Injectable Silk-Based Hydrogel as an Alternative to Cervical Cerclage: A Rabbit Study

Abstract

Background: Preterm birth is a common cause of morbidity and mortality in newborn infants. Cervical insufficiency (CI) is a significant cause of preterm birth. The treatment for CI is cerclage, which is a suture placed around the cervix to provide mechanical support. Cerclage, however, is associated with limited efficacy. Here we present an injectable, silk-based hydrogel as an alternative to cerclage.

Objective: Pregnant rabbits were used as an animal model of pregnancy to study the mechanical properties, biocompatibility, and degradation of the hydrogel after cervical injection.

Study Design: Silk hydrogel (200 μL volume) was injected into the cervix. Controls were either injected with saline or treated with cerclage (5-0 polyethylene terephthalate suture). To study the effect on mechanical properties, the cervix was tested in compression. Biodegradation of the hydrogel was followed over 6 weeks. For biocompatibility, expression levels of proinflammatory mediators were studied.

Results: Hydrogel injection resulted in significant tissue augmentation—the cross-sectional area of the cervix increased 46.3 ± 3.0%. The modulus of the uninjected and hydrogel-injected tissues was 3.3 ± 0.7 and 3.2 ± 0.5 kPa at 5–10% strain, respectively (p = 0.8). Histology showed a mild inflammatory response surrounding the hydrogel. Biodegradation of the hydrogel showed 70% volume loss over 6 weeks. Hydrogel-injected tissue showed similar inflammatory response compared with cerclage.

Conclusions: In pregnant rabbits, cervical injection of the silk-based hydrogel was biocompatible and naturally degraded. No adverse effects on timing of delivery and pup viability were seen. Silk-based hydrogels show promise for tissue augmentation during pregnancy.

Impact Statement

This research describes the use of injectable silk-based hydrogel for augmenting cervical tissue in vivo in a pregnant rabbit model. Further preclinical development of the methods and insights described in this article can lead to therapeutic use of this hydrogel as an alternative to cerclage in preterm birth due to cervical insufficiency.

Introduction

Cervical insufficiency (CI) is an important and treatable cause of preterm birth. Epidemiological data suggest the incidence of CI is 0.5%.¹ When preterm birth occurs between 23 and 28 weeks, CI is the cause in 5% of cases.^2,3 Cervical cerclage is the treatment for CI. Natality records document that cerclage procedures are performed in 0.3–0.4% of pregnancies, which translates to 15,000 annual procedures in the United States.^3,4 In most cases of CI, the patient's obstetrical history includes a history of prior preterm birth. For these women, current recommendations are to offer supplemental progesterone in addition to cerclage.⁵ Also, cervical pessary is being studied in research protocols for women with cervical dysfunction.⁶

Though cervical cerclage can prevent preterm birth in women with CI, cerclage has important limitations. Cerclage surgery and cerclage removal can be challenging. The risk of a procedure-related complication is 0.3% and 0.9% for an ultrasound-indicated and physical examination-indicated cerclage, respectively.⁷ After the cerclage is placed, the risk of subsequent cervical laceration is increased, especially if a cerclage is present during labor.^8–10 Furthermore, cerclage does not address the pathogenesis of CI, which is the impaired mechanical properties of fibrous connective tissue of the cervical stroma.¹¹ The development of an effective alternative treatment for CI would have a significant impact in clinical obstetrics.

In prior reports, we presented an injectable silk-based hydrogel for cervical tissue augmentation as an alternate treatment for CI.^12–14 We hypothesized that injection of a hydrogel could improve cervical function in two ways. First, hydrogel injection would increase tissue volume and increased tissue volume could prevent membrane funneling and retain cervical mucus. Second, in the setting of excessively soft tissue, injection of a stiffer hydrogel would create a composite tissue with improved ability to resist stresses acting to dilate the cervix.

