Hydraulic expansion facilitates remodeling of arteriovenous fistulas without increasing venous intimal hyperplasia in rabbits

The present study was conducted according to the Regulations of the Administration of Laboratory Animals formulated by the People’s Republic of China National Science and Technology Commission, the Guiding Opinions on the Good Treatment of Laboratory Animals issued by the Ministry of Science and Technology, and National Laboratory Animal Management Regulations (1988). The animal experimental protocols were approved by the Medical Ethics Committee of Jinan Central Hospital (2016.3.1). We followed the ARRIVE 2.0 reporting guidelines [11].

A model of jugular autogenous AVF in rabbit was followed [12]. Healthy male New Zealand white rabbits (provided by the Jinan Xilingjiao breeding center, laboratory animals quality certificate No. 0001084) with weight of 2.0–2.5 kg (age 3–4 months) were chosen for the present study. The rabbits were fed with a freely available diet and housed in single cages at ambient room temperature and humidity of 40%–60%. Sample size was set to 6 using a resource equation method. We used 12 New Zealand white rabbits numbered from 1 to 12, and 12 continuous random numbers were obtained from a random digit table corresponding to the rabbit number. On the basis of random numbers, rabbits were divided into a control group (vein without special handling and anastomosis with the artery directly, n = 6) and a hydraulic expansion group (vein was dilated with hydraulic pressure at about 8.41 kPa before anastomosis with the artery, n = 6).

New Zealand white rabbits were fed ad libitum for 3 days. Appetite and activities of rabbits were observed. A healthy rabbit having good appetite and behavioral activities was selected and then fasted for 8 h before surgery. An operator possessing relevant qualifications performed the procedures under license from the Shandong Experiment Animal Society of China. A dose of 1% sodium pentobarbital was used for anesthesia was based on the preoperative weight (30 mg/kg). The sodium pentobarbital was infused through an ear vein as the ear veins of rabbits are relatively superficial and easy to access. A sign of a suitable plane of anesthesia was that the rabbit’s limbs were weakened and did not respond to a pinch of the skin and respiration was deep. The limbs of the rabbit were fixed with the head supine and its jugular vein exposed by removing cervical fur. The internal jugular vein and carotid artery were scanned before and after surgery and the diameter of vein was recorded. The skin of neck was lifted slightly with forceps, and then the skin was incised with general tissue scissors along the right of the center of the neck. If the rabbit was observed to respond to the incision, 1% lidocaine was added additionally by local infiltration so that anesthesia was maintained for up to 2.5 h. The confluence of the internal and external jugular vein is the common jugular vein, so we could clearly identify the vein. We anastomosed the right common jugular vein and right common carotid artery. The arterial pulse between the vein and trachea was palpated, and deep layer muscles were bluntly dissected along the edge of the thyroid so as to find the internal carotid artery (Figure 1A). The artery was separated from the vein. The bleeding point and branch were ligated and the distal end of vein was ligated.

Figure 1

In rabbits of the hydraulic expansion group, the proximal part of vein was expanded by hydraulic pressure of about 8.41 kPa from heparin saline in the tube of a disposable venoclysis needle attached to a syringe, and the degree of expansion was such that the duct of vein was transparent (Figure 1B and 1C) and maintained for 30 s. Heparin saline was injected after expansion. In the control group, heparinized saline was injected into the vein directly. The distal end of artery was ligated and heparinized saline was injected into the proximal artery. The diameter of vein was about 2 mm and the diameter of artery was about 1.6 mm. The venous and arterial proximal parts were occluded using a vessel clamp (hemostat). The clamp was placed 2 cm above the anastomotic stoma. The proximal artery and vein were anastomosed in the way of the end-to-end gap with 6-0 suture silk to ensure a good flexibility of the blood vessel after anastomosis, less angular deformity, and less impact on the laminar flow of blood. The “sling with four thread” method was used to suture at 12, 3, 6, and 9 o’clock points at first, then the four threads were tightened, and the artery and venous gap was anastomosed until the anastomosis was complete. The fistula was created successful with tremor and bruit after releasing the clamp (Figure 1D). It was thus evident that the anastomotic stoma was in a good apposition without hemorrhage or severe angular deformity. Blood filled the vein, without high tension and tremor could be felt by palpation. The vascular murmur bruit was continuous and soft during auscultation. If there was hemorrhage from the stoma, the anastomosis was reinforced with 7-0 or 10-0 suture silk. If there was no active bleeding, the skin was sutured. The diameter of the internal jugular vein was measured using ultrasound. The respiration of rabbit was observed and if not labored the rabbit was returned to its cage. The rabbits were put in a quiet environment after the surgery, and buprenorphine was used for analgesia, 0.05–0.1 mg/kg, subcutaneously. Based on the rabbit’s appetite and activity, the analgesia was administered every 8–12 h, for 1–2 days. B-mode ultrasonic diagnostic equipment (mSonics MU1L; with Probe L10-5; Chengdu Youtou Technology, Sichuan, China) was used to measure venous diameter of rabbit before surgery, during surgery, and 1 day, 2, 3, 7, 14, and 21 days after surgery.

