The Controversial Role of Folic Acid on Diabetic Auditory Neuropathy

Purpose Diabetic auditory neuropathy(DAN) is a common complication of diabetes that seriously affects the quality of life in patients. In this study, we investigate the role of folic acid in the treatment of DAN in an experimental rat model. Methods Thirty-two Sprague-Dawley rats were equally divided into 4 groups: group 1, normal; group 2, diabetic rats; group 3and 4, rats treated with folic acid (40 and 80 mg/kg, respectively). The tools we used in this study to investigate the effect of folic acid on DAN were auditory brain stem response, stereology methodfor estimation ofthe volume and number ofspiral ganglion,volume of stria vascularis, and spiral ligament, and measurement of homocysteine (HCY), malondealdehyde(MDA) and superoxide dismutase. Results Our study showed that folic acid treatment was not signicantly effective in improving structural and functional disorders in DAN, despite its effect in reducing HCY and MDA as oxidative biomarkers. acid morphological changes shown in our study following the induction of diabetes in rats included a decrease in the volume of the stria vascularis, spiral ganglion, spiral ligament and a decrease in spiral ganglion neurons. The stereological method we used in this study allows us to more quantitatively and more accurately examine structural and cellular changes in DAN. The structural decits in DAN are consistent with the functional ndings of ABR. In this study, the preventive effect of folic acid on structural changes at different doses(40mg/kg,80mg/kg) was not shown.


Introduction
Diabetic neuropathy is a common complication of diabetes mellitus(DM), which affects the sensory, motor and autonomic nerves and in some cases the central nervous system [1][2][3]. Ischemia caused by damage to the feeding vessels of the nerves in diabetic patients causes neural bers demyelination,degeneration and the resulting neuropathy [2,4].
The major metabolic disorders that cause vascular damage in diabetic patients, followed by neuropathy, are mainly due to polyol pathway and oxidative stress, which increase free radicals, especially reactive oxygen species in the tissues [4][5][6]. Impaired neural microcirculation in the auditory system can cause morphological and functional disorders. Diabetic auditory neuropathy (DAN) ismainly accompanied by debilitating symptoms such as hearing loss, tinnitus and dizziness [7,8].
Due to the signi cant occurrence of neuropathy, as well as its auditory type in diabetic patients, various agents have been evaluated for prevention and treatment of neural damage inexperimental and clinical studies. Among the drugs that have been investigated for diabetic neuropathy, rutin [9], folic acid [10,11],alpha-lipoic acid [12] and benfotiamine(a vitamin B1 derivative) can be mentioned [13]. The results of these investigations are not su cient for clinically acceptable applications and results.
In a study conducted by the authors of this article on rutin in the prevention and treatment of diabetic auditory neuropathy in rats, theuseful effect of this drug in terms of morphological and functional improvement was de ned to some extent [9].
One of the drugs under investigation which is relatively more clinically used than others in diabetic neuropathy isfolic acid. Due to the discrepancies in animal and clinical studies and clinical observations that in some cases show the ineffectiveness of this drug, we set out to do an investigation about folic acid effectiveness against diabetic auditory neuropathy in rats, following our previous study on rutin.
In this study, auditory brain stem response(ABR), stereology of spiral ganglion,homocysteine (HCY), superoxide dismutase(SOD), and malondialdehyde(MDA)levels were used as de ning criteria toinvestigate the effect of folicacid treatment in DAN.
Folic acid (vitamin B9) is categorized as a water-soluble vitamin which is involved in the synthesis of genetic materials, the production of red blood cells and fetal health. Folic acid is asynthetic form of folate that is better absorbed than folate. The main sources of folate include leafy vegetables, citrus fruits, mushrooms, grains and liver [14,15].
Folate, as a donor of single carbon units, is involved in the synthesis of serine from glycine, nucleotides from purine precursors, and in the formation of deoxy thymidylate monophosphate. In addition to its role in the synthesis of DNA and messenger RNA, it acts as a methyl donor to produce methyl cobalamin, which is used in the remethylation of HCY to methionine [16,17].
Folic acid seems to have the potential ability for growth and differentiation of neural cells and also hasa neuroprotectiveproperty due to its effects on HCY metabolism as well as decreasing malondialdehyde levels and increasing the expression of nerve growth factors [18,19].
ABR is one of the best tools for evaluating auditory neuropathy functionally. In this method, the evoked electricalactivities in the auditory nerve and its brainstem connections are measured. Diabetes mellitus can cause changes in ABR potentials latency and waveforms by disrupting the microcirculation of the auditory pathway [7,20]. Impaired microcirculation in DAN and subsequent cochlear ischemia can cause morphological changes in the hair cells, spiral ganglion, spiral ligament and stria vascularis [7]. In this study, we used the stereology method in the spiral ganglion, spiral ligament,and stria vascularis to examine morphological changes in the DAN and the response to treatment by measuring the volume and number of cells.
HCY is a non-proteinogenic amino acid made from the amino acid methionine following a transmethylation reaction. Many studies have shown an association between high levels of HCY in diabetic patients and its vascular and neurological complications [21][22][23].
SOD is an antioxidant enzyme that catalyzes the dismutation of superoxide radicals into oxygen and hydrogen peroxide. This enzyme acts as a biomarker in antioxidant activity in controlling oxidative stress [24].
MDA is a product of polyunsaturated fatty acids peroxidation. MDA is considered as a reliable biomarker in oxidative stress measurement levels in many experimental and clinical studies [25,26].
In this study, ABR was used to evaluate the hearing function, the stereology method for morphological evaluation in the spiral ganglion and oxidative biomarkers of HCY, MDA, and SODlevels as tools to evaluate the e cacy of folic acid treatment on DAN.