Silk-based hydrogels are being studied as therapeutic agents for soft tissue regeneration in heart, muscle, and skin.^15–17 In the cervix, we previously showed the mechanical properties of the hydrogel could be modulated to mimic the properties of cervical tissue through control of enzymatic cross-linking reaction.¹² More recently, injection of the silk-based hydrogel in vitro increased the volume of human cervical tissue without a significant increase in tissue stiffness.¹⁸ However, prior studies did not assess the efficacy of silk-based hydrogel with respect to cervical augmentation or biodegradation in vivo.

The purpose of this study was to evaluate cervical augmentation and biodegradation of the silk-based hydrogel in a rabbit model of pregnancy. The present report shows the potential for an injectable silk-based hydrogel as an alternate treatment for CI.

Materials and Methods

Preparation of silk fibroin solution and enzymatically cross-linked hydrogels

Silk fibroin protein was purified as we have previously described.¹² Briefly, Bombyx mori cocoons were boiled in 20 mM sodium carbonate solution for 60 min to remove the sericin coating from the fibroin fibers. The fibers were dissolved in a 9.3 M lithium bromide solution and dialyzed against water for 72 h to remove the lithium bromide salt. Enzymatically cross-linked hydrogels were prepared as previously described.^12,18 Silk solutions were diluted to 5% concentration (w/v). Horseradish peroxidase (HRP), type VI lyophilized powder (Sigma-Aldrich, St. Louis, MO) was reconstituted in deionized water and added to the silk solution in a ratio of 10 U HRP to 1 mL of silk solution. Gelation was initiated by adding 10 μL of a 0.5% (v/v in water) hydrogen peroxide solution per 1 mL of silk-HRP (final H₂O₂ molarity 1.63 mM). Gels were allowed to set at room temperature for 2 h. The following silk gel parameters were selected so that the stiffness of the gel matched the stiffness of cervical tissue during pregnancy: silk extraction time 60 min, silk concentration 5%, HRP concentration 10 U/mL, and H₂O₂ molarity 1.63 mM. All materials were sterilized by filtration through a 0.22 μm filter before gelation.

Animals

Pregnant New Zealand white rabbits, obtained from Charles River Laboratories (Wilmington, MA), were used in all experiments. All rabbits were maintained on a normal chow diet. Animal care and the experimental procedures were carried out in accordance with National Institutes of Health (NIH) guidelines and approved by the Tufts Institutional Animal Care and Use Committee. Of note, pregnant rabbits were selected for the model of pregnancy because pilot studies showed hydrogel augmentation of the cervix of pregnant rats was not feasible in our hands.

A timeline showing the timing of experiments and experimental controls is shown in Figure 1. Surgery was performed on gestational day 16 or 17. General anesthesia was induced with injection of ketamine 125 mg and xylazine 18 mg intramuscular for sedation. General anesthesia was maintained with inhalation of 1.5% isoflurane in oxygen. After the rabbit was fully anesthetized, a midline laparotomy was performed in the lower abdomen. The uterus and vagina were brought to the abdomen surface. A 1.5 cm incision was made in the posterior wall of the vagina to allow direct visualization of the two cervices (Fig. 2A). Three experimental groups were studied (Fig. 1 shows the number of animals undergoing the procedure for each experiment).

FIG. 1.

Schematic diagram shows the experimental protocol. Color images are available online.

FIG. 2.

Cervical augmentation with injectable silk hydrogels. (A) The rabbit has two separate cervices. (B) One cervix was injected, which caused gross swelling; the other cervix was not injected. (C) A cerclage was placed in one cervix. (D) Histology reveals multifocal deposits of silk gel. Color images are available online.

• Silk hydrogel injection: approximately 200 μL of silk hydrogel was injected directly into four locations of each cervix using a thin-wall 23G needle (Fig. 2B). “Injections were at 12, 3, 6, and 9 O'clock.”

• Cerclage control: the most common suture for cerclage is Mersilene tape (Ethicon, Somerville, NJ). Mersilene tape is a braided, nonabsorbable 1 × 5 mm suture composed of polyethylene terephthalate (PET). As a positive control, a 5-0 PET suture was placed in the cervix (Dacron; Alcon Surgical, Inc., Fort Worth, TX) (Fig. 2C).