The investigators who conducted specimen collection, hematoxylin and eosin (HE) staining, and the immunohistochemistry were blinded to the grouping. Venous samples of about 5 mm with and without hydraulic expansion were obtained during surgery. Samples were fixed in 4% paraformaldehyde for 12–24 h and stored at room temperature before use. At 21 days after surgery, rabbits were anesthetized with 1% pentobarbital sodium (30 mg/kg) again and 10–20 mL of air was directly injected into the internal iliac vein during anesthesia causing air embolism used for euthanasia [13, 14]. The vessels of the fistula were found along the previous cut, they were separated, and a 2 cm venous sample was obtained after the anastomotic stoma. Samples were fixed in 4% paraformalde-hyde for 12–24 h and stored for use.

The venous tissue sample was rinsed with distilled water for 20 min and then dehydrated with alcohol for 30 min. The sample was embedded in molten wax at 58 °C with the venous lumina arranged verticality and a paraffin wax block was formed. The bulk paraffin was sliced off and the thickness of the slice was adjusted to 4 μm, with uniform thickness, integrity, and no defect in the tissue. Then the tissue sections were placed into 45 °C constant temperature waterbath, and attached to a microscope slide. A pencil was used for labeling. The slide was placed into an oven, baked at 65 °C for 30 min, and prepared for staining.

Tissue sections were routinely deparaffinated with xylene and alcohol. Hematoxylin staining using hematoxylineosin staining solution (catalog No. BSBA-4027; Zsgb-Bio, Beijing, China) was performed for 5 to 10 min, and the slide was rinsed with tap water for 1 min before differentiation with 0.7% hydrochloric acid in ethanol for about 5 s. The section was rinsed with water for 5 min. Eosin staining was performed for 1 min, and rinsed with tap water for 2 min. The sections were dehydrated though a graded series of alcohol concentrations at 75%, 85%, 95%, dehydrated alcohol, and dehydrated alcohol II (each for 3 min), and superfluous fluid was absorbed with absorbent paper between the different dehydration steps. The sections were vitrified with xylene I for 5 min and xylene II for 5 min. They were mounted under glass coverslips with neutral gum, and observed using an optical microscope.

