Altered Amniotic Fluid Levels of Hyaluronic Acid in Fetal Rats with Myelomeningocele: Understanding Spinal Cord Injury

Abstract

Myelomeningocele (MMC) is a devastating congenital neural tube defect that results in the exposure of spinal cord to the intrauterine environment, leading to secondary spinal cord injury and severe impairment. Although the mechanisms underlying the secondary pathogenesis are clinically relevant, the exact cause of in utero–acquired spinal cord damage remains unclear. The objective of this study was to determine whether the hyaluronic acid (HA) concentration in amniotic fluid (AF) in the retinoic acid-induced model of MMC is different from that in normal controls and whether these differences could have an impact on the viscosity of AF. Our data shows that the concentration of HA in AF samples from fetuses with MMC (MMC-AF) and normal control samples (Norm-AF) were not significantly different at embryonic day 18 (E18) and E20. Thereafter, the HA concentration significantly increased in Norm-AF but not in MMC-AF. Compared with Norm-AF, the concentration of HA in MMC-AF and the viscosity of MMC-AF were significantly lower at E21. Agarose gel electrophoresis confirmed a significant reduction in the HA level of MMC-AF compared with Norm-AF at E21. No HA-degrading activity was detected in MMC-AF. In summary, we identified a deficiency in the AF level of HA and the viscosity of AF in fetal rats with MMC. These data are discussed in relation to a potential role the reduction in the AF viscosity due to the low level of HA may play in the exacerbating effects of mechanical trauma on spinal cord damage at the MMC lesion site.

Introduction

Myelomeningocele (MMC), the most common and severe form of spina bifida, is a devastating congenital defect leading to severe impairment.^1,2 It is characterized by the protrusion of meninges and spinal cord through a defect in the vertebral arch and an open wound in the skin.³ MMC results in significant and life-long physical disabilities including leg paralysis, sensory loss, bowel and bladder dysfunctions, skeletal deformations, and Arnold-Chiari II malformation with secondary hydrocephalus.^4–6 Children affected by MMC often require life-long therapeutic interventions and medical support. Treatment and management of patients with these defects continues to have a huge economic burden on the health care system.^7,8

Currently, the initial treatment for MMC consists of surgical closure of the defect soon after birth, or in select cases, before birth. The early post-natal treatment is limited to preventing infections by surgical coverage of the exposed spinal cord, but these children usually require life-long support, rehabilitation, and institutional care.^9–11 Since in MMC the exposed spinal cord tissue is progressively damaged leading to irreversible loss of neurological function at birth, closing the defect in utero might prevent continuing damage and improve the neurological outcome.^12–16 To minimize in utero damage to the exposed spinal cord, prenatal surgical closure of the MMC defect has been undertaken and validated in select cases of MMC. Prenatal surgical closure is more successful in restoring neurological function and reducing the rate of ventriculoperitoneal shunt placement than post-natal closure; however, the surgical procedure can only be performed in a fraction of patients and restoration of neurological function is limited in many children.^17,18 Recently, tissue engineering approaches have emerged in experimental studies as potential therapeutic strategies for prenatal closure of MMC lesions.^19–26 However, the pathogenesis of MMC is not well understood and MMC still poses a diagnostic and therapeutic challenge.

Although MMC is primarily caused by the failure of the neural tube closure early during gestation, multiple studies demonstrate that the secondary damage to the openly exposed spinal cord acquired during the later phase is mainly responsible for the loss of neurological function in fetuses with MMC.^{12,14–16,27–29} These studies also indicate that the secondary spinal cord damage in MMC results from the exposure of unprotected spinal cord to the hostile intrauterine environment^12–16; however, the exact cause of this secondary pathogenesis remains unclear. Therefore, further research is needed to better understand the role of the intrauterine environment in development of MMC by identifying and investigating the amniotic fluid (AF) factors that trigger and/or contribute to the progression of in utero acquired spinal cord damage.

