Sex hormones impact early maturation and immune response in the arteriovenous fistula mouse model

Abstract

Arteriovenous fistulae (AVF) fail to mature more frequently in female patients compared with male patients, leading to inferior outcomes and decreased utilization. Since our mouse AVF model recapitulates sex differences in human AVF maturation, we hypothesized that sex hormones mediate these differences during AVF maturation. C57BL/6 mice (9–11 wk) were treated with aortocaval AVF surgery and/or gonadectomy. AVF hemodynamics were measured via ultrasound (days 0–21). Blood was collected for FACS and tissue for immunofluorescence and ELISA (days 3 and 7); wall thickness was assessed by histology (day 21). Inferior vena cava shear stress was higher in male mice (P = 0.0028) after gonadectomy, and they had increased wall thickness (22.0 ± 1.8 vs. 12.7 ± 1.2 µm; P < 0.0001). Conversely, female mice had decreased wall thickness (6.8 ± 0.6 vs. 15.3 ± 0.9 µm; P = 0.0002). Intact female mice had higher proportions of circulating CD3⁺ T cells on day 3 (P = 0.0043), CD4⁺ (P = 0.0003) and CD8⁺ T cells (P = 0.005) on day 7, and CD11b⁺ monocytes on day 3 (P = 0.0046). After gonadectomy, these differences disappeared. In intact female mice, CD3⁺ T cells (P = 0.025), CD4⁺ T cells (P = 0.0178), CD8⁺ T cells (P = 0.0571), and CD68⁺ macrophages (P = 0.0078) increased in the fistula wall on days 3 and 7. This disappeared after gonadectomy. Furthermore, female mice had higher IL-10 (P = 0.0217) and TNF-α (P = 0.0417) levels in their AVF walls than male mice. Sex hormones mediate AVF maturation, suggesting that hormone receptor signaling may be a target to improve AVF maturation.

NEW & NOTEWORTHY After arteriovenous fistula creation, females have lower rates of maturation and higher rates of failure than males. In a mouse model of venous adaptation that recapitulates human fistula maturation, sex hormones may be mechanisms of the sexual dimorphism: testosterone is associated with reduced shear stress, whereas estrogen is associated with increased immune cell recruitment. Modulating sex hormones or downstream effectors suggests sex-specific therapies and could address disparities in sex differences in clinical outcomes.

INTRODUCTION

Surgical creation of an arteriovenous fistula (AVF) is a commonly performed vascular procedure to enable hemodialysis and is more recently being tested as a potential treatment for refractory arterial hypertension (1–6). Despite the relative predominance of this procedure in the practice of vascular surgery, the rate of AVF failure remains high (7, 8). Some 30–50% of AVF fail to successfully mature by 6 mo as a result of “early failure” in which the vein wall fails to thicken and the lumen fails to dilate to a usable diameter, requiring interventions to promote maturation (8–11). In those that do mature as anticipated, 35–40% fail within the first year because of vigorous neointimal hyperplasia leading to excessive thickening of the vein wall, termed “late failure” (9), and ultimately stenosis or occlusion of the AVF (10–12). Female patients have substantially worse outcomes than male patients after AVF creation, with worse rates of maturation (31% vs. 51%) and higher rates of late failure (up to 78%), and comparably require more fistula salvage interventions than male patients (42% vs. 23%) (13–15).

Inflammation in the maturing AVF has classically been characterized by an innate immune response (16). The innate immune response is formed by nonspecific effector cells including neutrophils and macrophages that recognize damage-associated molecular patterns and release proinflammatory cytokines and chemokines. Although innate immunity has been implicated in the pathogenesis of AVF failure (17–20), broad inhibition of innate immunity leads to failure as well (16). These observations suggest that the innate immune response is a mechanism of successful venous adaptation to the fistula environment but is likely subject to complex regulation. We have recently shown that the adaptive immune response, consisting of tightly regulated interactions between antigen-presenting cells and T and B lymphocytes, is also a mechanism of successful venous remodeling (21, 22). Both innate and adaptive immune responses have multiple subsets of cells that may differentially regulate both early and late phases of venous adaptation to the fistula environment. In addition, there are several sex differences in both the innate and adaptive immune responses that lead to sex differences in autoimmunity, malignancy, and susceptibility to infection and may contribute to differences in outcomes after AVF creation (23). Importantly, these sex differences in immunity are variable across the life span and suggest differential hormone expression as a potential driver.

Although AVF have reduced maturation and increased failure in females, it is unknown whether these differences are associated with sex differences in the innate or adaptive immune responses. We previously characterized a mouse model of AVF that shows sexual dimorphism in hemodynamics and reduced patency in female mice (24). Since the impact of sex hormones on inflammation during early AVF maturation has not been well characterized, we determined the role of sex hormones during venous remodeling. We hypothesized that sex hormones modulate sex differences in immune responses during early venous remodeling in the fistula environment.

