An Injectable Reverse Thermal Gel for Minimally Invasive Coverage of Mouse Myelomeningocele

j o u r n a l o f s u r g i c a l r e s e a r c h  m a r c h 2 0 1 9 ( 2 3 5 ) 2 2 7 e2 3 6

Available online at www.sciencedirect.com

ScienceDirect journal homepage: www.JournalofSurgicalResearch.com

An Injectable Reverse Thermal Gel for Minimally Invasive Coverage of Mouse Myelomeningocele James Bardill, MS,a,b Sarah M. Williams, BS, AAS,b Uladzimir Shabeka, MD,b Lee Niswander, PhD,c Daewon Park, PhD,a,** and Ahmed I. Marwan, MDa,b,* a

Department of Bioengineering, University of Colorado Denver Anschutz Medical Campus, Aurora, Colorado Division of Pediatric Surgery, Department of Surgery, University of Colorado Denver Anschutz Medical Campus, Aurora, Colorado c Department of Molecular, Cellular and Developmental Biology, University of Colorado Boulder, Boulder, Colorado b

article info

abstract

Article history:

Background: Myelomeningocele (MMC) results in lifelong neurologic and functional deficits.

Received 16 May 2018

Currently, prenatal repair of MMC closes the defect, resulting in a 50% reduction in post-

Received in revised form

natal ventriculoperitoneal shunting. However, this invasive fetal surgery is associated with

24 September 2018

significant morbidities to mother and baby. We have pioneered a novel reverse thermal gel

Accepted 25 September 2018

(RTG) to cover MMC defects in a minimally invasive manner. Here, we test in-vitro RTG

Available online 31 October 2018

long-term stability in amniotic fluid and in vivo application in the Grainy head-like 3 (Grhl3) mouse MMC model.

Keywords:

Materials and methods: RTG stability in amniotic fluid (in-vitro) was monitored for 6 mo and

Reverse thermal gel

measured using gel permeation chromatography and solutionegel transition temperature

Neural tube defects

(lower critical solution temperature). E16.5 Grhl3 mouse fetuses were injected with the RTG

Myelomeningocele

or saline and harvested on E19.5. Tissue was assessed for RTG coverage of the gross defect

Prenatal repair

and inflammatory response by immunohistochemistry for macrophages.

Biomimetic patch

Results: Polymer backbone molecular weight and lower critical solution temperature remain stable in amniotic fluid after 6 mo. Needle injection over the MMC of Grhl3 fetuses successfully forms a stable gel that covers the entire defect. On harvest, some animals demonstrate >50% RTG coverage. RTG injection is not associated with inflammation. Conclusions: Our results demonstrate that the RTG is a promising candidate for a minimally invasive approach to patch MMC. We are now poised to test our RTG patch in the large preclinical ovine model used to evaluate prenatal repair of MMC. ª 2018 Elsevier Inc. All rights reserved.

Introduction Myelomeningoceles (MMC) are one of the most common global congenital birth defects of the spinal cord and result in

lifelong neurologic and functional deficits.1-4 This open neural tube defect occurs when the developing neural plate fails to form the neural tube during development, resulting in exposure of the nerve elements with or without a cystic sac

* Corresponding author. Division of Pediatric Surgery, Department of Surgery, University of Colorado Denver, Anschutz Medical Campus, Building: Research 2 (Building P15), Room: 6400A, Aurora, CO 80045. Tel.: þ1 720 777-6549; fax: 303-724-6330. ** Corresponding author. University of Colorado Denver Anschutz Medical Campus, Building: Research 1 North (Building P18), Room: 4118, Aurora, CO 80045. Tel.: þ1 303 724-6947; fax: 303-724-6330. E-mail addresses: [email protected] (D. Park), [email protected] (A.I. Marwan). 0022-4804/$ e see front matter ª 2018 Elsevier Inc. All rights reserved. https://doi.org/10.1016/j.jss.2018.09.078

228

j o u r n a l o f s u r g i c a l r e s e a r c h  m a r c h 2 0 1 9 ( 2 3 5 ) 2 2 7 e2 3 6

overlying the defect area.5-7 Subsequently, the developing neural tissue is exposed to mechanical trauma and chemical irritation within the amniotic environment, causing progressive neurodegeneration.8 Patients living with MMC experience functional and emotional deficits, including hydrocephalus requiring shunt placement, paralysis, musculoskeletal deformities, bladder and bowel incontinence, and psychosocial concerns. Moreover, infants born with MMC have a postnatal death rate of up to 10% secondary to brainstem dysfunction from the Chiari II malformation and hindbrain herniation.9-11 Although failure of neural tube closure currently cannot be corrected, the secondary effects of the progressive neurodegeneration can be targeted by covering the MMC defect before birth. A landmark clinical trial, referred to as the Management of Myelomeningocele Study (MOMS trial), investigated an early, prenatal surgical repair of MMC and compared it with traditional postnatal repair.12 This prenatal surgery uses surrounding skin, myofascial flaps and/or biological patches to cover the MMC, thus protecting the developing spinal cord from further exposure to the amniotic environment.12,13 Results demonstrated that fetal surgery provided significant benefits to the infant, including a 50% reduction in the need for postnatal ventriculoperitoneal shunting, reversal of hindbrain herniation, a tendency toward improved neurologic and functional outcomes, and a decreased risk of death before 1 y of age.14,15 Despite the improvements that prenatal surgery offers, the invasive nature of the surgery constitutes significant risks to mother and fetus, including a 50% risk of premature rupture of membranes, a 30% risk of preterm birth (often early enough to lead to major morbidity and mortality), and a 30% risk of uterine scar complications (partial or complete dehiscence and risk of uterine rupture with subsequent pregnancies).14 In addition, currently, prenatal surgery for MMCs is usually performed relatively late in gestation, at 22-24 wk of gestation relative to the development of a closed neural tube (4 wk), which may be too late to prevent significant damage to the developing spinal cord. To decrease these complications and improve outcomes even further, the future of fetal intervention for MMC is in the successful application of a minimally invasive approach to close or cover the defect earlier in gestation.16 Current research strategies have either incorporated biomaterial patches as a means to cover MMC defects for the duration of the pregnancy or used a conventional laparoscopic repair using multiple trocar insertion sites.17-24 Despite promising preliminary results, these approaches are still considered experimental and are done relatively late in gestation, at 22- to 24-wk gestation. Moreover, they are associated with chorioamniotic membrane complications, require two to three trocar insertion sites that create multiple membrane violations/fixations, require a significant operative time with CO2 insufflation that can lead to significant hypercarbia and acidosis, and fetal exposure to inhalational anesthetic agents used for the mother. The associated increased operative times and hypercarbia are concerning, as there is growing evidence that prolonged exposure to high-dose inhalational anesthetic agents used for fetal surgery may cause dose- and durationdependent neuroapoptosis, therefore, potentially worsening the neurologic outcomes associated with open repairs.25-28