Hyaluronic acid (HA), a glycosaminoglycan with remarkable viscoelastic, lubricating, and regenerative properties is an important component of AF, which surrounds and protects the fetus during growth.^30–38 In joint synovial fluid and artificial tears, HA is well known to increase the viscosity and provides lubrication to prevent cartilage deterioration and corneal abrasion.^{35,36,39–41} Using the retinoic-acid induced rat model of MMC, we previously identified clusters of neural cells within the AF of MMC fetuses, which are most abundant at later gestational stages, when injury of the openly exposed spinal cord is markedly increased.⁴² Thus, it is plausible that reduced viscosity of AF due to lower AF level of HA in MMC may exacerbate damage to the spinal cord, causing shedding of cells from the openly exposed spinal cord into the surrounding AF, particularly later in gestation when the fetal size and movement increases. Therefore, the present study has focused on an attempt to address the question whether the AF concentrations of HA in MMC are different from those in age-matched normal controls, and if so, whether these differences could have an impact on the viscosity of AF. For this study, MMC was established using the retinoic acid-induced rat model, which is developmentally and anatomically analogous to human MMC and provides an excellent translational model for studying MMC in rats.^43,44

Methods

Retinoic acid-induced animal model of MMC

All experiments were conducted under the guidelines of Temple University's Institutional Animal Care and Use Committee and the National Institutes of Health Guide for Care and Use of Laboratory Animals. Time-dated Sprague-Dawley pregnant rats obtained from Charles River Laboratories (Wilmington, MA) were placed on a standard dark:light schedule. An MMC defect was induced in the fetuses of time-dated pregnant rats by gavage of a single dose of 50 mg/kg of all trans-retinoic acid (Sigma-Aldrich Chemical, Saint Louis, MO) dissolved in olive oil on embryonic day 10 (E10), as described previously.^42,45 Normal control mothers were gavage fed with olive oil. AF samples from MMC fetuses (n = 168) and normal fetuses (n = 122) were collected at E18, E20, and E21 (term = E22). Dams were euthanized using chamber inhaled CO₂, and the total number of fetuses were determined following the midline laparotomy and exposure of the uterus. AF samples were collected using aseptic techniques into 1.5 mL Eppendorf tubes. After collection of AF, the uterus and gestational membranes were removed and each fetus was examined for the presence of lumbosacral MMC defect. The incidence of isolated MMC defects was observed in 79.2% (168/212) of fetuses. After harvesting, fetuses were collected and euthanized by decapitation according to standard procedures.

Measurement of HA concentration in AF

Amniotic fluid HA concentrations were analyzed using a competitive Hyaluronan Enzyme Linked Immunosorbent Assay (HA ELISA; K-1200; Echelon Biosciences Inc, Salt Lake City, UT) according to the manufacturer instructions.⁴⁶ All AF samples collected from fetuses at different embryonic ages (E18, E20, and E21) were diluted 100 × in provided diluent and analyzed in duplicate. Absorbance of each well was measured at 405 nm after 30 min of incubation with working substrate solution (K-1208). When high variation in duplicate reading was observed, samples were re-assayed in triplicate. A total of 18 randomly selected AF samples isolated from E18, 18 AF samples from E20, and 92 AF samples from E21 MMC fetuses as well as 18 AF samples isolated from E18, 18 AF samples from E20, and 66 AF samples from E21, normal controls were analyzed for the amniotic fluid HA levels. The HA concentration was normalized to the volume of amniotic fluid.