METHODS

Analgesia

Anesthesia for surgery was performed with isoflurane (30% isoflurane in propylene glycol, continuous administration). Postoperatively, animals received injectable analgesic (buprenorphine, 0.6 mg/kg) for 48 h. They were examined daily by the laboratory team as well as by the animal facility staff to detect, in a timely fashion, any potential animal discomfort, distress, pain, or injury.

Mouse Aortocaval Fistula Model

All animal experiments were performed in strict compliance with Federal guidelines and with approval from the Institutional Animal Care and Use Committee of Yale University. Briefly, female and male C57BL/6 mice aged 9–11 wk (The Jackson Laboratory) were used. With 4% isoflurane in 0.8 L/min oxygen for induction followed by 2–3% isoflurane, mice were anesthetized before the surgical procedure. A midline laparotomy was done under anesthesia. The bowel was moved aside and protected, and the aorta and inferior vena cava (IVC) were exposed. The distal aorta was then dissected free from surrounding tissue for needle puncture, and the IVC was not dissected away from the aorta. A clamp was placed just below the left renal artery. A 25-gauge needle was used to puncture from the aorta into the IVC (Supplemental Fig. S1A; all Supplemental Material is available at https://doi.org/10.6084/m9.figshare.21951437). The surrounding tissue was used for hemostatic compression. The operators were not blinded to the sex of the mice. Visible pulsatile arterial blood flow in the IVC along with hemostasis were used to indicate successful AVF creation. The skin and muscle were closed in a single layer with simple, interrupted suture. Animals were allowed to recover on a warming mat. Sham AVF models underwent midline laparotomy, bowel manipulation, and dissection of the aorta without any puncture. They were then closed and allowed to recover in a similar fashion.

Mouse Gonadectomy

All surgeries were performed in strict compliance with Federal guidelines and with approval from the Institutional Animal Care and Use Committee of Yale University. Female and male C57BL/6 mice aged 9–11 wk were used. Mice were anesthetized before the surgical procedure with isoflurane.

For female mice, hair over both flanks of the animal, between the rib cage and hip surgical area, was removed with Nair. A 0.5-cm incision over the flank just above the hip joint and below the rib cage was made. Muscle was then incised with scissors to enter the peritoneum. Each ovary was located by visualizing a white spot under the muscle on the flanks of the animal. This is the fat pad covering the ovary. Fine forceps were used to grab the fat pad and gently pull it out of the incision. The ovary was identified as a small, pink tortuous structure. Hemostats were placed both above and below the ovary, and sutures were tied around these. Scissors were used to remove the ovary. The hemostat was released, and hemostasis was ensured. One suture was used to close muscle and a second suture for skin. This was repeated with the second ovary on the contralateral side. Animals recovered in a clean cage with supplemental heat for ∼30 min and were monitored.

For male mice, a 1-cm midline incision was made under anesthesia just above the bladder. The testicular fat pad on one side was identified and pulled through the incision with blunt forceps. A hemostat was placed below the testes and epididymis across the testicular cord. Absorbable suture was tied below the hemostat. The testes and epididymis were then removed with scissors. The hemostat was released, and hemostasis was ensured. One suture was used to close muscle and skin. Animals recovered in a clean cage with supplemental heat for ∼30 min and were monitored.

Overall Experimental Design

AVF and sham procedures were performed on day 0. Gonadectomies and sham procedures were performed 1 wk before AVF (day −7). Ultrasound measurements were performed on days −7, 0, 3, 7, 14, and 21. Blood was collected via left ventricular puncture before perfusion on day 21 to measure sex hormone concentrations. Animals were euthanized, and the AVF was removed en bloc for sectioning and staining to measure wall thickness.

The same experimental groups were used for immune system assessment. Blood was obtained from retroorbital collection on days −7, 0, 3, and 7. This was stained for immune cells of interest and measured by FACS. Additionally, animals were euthanized and AVF was removed en bloc for sectioning on days 0, 3, and 7 for immunofluorescence. Finally, tissue from intact female and male mice was isolated on day 3 for cytokine array and ELISA analyses (Supplemental Fig. S1B).

Ultrasound

Doppler ultrasound was used to measure preoperative and postoperative venous and arterial diameters in transverse view and velocities in longitudinal view. Mice were anesthetized for procedures with isoflurane. The Vevo 770 High-Resolution Imaging System (VisualSonics, Toronto, ON, Canada) with probe RMV704 (20-0 MHz) was used in all experiments. To confirm the presence of the AVF, the waveforms in the IVC and aorta were recorded in the pulse wave mode in longitudinal view. In patent AVF, presence of the arterial or turbulent waveforms in the IVC and increased end-diastolic velocity in the aorta were used as markers of patency (24). Diameters of the aorta and IVC just below and rostral to the left renal vein were measured. Relative diameter was defined as the ratio of the vessel diameter at the specified time of measurement to the diameter measured at baseline (day 0). Ultrasound measurements were also collected on postoperative days 3, 7, 14, and 21. Velocities were obtained at the same anatomic location but in longitudinal view. Pulse wave mode was used to measure the waveforms of the aorta and IVC, and the average of the waveform was calculated on the Vevo 770 to determine overall mean velocity. With the use of IVC and aortic diameters and velocities, shear stress and flow rate were then calculated. The formula for flow rate was cross-sectional area times velocity: F = πr² × V, where πr² is the cross-sectional area formula with r as radius and V is velocity (in cm/s). The shear stress was calculated by the Hagen–Poiseuille formula: T = 4ηV/r, where T is shear stress, η is blood viscosity, V is velocity (in cm/s), and r is the radius (in cm). Blood viscosity was assumed to be constant at 0.035 P.