Animals preparation
Thirty-two male Sprague-Dawley rats (250-300 g) wereobtained from the Center of Comparative and Experimental Medicine (Shiraz, Iran). The rats were kept under standard lighting (12 hour-light/dark cycles), humidity (25-35 %), temperature (22-26 ˚C) per day and had free access to food and water. The rats were randomly allocated to four groups of eight rats, including non-diabetic normal rats with distilled water (Normal group), type 1 diabetic rats with distilled water without folic acid treatment (DM group), type 1 diabetic rats that received folic acid (40 mg/kg/day) (DM+FA40group), and type 1 diabetic rats that received folic acid (80 mg/kg/day) (DM+FA80group).
The experimental procedures were approved by the Ethics Committee of Shiraz University of Medical Sciences (Ethic code: IR.SUMS.REC.1397.459).

Induction of type 1 diabetesmellitus
To induce type 1 diabetes mellitus, a single intraperitoneal administration of STZ (60 mg/kg) was dissolved in cold citrate buffer andinjected into animals. Three days after induction, fasting blood sugar (FBS) levels were checked using a Glucometer (Accu-check®, Germany), and rats with FBS levels greater than300 mg/dL were considered as type 1 diabetic animals [27]. The con rmation day was considered 3daysafter STZ in jection.

Treatment
After the con rmation of diabetes, two groups received an oral solution of folic acid (40 mg/kg/day, DM+FA40group) and (80 mg/kg/day, DM+FA80group) by gavage once a day for 8 weeks. Normal and Diabetic groups received an equal amount of distilled water.
After 8 weeks of treatment, rats wereprepared for auditory assessment by ABR and histological stud ies.

ABR assessment
At the end of treatments,Xylazine (10 mg/kg) and Ketamine (75-100 mg/kg) were administered to the rats via intraperitoneal injection [9]. Thermal blanketand heat lampwere used to avoid hypothermiaduring anesthesia. In the ABR recording,groundelectrode (ECG Ambu Blue Sensor, Penang, Malaysia) was xed on the dorsal neck of rat. Also, active electrode (+) and the refer ence electrode (-) were attached to the forehead and post auricular area of right and left ears, respectively. Evoked potentials were achieved by intra acoustic EP25 (Co penhagen, Denmark) system with OtoAccess software (Middel fart, Denmark) after releasingclicks to the right ear via embedded earphones. Minimum 700 stimuli at a rate of 11.1 Hz were distributed to right ear and electrode impedance was below 3 KΩ. ABR evaluations were accomplishedusing band pass lter of 100-3,000 Hz and time window of 10 ms [28]. Evoked potentials were estimated according to the following parameters: absolute latencies and wave morphology of wave's II and Vand hearingthreshold. Hearing threshold was detected in 5 dB steps decliningfrom maximum stimulus of 70 dB until wave patterns morphology disappeared.