Agarose gel electrophoresis in TAE buffer

Electrophoresis of HA on agarose gels in Tris–acetate–EDTA (TAE) buffer (40 mM Tris, 5 mM acetate [CH₃COONa], and 0.9 mM EDTA, pH 7.9) was performed as previously described.^47,48 In brief, agarose gels were prepared by dissolving 1.5 g of agarose in 300 mL of TAE buffer for final concentrations of 0.5%. Dissolved and equilibrated agarose solution was poured into a gel casting tray to form a 6.5-mm thick gel. For electrophoresis, the gel was transferred to a Kodak BioMaxMP1015 electrophoresis unit filled with TAE buffer. Samples of amniotic fluid were prepared by adding to 15 μL of amniotic fluid, 35 μL of dd water and 10 μL of loading buffer (0.02% bromophenol blue and 2 M sucrose in 1XTAE buffer) to load 1.5-2.5 μg of HA. To determine the specificity of HA, 15 μL of each amniotic fluid sample was digested with Streptomyces hyaluronidase (Sigma-Aldrich Chemical, St. Louis, MO) as described in section “Enzymatic assay of hyaluronidase.” HA standards (MEGA Ladder and High Ladder; HYALOSE, Oklahoma City, OK) were prepared in water at concentrations of 0.04–0.1 μg/μL. HA standards, analyzed AF samples, high molecular weight (HMW) HA (1,630 kDa, Sigma-Aldrich) and AF samples after treatment with hyaluronidase were loaded onto gels and subjected to electrophoresis. Electrophoresis was carried out at room temperature at a constant voltage of 30 V for 30 min and then 50 V for 6h. The gels were stained with 0.005% StainsAll (Sigma Aldrich) in 50% ethanol overnight under light-protective cover on a shaker at room temperature. The gels were de-stained in water for 4 h and then scanned using Bio-RAD Gel Doc^™ EZ Imager (Bio-RAD laboratories, USA). Densitometric analysis was performed using ImageJ software.

Enzymatic assay of hyaluronidase

In order to examine the effect of MMC-AF on degradation of HA, 20 μg of commercially available HMW HA (Sigma-Aldrich) was added to 100 μL MMC-AF and incubated overnight at 37°C. Positive control samples were prepared by adding 20 μg of HMW HA (Sigma-Aldrich) to 100 μL MMC-AF and 80 U/mL of Streptomyces hyaluronidase (Sigma-Aldrich Chemical, Saint Louis, MO), which specifically degrades HA,⁴⁹ and incubated overnight at 37°C. After overnight incubation, hyaluronidase was denatured by boiling at 90°C for 10 min and concentration of HA in analyzed samples was measured using HA ELISA assay as described in section “Measurement of HA concentration in AF.” A total of six randomly selected MMC-AF samples were analyzed in duplicates. An aliquot of each reaction solution was also electrophoresed on agarose gel and stained with Stains-all as described in section “Agarose gel electrophoresis in TAE buffer.”

Determination of viscosity

The viscosity of amniotic fluid was determined using a Black Pearl Viscometer (ATS Rheo System). The viscometer is a cone and plate type with a diameter of 30 mm and an angle of 2°. The viscometer was set to a shear rate of 2000 (1/sec) to ensure sufficient torque for measurement. The viscometer allows measurement of a small volume of liquids. The device was calibrated to the known viscosity of pure water 0.001 Pa·sec. 500 μL of each AF sample was placed on thin sensor plate cone brought to gap height, and the viscosity was then immediately measured at 25°C. In order to examine the effect of HA on the viscosity of MMC-AF, 150 μg of commercially available HMW HA (Sigma-Aldrich) was added to 500 μL MMC-AF collected from E21 fetuses and the viscosity was measured as described above.

Protein concentration

Protein concentration in amniotic fluid was measured using colorimetric detection and quantitation Pierce^™ BCA Protein Assay Kit (Thermo Scientific, USA) according the manufacturer protocol. Absorbance was measured at 570 nm using BioTek ELx808 Absorbance Reader and analyzed using Gen 5.2 Plate Reader Software.

Statistical analysis

Comparison between the groups after measuring concentration of HA in AF was performed using one-way analysis of variance with Tukey's multiple comparison test. For statistical analysis of viscosity and protein concentration Student's t-test was used. All data are presented as mean ± standard deviation (SD). Differences in p values of <0.05 were considered to be significant.

