml/min; n = 19). Histology. Scanning electron microscopy showed a confluent layer of cells lining sodded grafts. There were no confluent cells lining unsodded grafts (Fig. 1). Immunohistochemistry. In general, none of the sections stained in the absence of primary antibodies. In all sections from sodded grafts a layer of cells staining positive for heaw-chain myosin (Fig. 2a) was observed between the cells lining the lumen and the graft material. A general feature of the unsodded grafts was white cells infiltrating the graft from the graft/adventitia boarder (Fig. 2b, arrow). In sodded grafts positive staining for CNP (Fig. 2c, d) was observed in the cells lining the lumen, cells in the underlying layer, and those infiltrating the ePTFE and adventia. In the unsodded graft positive staining for CNP was observed only in cells of the adventia. In sodded grafts positive specific staining for endothelin-1 (Fig. 2e) was observed only in ceils lining the lumen and the underlying layer but not in the ePTFE and adventia. Cells stained for endothelin-1 along the lumenal border but not in the ePTFE or adventia in unsodded grafts (Fig. 2f). Positive staining for ecNOS appeared along the lining cells/ePTFE border in

JOURNAL OF VASCULAR SURGERY Volume 25, Number 1

Lewis et al.

193

¢0

|m

mS

¢-

-10 C

I

-15,

° ]

Sodded (n=12) I Unsodded (n=12)

-20 ] -25

"k

-30 Fig. 4. Relaxations of bioassay rings to effluent from sodded and unsodded ePTFE grafts stimulated with acetylcholine (1 × 10-6 tool/L; mean -+ SEM). Values are expressed as percent change in tension after contraction with phenylephrine (1 × 10 - 6 or S tool/L). *Represents significant difference between sodded and unsodded grafts, p < 0.05. sodded grafts (Fig. 2g). The unsodded grafts (Fig. 2h) showed diffuse nonspecific staining throughout the graft. Bioassay. Bioassay rings (7 of 12) relaxed on exposure to superfusate from sodded grafts in the absence of agonist stimulation (Fig. 3); only 3 of 12 rings relaxed when exposed to effluent from the unsodded grafts in the absence of agonist stimulation. Bioassay rings superfuscd with the effluent of sodded grafts stimulated with acetylcholine (1 × 10 -6 mol/L) relaxed significantly more than rings superfused with effluent from the unsoddcd grafts stimulated with acetylchofine (Fig. 4). Relaxations of the bioassay tings to adenosinc 5-diphosphate, thrombin, and the calcium ionophore 3223187 were not statistically different between sodded and unsodded grafts (Table I). Acetylcholine, adenosine-5-diphosphate, thrombin, and A23187 did not cause relaxation ofbioassay rings when applied through the direct line. Biochemical assays. 6-keto prostaglandin FI~ was measurable in effluent from both sodded and unsodded grafts stimulated with any of the agonists. After perfusion with A23187 was performed, significantly more 6-keto prostaglandin FI~ was released from the sodded as opposed to the unsodded grafts (Fig. 5). The amount of thromboxane ]32 released was significantly less in sodded grafts compared with the unsodded grafts after perfusion with acetylcholine or 3,23187 was performed. Endothelin-1 and

Table I. Responses of paired canine femoral arteries (without endothelium) to effluent from sodded or unsodded ePTFE grafts stimulated with adenosine-5-diphosphate, thrombin, or A23187

Adenosine- 5- Diphosphate (10 4 t o o l / L )

Sodded ePTFE grafts

UnsoddedePTFE grafts

- 3 4 . 6 + 11.0

- 9 . 4 -+ 5.6

- 1 6 . 9 ± 11.0

3.2 _+ 6.2

- 1 7 . 2 ± 12.0

3.3 ± 2.7

(n = 6) Thrombin (0.1 U / m l )

(n = s) A23187 (10 -6 m o l / L ) (n = 9)

Data are paired and represent percent change in tension from the contraction with phenylephrine (1 × 10 -6 or --S t o o l / L ) and are presented as mean + SEM; n equals the number of individual dogs.

C-type natriuretic peptide were above dctectable limits when stimulated with either acetylcholine (3 of 11 samples), thrombin (2 of 7 samples), adenosine 5-diphosphate (2 of 6 samples), or the calcium ionophore (3 of 13 samples) for both sodded and unsodded grafts (Table II). DISCUSSION

Results of this study extend previous observations beyond morphologic and histologic characteristics of

194

JOURNAL OF VASCULAR. SURGERY ]~anuary 1997

Lewis et aL

250.