Sex Hormone Level Quantification

Whole blood was collected from left ventricular puncture on day 21 and compared with serum collected before gonadectomy. Samples were allowed to clot at room temperature for 90 min. They were then centrifuged at 2,000 g for 15 min. The serum was collected and stored at −20°C.

Estradiol was measured by ELISA with commercial kits (E2; ALPCO, Salem, NH, catalog no. 11-ESTHU-E01) according to the instructions for use. Estradiol assay characteristics were as follows: sensitivity = 5 pg/mL; intra-assay coefficient of variation (CV) = 7.0%; interassay CV = 11.1%. Briefly, 50 µL of each sample was placed on coated wells for 1 h, stained with TMB substrate, and measured on a microwell plate reader at 450 nm.

Testosterone was also measured by ELISA with commercial kits (IBL Inc., Minneapolis, MN, catalog no. IB79106) according to the instructions for use. Sensitivity was 10 ng/dL; intra-assay CV = 6.4%; and interassay CV = 9.7%. Briefly, 25 µL of each sample was placed on coated wells for 1 h, stained with substrate, and measured on a microwell plate reader at 450 nm.

Euthanasia

Euthanasia for all animals was exsanguination under anesthesia, consistent with recommendations of the American Veterinary Medical Association (AVMA) Guidelines on Euthanasia and complying with local Veterinary Clinical Services policies. As with the survival surgery, the induction agent was isoflurane at twice the stated induction doses. When adequate anesthetic level was achieved, the thorax was opened and 10% formalin was perfused through the heart. After removal of the fistula, the animal was allowed to exsanguinate until satisfactory evidence of death had occurred by evidence of absence of cardiovascular and pulmonary function.

Histology

After euthanasia, the AVF was extracted en bloc. The tissue was then embedded in paraffin and cut into 5-μm cross sections. Animals euthanized at day 21 were used to measure intima-media thickness, and elastin Van Gieson staining was used. With a microtome, sections were obtained 50–100 μm cranial to the fistula. Twenty equidistant points around the IVC wall were averaged in each cross section to obtain the mean AVF outer wall thickness. Additional unstained cross sections obtained at days 3 and 7 in this same region were used for immunofluorescence microscopy.

Fluorescence-Activated Cell Sorting

Approximately 100 µL of blood was obtained through retroorbital collection with heparin-coated capillary tubes; 100 µL of staining cocktail [primary antibodies at 1:500, live-dead stain at 1:1,000, and Fc block (BD Biosciences, No. 553142, 1:400) in fluorescence-activated cell sorting (FACS) buffer] was added to 100 µL of blood and incubated for 30 min in the dark. RBC lysis buffer was then added, and the mixture was incubated for another 10 min in the dark. The samples were centrifuged at 500 g and resuspended three times in 500 µL. They were then resuspended in 200 µL and stored at 4°C. The samples were read ≤24 h after staining. Primary antibodies are listed in the major resources table (Supplemental Table S1). Cells were analyzed in a FACS-Aria flow cytometer (BD Biosciences, Franklin Lakes, NJ). FlowJo (V 10.8.1) Software (BD Biosciences) was used to analyze the data. Samples were normalized to IgG isotype controls.

Immunofluorescence

Tissue sections were deparaffined with xylene and a graded series of alcohols. Sections were then heated in citric acid buffer (pH 6.0) at 100°C for 20 min for antigen retrieval. They were blocked with 5% bovine serum albumin in PBS containing 0.05% Triton X-100 (T-PBS) for 1 h at room temperature before incubation overnight at 4°C with the primary antibodies diluted in T-PBS. Primary antibodies used are listed in the major resources table (Supplemental Table S1). Sections were then treated with secondary antibodies at room temperature for 1 h: goat anti-rat Alexa Fluor 488 (Thermo Fisher) and donkey anti-rabbit Alexa Fluor 647 (Thermo Fisher). Sections were stained with DAPI (Life Technologies), and a coverslip was applied. Digital fluorescence images were captured, and positive cells were counted with ImageJ software. Negative controls for antibodies were isotype-matched antibodies and positive controls used lymph node, as previously validated (21, 22).