Results

Amniotic fluid levels of HA

MMC was induced in fetal rats by exposure of pregnant mothers to all-trans retinoic acid. This model is similar to human MMC and develops the entire spectrum of the disease including loss of skin and underlying soft tissues, defects in the spinal column and progressive spinal cord damage at the level of MMC lesion with advancing gestational age.^42,50 Analogous to previous studies, all MMC fetuses collected at E18, E20, and E21 presented with lumbosacral MMC defects (Fig. 1A-C), showing progressive severity of injury with age. To determine whether the HA concentration in MMC-AF is different from that in Norm-AF, we quantitatively assessed the HA concentration in MMC-AF and age-matched Norm-AF samples. As illustrated in Figure 1D, no significant difference in HA concentration was found between the Norm-AF and MMC-AF samples at E18 and E20, corresponding to 21.22 ± 4.37 μg/mL and 21.25 ± 4.09 μg/mL at E18, respectively, and 32.69 ± 6.25 μg/mL and 24.50 ± 4.31 μg/mL at E20, respectively. Thereafter, the HA concentration significantly (p < 0.001) increased in Norm-AF samples from 32.69 ± 6.25 μg/mL at E20 to 143.60 ± 52.83 μg/mL at E21. Although the HA concentration in MMC-AF increased from 24.50 ± 4.31 μg/mL at E20 to 43.64 ± 18.29 μg/mL at E21, the difference was not significant. Thus, a significantly lower HA concentration (p < 0.001) was found in MMC-AF compared with Norm-AF at E21. The data indicated that MMC-AF resulted in greater than a 3-fold lower concentration of HA in MMC-AF compared with Norm-AF at E21.

FIG. 1.

(A-C) Photomicrographs showing external views of retinoic acid–induced myelomeningocele (MMC) defect in fetal rats at embryonic Day 18 (E18), E20, and E21, respectively. (A) At E18, the MMC placode is exposed dorsally and represents the most superficial element of the fetus. (C) The MMC placode demonstrates damage and excessive flattening of dorsal tissue at E21. Arrows indicate the beginning of the exposed spinal cord. (D) Histogram illustrates hyaluronic acid (HA) concentrations in MMC-amniotic fluid (AF) and normal control (Norm)-AF samples at different stages during gestation. No significant differences in concentrations of HA were found between MMC-AF and Norm-AF samples at E18 and E20. The concentration of HA in MMC-AF was significantly lower compared with Norm-MMC-AF at E21. Data presented as mean ± standard deviation of 18 randomly selected MMC-AF and Norm-AF samples at E18 and E20, 66 Norm-AF samples, and 92 MMC-AF samples at E21. ***p < 0.001.

HA analysis

Next, we assessed the molecular weight (MW) distribution of HA in MMC-AF and Norm-AF samples by comparison with HA standards and commercially available HMW HA electrophoresed on agarose gels and stained with Stains-all. Figure 2A illustrates the MW distribution of HA from two representative MMC-AF samples and two representative Norm-AF samples collected at E21. The results showed that unlike the HA standards, which run as separate bands ranging in size from 6100 to 509 kD (Fig. 2A, lane: 1), MMC-AF and Norm-AF samples run as HA smears containing a range of sizes indicating the natural polydispersity of HA in AF (Fig. 2A, lanes 2, 3, 4, and 5) similar to commercially available HMW HA (Fig. 2A, lane 7). AF samples showed a wide MW distribution with a visibly lower intensity of HA smears between MMC-AF and Norm-AF, especially in the HMW range (Fig. 2A; compare lanes 2, 3 to 4, 5). To confirm that the smears on the agarose gel were HA specific, AF samples were digested with Streptomyces hyaluronidase prior to loading as a negative control (Fig. 2A, lanes 6 and 8). Densitometric analysis of respective HA smears showed significantly lower levels of HA in MMC-AF compared with Norm-AF samples (Fig. 2B). These results are consistent with the significantly lower HA concentration in MMC-AF compared with Norm-AF samples collected at E21 determined by ELISA assay shown in Figure 1D.

FIG. 2.

(A) Representative myelomeningocele (MMC)-amniotic fluid (AF) and normal control (Norm)-AF samples electrophoresed on agarose gel and stained with Stains-all show molecular weight (MW) distribution of hyaluronic acid (HA). Lane1, Mega HA ladder and Hi HA ladder; lanes 2 and 5 represent MMC-AF samples at E21; lanes 4 and 5 represent Norm-AF samples at E21; lane 6, Norm-AF sample digested with Streptomyces hyaluronidase (HYAL); lane 7, high MW HA sample generated by streptococcal bacteria (Sigma-Aldrich), and lane 8, high MW HA sample digested with Streptomyces hyaluronidase. Results are representative of three independent experiments. (B) Relative density of HA smears from MMC-AF and Norm-AF reflects the intensity of respective HA-bound dye. Densitometries were performed using ImageJ Software. The relative density is expressed as the percent of the total amount of HA-bound dye. Data presented as mean ± standard deviation of 12 randomly selected samples. ***p < 0.001.