Therefore, based on the limitations of current MMC repair techniques, we hypothesize that earlier repair using needlebased/single-access fetoscopy and a bioengineered reverse thermal gel (RTG) will combine the advantages of reduced port access and minimal operative times, thus potentially improving overall outcomes. The development of an RTG poses a unique approach to prenatal MMC coverage. RTGs exist as a solution at room temperature (RT) that quickly forms a stable, solid scaffold in physiological environments. RTG can be delivered by small gauge needle injection. The injectable properties of an RTG present an opportunity to deliver a patch to cover MMCs by fetoscopic injection that would require less invasive in-utero surgical interventions and decreased operative time.29 Furthermore, a fetoscopic-guided needle-based procedure offers an opportunity to patch the MMC at an earlier gestational age that is not yet achievable with current in-utero surgery methods in addition to the promising scaffolding properties that may allow cellular and/ or growth factor therapy. Our previous work demonstrated that the RTG, poly (serinol hexamethylene urea) poly (N-isopropylacrylamide) (PSHU-PNIPAAm) has in-vitro biocompatibility, limited permeability, and stability in amniotic fluids. In addition, an ex-utero culture using Grlh3 mouse embryos cultured with PSHU-PNIPAAm demonstrated no toxic effect of the RTG on fetal viability and development up to 15 h in culture.30 In the present study, we aim to expand the exploration of our PSHU-PNIPAAm RTG as a minimally invasive approach to patch MMC by assessing its long-term stability in amniotic fluid and by developing an in-utero application in the open MMC Grhl3 mouse model followed by testing for cellular inflammatory response.

Methods and materials Synthesis of reversible thermal gel: PSHU-PNIPAAm PSHU-PNIPAAm was synthesized and characterized as previously described.31 Briefly, serinol (1.959 g, 21.50 mmol; SigmaeAldrich, St. Louis, MO) was dissolved in 20 mL of ethanol. Di-tert-butyl dicarbonate (5.973 mL, 26 mmol; Alfa Aesar, Ward Hill, MA) was dissolved in 15 mL ethanol (Decon Labs, Inc, King of Prussia, PA) and added dropwise to the serinol solution at 4 C with stirring. This solution was then stirred at 37 C for 1 h, followed by rotary evaporation of ethanol. A 1:1 mixture of ethyl acetate (Alfa Aesar) and hexane (EMD Millipore, Billerica, MA) was used to recrystallize the NBOC serinol, and the purified crystals were collected by vacuum filtration. The resulting white crystals were stored at RT. PSHU was synthesized with a 1:1:2 M ratio of urea (SigmaeAldrich), N-BOC-serinol, and hexamethylene diisocyanate (SigmaeAldrich), respectively, dissolved in 6 mL anhydrous dimethylformamide (DMF; EMD Millipore, ) and reacted at 90 C under nitrogen atmosphere for 7 d. The reaction mixture was washed in diethyl ether (Fisher Scientific, Pittsburgh, PA), water, and solvents were removed with rotary evaporation. These washing steps were repeated several times to remove remaining DMF solvent. The resulting white PSHU powder product was stored at RT. Next, carboxylic acideterminated

b a r d i l l e t a l  r e v e r s e t h e r m a l g e l f o r s p i n a b i fi d a c l o s u r e

PNIPAAm (PNIPAAm-COOH) was synthesized. N-isopropylacrylamide (NIPAAm, 44 mmol; SigmaeAldrich) and N, and 4, 40 -azobis (4-cyanovaleric acid) (0.2 mmol; SigmaeAldrich) were dissolved in 25 mL of anhydrous methanol (Alfa Aesar) and purged with nitrogen gas for 30 min. A polymerization reaction was then carried out at 68 C for 3 h with refluxing. The reaction mixture was precipitated in warm water (45 C), followed by purification in dialysis tubing with molecular weight cutoff 12,000-14,000 Da (Da). The purified PNIPAAm-COOH was lyophilized at 45 C for 72 h and stored at RT. To conjugate PNIPAAm-COOH to the PSHU backbone, the BOC protecting groups on the PSHU backbone were removed using a 1:1 (vol/vol) mixture of dichloromethane (JT Baker, Phillipsburg, NJ) and trifluoroacetic acid (Alfa Aesar) for 45 min to generate free amine groups on the PSHU backbone. The resulting product was dissolved in trifluoroethanol (TFE; SigmaeAldrich) precipitated in ether, and stored at RT. Next, PNIPAAm-COOH (0.2 mmol) was activated with N-(3 Dimethylamino-propyl)-N0 -ethylcarbodiimide hydrochloride (1.0 mmol; Alfa Aesar) and N-hydroxysuccinimide (1.0 mmol; Alfa Aesar) in anhydrous DMF for 24 h at RT. Five hundred milligram of dPSHU was dissolved in anhydrous DMF and added dropwise to the activated PNIPAAm solution. The conjugation reaction took place for 24 h. The product mixture was precipitated in ether and then purified by dialysis (molecular weight cutoff: 12,000-14,000 Da) for 7 d at RT. The purified PSHU-PNIPAAm was then lyophilized for 3 d and stored at RT.