Cytokine Array

The infrarenal IVC was removed with care taken to avoid surrounding arterial and connective tissues. Proteins were extracted with RIPA lysis buffer containing protease inhibitors from single IVC. Lysates were examined with a cytokine array with the Mouse Neuro Antibody Array kit (Abcam, catalog no. ab211069) according to the instructions for use. Briefly, after the wells were blocked, the samples were added and incubated overnight. Horseradish peroxidase (HRP)-streptavidin and biotin were used for detection. The membranes were then measured with a chemiluminescence imaging system. Results were analyzed on ImageJ and normalized to the positive and negative control spots to normalize to amount of protein loaded.

ELISA

The infrarenal IVC was removed with care taken to avoid surrounding arterial and connective tissues. Proteins were extracted with RIPA lysis buffer containing protease inhibitors from a single IVC. Protein concentrations were assessed with a colorimetric assay (Bio-Rad) and noted. For the IL-10 ELISA, the Mouse IL-10 ELISA kit (Abcam, catalog no. ab108870) was used. Briefly, 50 µL of sample was added to each well for 2 h. Biotinylated IL-10 antibody was added, followed by SP Conjugate and Chromogen Substrate.

For the TNF-a ELISA, the Mouse TNF-α DuoSet ELISA (R&D Systems, catalog no. DY410) was used. Briefly, 100 µL of sample was added to each well for 2 h. Detection antibody and streptavidin-HRP were then added.

All plates were measured on a microwell plate reader at 450 nm. The substrate amount detected was then divided by the protein concentration of the sample determined beforehand to find the microgram concentration of the cytokines per unit volume.

Statistics

Data are represented as means ± SE. All data were analyzed with Prism 9 software (GraphPad Software, La Jolla, CA). Normality was confirmed with the Shapiro–Wilk test. Statistical significance was determined with Student’s t test or ANOVA with Sidak’s post hoc correction. The Mann–Whitney U test was used for nonparametric analysis or if the sample size was <6. P < 0.05 was considered significant.

RESULTS

Baseline Hemodynamics Demonstrate Sex Differences

We first examined baseline sex differences in vessel hemodynamics in intact female and male mice, before AVF creation and without gonadectomy. Female mice had significantly lower weight, arterial diameter, velocity, and shear stress (Fig. 1, A–D) compared with male mice. In the inferior vena cava, female mice had equivalent venous diameter (Fig. 1E) but lower velocity (Fig. 1F) and shear stress (Fig. 1G). This difference was also visually discernable on ultrasound (Supplemental Fig. S2). Baseline venous wall thickness of female and male mice were similar (Fig. 1H). This data value is consistent with some sex differences in hemodynamics, but not in wall thickness, at baseline.

Figure 1.

Baseline hemodynamics for female and male mice. A: male (n = 14) vs. female (n = 15) weight; ****P < 0.0001. B: male (n = 27) vs. female (n = 27) arterial diameters; P = 0.2133. C: male (n = 19) vs. female (n = 19) arterial velocity; *P = 0.0139. D: male (n = 19) vs. female (n = 19) arterial shear stress; *P = 0.0207. E: male (n = 24) vs. female (n = 25) inferior vena cava (IVC) diameter; P = 0.6114. F: male (n = 19) vs. female (n = 20) IVC velocity; ***P = 0.0001. G: male (n = 19) vs. female (n = 20) venous shear stress; ***P = 0.0008. H: wall thickness of male (n = 10) and female (n = 10) mice, day 21; P = 0.9832.

Hemodynamics after AVF creation were assessed in both female and male intact mice (Fig. 2, A–D, and Supplemental Fig. S3, A–D). In the IVC, female mice had increased absolute (Fig. 2A) and relative (Fig. 2B) diameter at all time points despite similar baseline diameter (Fig. 1E); however, IVC velocity (Fig. 2C) and shear stress (Fig. 1D) were similar. Wall thickness was assessed on postoperative day 21 and was similar between female and male mice (12.7 ± 1.2 vs. 15.3 ± 0.9 µm; P = 0.1027) but increased compared with baseline (Fig. 2J and Fig. 1H), suggesting adaptive remodeling in all mice.

Figure 2.

Arteriovenous fistula (AVF) hemodynamics and wall thickness among female and male mice without and with gonadectomy. A: absolute inferior vena cava (IVC) diameter after AVF in intact male (n = 7–14) and female (n = 7–15) mice; ***P = 0.0006 (ANOVA). B: relative IVC diameter after AVF; ***P = 0.0004 (ANOVA). C: IVC velocity after AVF; P = 0.5257 (ANOVA). D: IVC shear stress after AVF; P = 0.9135 (ANOVA). E: absolute IVC diameter after AVF with gonadectomy; ***P = 0.0007 (ANOVA). F: relative IVC diameter after AVF with gonadectomy; ***P = 0.0006 (ANOVA). G: IVC velocity after AVF with gonadectomy; P = 0.0848 (ANOVA). H: IVC shear stress after AVF with gonadectomy; **P = 0.0028 (ANOVA). I and J: IVC wall thickness (day 21) (n = 8); ***P < 0.001 (ANOVA). J: summary bar graph; gonad, gonadectomy. I: representative images stained for elastin. Scale bars, 50 μm. Arrowheads denote intima-media thickness.