• Saline injection: as a negative control, saline injections were performed in a manner similar to the silk hydrogel injections.

After the cervical procedure, the vagina was closed. The vagina, cervix, and uterus were returned to the abdominal cavity and the abdomen was closed. The rabbits were monitored daily for signs of preterm delivery.

At harvesting, cervical excision was performed using general anesthesia and a laparotomy. The animal was sacrificed after the cervix was excised. The cervix was either snap frozen in liquid nitrogen for mechanical testing or protein/RNA assays, or fixed in 10% formalin for hematoxylin and eosin (H&E) histology.

Effect on labor

To determine whether the hydrogel caused intrauterine demise, preterm birth, or labor dystocia, the hydrogel injection (n = 3 animals) was compared to saline injection (n = 3 animals). At surgery, the number of gestational sacs was recorded. The rabbits were monitored daily for signs of preterm delivery. At parturition, the animals were monitored for signs of dystocia. The liveborn kits were counted and weighed.

Mechanical analysis of cervical tissue

To determine the stiffness of augmented cervical tissue, one cervix was injected with silk hydrogel as described above. The uninjected cervix in the same animal served as the control. Five animals were studied. Cervical specimens were prepared following a previously published protocol for mechanical testing of mouse cervix.¹⁹ Briefly, the cervices were excised immediately after surgery and snap frozen in liquid nitrogen. The day before testing, tissues were equilibrated overnight in phosphate-buffered saline (PBS) at 4°C. The cervix was trimmed to the following dimensions: 8 mm diameter, 5 mm height. The cervix to be tested was placed between stainless steel parallel plates in an immersion chamber containing fresh PBS (Fig. 3A). The tissue was tested in unconfined compression with an Instron Universal Testing System (Model 3366; Instron, Norwood, MA) using a 5 N load cell. Compression occurred along the axis of the cervical canal (Fig. 3A). Each sample was initially loaded to a preload of 0.02 N. The sample was compressed to 20% axial strain and unloaded at a strain rate of 1 mm/min. Three cycles were performed and data were reported for the final cycle. The modulus was measured as the slope between 5% and 10% axial strain, which was a linear region of the stress–strain curve.

FIG. 3.

Mechanical tests show that the mechanical features are no significant differences between non-injected control cervical tissue and silk-injected cervical tissue (n = 5). (A) Samples (8-mm diameter, 5-mm height) were loaded onto an Instron Universal Testing System between stainless steel parallel plates in an immersion chamber containing fresh phosphate-buffered saline. A preload of 0.02 N was used to ensure proper contact with the sample. (B) Representative stress–strain curves between control and silk. (C) Paired data are shown. One rabbit yielded two cervices: one control and one with silk hydrogel. Modulus was taken as the slope between 5% and 10% region that is linear. Color images are available online.

Determination of in vivo degradation of silk hydrogel

Characterization of hydrogel degradation is critical for predicting hydrogel efficacy and safety. To study the degradation of silk hydrogel in vivo, the cervices were harvested at postoperative day 0 (n = 4), day 5 (n = 2), day 14 (n = 4), day 28 (n = 6), and day 42 (n = 4). Formalin-fixed whole cervices were embedded in paraffin, cut in cross section (5 μm thickness) at 750 μm intervals (eight sections per cervix) and stained with H&E. To capture an entire cervical specimen at high-magnification, image stitching module of Keyence Microscope BZ-X700 (Keyence Corporation of America, Itasca, IL) was used and a single high-resolution image was created. ImageJ (NIH) was used to quantify silk gel areas in cervical histological samples.