Tissue preparation& stereological analysis Preparation of inner ear in rat
At the end of study, all of the animals were sacri ced after induction of deep anesthesia by intraperitoneal injection of 100 mg/kg Ketamine and 10 mg/kg xylazine [9].Thetemporal bone was carefully dissected and separated from skull bones, decalci cation of temporal bone was performed by use of 8% HCL and 8% formic acid for 3days [29]. All temporal bones underwent tissue processing and after that were embedded in para n wax. The para n blocks were cut into25 µm thickness serial section together with the external acoustic meatus axis by a microtome (Microme, Germany). All of the serial sections of cochlea were collected and separated from apical to basal cochlear.
Thirty-two to thirty-six sections of cochlear tissue per rat were obtained and all slides were stained with Hematoxylin and eosin. Afterward, 8-12 sections of the 32-36 sections of cochlea in a systematic uniform random sampling were selected. The rst section was randomly selected and the next sections were chosen at equal intervals.
Stereological techniques were applied to estimate the volumes of the spiral ganglion,spiral ligament and striavascularis and the number of neurons in spiral ganglion. The cochlear sections evaluation was done by an examiner who was blind to the animal groups.
Estimation of the spiral ganglion, spiral ligament and striavascularis volumes The live gure of each cochlea section was evaluated by a video microscopy system and the stereology software (StereoLite, SUMS, Shiraz, Iran).
Regions of the selected structure (spiral ganglion, spiral ligament, and stria vascularis) were recognized in each cochleasection (Fig. 1a). The volume of the selected structure was analyzed by the point-counting methods based on "Cavalieri's principle" at the nal magni cation of 20 on 8-12 sections per animal [30][31][32].
The probe of stereology (a grid of points) was overlaid on thecochlear images (Fig. 1b), the crosssectional area of the selected structure "∑A" was calculated by the stereology software. Then the crosssectional area of the selected structure "∑A" was multiplied by the distance between the sections(T). The volume of selected structure was estimated by the following formula: where V is the volume of selected structure; T is distance between two selected sections; a/p is area per point; and ∑P is total points hitting selected structure. The area per point (a/p) was 722500 (850×850) µm 2 and an average of 121 points were counted per animal.

Estimation of the neuronnumbers in the spiral ganglion
The spiral ganglion is a collection of bipolar neuron cell bodies in the modiolus of inner ear, the lemon-like shaped central axis of the cochleawhose bers innervate the corti organ [30].One of the stereological techniques for spiral ganglion neuronscounting inahistological section thickness is the optical dissector[32 -34].
The site of the microscopic eld was chosen by moving the stage in directions (x and y) at the same distances pursuant to systematic uniform random sampling order. Objective lens (40×, NA 1.30) was employed by oil immersion.
The optical disector is made of several parts.An Eclipse microscope (E200, Nikon, Tokyo, Japan) with a large numerical aperture (NA=1.30) ×40 oil-immersion objective was linked to a video camera, which conveyed the live image of microscope to a computer monitor, and an electronic microcator with digital readout (MT12, Heidnehain, Traunreut, Germany) for estimating the movements in the Z-axis with a precision of 0.5μm.
One of the stereological probes isunbiased counting frame that is employed to count the number of select cells, by a stereology software system (Stereolite, SUMS, Shiraz, Iran). Theunbiased counting frame, composed of two exclusion lines (the lower and left borders and their extensions) and two inclusion lines (the upper and right borders), was overlaid on the live image at nal magni cation of 1500× (Fig. 2a).
The guard areas are zones at the superior and inferior part the histological sections. These regions were applied to avoid cutting tissue artifacts that occur throughout tissue processing on these areas of the sections. Any counting incident in focus within the up (the rst 3.5 μm) or down guard zones was deleted.
Each nucleolus cell that came into focus within the guard areas was not counted. The "height of disector" was the distance between the guard areas which was 18µm here.Every nucleolus cell appearing in the focus inside the later focal sampling plane was selected if it was located completely or partially inside the counting frame (Fig. 2b)and did not contact the forbidden lines[32-34].
The numerical density (NV) of the spiral ganglion neurons were evaluated using the following formula: Where is the number of the spiral ganglion neuronscoming into high quality focus in disector height, "ΣP" is the total of the counting unbiased frames in all elds of microscopic, and "h" is the height disector. Where "a/f" is the area of counting frame, "t" was the true section thickness calculated with the microcator, and "BA" is the microtome block advance set. The thickness of the internal ear section was calculated in the whole microscopic elds of view with uniform random sampling order from each section.
On average, 100-200 neurons were counted in 26 disectors in 8-12 sections per animal. The area of the counting frame (a/f) was 1082.41 (32.9×32.9) μm 2 .The real average thickness of the sections was 22 µm (t) and "BA" was 25 μm. The total number of the neurons was estimated by multiplying the numerical density (Nv) by V (submucosa or muscularis layers).