RTG long-term in-vitro stability in amniotic fluid Stability of PSHU polymer backbone: gel permeation chromatography Human amniotic fluid was collected at the time of an MMC case after obtaining consent from the mother (Children’s Hospital ColoradoeCOMIRB # Protocol #: 15-1141). Amniotic fluid was centrifuged at 2000 rpm for 10 min, and the supernatant was removed, and stored at 80 C. Only one amniotic fluid donor was used for in-vitro studies. PSHU was dissolved in TFE to a final concentration of 20 mg/mL, and 1 mL aliquots were added to 1.8 mL glass vials. To obtain an even PSHU coating in each vial, the TFE was evaporated overnight. Individual vials received 500 mL of phosphate-buffered saline (PBS; Thermo Fisher Scientific, Waltham, MA), 0.25% Trypsin (Thermo Fisher Scientific), or human amniotic fluid, ensuring the entire polymer coating was submerged in the fluid. The vials were placed in a 37 C, 5% CO2 incubator for the duration of each time point. After each time point, the supernatant was removed, followed by three washes ultrapure water, and each vial was lyophilized for 24 h. Each vial was then dissolved in 1 mL of DMF and then filtered through a 0.2 mm polytetrafluoroethylene filter into a new 1.8-mL glass vial. Gel permeation chromatography (GPC; Malvern Instruments, Houston, TX) analysis was performed on each sample using a 100 mL injection into a single Viscotek D6000M column and 270 Dual Detector with right-angle light scattering with DMF as the system solvent. The column and detector temperatures were kept constant at 45 C. The instrument was calibrated with polystyrene standards (MW:,105,000, dn/dc: 0.185 mL/g). Each

229

exposure fluid type at each time point had three replicates per fluid group that were averaged.

PSHU-PNIPAAm solution-gel transition temperature To verify the RTG retains its solution-gel transition temperature within the amniotic environment over time, the RTG was exposed to amniotic fluid for 2, 3, and 6 mo, and its lower critical solution temperature (LCST) was determined afterward. First, a 10% (g/mL) polymer stock solution of PSHUPNIPAAm was prepared in PBS, and 100 mL (10 mg of polymer) of this polymer stock solution was injected into a vial containing 1 mL of warm PBS, 0.25% Trypsin, or human amniotic fluid. The gelled polymer was exposed to the warm fluid for 2, 3, and 6 mo. After each time point, the amniotic fluid was removed, and the gelled polymer was washed with warm ultrapure water three times and then freeze-dried. LCST was determined using a temperature sensitive UV-VIS spectrometer (Varian Cary 100; Agilent, Santa Clara, CA). A 1% polymer solution is required for this instrument to best visualize the phase transition from a translucent solution to an opaque gel. The freeze-dried polymer samples were redissolved in PBS and placed in a cuvette for LCST testing in the spectrometer, using a constant wavelength of 500 nm. Samples were tested over a temperature range of 25 C-45 C, with a rate of 2 C/min. Each exposure fluid type at each time point had three replicates per fluid group that were averaged.

Animal procedures All animal procedures were carried out in accordance with the National Institutes of Health guide for the care and use of laboratory animals and were approved by the Institutional Animal Care and Use Committee at the University of Colorado Anschutz Medical Campus (Protocol Number 00181, Approval date: October 24, 2016). Mice were maintained on a lightedark (14-h light, and 10-h dark cycle) with access to food and water ad libitum.

Mouse model Grainy headlike-3 (Grhl3Cre) mouse line was created by the Coughlin laboratory and supplied by Dr Niswander’s laboratory.32 The Grhl family of transcription factors has proven to play critical roles in neural tube closure and epidermal development.33-36 The Grhl3Cre mutant strain carries a “knockin” of Cre-recombinase into the Grhl3 locus, causing a loss of function allele. Homozygous mutant fetuses fail to undergo spinal neural tube closure, leading to a neural tube defect that extends over the caudal spinal region. This mutant mouse model is ideal, as it has been previously shown to recapitulate a phenotype of nonsyndromic human neural tube defects.

In-utero injections A colony of Grhl3 mice was established. Grhl3Cre/þ males and females were timed mated and on embryonic day (E) 16.5 the pregnant dams were anesthetized using isoflurane then placed on a heated pad of a surgical platform. Bupivacaine was administered subcutaneously at the incision site, and a midline laparotomy was performed, and the gravid uterine horns were delivered gently onto a sterile surgical drape. MMC-defect fetuses (Grhl3Cre/Cre) were identified visually. All

230

j o u r n a l o f s u r g i c a l r e s e a r c h  m a r c h 2 0 1 9 ( 2 3 5 ) 2 2 7 e2 3 6

MMC fetuses were treated with either PSHU-PNIPAAm or saline; 30-50 mL per injection. Fetuses without defects were also treated with 30-50 mL of PSHU-PNIPAAm or saline. The fetuses in each dam all received the same injection type. The needle was inserted into the yolk sac over the caudal region, and the solution was injected directly onto the caudal defect region of the fetus, using enough PSHU-PNIPAAm to completely cover the defect. Warm saline was immediately applied on top of the yolk sac over the injection site to allow the polymer to form a stable gel. Each fetus was monitored for 2 min after injection and kept warm with saline. On completion of the injections, the uterine horns were returned to the abdominal cavity. Muscle and skin layers were sutured separately. Buprenorphine was administered subcutaneously, and Ibuprofen was added to the water bottle for pain management, along with subcutaneous saline for fluid replacement. Pregnant females were recovered on warm surgical platform until mobile and returned to a new cage. All animals were checked each day for signs of surgical complications. On E19.5, pregnant females were euthanized before tissue harvest. The MMC defects are large and can be identified visually, and RTGinjected mice show an apparent patch at the defect site. Fetal mice were photographed to document defect coverage, and the defect area was harvested in block and either placed in optimal cutting temperature (Torrance, CA) or formalin (JT Baker) for analysis. RTG percentage coverage was calculated by tracing the area of RTG covering the MMC and dividing this area by the total traced area of the defect. Tissue for morphologic analysis was fixed in formalin overnight, embedded in paraffin, cut into 4 mm transverse sections, and stained for hematoxylin and eosin (H&E) by the University of Colorado Anschutz Medical Campus Morphology and Phenotyping Core. Tissue for immunohistochemistry (IHC) was fixed overnight in formalin, cryoprotected in 10%, 20%, and 30% sucrose (dissolved in PBS), embedded in optimal cutting temperature, and cut into 10-mm thick transverse sections on a cryostat (ThermoFisher Cyrostat NX70).