2OO

Sodded Grafts E ~ Unsodded GraRs

!

1501

Y

5o1 0

C "

ol Thromboxane

A23187

B2

450, 400, 350. N m

300.

~" )< .~

200.

E

loo.

O~

50.

I-,-

0

Sodded Grafts i '" I Unsodded Grafts

250. 150.

Acety~lcholine ~ADI

Thrombin

| A23187

Fig. 5. Biochemical analysis of effluent with radioimmunoassay for 6-keto prostaglandin FI~ (top panel) and thromboxane B 2 (bottom panel; mean ± SEM). Grafts were stimulated with either acetylcholine (1 × 1 0 - 6 tool/L; n = 11), thrombin (0.1 U/ml; n = 7), adenosine 5-diphosphate (1 × 1 0 - 4 mol/L; n = 6), or calcium ionophore A23187 (1 × 10 6 mol/L; n = 13). *Represents significant difference in production of prostacyclin and thromboxane in sodded grafts compared with unsodded grafts. sodded grafts i°-16 to indicate that cells lining sodded ePTFE grafts release vasoactive substances in response to mechanical and receptor activation. This conclusion is based on functional responses o f bioassay tissue, direct biochemical analysis o f graft effluent, and immunostaining o f the grafts. It was not possible to identify the cells lining the grafts as endothelium by use o f cell-specific antibody for yon Willibrand factor. However, even with this limitation cells lining the grafts have the morphologic characteristics of endothelial cells and respond to mechanical and agonist stimulation as do endothelial cells o f arteries. For example, before the first agonist was applied to the cells, relaxations were observed in many o f the bioassay tings after they were placed under the sodded grafts. Thus flow o f perfusate through the sodded graft induced release o f relaxing factors from cells lining the grafts as

occurs in perfused arteries with endothelium. 17 One o f the most interesting results o f this study is that cells from sodded grafts produce relaxing factors in response to the agonist acetylchollne as do arterial endothelial cells? 8 This finding is important, because damaged endothelium and endothelial cells in culture do not release endothelium-derived relaxing factors when stimulated with acetylcholine. 19-22 Relaxation o f bioassay rings in response to stimulation o f grafts with adenosine 5-diphosphate, thrombin, and the calcium ionophore were not significantly different between graft types. The reasons for this are not clear. The duration of the bioassay experiments was relatively short; however, it is possible that constant flow through the grafts depleted substrate or other cofactors or receptor-coupling mechanisms necessary for the production/release o f endothelium-derived relaxing factors. It is also possi-

JOURNAL OF VASCULAR SURGERY Volume 25, Number l

Lewis et ak

195

Table ii. Endothelin-1 and C-type natriuretic peptide values from the effluent of sodded and unsodded ePTFE grafts explanted from dogs and stimulated with acetylcholine, adenosine- 5- diphosphate, thrombin, or A2 318 7 Endothelin-1 (pg/ml)

Ace@choline 10 -6 ~ m o i / L

C-type natriuretic peptide (pg/ml)

Sodded ePTFEgrafts

UnsoddedePTFEgrafts

Sodded. ePTFEgrafts

UnsoddedePTFEgrafts

14.33 ± 2.84

13.97 +_ 1.53

56.67 _+ 3.93

38.33 + 6.36

13.15 + 4.45

10.65 -+ 1.95

10.55 ± 1.45

16.20 _+ 7.10

10,60 ~+ 1.97

11.10 ± 2.20

(n = 3) Adenosine-5-Diphosphate 10 .4 b~mol/L (n = 2) Thrombin o.1 u / M (n = 2) A23187 10 -6 ~ m o l / L (n = 3 )

115.00 -+ 76

71.50 ± 4.50

135.00 _+ 37

100.50 ± 36.50

62.50 ± 15.50

90.00 --- 43.40

Data are paired and presented as mean + SEM.