Male Mice Have Increased Shear Stress and Wall Thickness after Gonadectomy

Since we hypothesize that sex hormones may impact AVF remodeling in intact mice, we determined the effects of gonadectomy on vessel hemodynamics (Supplemental Fig. S4). The effects of gonadectomy were confirmed by assessment of serum estrogen levels in female mice and testosterone levels in male mice; after gonadectomy, male mice had reduced testosterone (P < 0.0001; Supplemental Fig. S4A) and female mice had reduced estrogen (P = 0.0363; Supplemental Fig. S4B). The effects of gonadectomy on vessel hemodynamics were assessed after 1 wk, just before AVF creation. In the IVC, female mice had 31.1% increase in diameter (Supplemental Fig. S4, G and H) and concomitant decreased shear stress (Supplemental Fig. S4J). This may reflect the established role for estrogen in regulation of vessel tone (25–27).

Mice with both gonadectomy and AVF were followed for 21 days (Fig. 2, E–H, and Supplemental Fig. S3, E–H). Both aortic flow velocity (Supplemental Fig. S3C) and shear stress (Supplemental Fig. S3D) were reduced in female mice, consistent with reduced inflow to the fistula in female mice after gonadectomy. In the IVC, female mice had higher absolute (P = 0.0007; Fig. 2E) and relative (P = 0.0006; Fig. 2F) diameters after day −7, a trend toward reduced velocity (P = 0.0848; Fig. 2G), as well as reduced shear stress (P = 0.0028; Fig. 2H). Direct comparison of the shear stress in the AVF of intact mice (Fig. 2D) to mice with gonadectomy (Fig. 2H) showed no difference in female mice (P = 0.7202) but increased shear stress in male mice with gonadectomy (P < 0.0001; Supplemental Fig. S5).

Fistula wall thickness was measured on postoperative day 28 after gonadectomy. Control mice, without an AVF but with a gonadectomy, showed increased wall thickness in both male (P < 0.0001) and female (P = 0.0009) mice, with slightly thicker walls in the male mice (P = 0.0045; Supplemental Fig. S6). Interestingly, male mice with both an AVF and gonadectomy had significantly increased wall thickness compared with intact males (22.0 ± 1.8 vs. 12.7 ± 1.2 µm; P < 0.0001), whereas female mice with both an AVF and gonadectomy had significantly decreased wall thickness compared with intact females (6.8 ± 0.6 vs. 15.3 ± 0.9 µm; P = 0.0002; Fig. 2J). Reduced wall thickness in female mice after gonadectomy was observed despite a lack of change in shear stress (Fig. 2J and Supplemental Fig. S5), suggesting that the influence of sex hormones during AVF remodeling in female mice may not be solely via regulation of hemodynamics.

Females’ Higher Immune Activation from AVF Creation Disappears after Gonadectomy

Although gonadectomy in male mice was associated with increased AVF wall thickness, gonadectomy in female mice was associated with decreased AVF thickness with no changes in hemodynamics (Fig. 2J), suggesting alternative mechanisms of reduced AVF maturation in female mice (24). Since the immune response is active during venous remodeling (21, 28) and the immune response shows sexual dimorphism (29, 30), we determined whether sex differences are present in the immune response after AVF creation. Circulating immune cells were evaluated on postoperative days 0, 3, and 7 after AVF creation (Supplemental Fig. S7); intact female mice showed higher proportions of CD45⁺CD3⁺ T cells on day 3 (Fig. 3A and Supplemental Fig. S8A) and higher proportions of CD45⁺CD4⁺CD3⁺ T cells on day 7 (Fig. 3B and Supplemental Fig. S8B), as well as CD45⁺CD8⁺CD3⁺ T cells (Fig. 3C and Supplemental Fig. S8C), than male mice. CD45⁺CD11b⁺ monocytes were similarly increased in female mice on day 3 (Fig. 3D and Supplemental Fig. S8D), whereas CD45⁺CD19⁺ B cells were reduced in female mice on day 7 (Fig. 3E and Supplemental Fig. S8E) relative to males. However, after gonadectomy these differences were no longer present (Fig. 3 and Supplemental Fig. S8, F–J), confirming the effect of sex hormones on the immune response after AVF creation. As a control for the stress of the gonadectomy surgery, we assessed circulating immune cells in mice that had a sham gonadectomy and AVF; circulating cells (Supplemental Fig. S9) were similar to those in intact mice with AVF (Supplemental Fig. S8, A–E), consistent with the presence of sex differences in the immune response during AVF remodeling.

Figure 3.