Morphology: H&E staining H&E staining was performed on paraffin-embedded tissues by the University of Colorado Anschutz Medical Campus Morphology and Phenotyping Core. Brightfield images were obtained with a Nikon Ti Eclipse microscope.

IHC staining for CD68- and F4/80-positive cells Tissue sections were fixed in neutral buffered formalin for 10 min and washed three times with wash buffer (0.1% Tween 20 [Sigma, St. Louis, MO], PBS) for 5 min each. Sections were permeabilized with 0.5 % Triton X-100 (Sigma) in PBS for 10 min. Sections were washed three times with wash buffer for 5 min each. Blocking buffer consisting of 0.25% Triton X100, 2% bovine serum albumin (Sigma), 4% bovine gamma globulins (Sigma), and PBS was used for 60 min on the sections at RT. Sections were stained with primary antibodies (diluted in blocking buffer) F4/80 (Thermo Fisher Scientific; 1:500) and cluster of differentiation 68 (CD68; Abcam 125212, Cambridge, UK; 1:250) overnight at 4 C and washed three times in wash buffer for 5 min each the next day, followed by staining with

secondary antibodies Alexa Fluor 488 (Thermo Fisher Scientific; 1:500) for F4/80 and Alexa Fluor 594 (Thermo Fisher Scientific; 1:500) for CD68 at RT for 1 h. The sections were washed three times in wash buffer for 5 min each, followed by three washes in ultrapure water for 5 min each. Sections were mounted to glass slides with 40 ,6-diamidino-2-phenylindole Fluoromount-G (Electron Microscopy Sciences, Hatfield, PA). Negative control slides had no secondary antibody. Fluorescent images were obtained using Zeiss LSM 780 confocal microscope. For defect and nondefect fetuses, the images were taken of the defect area where the injection took place. F4/80- and CD68-positive cells were quantified by counting all F4/80- or CD68-positive cells, which overlapped with 40 ,6-diamidino-2-phenylindole within the entire field of view. Two sections were counted per fetus, with four to five fetuses counted per injection group for both defect and nondefect fetuses.

Statistical analysis All experimental groups have four to six animals evaluated against an identical number of controls. Results are expressed as mean  standard error (SE) of the mean. Comparisons were made between experimental groups using a two-tailed t-test assuming unequal variances. A difference was accepted as significant when the statistical test yielded a P < 0.05.

Results Long-term in-vitro stability in amniotic fluid To monitor possible degradation of the polymer backbone after long-term exposure to amniotic fluid, the PSHU polymer was exposed to PBS, 0.25% Trypsin, and human amniotic fluid for 2, 3, and 6 mo. We measured polymer chain length of three samples per medium group using GPC and found no significant changes to the molecular weight (Mn) of PSHU after 2, 3, or 6 mo in any of the three fluids relative to nonexposed PSHU controls (Fig. 1). The gelling properties of PSHU-PNIPAAm were measured using LCST after long-term exposure to PBS, 0.25% Trypsin, and human amniotic fluid. At RT, PSHUPNIPAAm exists as an aqueous solution because of the hydrophilic components of the polymer dominating the chemical properties (Fig. 2A). As the temperature increases, the hydrophobic groups begin to come together to form a physical gel that turns opaque (Fig. 2B). Transmission measurements record the temperature at which this transition from solution to opaque gel occurs (referred to as the LCST). The LCST for PSHU-PNIPAAm occurs near 34 C and occurs quickly, as indicated by the steepness of the slope of the LCST curves. Once gel formation occurs, a small decrease in the temperature below the LCST will cause a shift back to the solution state. We measured the LCST of three samples per medium group using UV-VIS spectroscopy, and no changes to the LCST are observed after PSHUPNIPAAm was exposed to PBS, 0.25% Trypsin, or human amniotic fluid for 6 mo, with the LCST remaining near 34 C-35 C (Fig. 2C).

b a r d i l l e t a l  r e v e r s e t h e r m a l g e l f o r s p i n a b i fi d a c l o s u r e

231

Fig. 1 e Gel permeation chromatography (GPC) results: Long-term PSHU exposure to PBS, 0.25% Trypsin, and human amniotic fluid. No significant decreases in PSHU molecular weight (Mn) are observed. Error bars: standard error of the mean.

Grhl3 mouse model for PSHU-PNIPAAm MMC coverage Grhl3Cre/Cre mice exhibit MMC and/or curly tail phenotypes with a lower penetrance of exencephaly or craniorachischisis on dissection on embryonic stage day (E19.5; Supplemental Fig. 1). As the neural tube defect is a recessive trait, w25% of the litter will be Grhl3Cre/Cre embryos.

harvested on E19.5, and the coverage of the defect was assessed. Saline injections were selected as a vehicle control and are considered as 0% defect coverage. The percentage of coverage is variable. Of six MMC fetuses, three exhibit a mean 60.4% coverage (SE ¼ 1.9%), and three exhibit a mean 11.5% coverage (SE ¼ 7.1%; Fig. 3).

Histology Grhl3 in-utero PSHU-PNIPAAm injections Morphology: H&E staining The RTG quickly forms a stable gel on transuterine needle injection of Grhl3 E16.5 fetuses and covers the MMC defects (Supplemental Fig. 2, Supplemental Video 2). Fetuses were

We next assessed the morphology of the Grhl3 defect and nondefect fetal tissue at the site of RTG or saline injection using H&E staining (Fig. 4). Nondefect Grhl3 fetuses injected

Fig. 2 e Solution-gel transition temperatureeLCST: (A) PSHU-PNIPAAm is a flowing solution in a vial at RT that forms a stable gel at 37 C with the vial inverted (B). (C) PSHU-PNIPAAm LCST results after 6 mo of exposure to PBS, 0.25% Trypsin, and human amniotic fluid (transmission at 500 nm). The LCST remains at 34 C-35 C, maintaining the sharp transition from solution to gel, as indicated by the steep curve for each sample. Error bars: standard error of the mean. (Color version of figure is available online.)