ble that the bioassay tissue, in this case canine femoral artery, may not have been the most sensitive tissue to respond to all the factors being released from the cells on the grafts (for example, hyperpolarizing factor). Finally, because there was most likely a simultaneous release of both relaxing and contracting factors, the response of the smooth muscle represented the sum of these opposing effects. Arterial endothelial cells release nitric oxide in response to stimulation with acetylcholine.23 Whether nitric oxide is released in response to ace@choline in sodded endothelial cells is unknown at this time. Direct measurement of nitric oxide/total NO x in the effluent of sodded compared with superfusate through the direct line (stainless steel tube) was equivocal with chemilumenscence (preliminary unpublished observations from our laboratory). Nonspecific staining of unsodded grafts for ecNOS was observed and may reflect problems with specificity of the antibody. Four other endothelium-derived vasoactive factors were measured directly from the effluent of sodded and unsodded grafts: CNP, endothelin-1, 6-keto prostaglandin Fl~ (the stable metabolite ofprostacydin), and thromboxane B 2 (the stable metabolite of thromboxane A2). C-type natriuretic peptide has been implicated as a possible endothelium-derived hyperpolarizing factor. 24 Arteries are less responsive to CNP than veins, which may account for less robust relaxations of the arterial bioassay rings to graft effluent. In those grafts in which it was possible to detect CNP in the effluent, there was increased release in the sodded versus the unsodded grafts. There was also more positive stain-

ing for CNP in the sodded grafts. Endothelin-1 was produced by cells of sodded and unsodded grafts. Although there was little evidence of a continuous lining of cells on the unsodded graft, there may have been enough cells of endothelial or other origin trapped within the nodes of the graft material to respond to the agonists used in the bioassay experiments. Not only were CNP and endothelin-1 measured in the effluent, but they also could be identified by immunohistochemistry in cellular elements differentially distributed throughout the sodded and unsodded grafts. Positive staining for CNP, endothelin-1, and ecNOS in the cells lining the lumen of the sodded grafts supports the suggestion that the sodded grafts have a viable endothelial layer adjacent to the lumen. Nonspecific staining of graft material to CNP and endothelin-1 could reflect the presence of peptide released from cellular elements during the bioassay experiments and their subsequent binding to the graft material. Effluent from sodded and unsodded grafts showed significant amounts of prostanoids. Whereas the functional studies with bioassay rings showed relaxations only to acetylcholine in the sodded grafts, only stimulation of the sodded graft with the calcium ionophore A 2 3 1 8 7 showed a significant increase in 6-keto prostaglandin FI~. It is possible that acetylcholine stimulates release of relaxing factors other than prostacyclin. 2s,26 Positive staining for smooth-muscle heavy-chain myosin is characteristic of a differentiated layer of smooth-muscle cells27 under the cells lining the lu; men of sodded but not unsodded grafts. This result

196

JOURNAL Ot: VASCULARSURGERY January 1997

Lewis et al.

suggests t h a t a "neo-vessel" is f o r m i n g in the l u m e n o f the s o d d e d e P T F E graft. I t remains to be seen w h e t h e r the d i m e n s i o n s o f this "neo-vessel" w o u l d be l i m i t e d by flow characteristics in the graft as in the n e o i n t i m a o f u n s o d d e d p r o s t h e t i c a n d vein grafts. 2s-a° C N P , e n d o t h e l i n - 1 , a n d nitric oxide m a y influence differentiation a n d g r o w t h o f s m o o t h - m u s cle cells. 3133 S o d d e d grafts have n u m e r o u s advantages t h a t m a y p r o m o t e l o n g - t e r m patency. S o d d e d grafts contain a lining o f cells w i t h a n a t o m i c characteristics o f e n d o t h e l i a l cells. I n a d d i t i o n , e n d o t h e l i a l ceils lack constitutive expression o f a cell surface l i g a n d for platelet receptors 34 a n d secrete e n d o t h e l i u m - d e r i v e d relaxing factors like prostacyclin a n d perhaps nitric oxide, w h i c h w o u l d reduce platelet a g g r e g a t i o n . 35-a7 U n s o d d e d grafts, in contrast, p r o d u c e d significantly m o r e t h r o m b o x a n e 8 2 t h a n the cells o n the s o d d e d grafts, w h i c h m a y be associated w i t h increased platelet a g g r e g a t i o n . 3<38 T h u s decreased p a t e n c y in s m a l l - d i a m e t e r grafts m a y n o t only be d u e t o the mechanical surface b u t also to release o f factors by cells infiltrating u n s o d d e d grafts, w h i c h p r o m o t e platelet a g g r e g a t i o n . 34 I n conclusion, this study confirms observations o f others l°a~ t h a t s o d d i n g o f a u t o g e n o u s e n d o t h e l i a l ceils o n t o e P T F E grafts p r o d u c e s a c o n f l u e n t layer o f cells lining the grafts 4 to 6 weeks later. This study extends previous observations to indicate t h a t t h e cells lining the s o d d e d grafts p r o d u c e factors t h a t have t h e p o t e n t i a l to influence t o n e a n d g r o w t h o f u n d e r l y i n g cellular e l e m e n t s a n d affect a g g r e g a t i o n a n d a d h e s i o n o f b l o o d b o r n e elements. The authors acknowledge Dr. Gary Owens of the University of Virginia, who generously supplied antibodies for the ct-actin and heavy-chain myosin, and the technical help of Mr. William Anding, Mrs. Arlene Morris, Mr. Steve Ziesmer, and Mr. Kevin Rud.