Circulating immune cell concentrations among female and male mice without and with gonadectomy. A: circulating CD3⁺ T cells on day 3 after arteriovenous fistula (AVF) in male (n = 11) and female (n = 10) mice without gonadectomy (P = 0.0043) and in male (n = 6) and female (n = 8) mice with gonadectomy (P = 0.3623). B: CD3⁺/CD4⁺ T cells in male (n = 7) and female (n = 7) mice on day 7 after AVF without gonadectomy (P = 0.0003) and males (n = 5) and females (n = 10) with gonadectomy (P = 0.8891). C: CD3⁺/CD8⁺ T cells in male (n = 10) and female (n = 11) mice on day 0 (baseline) without gonadectomy (**P = 0.005) and males (n = 18) and females (n = 13) with gonadectomy (P = 0.7257). D: CD11b⁺ monocytes in male (n = 11) and female (n = 11) mice on day 3 after AVF without gonadectomy (**P = 0.0046) and males (n = 6) and females (n = 8) with gonadectomy (P = 0.9413). E: CD19b⁺ B cells in male (n = 7) and female (n = 7) mice on day 7 after AVF without gonadectomy (P = 0.0046) and males (n = 5) and females (n = 10 with gonadectomy (P = 0.2967).

We next directly evaluated the presence of immune cells in the fistula wall. Control mice without an AVF showed that female mice had significantly more CD45⁺ leukocytes in the IVC wall at baseline (Supplemental Fig. S10A) but not after gonadectomy (Supplemental Fig. S10B), indicating higher immune activation in females at baseline. In intact female mice, immune cells, including CD45⁺CD3⁺ T cells, CD45⁺CD4⁺ helper T cells, CD45⁺CD8⁺ cytotoxic T cells, and CD68⁺ macrophages, increased significantly in the fistula wall, both on day 3 (Fig. 4, A and C, and Supplemental Fig. S11A) as well as day 7 (Fig. 4, B and C, and Supplemental Fig. S12A), and these increases were not present in female mice treated with gonadectomy (Fig. 4 and Supplemental Figs. S11B and S12B). Intact male mice had less immune cell activation than female mice on day 3 (Fig. 4, A and C, and Supplemental Fig. S11A) and day 7 (Fig. 4, B and C, and Supplemental Fig. S12A). However, after gonadectomy an increase in all cell types were seen on day 3, but only CD45 and CD3 increases were sustained at day 7 (Fig. 4 and Supplemental Figs. S11B and S12B). When male mice with gonadectomy were directly compared with intact male mice, they had higher numbers of all cell types on day 3 (Fig. 4A). Conversely, female mice with gonadectomy had reduction of all cell types compared with intact female mice on day 3 (Fig. 4A). These data are similar to the patterns of circulating immune cells (Fig. 3) and suggest a role for immune system sexual dimorphism during AVF remodeling.

Figure 4.

Immune cell numbers in the inferior vena cava (IVC) wall among female and male mice without and with gonadectomy after arteriovenous fistula (AVF). A: bar graphs show number of immune cells within the IVC wall on day 3 after AVF of intact male (n = 7) and female (n = 5) mice and male (n = 4) and female (n = 4) mice pretreated with gonadectomy. B: bar graphs show number of immune cells within the IVC wall on day 7 after AVF of intact male (n = 4) and female (n = 6) mice and male (n = 4) and female (n = 4) mice pretreated with gonadectomy. C: immunofluorescence of IVC walls of male and female mice, DAPI/CD45⁺. From left to right, intact mice, day 3 after AVF; mice pretreated with gonadectomy, day 3 after AVF; intact mice, day 7 after AVF; and mice pretreated with gonadectomy, day 7 after AVF. Scale bars, 50 μm. Arrowheads denote dual-positive cells. hpf, High-power field.

Since immune cells can activate both the innate and adaptive immune systems, we determined whether there were sex differences in the innate and/or adaptive immune responses after AVF creation. A cytokine array analysis of the AVF wall showed that female- and male-derived AVF show several differences in both adaptive and innate cytokines (Supplemental Fig. S13, A and B); IL-10 showed the most significant sex difference among cytokines associated with the adaptive immune system (Fig. 5A and Supplemental Fig. S13C), and TNF-α similarly showed a trend toward a difference among cytokines associated with the innate immune system (Fig. 5B and Supplemental Fig. S13D). To validate sex differences in the expression of these cytokines, we quantified expression with a more sensitive ELISA immunoassay. Both male (P = 0.0594) and female (P < 0.0001) mice with AVF showed increased IL-10 expression in their AVF walls, with female mice showing higher IL-10 levels (P = 0.0217; Fig. 5C). Similarly, female mice showed increased TNF-α expression in their AVF walls compared with male mice (P = 0.0417; Fig. 5D). These data show the presence of both the innate and the adaptive immune response in the AVF, with increased cytokine expression in female mice, suggesting increased immune response in female mice. In total, these data suggest that sex hormones play a role in hemodynamic changes that influence wall remodeling during fistula maturation in male mice, whereas sex hormones may more strongly influence immune responses that impact wall remodeling and fistula maturation in female mice.