232

j o u r n a l o f s u r g i c a l r e s e a r c h  m a r c h 2 0 1 9 ( 2 3 5 ) 2 2 7 e2 3 6

Fig. 3 e Grhl3 knock-out Fetuses 3 d after in-vivo injections: Grhl3Cre/Cre fetuses after harvest on E19.5. (A) Saline injection, (B) PSHU-PNIPAAm injection with <50% coverage (n [ 3, mean: 11.5%, standard error [SE]: 7.1), (C) PSHU-PNIPAAm injection with >50% coverage (n [ 3, mean: 60.4%, SE: 1.9). Overall percentage coverage: n [ 6, mean: 36%, SE: 11.4. (Color version of figure is available online.)

with the saline (A) or RTG (B) both demonstrated normal skin, muscle, and spinal cord morphology. In fetuses with an MMC defect, both the saline injected (C) and RTG injected (D) fetal

tissue demonstrated significantly damaged skin, muscle, and spinal cord tissue. An almost complete loss of spinal cord tissue is seen in each Grhl3Cre/Cre fetus at the defect site.

Fig. 4 e Immunohistochemistry and histologic morphology: IHC to assess the immune response using F4/80 and CD68 macrophage markers and H&E to assess morphology of RTG and saline-injected fetuses with defects (C and D) or without defects from the Grhl3 matings (A and B), harvested at E19.5. Images are taken of the caudal region at the level of the defect or comparable region in nondefect fetus where the injection occurred. IHC white scale bar [ 200 mm. H&E black scale bar [ 500 mm. (Color version of figure is available online.)

b a r d i l l e t a l  r e v e r s e t h e r m a l g e l f o r s p i n a b i fi d a c l o s u r e

IHC: cellular inflammatory response CD68 and F4/80 are both commonly used macrophage markers to assess the inflammatory response of applied biomaterials.37-40 The RTG injection into nondefect mouse fetuses shows a significant increase in F4/80-positive cells (mean: 55.1 cells [SE ¼ 10.9], P < 0.001) when compared with saline injected nondefect mouse fetuses (mean: 31.2 [SE ¼ 7.7]). CD68-positive cells are also significantly increased in RTG-injected nondefect mouse fetuses (mean: 36.4 [SE ¼ 4.8], P < 0.001) when compared with saline-injected nondefect mouse fetuses (mean: 3.9 [SE ¼ 0.68]). However, RTG injections into defect Grhl3 mouse fetuses demonstrate a nonsignificant increase in F4/80 macrophages (mean: 128.8 [SE ¼ 23.4], P ¼ 0.16) compared with the saline-injected Grhl3-defect mouse fetuses (mean: 103.5 [SE ¼ 17.9]). CD68-positive cells also demonstrate a nonsignificant increase in RTG-injected Grhl3-defect mouse fetuses (mean: 54.8 [SE ¼ 6.8], P ¼ 0.13) when compared with saline-injected Grhl3-defect mouse fetuses (mean: 41.8 [SE ¼ 11.6]; Figs. 4 and 5).

Discussion RTG demonstrates in vitro and in vivo characteristics that are conducive to biological application as an alternative approach to patch MMC in utero. RTG is stable in amniotic fluid up to 6 mo in-vitro with no evidence of degradation. Using simple needle injection, the RTG can partially patch a mouse open neural tube defect with minimal macrophage response. Open prenatal repair of MMC has become the standard of care. However, the size and nature of the hysterotomy have resulted in significant maternal and fetal concerns. Therefore, a minimally invasive or the use of alternative approaches is warranted. Fetoscopic multiport repair of MMC has undergone

233

tremendous progress since it was first reported by Bruner et al. in 2000.41 Initially, the results were disappointing but eventually demonstrated a remarkable improvement with the introduction of partial CO2 insufflation by Kohl et al. in 2010.42 Subsequently, this technique has been used by researchers performing both experimental and human studies. However, fetoscopic multiport techniques have been associated with premature rupture of membrane, premature delivery, and extended operating times with a concurrent risk of hypercarbia, acidosis, and neurotoxicity potentially affecting fetal neurodevelopment. Therefore, there is a persistent need to develop novel approaches to overcome these concerns. Biomaterial patching approaches for MMC closure have demonstrated promising results in rodent and sheep experimental studies because of their tunable properties, including the ability to act as cellular scaffolds and growth factor delivery devices.19,20,43 A particular patch derived from human umbilical cord promoted native cellular migration with minimal inflammatory response.44 Other investigators have focused on the use of stem cells as an approach to facilitate/expedite skin closure. Intra-amniotic injections of amniotic or placental fluid derived mesenchymal stem cells induced partial or complete coverage of MMC defects and a reduction in Chiari II formation in a retinoic acideinduced rat model of MMC.24,45-47 These promising results demonstrate the potential of tissue-engineered approaches compared with traditional open repair approaches for MMC repair. The use of an injectable RTG (PSHU-PNIPAAm) to patch MMC has key advantages including minimally invasive application using a small gauge needle, short operative times, and the possibility of using the RTG scaffolding properties for potential cellular transplantation or growth factor delivery approaches.

Fig. 5 e F4/80 and CD68 cellular inflammatory response: RTG injection of nondefect mouse embryos shows a significant increase in F4/80-positive cells (average: 55.1 [SE: 10.9], *P < 0.001) when compared with saline-injected nondefect mouse embryos (average: 31.2 [SE: 7.7]), as well as a significant increase in CD68-positive cells in RTG injected nondefect mouse embryos (average: 36.4 [SE: 4.8], *P < 0.001) when compared with saline control (average: 3.9 [SE: 0.68]). RTG injections of MMC-defect Grhl3 mouse embryos demonstrate a nonsignificant increase in F4/80 macrophages (average: 128.8 [SE: 23.4], P [ 0.16) compared with the saline control (average: 103.5 [SE: 17.9]). CD68-positive cells also demonstrate a nonsignificant increase in RTG-injected Grhl3-defect mouse embryos (average: 54.8 [SE: 6.8], P [ 0.13) when compared with saline-injected Grhl3-defect mouse embryos (average: 41.8 [SE: 11.6]).