Tissue sections were attached to glass slides with poly-L-lysine and baked at 65 °C for 4.5 h. Then the section was soaked in xylene for 30 min, rinsed with 100% alcohol twice, with 95% alcohol once, and with 80% alcohol once, followed by tap water. We added 4.3% methanol perhydrol at room temperature for 20 min, rinsed with distilled water, and then rinsed with 0.1 M phosphate-buffered saline (PBS). Antigen retrieval was performed by incubating the sections with antigen retrieval buffer (EDTA buffer pH 9.0; catalog No. ZLI-9086, Zsgb-Bio) in a beaker placed in a waterbath at a sub-boiling temperature for 20 min, before cooling for 3–5 min with tap water. Wax was painted on the slide to confine droplets of goat serum applied and incubated with the section at room temperature for 20 min to block nonspecific binding sites. Unnecessary liquid was discarded without rinsing. Primary antibody against connective tissue growth factor (CTGF) (polyclonal rabbit anti-CTGF antibody, product No. bs-0743R; Beijing Bioss Biotechnology Co.; RRID: AB_10856885) or (anti-TGF β1 monoclonal antibody from mouse, clone number 12E5, 1:100, product No. bsm-33287M; Beijing Bioss Bio-technology Co.) was added at 4 °C overnight, rinsed with 0.1 M PBS for 5 min/3 times. Secondary antibody (biotin-labeled goat anti-rabbit IgG polymer, product Nos. SP-9001, SP9002, SP kit, rabbit streptavidin–biotin method detection system, Beijing Zsgb-Bio; RRID: AB_2758396) was added at 37 °C for 20 min, rinsed with 0.1 M PBS for 5 min/3 times. Cochlearia enzyme (horseradish peroxidase (HRP)) labeled strepto-albumin fluid (S-A/HRP, SP kit SP9001, rabbit streptavidin-biotin method detection system, Beijing Zsgb-Bio) was added at 37 °C for 20 min, rinsed with 0.1 M PBS for 5 min/3 times. A diaminobenzidine (DAB) chromogen kit (HeBei Bohai Bioengineering Exploitation Co.) was used to develop staining and the section was rinsed with water. Hematoxylin counterstaining of the nucleus was performed for 1 min and rinsed with water sufficiently. The staining was differentiated by removing excess dye with 1% HCl in ethanol and rinsing the section with water sufficiently. Then 1% ammonia water was added to differentiate the staining by bluing the color of the hematoxylin-stained nuclei, then the sections were rinsed with water sufficiently, and dehydrated through a graded series of alcohols 80%, 95%, and 100% alcohol twice. The sections were vitrified with xylene for at least for 30 min, and mounted under glass coverslips with neutral resin. The image was analyzed with a CellSens Standard pathology image analysis system (version 1.12; Olympus, Japan). The section was observed with a light microscope (×200). The venous wall was observed and the thickness of the tunica intima (intimal thickness was the distance from endothelium to internal elastic lamina) was recorded. We measured at 4 places from every section to obtain a mean value. The corresponding immunohistochemistry score was calculated. Standardized score: 10 nonrepetitive visual fields under ×200 were selected without preference in each slice, and percentage of positive cell and staining degree in each visual field were determined. The scores for positive cell percentage were: 0: stained area <1%; 1: stained area 1%–10%; 2: stained area 10%–50%; 3: stained area 50%–80%; 4: stained area >80%, diffuse staining. The degree of staining scores were: 0: no staining; 1: light yellow; 2: yellowish-brown; 3: brown [15].

Data were analyzed using SPSS for Windows (version 16.0; SPSS Inc). A repeated measures analysis of covariance (ANCOVA) was used for adjusted baseline (weight, dose of pentobarbital, operating time, preoperative diameter). P < 0.05 was considered significant.

The mean weight of the rabbits between the 2 groups was not significantly different. The dose of 1% sodium pentobarbital in hydraulic expansion group was not significantly different from that used in the control group. The operating time in hydraulic expansion group was significantly longer than in the control group (P = 0.013). The preoperative arterial and venous diameters measured by B-mode ultrasonic diagnostic equipment were not significantly different between rabbits in the hydraulic expansion and control groups (Table 1).

The diameter of the fistula vein was increased significantly after surgery. After adjusting for weight, dose of pentobarbital, operating time, and preoperation diameter, the diameter of the fistula vein in hydraulic expansion group immediately after surgery (baseline) was significantly greater than it was in the control group (F_4,11 = 6.79, P = 0.02). At 1 day after surgery, the diameter of the fistula vein in hydraulic expansion group was also significantly greater than it was in those rabbits of the control group (F_4,9 = 7.87, P = 0.03). However, the diameter of fistula vein between the groups at 2, 3, 7, 14, or 21 days after surgery was not significantly different (Table 2). Vasospasm occurred in one rabbit in the hydraulic expansion group, and reduced after spasmolysis with topical lidocaine. One rabbit had an infection at the incision site that healed after changing dressings. Death as a result of anesthesia overdose occurred in another rabbit that was replaced with another to maintain group size.