Amniotic fluid analysis

One explanation that MMC-AF samples at E21 do not show normal HA concentration might be that HA released into the MMC-AF is rapidly degraded by hyaluronidases present in the MMC-AF that are not present in the Norm-AF. Therefore, the MMC-AF was assessed for HA degrading activity. MMC-AF was collected from fetuses at E21 and incubated with HMW HA used as a substrate or HMW HA and Streptomyces hyaluronidase used as a positive control. As demonstrated in Figure 3A, HA concentration in MMC-AF samples incubated with HMW HA remained at 256.39 ± 42.66 μg/mL after overnight incubation, whereas, HA concentration in MMC-AF samples incubated with HMW HA and Streptomyces hyaluronidase was significantly decreased to 28.69 ± 3.13 μg/mL. To further confirm that the MMC-AF does not degrade HA to a low molecular weight, an aliquot of each reaction solution was electrophoresed on agarose gel and stained with Stains-all. MMC-AF samples incubated with HMW HA showed a wide MW distribution of HA on agarose gel, similar to HA control in Figure 2, while the smear was not detected in control MMC-AF samples incubated with HMW HA and Streptomyces hyaluronidase (Fig. 3B; compare lane 3 to 4). Overall, these results indicate that the low concentration of HA in MMC-AF at E21 is not caused by its degradation.

FIG. 3.

(A) Histogram illustrates hyaluronic acid (HA) concentration in myelomeningocele (MMC)-amniotic fluid (AF) samples at embryonic Day 21 (E21) and after an overnight incubation of MMC-AF with 200 μg/mL of exogenous high molecular weight (HMW) HA used as a substrate or with 200 μg/mL of exogenous HMW HA and Streptomyces hyaluronidase (80 U/mL) used as a positive control. Data presented as mean ± standard deviation of three independent experiments involving six randomly selected MMC-AF samples. ***p < 0.001. (B) Representative analysis of HA staining intensity in MMC-AF samples incubated with HMW HA overnight and control samples electrophoresed on agarose gel and stained with Stains-all. Lane1, Mega HA ladder and Hi HA ladder; lane 2, MMC-AF; lane 3, MMC-AF after an overnight incubation with HMW HA; lane 4, MMC-AF after an overnight incubation with HMW HA and Streptomyces hyaluronidase (HYAL). Results are representative of three independent experiments.

Viscosity of amniotic fluid

The presence of HA in the aqueous environment such as synovial fluid is associated with synovial fluid viscosity.^35,51,52 Next, we measured the viscosity of MMC-AF and Norm-AF collected at E21. Figure 4A shows that the viscosity value for Norm-AF was 0.0023 ± 0.0002 Pa·s and was significantly higher (p < 0.001) than the viscosity value for MMC-AF, corresponding to 0.0018 ± 0.0003 Pa·s. Thus, in our study the lower concentration of HA in MMC AF was accompanied by a lower average viscosity of MMC-AF. In addition to water and HA, proteins represent a large fraction of AF components.³⁸ To determine whether the difference in the MMC-AF and Norm-AF viscosity may be associated with different protein concentrations, we next measured the total protein concentrations in MMC-AF and Norm-AF samples collected at E21. We found no significant difference in the total protein concentrations between the MMC-AF and Norm-AF samples. Figure 4B shows that the total protein concentration in MMC-AF from fetuses at E21 was 3.78 ± 0.67 μg/μL and in Norm-AF was 3.22 ± 0.39 μg/μL. Thus, a significantly lower concentration of HA in MMC-AF from E21 fetuses was accompanied by a significantly lower viscosity value, but not a significantly lower total protein concentration in MMC-AF compared with Norm-AF.

FIG. 4.