234

j o u r n a l o f s u r g i c a l r e s e a r c h  m a r c h 2 0 1 9 ( 2 3 5 ) 2 2 7 e2 3 6

For this in-utero application, the RTG patch must remain stable in the amniotic environment for the duration of pregnancy. Using GPC, the PSHU backbone of the RTG demonstrates little to no change in polymer chain length in an in-vitro amniotic environment after 6 mo. The observed minor differences are not considered a sign of degradation of the polymer because significant degradation of PSHU would be manifested as an Mn value decrease in the order of thousands. The solution-gel transition temperature (LSCT) of the RTG remains unchanged after 6 mo of exposure in an in-vitro amniotic environment. Degradation of PSHU-PNIPAAm would cause the LCST to shift to the right. If this shift occurs above 37 C, it would be an indication of gel instability on exposure to physiological temperatures. For this application, the rapid transition from solution to stable gel is necessary for a stable patch to form on injection into the intrauterine environment. These results demonstrate long-term in-vitro stability in an amniotic environment of the polymer backbone molecular weight and the maintenance of an LCST below 37 C. The Grhl3 mouse model generated fetuses with MMC defects that were then used to demonstrate the minimally invasive application of the RTG into an in-utero environment. We selected the Grhl3 mouse model over the retinoic acid induced rat model that is commonly used in preliminary MMC repair studies.48 The Grhl3 is one of the most common mouse models used to study MMC, has the ability to generate large numbers of pregnant mice, and has been well established and made available at our institution with the assistance of Dr Niswander’s expertise. Furthermore, our mouse studies were planned as proof of technical success and principle before studying the RTG approach in the large ovine model that is more representative of the human gestation. RTG injections into Grhl3 fetuses with MMC defects demonstrate the minimally invasive application of a novel MMC patch using a small gauge needle without the use of sutures therefore reducing operative time. Partial coverage of the MMC defect by the RTG was observed at harvest. However, the large variability of RTG defect coverage observed in this work may be attributed to several reasons including (1) the dynamic nature of the phenotypic development of the Grhl3 mouse defect, (2) the associated ectodermal defect because of loss of Grhl3 function, or (3) lack of physical adhesive properties of the RTG. These results demonstrate that while the RTG can be applied quickly in a minimally invasive manner, there are issues with long-term RTG localization to the defect area in this Grhl3 mouse model. Although the mouse model has limited gestation duration, previous studies have used rodent models to begin testing new biomaterials as MMC patches that were eventually taken to large ovine animal models.19,21 The Grhl3 mouse model used in this work has provided promising in-utero preliminary results for this novel MMC patching approach. However, the long-term effects of the RTG on issues such as spinal tethering, cerebrospinal fluid leakage, and hindbrain herniation will need to be studied in large ovine surgical neural tube defect model. The developing spinal cord becomes progressively damaged in patients with MMC in part by chemical and mechanical inflammation caused by open exposure to the surrounding amniotic environment. Therefore, minimizing inflammation in prenatal MMC patches should be considered

when using alternative approaches.49 RTG injections into Grhl3-defect mice showed no significant difference in macrophage response compared with saline-injected Grhl3-defect mice. These findings demonstrate that the RTG will induce a cellular inflammatory response, but the nonsignificant F4/80 and CD68 macrophage responses in defect fetuses after RTG and saline injections may be an indication of the inherent inflammation associated with the Grhl3 defect itself. However, the RTG is a foreign material that does cause a significant F4/ 80 and CD68 macrophage response in nondefect Grhl3 mouse fetuses when compared with saline. Macrophages play a critical role in the formation of granulation tissues in wound healing. Despite the tissue being harvested after only 3 d, the presence of macrophages within the Grhl3 defect at the RTG injection site could contribute to the formation of granulation tissue. Over time, the formation of granulation and epithelial tissues around and within the RTG could lead to complete coverage of the defect, with the goal of creating a water-tight tissue barrier to protect the spinal cord from the amniotic environment. However, longer study durations using an ovine animal model of MMC are needed to fully elucidate the effect of the RTG on granulation tissue formation. This study has the following limitations: we observed only partial MMC coverage of the Grhl3 mouse defect at the time of harvest that may be because of the dynamic phenotypic nature of this model. Also, we were not able to assess long-term RTG coverage in-utero because of the short gestational application in the mouse model. Moreover, it is possible that an ectodermal defect in Grhl3 mutants affects the ability of the RTG to remain in place. In addition, as seen in Figure 5, the RTG is not observed over the defect area after processing for H&E or IHC. The RTG is water soluble and easily washed away during the histologic processing steps. Because of these limitations, we are currently exploring methods to chemically modify the PSHU polymer backbone for enhanced adhesive properties to improve defect coverage. Previous work has demonstrated this RTG can be modified for multifunctional capabilities in other tissue engineering applications, both invitro and in-vivo. The PSHU polymer backbone can be chemically modified for different functional applications, such as RGD peptide conjugation to mimic the extracellular matrix, and sulfonation to mimic heparin function for controlled drug/growth factor delivery.50-55 Modification of the RTG for growth factor delivery will be particularly useful for MMC patching materials, as incorporation of growth factors such as basic fibroblast growth factor into other MMC patches has been found to enhance the formation of granulation and epithelial tissue that leads to complete MMC defect coverage in experimental animal models.43

Conclusion In this study, we demonstrate the in-utero applicability of the RTG, PSHU-PNIPAAm, as a minimally invasive alternative to patch MMCs in a Grhl3 mouse model. The development of a needle-based approach to patch MMC has the potential to shorten operating time, reduce invasive surgical procedures, and offer earlier gestational application compared with current open surgical repair. Overall, the RTG, PSHU-PNIPAAm, is

b a r d i l l e t a l  r e v e r s e t h e r m a l g e l f o r s p i n a b i fi d a c l o s u r e

a promising candidate for a minimally invasive, bioengineered approach to patch MMC in-utero.