After hydraulic expansion and before construction of the fistula, the wall of the vein was thinner in macroscopic observation and the anterior and posterior wall were difficult to distinguish during surgery in the hydraulic expansion group. Saline seepage was caused by extreme dilation, but blood did not leak. At 3 weeks after surgery, the incisions in rabbits of both groups healed well. The blood vessel of the fistula could be felt, and the wall of vessel was thickened and there was slight adhesion to surrounding tissue, which could be separated easily.

The changes of the venous wall are shown before and after dilation (Figure 2A and B). Before the hydraulic expansion, the venous endothelium had integrity without dilation. The internal elastic lamina was clear. There were a few circular smooth muscle cells in the tunica media. The tissue was comparatively compact. There were longitudinal smooth muscle cells in the tunica externa. The external elastic lamina was not clear and boundary was obscure (Figure 2A). After hydraulic expansion, the structure of venous wall was deranged, the endothelium was disrupted and then sloughed off, circular smooth muscle cells under the epithelium were exposed, elastic fibers were disrupted and deformed. A large number of erythrocytes had entered the tunica media and tunica externa, the wall was compact and thinner, and the boundary of tunica externa was clear (Figure 2B).

Figure 2

At 3 weeks after surgery, the venous wall of fistula had thickened in control group, intimal hyperplasia was obvious, endothelial cells protruded into the lumina, smooth muscle cells had proliferated, cell number was increased in each layer, and the density of tissue was greater than it was before surgery (Figure 2C). In the hydraulic expansion group, the venous wall of fistula was thickened, intimal hyperplasia was obvious, most of cells were smooth muscle cells, and cells were arranged chaotically. There were fewer elastic fibers than seen in the control group. The smooth muscle cells in the tunica externa had proliferated. The tissues from the endothelium to the tunica externa in hydraulic expansion group were rarefied compared with tissues in the control group. There was less extracellular matrix in the hydraulic expansion group than in the control group (Figure 2D).

The expression of transforming growth factor-β1 (TGF-β1) and CTGF (polyclonal rabbit anti-CTGF antibody, product No. bs-0743R; Beijing Bioss Biotechnology Co.; RRID: AB_10856885) were mainly located in cellular cytoplasm (Figure 3). In the vein before surgery, TGF-β1 immunoreactivity was expressed in endothelial cells, but there was only weak expression (yellowish-brown staining) in the tunica media and tunica externa (Figure 3A). Three weeks after surgery TGF-β1 immunoreactivity was expressed in the tunica intima, tunica media, and tunica externa, and appeared as yellowish-brown staining in the control group (Figure 3B). TGF-β1 immunoreactivity was expressed in the tunica intima and tunica externa, but only slightly in the tunica media, and appeared as yellowish-brown staining in the hydraulic expansion group (Figure 3C). CTGF immunoreactivity was expressed clearly in endothelial cells, but was weak in the tunica media and tunica externa appearing as yellowish-brown staining (Figure 3D). Three weeks after surgery, CTGF was expressed clearly in tunica intima and tunica externa, but was stronger in the tunica intima, and appeared as yellowish-brown staining in the control group (Figure 3E). CTGF immunoreactivity was expressed in tunica externa, but weakly in the tunica intima and tunica media, and appeared as yellowish-brown staining in the hydraulic expansion group (Figure 3F). After adjusting for the baseline (weight, dose of pentobarbital, operating time, pre-operative diameter), there was no difference in intimal thickness between 2 groups 3 weeks after surgery (F_4,11 = 0.95, P = 0.51). Immunohistochemistry scores for TGF-β1 in hydraulic expansion group were less than those in the control group, but the difference was not significant (F_4,11 = 0.60, P = 0.71). Immunohistochemistry scores for CTGF in the hydraulic expansion group were significantly less than those in the control group (F_4,11 = 7.87, P = 0.01) (Table 3).

Figure 3