Biocompatibility: isolation of RNA and quantitative real-time polymerase chain reaction

Gene expression of inflammation-related proteins was compared between saline-injected, hydrogel-injected, and cerclage groups (n = 4 cervices per group). Cervices were harvested 7 days after surgery and snap frozen in liquid nitrogen. RNA was isolated using an RNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. One microgram of RNA sample was used for reverse transcription and synthesis of cDNA using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA). Quantitative real-time polymerase chain reaction (RT-PCR) was performed on a QuantStudio 7 Flex (Applied Biosystems) using SYBR Green PCR Master Mix (Qiagen). All genes were normalized by the housekeeping gene β-actin. Quantification was expressed as fold change relative to gene expression of the saline-injected control group using the 2^−(ΔΔCt) method. Predesigned RT-PCR primers for rabbit inflammatory target genes were purchased from Sigma-Aldrich. The following genes were studied: interferon gamma (IFNγ), chemokine (C-C motif) ligand 2 (CCL2), tumor necrosis factor alpha (TNFα), interleukin 8 (IL8), interleukin 1 beta (IL1β), C-X-C motif chemokine receptor 2 (CXCR2), C-C chemokine receptor type 3 (CCR3), and C-C chemokine receptor type 1 (CCR1).

Biocompatibility: protein assay

Native proteins from the cervix were isolated using Total Protein Extraction Kit (Biochain Institute, Inc.). The protein levels of IL1β and IL8 in the tissue lysate was determined using commercially available enzyme-linked immunosorbent assay (ELISA) kits (Sigma-Aldrich).

Statistical analysis

Data are expressed as scatterplots. Comparisons between groups were performed using nonparametric statistics. Kruskal–Wallis test was used for statistical significance between groups and the Dunn's multiple comparison test was used if the Kruskal–Wallis test was significant (GraphPad Prism ver. 5.04; San Diego, CA). Results were considered significant at p < 0.05.

Results

Effect on labor

All animals tolerated the surgery well. In the postoperative period, however, one of the hydrogel-injected animals was diagnosed with a vertebral fracture at L6/L7. She was completely paralyzed below the fracture and euthanized for humane reasons. The cause of the vertebral fracture was thought to be a startle reaction as she was coming out of sedation. A startle reaction causing a vertebral fracture in rabbits is a known complication of the rabbit model due to the combination of their strong hind legs and low density of their bones.²⁰ Otherwise, there were no complications from the surgery. None of the animals delivered preterm. The surgical mortality rate for the kits was not significantly different between the saline and hydrogel groups. The mortality rate was defined as the fraction of gestational sacs seen at surgery that did not result in a live kit (e.g., if 10 sacs were seen at surgery and 5 kits were seen at parturition, the mortality rate was 50%). The mortality rate was 37.1% and 40.9% for the hydrogel and saline groups respectively (p = 0.78).

Mechanical testing of silk hydrogel-injected cervical tissue

Silk hydrogel injection into the cervix resulted in significant tissue augmentation both grossly and histologically (Fig. 2B, D). The stress–strain profile of the silk hydrogel-injected tissue was similar to uninjected controls (Fig. 3B). The modulus of the uninjected and hydrogel-injected tissues was 3.3 ± 0.7 and 3.2 ± 0.5 kPa at 5–10% strain, respectively (Fig. 3C, p = 0.8). This result is important because it suggests the injectable hydrogel does not cause excessive stiffening of the tissue.

In vivo biodegradation of silk hydrogel

Silk hydrogel injections were localized close to the external opening of the cervix into the vagina (external os). The depth of silk hydrogel was seen in the 2 mm of the cervical stroma closest to the external os (total cervical length was 5 mm). The ratio of maximum silk area/tissue area (%) was used to measure the degradation rate. At time = 0, the silk gel comprised 46.3 ± 3.0% of the cross-sectional area. The silk hydrogel gradually degraded significantly over the 7-week study period (Fig. 4). By day 42, the silk hydrogel comprised 12.2 ± 7.1% of the cross-sectional area. Compared to day 0, statistically significant silk degradation was seen at day 28 and 42. Compared to day 5, significant degradation was seen at day 42.

FIG. 4.