(A) Histogram illustrates significantly lower viscosity of myelomeningocele (MMC)-amniotic fluid (AF) compared with normal control samples (Norm)-AF at embryonic day 21 (E21). Data presented as mean ± standard deviation of 14 randomly selected MMC-AF and Norm-AF samples. ***p < 0.001. (B) No significant differences in the total AF protein concentration were found between MMC-AF and Norm-AF at E21. Data presented as mean ± standard deviation of 12 randomly selected MMC-AF and Norm-AF samples.

To assess whether the lower viscosity of MMC-AF was associated with reduced concentration of HA in MMC-AF compared with Norm-AF at E21, we supplemented MMC-AF collected from E21 rat fetuses with HMW HA and determined the correlative viscosity. In vitro supplementation of HA to increase the HA concentration in MMC-AF from 33.82 ± 6.98 μg/mL to 351.28 ± 148.20 μg/mL resulted in a statistically significant increase in the viscosity of MMC-AF from 0.0018 ± 0.0003 PA·s in MMC-AF to 0.0025 ± 0.0002 PA·s (Table 1). Overall, these results indicate the causal relationship between increased concentrations of HA and increased viscosity of AF.

Table 1.

Results of Viscosity and HA Concentration before and after in vitro Supplementation of MMC-AF with HMW HA

	E21 MMC-AF	E21 MMC-AF + HA
AF viscosity [Pa·s]	0.0018 ± 0.0003	0.0025 ± 0.002***
HA concentration [μg/mL]	33.82 ± 6.98	351.28 ± 148.20***

^***p < 0.001

Data presented as mean ± standard deviation of 20 randomly selected MMC-AF samples. HA, hyaluronic acid; MMC, myelomeningocele; AF, amniotic fluid; HMW, high molecular weight; E21, embryonic day 21.

Discussion

MMC is characterized by both a primary and secondary insult. The primary insult is caused by incomplete closure of the neural tube early during gestation; whereas, the secondary insult causes in utero damage to the exposed spinal cord at later gestational stages, leading to deterioration of neurological function in MMC fetuses.^{12,14,16,29,53–55} Experimental and clinical evidence indicate that secondary in utero damage results from the exposure of spinal cord to the hostile intrauterine environment.^{12,14–16,28,29,50,53} It was postulated that chemical and/or mechanical insults to the unprotected fetal spinal cord abnormally exposed to the AF and the uterine wall are involved in the secondary in utero damage to the spinal cord.^12–15 The presence of fetal urine was indicated as a possible factor for toxicity of AF.^12–15 In vitro analysis of the toxicity of AF to the spinal cord tissue was primarily based on an assay for lactate dehydrogenase efflux by organotypic rat spinal cord cultures treated with human AF; however, the histological studies failed to reveal any evidence of tissue damage.⁵⁶

In a separate study of fetal delayed splotch mice, McClone and colleagues⁵⁷ found no apparent evidence of toxicity of AF on exposed neural tissue. More recently, several studies including our own have identified a multitude of live astrocytes and neural stem cells within the AF of MMC fetuses, indicating little, if any, neurotoxicity of the AF.^42,58–60 In addition, studies using fetal lambs reported that MMC promoted stem cell phenotypes and proliferation of mesenchymal stem cells in the AF,⁶¹ suggesting that AF in the MMC provides a permissive environment for proliferation of stem progenitor cells. As mentioned above, multiple studies indicate that most of neural damage observed in MMC occurs later in gestation as a result of spinal cord exposure to the hostile intrauterine environment.^{12–16,29,50} However, how exactly the AF affects MMC progression remains an open question and the actual pathogenesis of the injury remains unclear.

In the present study, for the first time, we assessed HA concentration in AF samples from fetuses with retinoic acid induced MMC for comparison to AF samples from age-matched normal controls using ELISA. Quantitative comparisons of HA concentration in MMC-AF and Norm-AF samples showed similar mean HA concentration in Norm-AF and MMC-AF at E18. Compared with MMC-AF, the trend towards a higher HA level was observed in Norm-AF at E20 and the mean concentration of HA was significantly higher than that found in MMC-AF at E21. Agarose gel electrophoresis of MMC-AF and Norm-AF samples from E21 fetuses confirmed a significant difference in the intensity of HA smears between MMC-AF and Norm-AF controls. We also identified a significantly lower viscosity of MMC-AF compared with Norm-AF at E21.