Biochemical assays
At the end of treatments, SOD activity, MDA serum levels, and serum activity of HCYwere measured in all groups, using the SOD ELISA kit (Cayman, Ann Arbor, MI, USA), Rat MDA ELISA kit (CusabioBiotech, China), and HCY ELISA kit (Axis Homocysteine Enzyme Immunoassay,IBL, Germany), respectively.

Statistical analysis
Data were presented asmean for stereological resultsandmean ± SEM forother results. Statistical analysis was determined using one-way analysis of variance (ANOVA) for parametric data such as latency of wave II, hear ing threshold, volume of spiral ganglion, spiral ligament and striavascularis,number of spiral gan glion neurons, SOD, MDA,and HCY.Also,Kruskal-Wallis test was appliedfor non parametric data such as different forms of wave V. GraphPad Prism software (Version 6) was used to compare data. If a signi cant difference was obtained, the source of difference was located by Tukey or Dunn post-hoc test. P<0.05 was considered as statistically signi cant.

Results
Outcome of folic acidon the function ofABR After 8 weeks ofDM induction, theeffect of diabetes on ABR wave forms was observed.
The negativechanges were signi cantly on the latency of wave II, hearing threshold, andthe presence of wave V (P<0.001). The latency of wave II,increasedhearing threshold, andthe rate of presence of wave V are shown in Fig. 3 and Table 1, respectively.Folic acid (40 and 80 mg/kg) had no signi cantrecovery effect onABRrecordingsat the end of study (about 8 weeksafter the induction of diabetes) (Fig. 3a, 3b andTable 1).

Stereological studies
Evaluation of spiral ganglion,spiral ligament, stria vascularisvolume, and Spiral ganglion neuron count There is a signi cant decline in the total volume of the spiral ganglion,spiral ligament,stria vascularis andthe count of spiral ganglion neurons in the rats with DM compared to the normal animals (P<0.05) (Fig. 4a, 4c, 4d, and 4b).
Biochemical assays Erythrocyte level ofSOD of diabetic group was signi cantly lower than normal group (P<0.001). On the other hand, these values for folic acidgroups (DM+FA40 and DM+FA80) were not higher than the diabetic group (Fig. 5a).
TheMDAserum level in the diabetic group was signi cantlyhigher than the normal group (P<0.01) but, serum levels of MDA for folic acid-treated groups (40 and 80 mg/kg/day) were signi cantly lower than that of the diabetic group (P<0.05) after the intervention (Fig. 5b).
Serum level of HCY was increased signi cantly in the diabetic group compared to the normal group (P<0.001; Fig. 5c). Administration of folic acid 40 and 80 mg/kg/dayshowed a signi cant decrease in serum level of HCY (P<0.001, and P<0.0001, respectively; Fig. 5c).