REFERENCES

1. VA Study Group 141. Comparative evaluation of prosthetic, reversed, and in Sim vein bypass grafts in distal popliteal and tibial-peroneal revascularization. Arch Surg 1988;123:434-8. 2. Londrey GL, Ramsey DE, Hodgson KJ, Barkmeier LD, Sumner DS. Infrapopllteai bypass for severe ischemia: comparison of autogenous vein, composite, and prosthetic grafts. J Vase Surg 1991;13:631-6. 3. Herring MB, Compton RS, LeGrand DR, Gardner AL, Madison DL, Glover JL. Endothelial seeding of polytetrafluoroethylene popllteal bypasses. A preliminary report [published erratum appears in J Vasc Surg 1987;6:547]. J Vase Surg 1987;6:114-8. 4. Ziila P, Fasol R, Deutsch M, Fischlein T, Minar E, Hammerle A, Krupicka O, Kadletz M. Endothelial cell seeding of poly-

5.

6.

7.

8.

9.

10.

11.

12.

13.

14. 15.

16.

17.

i8.

19.

20.

21.

22.

23.

tetrafluoroethylene vascular grafts in humans: a preliminary report. J Vasc Surg 1987;6:535-41. Williams SK. Microvascular perfusion and transport in health and disease. In: McDonagh PF, editor. Isolation and culture of microvessel and large-vessel endothelial cells: their use in transport and clinical studies. Basel, Switzerland: S. Karger, 1987:204-45. Rubanyi GM, Lorenz RR~ Vanhoutte PM. Bioassay of endothelium-derived relaxing factor(s): inactivation by catecholamines. Am J Physiol 1985;249:H95-H101. Miller VM, Aarhus LL, Vanhoutre PM. Modulation of endothelium-dependent responses by chronic alterations of blood flow. Am J Physiol 1986;251:H520-7. Miller VM, Vanhoutte PM. Enhanced release of endothelium-derived factor(s) by chronic increases in blood flow. Am l Physiol 1988;255:H446-51. Margulies KB, Hildebrand FL, Jr., Lerman A, Perrella MA, Burnett JC, Jr. Increased endothelin in experimental heart failure. Circulation 1990;82:2226-30. Williams SK, Schneider T, Kapelan B, Jarrell BE. Formation of a functional endothelium on vascular grafts, l Elect Micro Tech 1991;19:439-51. Williams SK, Carter T, Park PK, Rose DG, Schneider T, Jarrell BE. Formation of a multilayer cellular lining on a polyurethane vascular graft following endothelial ccll sodding. J Biomed Mat Res 1992;26:103-17. Kaufman BR, Fox PL, Graham LM. Platelet-derived growth factor production by canine aortic grafts seeded with endothelial cells. ]Vasc Surg 1992;15:699-707. Kent KC, Shindo S, Ikemoto T, Whittemore AD. Species variation and the success of endothelial cell seeding. J Vase Surg 1989;9:271-6. Williams SK, Jarrell BE. Thrombosis on endothelializable protheses. Ann NY Acad Sci 1987;516:145-57. Hussain SM, Long GW, Juleff RS, et al. Comparison of immediate seeding of endothelial cells with culture lining of small diameter ePTFE carotid interposition grafts. J Surg Res 1991;51:33-9. Shepard AD, Eldrup-Jorgensen J, Keough EM, et al. Endothelial cell seeding of small-caliber synthetic grafts in the baboon. Surgery 1986;99:318-25. Rubanyi GM, Romero JC, Vanhoutte-PM. Flow-induced release of endothelium-detived relaxing factor. Am J Physiol I986;250:Hl145-9. Furchgott RF, Zawadsld I. The obligatory role of endothelium cells in the relaxation of arterial smooth muscle by acetylcholine. Nature 1980;288:373-6. Miller VM, Reigel MM, Hollier LH, Vanboutte PM. Endothclium-dependent responses in autogenous femoral veins grafted into the arterial circulation of the dog. I Clin Invest 1987;80:1350-7. Tracey WR, Peach MI. Differential muscarinic receptor mRNA expression by ftcshly isolated and cultured bovine aortic endothelial cells. Circ Res 1992;70:234-40. Loeb A, Owens G, Peach M. Evidence for endotheliumderived relaxing factor in cultured cells. Hypertension i985; 7:804-7. Nal
JOURNAL OF VASCULARSURGERY Volume 25, Number 1