HA is a naturally occurring polysaccharide, which has a wide range of physical and biological functions.^38–41 Some of these functions include maintenance of elastoviscosity of liquid connective tissues such as joint synovial and eye vitreous fluid, thereby providing a viscoelastic protection of sensitive tissues.^36,39–41 HA is well known to increase the viscosity of synovial fluid and provides joint lubrication to prevent cartilage deterioration.^35,36,52,62 There also is evidence linking high HA levels in AF with its high viscosity properties.⁶³ In our study, the lower concentration of HA in MMC-AF was accompanied by a significantly lower average viscosity compared with Norm-AF controls, but not a significantly lower total protein concentration in MMC-AF compared with Norm-AF. Further, we showed that supplementation of MMC-AF with HMW HA resulted in a statistically significant increase in the viscosity of MMC-AF. These findings are consistent with the hypothesis that reduced viscosity of MMC-AF caused by the deficiency of HA in MMC-AF may play a role in the pathological development of MMC.

Normally, AF acts as a cushion surrounding the fetus, allowing fetal movement and providing protection from the external trauma during fetal development.³⁸ It is possible, however, that the AF of MMC fetuses loses this capacity. Previously, we observed an accumulation of neural cell clusters within the AF of MMC fetuses during later gestational ages.⁴² The emergence of neural cell clusters within the AF of MMC fetuses and the in utero injury associated with dorsally displaced spinal cord is consistent with contribution of mechanical trauma to this phenomenon.^42,50 Our results show that these pathological changes coincide with decreased levels of HA and low viscosity within the AF of late gestational age rat MMC fetuses. Based on these observations, we provide a working model for the mechanical abrasion of spinal cord tissue in utero (schematized in Fig. 5).

FIG. 5.

Proposed mechanism contributing to in utero spinal cord damage and abrasion of cells from the exposed spinal cord in myelomeningocele (MMC) fetuses. The layer of amniotic fluid surrounding the fetus present in the amniotic sac provides protection from external trauma or pressure and allows the free movement as the fetus grows. In fetuses with MMC the exposed spinal cord is vulnerable. Reduced hyaluronic acid level and viscosity of amniotic fluid in MMC may abrogate the ability of amniotic fluid to exert its protective action against mechanical trauma causing disruption of the dorsal spinal cord protruding through the MMC defect and shearing of cells from the dorsal surface of the spinal cord into surrounding amniotic fluid. These phenomena might occur particularly late in gestation, when fetal seize and movement increases.

According to this model, reduced HA levels in MMC-AF cause inadequate viscoelastic protection of the spinal cord against mechanical trauma, particularly, late in gestation when the size of the fetus and the pressure against the uterine wall increases. Under such conditions, fetal movement, which increases late in gestation,⁶⁴ would cause mechanical disruption of the dorsal spinal cord protruding through a defect in the vertebral arch and skin, shearing cells from the dorsal surface of the spinal cord into the surrounding AF. Interestingly, our data indicates a significant reduction in HA level and the viscosity of AF at E21, when loss of spinal cord tissue and the number of neural cells found in the AF is markedly increased in this model of MMC.^42,50 This data further supports our hypothesis that inadequate lubrication at the wound interface due to the deficiency of HA in AF may contribute to the exacerbation of MMC lesion and deterioration of the spinal cord caused by mechanically-induced abrasion of cells from the exposed spinal cord into the AF where they accumulate.

In addition to HA providing protection, numerous studies also have shown that AF enriched in HA demonstrates growth promoting activities including enhancement of bone and cartilage regeneration^65–67 and plays a central role in enhancing the healing capacity of fetal wounds.^30–34 As gestation proceeds, MMC leads to progressive injury of the openly exposed spinal cord tissue and associated pathology highlighted by a significant loss of the dorsal cord and extensive astrocytosis associated with increased glial fibrillary acidic protein immunoreactivity at the lesion site.^15,50,68 Astrogliosis is a common reaction in the central nervous system in response to spinal cord injury.^69–71 Because the MMC defect represents a failure in the establishment of skin and vertebral arch closure at the MMC site, closing the MMC defect in utero could be beneficial to enhancing neurological outcome.^12–16 Multiple studies have attempted to close the MMC site using biomaterial patches to cover the MMC lesion or delivery of stem cells to accelerate the wound healing process, showing variable, but some improvement in outcome measures.^19–25 Given the role of HA in stimulation of tissue regeneration and the enhanced healing capacity of fetal wounds in the presence of high AF levels of HA,^{32,37,38,72,73} it is possible that deficiency in HA within the AF creates a hostile environment that not only contributes to spinal cord damage, but also hampers healing of the MMC lesion.