Acknowledgment Authors’ contributions: All authors contributed to this article. J.B. conducted all aspects of the study, including polymer synthesis and characterization, degradation studies, animal experiments, histology, and data analysis. S.W. conducted animal experiments. U.S. aided in data analysis. L.N. provided the Grhl3 mouse model and provided critical review of the article. D.P. created the RTG used in this study and provided polymer synthesis and characterization guidance. A.M. conceived the use of an RTG for this application, designed experiments, provided direction and guidance for all aspects of this study, and participated in critical review of the article. The authors would like to thank the University of Colorado School of Medicine Morphology and Phenotyping Core for their assistance with histologic staining. This work was supported by the Nathaniel and Gabriel Stowell AwardeFetal Health Foundation, University of Colorado-Department of Surgery Academic Enrichment Fund and Dr Park’s startup funding.

Disclosure The authors report no proprietary or commercial interest in any product mentioned or concept discussed in this article.

Supplementary data Supplementary data to this article can be found online at https://doi.org/10.1016/j.jss.2018.09.078.

references

1. Mitchell LE. Epidemiology of neural tube defects. Am J Med Genet C Semin Med Genet. 2005;135C:88e94. 2. Dolk H, Loane M, Garne E. The prevalence of congenital anomalies in Europe. In: Posada de la Paz M, Groft SC, eds. Rare Diseases Epidemiology. Dordrecht: Springer Netherlands; 2010:349e364. 3. Horn F, Martinka I, Funakova M, et al. [Epidemiology of neural tube defects]. Rozhl Chir. 2011;90:259e263. 4. Seidahmed MZ, Abdelbasit OB, Shaheed MM, et al. Epidemiology of neural tube defects. Saudi Med J. 2014;35(Suppl 1):S29eS35. 5. Greene ND, Stanier P, Copp AJ. Genetics of human neural tube defects. Hum Mol Genet. 2009;18:R113eR129. 6. Copp AJ, Greene ND. Neural tube defects–disorders of neurulation and related embryonic processes. Wiley Interdiscip Rev Dev Biol. 2013;2:213e227. 7. Copp AJ, Greene ND. Genetics and development of neural tube defects. J Pathol. 2010;220:217e230. 8. Meuli M, Meuli-Simmen C, Yingling CD, et al. In utero repair of experimental myelomeningocele saves neurological function at birth. J Pediatr Surg. 1996;31:397e402.

235

9. Greene ND, Copp AJ. Neural tube defects. Annu Rev Neurosci. 2014;37:221e242. 10. Copp AJ, Adzick NS, Chitty LS, Fletcher JM, Holmbeck GN, Shaw GM. Spina bifida. Nat Rev Dis Primers. 2015;1:1500. 11. Detrait ER, George TM, Etchevers HC, Gilbert JR, Vekemans M, Speer MC. Human neural tube defects: developmental biology, epidemiology, and genetics. Neurotoxicol Teratol. 2005;27:515e524. 12. Adzick NS, Thom EA, Spong CY, et al. A randomized trial of prenatal versus postnatal repair of myelomeningocele. N Engl J Med. 2011;364:993e1004. 13. Marwan AI, Chatterjee D, Howley LW, Crombleholme TM. 118 - open fetal surgery A2 - Copel, Joshua A. In: D’Alton ME, Feltovich H, Grataco´s E, et al., eds. Obstetric Imaging: Fetal Diagnosis and Care. 2nd ed. Philadelphia, PA: Elsevier; 2018:499e505.e491. 14. Adzick NS. Fetal surgery for myelomeningocele: trials and tribulations. J Pediatr Surg. 2012;47:273e281. 15. Peranteau WH, Adzick NS. Prenatal surgery for myelomeningocele. Curr Opin Obstet Gynecol. 2016;28:111e118. 16. Watanabe M, Kim AG, Flake AW. Tissue engineering strategies for fetal myelomeningocele repair in animal models. Fetal Diagn Ther. 2015;37:197e205. 17. Verbeek RJ, Heep A, Maurits NM, et al. Fetal endoscopic myelomeningocele closure preserves segmental neurological function. Dev Med Child Neurol. 2012;54:15e22. 18. Fontecha CG, Peiro JL, Aguirre M, et al. Inert patch with bioadhesive for gentle foetal surgery of myelomeningocele in a sheep model. Eur J Obstet Gynecol Reprod Biol. 2009;146:174e179. 19. Watanabe M, Jo J, Radu A, Kaneko M, Tabata Y, Flake AW. A tissue engineering approach for prenatal closure of myelomeningocele with gelatin sponges incorporating basic fibroblast growth factor. Tissue Eng A. 2010;16(5):1645e1655. 20. Saadai P, Nout YS, Encinas J, et al. Prenatal repair of myelomeningocele with aligned nanofibrous scaffoldsda pilot study in sheep. J Pediatr Surg. 2011;46:2279e2283. 21. Watanabe M, Li H, Roybal J, et al. A tissue engineering approach for prenatal closure of myelomeningocele: comparison of gelatin sponge and microsphere scaffolds and bioactive protein coatings. Tissue Eng Part A. 2011;17(7-8):1099e1110. 22. Guilbaud L, Roux N, Friszer S, et al. Fetoscopic patch coverage of experimental myelomeningocele using a two-port access in fetal sheep. Childs Nerv Syst. 2017;33:1177e1184. 23. Kajiwara K, Tanemoto T, Wada S, et al. Fetal therapy model of myelomeningocele with three-dimensional skin using amniotic fluid cell-derived induced pluripotent stem cells. Stem Cel Rep. 2017;8:1701e1713. 24. Kabagambe S, Keller B, Becker J, et al. Placental mesenchymal stromal cells seeded on clinical grade extracellular matrix improve ambulation in ovine myelomeningocele. J Pediatr Surg. 2018;53:178e182. 25. Brambrink AM, Evers AS, Avidan MS, et al. Isoflurane-induced neuroapoptosis in the neonatal rhesus macaque brain. Anesthesiology. 2010;112:834e841. 26. Jevtovic-Todorovic V. General anesthetics and neurotoxicity: how much do we know? Anesthesiol Clin. 2016;34:439e451. 27. Joyeux L, Engels AC, Russo FM, et al. Fetoscopic versus open repair for spina bifida aperta: a systematic review of outcomes. Fetal Diagn Ther. 2016;39:161e171. 28. Arens C, Koch C, Veit M, et al. Anesthetic management for percutaneous minimally invasive fetoscopic surgery of spina bifida aperta: a retrospective, descriptive report of clinical experience. Anesth Analg. 2017;125:219e222. 29. Graves CE, Harrison MR, Padilla BE. Minimally invasive fetal surgery. Clin Perinatol. 2017;44:729e751. 30. Marwan AI, Williams SM, Bardill JR, Gralla J, Abdul-Aziz NM, Park D. Reverse thermal gel for in utero coverage of spina