In vivo biodegradation of silk gels in cervical tissues. The content of silk gel in cervical tissue is gradually decreased, losing about 70% after 42 days (*p < 0.05 vs. day 0, ^#p < 0.05 vs. day 5). Color images are available online.

Inflammatory response: histology

Histology of the cervical stroma showed a mild inflammatory response surrounding the silk gel from day 5 to 42 (Fig. 5). On day 5, there was a thin rim of multinucleated giant cells with scattered rare heterophils and macrophages. On day 14, which was several days postpartum, there was more inflammation with increased density of multinucleated giant cells, lymphocytes, plasma cells, and heterophils compared to day 5. On day 28, the inflammatory response was minimal. Scattered lymphocytes and heterophils were seen. In addition, an epithelial layer was visualized surrounding parts of the silk hydrogel. On day 42, the inflammatory response was minimal and the epithelial layer was visualized in the vicinity of the silk hydrogel.

FIG. 5.

Histology of the stroma showed a mild inflammatory response surrounding the silk gel. Macrophages, together with smaller numbers of neutrophils and eosinophils, were seen at the interface of the gel. At day 28 and 42, an epithelial layer was seen around parts of the silk gel (black arrow). Color images are available online.

Inflammatory response: gene expression

Expression levels of inflammation-related genes were compared in saline-injected, hydrogel-injected, and cerclage cervices 1 week after the surgery (Fig. 6A). Expression of inflammatory genes in hydrogel-injected cervices was not significantly different compared with saline-injected and cerclage cervices. Increased expression of inflammatory genes was seen in the cerclage group compared to the saline-injected groups in the following genes: IL8, IL1β, IFNγ, CCR3, CXCR2, and CCR1. For TNFα, and CCL2, no differences in gene expression were seen in any group.

FIG. 6.

Expression of inflammatory genes in the silk gel, cerclage, and saline controls. (A) Quantitative polymerase chain reaction. (B) Enzyme-linked immunosorbent assay for IL1β and IL8 (n = 4, *p < 0.05 vs. saline). IL, interleukin.

Inflammatory response: protein levels

ELISA showed protein levels correlated with gene expression for IL8 and IL1β (Fig. 6B). The protein concentration of IL8 and IL1β was lowest in the saline-injected cervices. Statistically significant increased concentrations of IL8 and IL1β were seen in the cerclage cervices compared with the saline-injected cervices. Hydrogel-injected cervices showed no differences in IL8 and IL1β concentrations compared with the other two groups.

Discussion

Injection of silk hydrogel significantly increased the volume of the rabbit cervix without an adverse effect on parturition or kit development. Immediately after injection, the hydrogel was seen in 46% of the cross-sectional area of the cervix. No change in the mechanical properties of the cervix was measured after hydrogel injection. Over the 6-week study period, the hydrogel degraded to 14% of the cross-sectional area. The expression of inflammatory markers after hydrogel injection was not different compared with cerclage controls or saline-injected controls. These results suggest an injectable hydrogel has promise as an alternative treatment for CI.

The rationale for the development of an injectable treatment for CI is to address the limitations of cerclage treatment. A cerclage is indicated after one or more mid-trimester losses caused by painless cervical dilation without evidence of an alternate cause (e.g., no evidence of labor, bleeding, infection, ruptured membranes).²¹ A cerclage exerts its therapeutic efficacy, in part, by providing load bearing support for cervical tissue that is unable to withstand the stresses of fetal growth.¹¹ A cerclage also protects the amniotic sac and retains the cervical mucus.

The current report demonstrates an alternate treatment approach for cases of CI. Rather than using a suture to provide load bearing support for cervical tissue, the hydrogel increases the volume of the cervix. Increased cervical volume could act to counter cervical shortening by acting as a stronger barrier between the vaginal and fetal environments. Also, an injectable treatment avoids surgical risks associated with cerclage and avoids the need to remove the cerclage. In addition, the likelihood of cervical laceration is less with an injectable hydrogel because the soft hydrogel will deform as the cervix effaces and dilates. Should a silk-based hydrogel prove to be efficacious, the hydrogel could be designed to release medication (e.g., progesterone) to improve efficacy.^22,23 In addition, a hydrogel injection could be used in conjunction with a pessary for added therapeutic effect. It is important to emphasize that the benefits of an injectable treatment are hypothesized at present. A rigorous analysis of cervical biomechanical function as it relates to different treatment options is not available. In addition, an animal model of CI has not been developed. Ultimately, the efficacy of an alternate treatment for CI will have to be tested in a clinical trial.