Discussion
This study revealed that folic acid does not have a signi cant effect on prevention and treatment of functional and structural de cits in DAN despite reduced MDA and HCY levels.
In this study, folic acid was administered in two doses of 40mg/kg and 80mg/kg, but neither of them showed a signi cant response in ABR recordings and spiral ganglion stereology.
In our opinion, spiral ganglion stereology and ABR recordings are indicators that are more closely related to clinical situation than MDA and HCY levels.
The advantage of our study is the histological examination through stereological method and its compliance with functional outcomes in diabetic neuropathies treated with folic acid.
ABR is an important tool for assessing hearing function. Numerous clinical and experimental studies have shown the effect of diabetes on ABR recordings. These changes include an elevation in the hearing threshold and an increase in waves latency as well as a decrease in the amplitude with changes in waveforms [35,36].
Our study showed that 8 weeks' treatment of diabetic rats with folic acid at different doses(40mg/kg,80mg/kg) did not signi cantly improve hearing threshold and also waves II and V latency as compared to control diabetic rats. There are some differences in rat ABR compared to humans that we have considered in this study. It seems wave II that was generated in the cochlear complex nucleus, is the most prominent in rats ABR recordings [37].
Morphological changes in diabetic auditory neuropathy mainly include atrophy of stria vascularis, thickening of the basilar membrane in the stria vascularis, and a decrease in the number of spiral ganglion neurons [7,38].
Similar to other studies, the histological changes shown in our study following the induction of diabetes in rats included a decrease in the volume of the stria vascularis, spiral ganglion, spiral ligament and a decrease in spiral ganglion neurons. The stereological method we used in this study allows us to more quantitatively and more accurately examine structural and cellular changes in DAN. The structural de cits in DAN are consistent with the functional ndings of ABR. In this study, the preventive effect of folic acid on structural changes at different doses(40mg/kg,80mg/kg) was not shown.
Our study revealed a signi cant decrease in HCY levels following folic acid administration in diabetic rats. At higher doses of folic acid(80mg/kg), this reduction is more signi cant. Its role and underlying mechanism in development of vascular complications in diabetes and reducing its levels on complication incidence following treatment are not exactly de ned [39,22].
The increased oxidative degradation of nitric oxide(a regulator of endothelial homeostasis) during the process of oxidative stress, is consideredthe presumptive mechanism of endothelial vascular injury in elevated HCY levels [40,41]. Folic acid through its bioactive form, 5-methyltetrahydrofolate, donates a methyl-group forHCY remethylation to methionine [17,42].
Our study showed that plasma MDA levels were signi cantly increased following folic acid treatment in diabetic rats. This nding is consistent with other ndings in experimental studies on the folic acid treatment of neuropathy [10]. Another important nding in this study was the lack of SOD increase following folic acid treatment. These antioxidant enzymes control and eliminate reactive oxygen species (ROS)during oxidative stress in cells [43]. Based on the ndings of this study, folic acid lowered plasma HCY and MDA levels signi cantly while it had no signi cant effect on SOD. It may imply that oxidative stress is reduced via a mechanism other than an increase in the amount or action of SOD.
Another perception that can be drawn from these ndings is that, due to the lack of improvement in structural and functional disorders in DAN and the lack of signi cant change in SOD levels, SOD may be a better measure of oxidative stress than HCY and MDA.
It can be deduced that the selected doses of folic acids were not high enough to produce meaningful action on diabetic neuropathy. However, based on our scienti c bibliography the selected doses werehigher thanthose applied in similar studies.

Conclusion
Diabetic auditory neuropathy is one of the serious complications of diabetes that has a serious impact on quality of life. Our study showed that folic acid is not effective in relieving structural and functional disorders in DAN, despite its effect in reducing HCY and MDA. Due to this issue, more studies are needed to nd potent drugs for preventing and treating this complication in diabetic patients.   The aligned dot plot of the volumes of the spiral ganglion(a), spiral ligament(c), striavascularis(d), and number of spiral ganglion neurons(b)in the normal, DM, DM+FA40, and DM+FA80 groups. Each dot represented an animal and the horizontal bar is the mean of the listed parameters. *indicates P < 0.05vs. diabetic group.