236

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

j o u r n a l o f s u r g i c a l r e s e a r c h  m a r c h 2 0 1 9 ( 2 3 5 ) 2 2 7 e2 3 6

bifida defects: an innovative bioengineering alternative to open fetal repair. Macromol Biosci. 2017;17:1e10. Yun D, Famili A, Lee YM, Jenkins PM, Freed CR, Park D. Biomimetic poly(serinol hexamethylene urea) for promotion of neurite outgrowth and guidance. J Biomater Sci Polym Ed. 2014;25:354e369. Camerer E, Barker A, Duong DN, et al. Local protease signaling contributes to neural tube closure in the mouse embryo. Dev Cell. 2010;18:25e38. Rifat Y, Parekh V, Wilanowski T, et al. Regional neural tube closure defined by the Grainy head-like transcription factors. Dev Biol. 2010;345:237e245. Auden A, Caddy J, Wilanowski T, Ting SB, Cunningham JM, Jane SM. Spatial and temporal expression of the Grainy headlike transcription factor family during murine development. Gene Expr Patterns. 2006;6:964e970. Pyrgaki C, Liu A, Niswander L. Grainy head-like 2 regulates neural tube closure and adhesion molecule expression during neural fold fusion. Dev Biol. 2011;353:38e49. De Castro SCP, Hirst CS, Savery D, et al. Neural tube closure depends on expression of Grainy head-like 3 in multiple tissues. Dev Biol. 2018;435:130e137. Kunisch E, Fuhrmann R, Roth A, Winter R, Lungershausen W, Kinne R. Macrophage specificity of three anti-CD68 monoclonal antibodies (KP1, EBM11, and PGM1) widely used for immunohistochemistry and flow cytometry. Ann Rheum Dis. 2004;63:774e784. Lee S-H, Starkey PM, Gordon S. Quantitative analysis of total macrophage content in adult mouse tissues. Immunochemical studies with monoclonal antibody F4/80. J Exp Med. 1985;161:475e489. Brochhausen C, Schmitt VH, Mamilos A, et al. Expression of CD68 positive macrophages in the use of different barrier materials to prevent peritoneal adhesionsdan animal study. J Mater Sci Mater Med. 2017;28:15. Spiller KL, Anfang RR, Spiller KJ, et al. The role of macrophage phenotype in vascularization of tissue engineering scaffolds. Biomaterials. 2014;35:4477e4488. Bruner JP, Tulipan NB, Richards WO, Walsh WF, Boehm FH, Vrabcak EK. In utero repair of myelomeningocele: a comparison of endoscopy and hysterotomy. Fetal Diagn Ther. 2000;15:83e88. Kohl T, Ziemann M, Weinbach J, Tchatcheva K, Gembruch U, Hasselblatt M. Partial amniotic carbon dioxide insufflation during minimally invasive fetoscopic interventions seems safe for the fetal brain in sheep. J Laparoendosc Adv Surg Tech. 2010;20:651e653. Watanabe M, Li H, Kim AG, et al. Complete tissue coverage achieved by scaffold-based tissue engineering in the fetal

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

sheep model of Myelomeningocele. Biomaterials. 2016;76:133e143. Snowise S, Mann L, Morales Y, et al. Cryopreserved human umbilical cord versus biocellulose film for prenatal spina bifida repair in a physiologic rat model. Prenat Diagn. 2017;37:473e481. Wang A, Brown EG, Lankford L, et al. Placental mesenchymal stromal cells rescue ambulation in ovine myelomeningocele. Stem Cell Transl Med. 2015;4:659e669. Dionigi B, Ahmed A, Brazzo III J, Connors JP, Zurakowski D, Fauza DO. Partial or complete coverage of experimental spina bifida by simple intra-amniotic injection of concentrated amniotic mesenchymal stem cells. J Pediatr Surg. 2015;50:69e73. Dionigi B, Brazzo III JA, Ahmed A, et al. Trans-amniotic stem cell therapy (TRASCET) minimizes Chiari-II malformation in experimental spina bifida. J Pediatr Surg. 2015;50:1037e1041. Danzer E, Schwarz U, Wehrli S, Radu A, Adzick N, Flake A. Retinoic acid induced myelomeningocele in fetal rats: characterization by histopathological analysis and magnetic resonance imaging. Exp Neurol. 2005;194:467e475. Marton E, Giordan E, Gioffre` G, et al. Homologous cryopreserved amniotic membrane in the repair of myelomeningocele: preliminary experience. Acta Neurochir (Wien). 2018;160:1625e1631. Pena B, Shandas R, Park D. A heparin-mimicking reverse thermal gel for controlled delivery of positively charged proteins. J Biomed Mater Res A. 2015;103:2102e2108. Yun D, Lee YM, Laughter MR, Freed CR, Park D. Substantial differentiation of human neural stem cells into motor neurons on a biomimetic polyurea. Macromol Biosci. 2015;15:1206e1211. Laughter MR, Ammar DA, Bardill JR, et al. A self-assembling injectable biomimetic microenvironment encourages retinal ganglion cell axon extension in vitro. ACS Appl Mater Inter. 2016;8:20540e20548. Lee DJ, Fontaine A, Meng X, Park D. Biomimetic nerve guidance conduit containing intraluminal microchannels with aligned nanofibers markedly facilitates in nerve regeneration. ACS Biomater Sci Eng. 2016;2:1403e1410. Penea B, Martinelli V, Jeong M, et al. Biomimetic polymers for cardiac tissue engineering. Biomacromolecules. 2016;17:1593e1601. Laughter MR, Bardill JR, Ammar DA, Pen˜a B, Calkins D, Park D. An injectable neurotrophic factor delivery system supports retinal ganglion cell survival and regeneration following optic nerve crush. ACS Biomater Sci Eng. 2018;4:3374e3383.