Consistent with this idea, supplementing AF with HA may be necessary to limit the secondary injury pathology in MMC fetuses by increasing the viscosity of AF to immediately reduce mechanical trauma of the exposed spinal cord, while stimulating regenerative healing of MMC lesion to cover the exposed spinal cord, thus further limiting spinal cord trauma and associated astrocytosis at the MMC lesion site. Previous studies have mostly focused on optimizing the closure of the MMC defect,^{20,20,25,74–77} while the mechanisms underlying the pathogenesis of the injury remained largely unclear. The study of AF properties is, therefore, not only essential to better understand the mechanisms underlying in utero spinal cord damage in MMC, but also the development of improved prenatal strategies for enhancing regeneration and closing the defect to protect the spinal cord from further damage.

The observed significant deficiency of HA in MMC-AF in fetal rats at E21 may be caused by a number of pathological factors active during this disease stage. One explanation for the low level of HA in MMC-AF might be degradation of HA released into the amniotic cavity by enzymes with hyaluronidase activity. In our studies, however, we observed high molecular weight forms of HA (6100 kDa-509 kDA) in Norm-AF and MMC-AF samples analyzed using agarose gel electrophoresis, indicating the absence of hyaluronidase activity in MMC-AF. This was confirmed by the lack of HA degradation when MMC-AF was incubated with HMW HA used as a substrate. Thus, the observation that AF levels of HA are lower in MMC fetuses at E21 than in age-matched normal controls, suggests the inhibition of HA release to MMC-AF in more mature fetuses. Studies in sheep have demonstrated that in more mature fetuses HA enters the AF from fetal lung secretions and fetal urine, which constitutes the major volume of AF as gestation proceeds.⁷³ Therefore, the lower level of HA in MMC-AF of more mature fetuses may be mediated to some extent through a lower urinary and / or lung contribution of HA.

Alternatively, the lower HA release to the extracellular environment of MMC-AF may be related to lower production of HA by MMC-AF cells and possibly other non-embryonic sources. Future studies of expression levels of major HA synthases in fetal and amniotic tissues and MMC-AF cells should help to elucidate whether decreased synthesis can contribute to the observed lower level of HA in MMC-AF. The etiology of neural tube defects is complex and in most MMC cases multi-factorial, involving genetic, teratogenic, and nutritional factors.^78–80 Curly tail mouse embryos developing neural tube defects have been shown to exhibit reduced amounts of HA in the posterior neuropore region during the time of neural tube closure suggesting its possible role in the etiology of these defects.⁸¹ HA is an important structural component of extracellular matrix⁸² and the retinoic acid induced MMC in fetal rats, which is anatomically analogous to human MMC^23,43,50 may offer an experimental model for investigating the role of extracellular matrix imbalance in development and treatment of MMC.

Summary

These studies addressed previously unexplored aspects of HA levels of AF in a rat model of MMC. We showed that the content of HA in AF was significantly lower in MMC than control fetuses during later gestational stage. Further, we demonstrated that the lower level of HA in MMC-AF is accompanied by a lower viscosity of MMC-AF. This data shows alterations of the AF environment in rat MMC fetuses. Currently, effective in utero treatment of MMC is limited by our lack of understanding the mechanisms underlying its pathogenesis. In particular, the cause of spinal cord injury that is acquired in utero has not been clearly defined. The newly proposed role for HA, and potentially other components of extracellular matrix, in pathological development of MMC opens future investigations to deepen our understanding of mechanisms underlying MMC pathogenesis and ultimately creating therapies for MMC repair directed at rebuilding the fetal environment.

Author Disclosure Statement

No competing financial interests exist.