Figure 5.

Inferior vena cava wall cytokine concentrations in female and male mice after arteriovenous fistula (AVF) creation. A: relative signal of IL-10 cytokine array at baseline (n = 4) and after AVF creation (n = 4) among male (P = 0.744) and female (*P = 0.048) mice. Con, control. B: relative signal of TNF-α cytokine array at baseline (n = 4) and after AVF creation (n = 4) among male (P = 0.688) and female (P = 0.332) mice. C: IL-10 protein determined by ELISA at baseline (n = 5 males, 6 females) and after AVF creation (n = 6 males, 8 females) in male (P = 0.0594) and female (*P < 0.0001) mice. Male vs. female with AVF, *P = 0.0217. ANOVA; P = 0.001. D: TNF-α protein determined by ELISA at baseline (n = 3) and after AVF creation (n = 3) in male (P = 0.7278) and female (*P = 0.0164) mice. Male vs. female with AVF, *P = 0.0417. ANOVA **P = 0.0086.

DISCUSSION

We show sex differences during venous remodeling in the fistula environment, both in hemodynamics and in the immune response, that are sex hormone dependent. In male mice, decreased testosterone was associated with increased wall thickening and shear stress (Fig. 2, H–J), suggesting accelerated stenosis and early failure; testosterone led to a slight increase in the number of immune cells in the AVF wall in male mice (Fig. 4). On the other hand, decreased estrogen was not associated with altered hemodynamics, but less thickening of the venous wall was observed (Fig. 2, H–J). The significantly reduced immune cell numbers in the venous wall after estrogen reduction (Fig. 4) suggest that reduced inflammation was associated with slower early maturation and less wall thickening. These data are consistent with previous data showing that reduced patency in female mice was preceded by lower flow velocities through the fistula, reduced magnitudes of shear stress, and less flow volume compared with male mice (24). Interestingly, some of the baseline differences in IVC velocity and shear stress between female and male mice were no longer observed after AVF creation, perhaps because of the increased variability in shear stress and velocity after fistula creation, similar to the variability in hemodynamics observed among patients with successful fistulae (31).

Our study found increased wall thickening and shear stress in male mice after gonadectomy during early maturation. During the early phase of fistula maturation, AVF adapt via outward wall thickening, leading to increased luminal diameter (9, 32). This is accompanied by higher shear stress, leading to extracellular matrix deposition and smooth muscle cell proliferation. Conversely, excessive proliferation of the intimal layer, often due to disturbed or turbulent flow, leads to late AVF failure secondary to neointimal hyperplasia (9, 32). Increased wall thickness in AVF of male mice was predominantly observed in the intimal layer, suggesting accelerated neointimal hyperplasia.

One proposed explanation for the sex-specific differences seen in fistula maturation is the differential expression of androgens and estrogens. After gonadectomy, male mice had a higher shear stress and intimal wall thickening (Fig. 2, H–J), with no change in the recruitment of immune cells (Fig. 4). Dorsett-Martin et al. (33) investigated the impact of sex hormones on arterial thickening at the anastomoses proximal and distal to the AVF. They found that testosterone administration increased wall thickness in the mouse artery; however, they did not report any data examining the venous outflow of the fistula (33). On the other hand, Wilhelmson et al. (34) showed that testosterone replacement in castrated mice was associated with decreased intimal hyperplasia in proportion to the degree of androgen receptor activation. Similarly, Wu et al. (35) showed that neointimal size in a femoral artery injury model increased in mice treated with gonadectomy, suggesting that androgens have a protective effect after arterial injury. These data suggest that androgens play both a receptor-independent and -dependent effect on the arterial intima. However, it is important to note that none of these studies reported the effect on the venous wall, and thus they are not directly comparable to our data.

After gonadectomy, female mice showed less venous wall thickness (Fig. 2J) and fewer immune cells (Fig. 4). Akishita et al. (36) showed that estrogen reduced intimal thickening of the rat femoral artery, and Dorsett-Martin et al. (33) similarly showed that estrogen administration decreased wall thickness in the arterial anastomoses proximal and distal to the AVF. Knaapen et al. (37) reported that increased smooth muscle cell expression in varicose veins was associated with increased estrogen expression in male patients. Estrogen may exert anti-inflammatory effects through nitric oxide (NO) but could be induced to have proinflammatory effects under stress conditions (38). In addition, estrogen may have increased inflammatory roles with aging (39, 40) and may repress the monocyte-macrophage axis (41). Estrogen and estrogen receptors afford cardioprotection by decreasing inflammatory cytokines and recruitment of leukocytes, modulating Th1 and Th2 responses with a shift toward Th2 differentiation, and promote Treg differentiation (42, 43). In addition, macrophages express sex hormone receptors; exogenous administration of estradiol decreases macrophage foam cell formation in atherosclerotic plaques (44). Sex differences have also been shown for endothelial nitric oxide (eNOS) expression, which mediates dilation of the remodeling vein upon exposure to the fistula environment (45). In our study, reduced estrogen was associated with fewer monocytes and macrophages in the remodeling venous wall (Fig. 4, A and B). However, the role of estrogen in the venous system is not fully understood and may underlie the differences between arterial and venous remodeling. In humans, before menopause females have approximately five times the amount of circulating estrogen as men, with a drop to near-equal levels following menopause (46). Sex differences have been described in eNOS expression, which mediates dilation of the remodeling fistula, with higher expression associated with increased diameter and less neointimal hyperplasia (45). In humans, there is greater eNOS expression and activation in female-derived cells with reduced arterial NO after menopause (47, 48), suggesting an estrogen-related mechanism. However, the effect of menopause on venous remodeling as well as on AVF maturation and patency has not been described.