Injectable, degradable hydrogels are used widely in medicine with applications in dermatology, orthopedics, hemostasis, and otolaryngology.^24–27 Silk-based hydrogels show promise for vocal cord augmentation²⁸ and regenerative medicine.²⁹ In previous work, we studied sonicated and chemically cross-linked silk gels for cervical augmentation.¹³ We found, however, that earlier prototypes had drawbacks such as excessive stiffness or the requirement for exogenous ethanol to initiate gelation. We recently showed enzymatically catalyzing cross-links between silk chains produces an elastomeric hydrogel and that the mechanical properties of silk hydrogels could be formulated to match the properties of normal cervical tissue in pregnancy.¹² It is important to emphasize that cervical tissue in women with CI is thought to be substantially weaker than normal cervical tissue. Thus, injection of a hydrogel with properties similar to normal cervical tissue will create a composite tissue. The composite tissue will have improved ability to resist stresses acting to dilate the cervix because part of the tissue will be composed of hydrogel with properties similar to normal tissue. Also, our goal was to not formulate a hydrogel that is stiffer than normal tissue because a stiffer gel could prevent cervical dilation in cases where delivery is indicated.

Another important feature of silk hydrogels is the time course of degradation. The time course of silk hydrogel degradation can be controlled by changing the purification parameters of the silk fibroin protein^30,31 and the contents of the hydrogel.³² In the present study formulation, a significant amount of hydrogel degradation was seen over the 6-week study period. Six weeks may be an appropriate time period for pregnancies that would have been treated with a mid-trimester ultrasound-indicated cerclage or a physical examination-indicated cerclage. However, 6 weeks is too short a time period for an alternative prophylactic intervention to be effective, as a prophylactic cerclage is placed at 12–14 weeks. Further work is needed to optimize the degradation profile for cervical treatment.

PET woven tape is the most common material used for cerclage. A cerclage composed of a nonabsorbable PET tape acts as a foreign body in two compartments: cervical stroma and vaginal canal. In the vagina, woven PET cerclage is exposed to the vaginal microbiome, and PET has been associated with vaginal microbiome dysbiosis compared to monofilament suture.³³ In cervical tissue, there have been no previous studies examining the inflammatory response of different suture materials. In the present study, the inflammatory response to the hydrogel was not different compared with PET and saline. Also, we noted the development of an epithelial layer at the border between the cervical stroma and the hydrogel at 4 and 6 weeks postsurgery. Whether the histological changes seen in the rodent cervix are relevant to the human cervix is not known and requires further study.

Several limitations in the current study are noted. The precise measurement of the mechanical properties of the rodent cervix is challenging due to its small size.¹⁹ Injection of the hydrogel into cervical tissue essentially created a heterogeneous composite tissue—the hydrogel was not uniformly distributed. Hence, measurement of tissue mechanical properties was not precise. A second limitation is the cervix was injected after a laparotomy, which does not mimic clinical use. For us, it was not feasible to visualize the cervix via a vaginal approach.

In summary, this study shows the feasibility of injectable hydrogel for tissue augmentation during pregnancy. The concept of an injectable hydrogel may offer an alternate treatment approach for cases of CI. Future work will focus on optimizing the mechanical properties and degradation profile for use in the cervix.

Disclosure Statement

M.H. reports stock ownership in Cx Therapeutics, Inc. The terms of this arrangement have been reviewed and approved by Tufts Medical Center in accordance with its policy on objectivity in research.

Funding Information

This project has been funded and supported by a grant from National Institutes of Health (R01EB021264).