24.

25.

26.

27.

28.

29.

30.

31.

oxide by vascular endothelial cells accounts for the activity of EDRF. Nature 1987;327:524-6. Wei C, Hu S, Miller VM, Burnett JC, Jr. Vascular actions of C-typc natriuretic peptide in isolated porcine coronary arteries and coronary vascular smooth muscle cells. Biochem Biophys Res Commun 1994;205:765-71. Chaudhari A, Pedram A, Kirschenbaum MA. Prostanoid biosynthesis in cultured rabbit renal microvascular smooth muscle cells. Effect ofarachidonic acid, calcium, andA23187. Lab Invest 1990;63:30-7. Gryglewsld RJ, Moncada S, Palmer RM. Bioassay ofprostacydin and endothelium-derived relazing factor (EDRF) from porcine aortic endothelial cells. Br J Pharmaco11986;87:68594. Rovner AS, Murphy RA, Owens GK. Expression of smooth muscle and nonmuscle myosin heavy chains in cultured vascular smooth muscle cells. J Biol Chem 1986;261:14740-5. Kohler TR, Kirkman TR, Kraiss LW, Zierler BK, Clowes AW. Increased blood flow inhibits neointimal hyperplasia in endothelialized vascular grafts. Circ Res 1991;69:1557-65. Cambria RA, Lowell RC, Gloviczki P, Miller VM. Chronic changes in blood flow alter endothelium-dependent responses in autogenous vein grafts in dogs. J Vasc Surg 1994; 20:765-73. Faullmer SC, Fisher RD, Conlde DM, Page DL, Bender HW, Jr. Effect of blood flow rate on subendothelial proliferation in venous autografts used as arterial substitutes. Circulation 1975;51(suppl I):163-72. Ohlstein EH, Arleth A, Bryan H, Elliott ID, Sung CP. The selective endothelin ETA receptor antagonist BQ123 antagonizes endothelin-1 mediated mitogenesis. Eur J Pharmacol 1992;225:347-50.

Lewis et aL

197

32. Porter JG, Catatano K, McEnroe G, Lewicld JA, Protter AA. C-type natriuretic peptide inhibits growth factor-dependnet DNA synthesis in smooth muscle cells. Am J Physiol 1992; 263:C1001-6. 33. Dubey RK, Overbeck HW. Culture of rat mesenteric arteriolar smooth muscle cells: effects of platelet-derived growth factor, angiotensin, and nitric oxide on growth. Cell Tissue Res 1994;275:133-41. 34. Rubin BG, Santoro SA, Sicard GA: Platelet interactions with the vessel wall and prosthetic grafts. Ann Vase Surg 1993;7: 200-7. 35. Bunting S, Gryglewstd R, Moncada S, Vane JR. Arterial walls generate from prostaglandin endoperoxides a substance (prostaglandin X) which relaxes strips of mcsenteric and coeliac arteries and inhibits platelet aggregation. Prostaglandins

1976;12:897-913. 36. Furlong B, Henderson A_H, Lewis MJ, Smith JA. Endotheliurn-derived relaxing factor inhibits in vitro platelet aggregation. Br J Pharmacol 1987;90:687-92. 37. Macdonald PS, Read MA, Dusting GJ. Synergistic inhibition of platelet aggregation by endothelinm-derived'relaxing factor and prostacyclin. Thromb Res 1988;49:437-49. 38. Nordestgaard AG, Buckels IAC, Wilson SE, Platelet antagonists eliminate thromboembolic complications of small-diameter polytetrafluoroethylene arterial prostheses. J Vasc Surg

1987;5:110-7. Submitted Dec. 13, 1995; accepted May 10, 1996.