The inflammatory response to the surgical trauma of AVF creation has been characterized by both innate and adaptive immune responses (16, 21, 22), and although the impact of sex-specific differences in immunity on AVF maturation has not been well described, there are well-described differences in immunologic responses between the sexes that may play a role. In humans, females have higher CD4⁺ T cell counts and CD4-to-CD8 ratios than males, and females have more robust CD4⁺ and CD8⁺ T cell activation and proliferation than males (23). We similarly showed higher CD4⁺ and CD8⁺ T cell recruitment in female than male mouse fistula walls (Fig. 4, A and B). Naive T cells from females preferentially produce IFNγ when stimulated, whereas male T cells produce more IL-17 (23). After fistula creation in mice, we also found increased expression of inflammatory cytokines in females, specifically TNF-α and IL-10 (Fig. 5, C and D). The increase in the IL-10 anti-inflammatory cytokine is often a marker of increased inflammatory response, as it increases in response to increased levels of IL-1, IL-2, IL-6, TNF-α, and others (49–52), and is further supported by the increase seen in TNF-α. Finally, human studies show increased Tregs in adult males compared with females (23). We previously showed that Tregs were upregulated during fistula remodeling, inducing polarization of M₂-type macrophages in the wall (22). Across the whole spectrum of age, females have greater antibody responses than males, higher baseline immunoglobulin levels, and higher B-cell numbers (23). These differences are influenced by a number of factors including differential gene regulation and hormone expression, all of which may regulate the inflammatory response to AVF creation.

Some limitations are present in this study. First, there exists no cell model that can recapitulate fistula physiology in a mechanistic way, limiting the in vitro experiments that can be performed. Furthermore, a limitation of the in vivo mouse aortocaval fistula is that this central AVF model may have differences in hemodynamics and remodeling compared with peripheral venous models; however, we have previously shown that this mouse model accurately recapitulates human venous remodeling, with similar dilation, thickening, and flow patterns, as well as rates of patency and sex differences, compared with human AVF (32).

Conclusions

Sex hormones may be a mechanism of the sexual dimorphism observed during arteriovenous fistula maturation. Testosterone may play a protective role during early remodeling and is associated with reduced shear stress, whereas estrogen may lead to early failure because of increased immune cell recruitment. These differential effects on venous remodeling are consistent with reduced maturation and increased failure of fistulae in females. It is possible that reducing inflammation in female patients may reduce early failure and improve maturation. Modulating sex hormones or their downstream effectors may be a useful strategy in reducing the disparity in sex difference seen in clinical outcomes.

DATA AVAILABILITY

The authors declare that all supporting data are available within the article.

SUPPLEMENTAL DATA

Supplemental Table S1 and Supplemental Figs. S1–S13: https://doi.org/10.6084/m9.figshare.21951437.

GRANTS

This work was funded by Society for Vascular Surgery Student Research Fellowships, 2021–2022 (to K.S.); Master of Health Science One-Year Research Fellowships, 2021–2022 (to K.S.); Richard K. Gershon, MD, Fund at Yale (to K.S.); and National Institutes of Health (NIH) Grants T35DK104689 (to K.S.) and R01-HL144476 and R01-HL162580 (to A.D.). The University of Virginia Center for Research in Reproduction Ligand Assay and Analysis Core is supported by the Eunice Kennedy Shriver NIH Grant R24HD102061.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

K.S. and A.D. conceived and designed research; K.S., Y.O., L.G., O.S., H.B., Y.A., Y.X., and W.Z. performed experiments; K.S., Y.O., L.G., and A.D. analyzed data; K.S., L.G., W.Z., B.Y., K.M., Y.C., and A.D. interpreted results of experiments; K.S. and A.D. prepared figures; K.S. and C.T. drafted manuscript; K.S., C.T., H.B., K.M., Y.C., and A.D. edited and revised manuscript; K.S., Y.O., C.T., L.G., O.S., H.B., Y.A., Y.X., W.Z., B.Y., K.M., Y.C., and A.D. approved final version of manuscript.