Treatment with Tang-luo-ning altered the microRNA expression profile in rats with diabetic peripheral neuropathy

ABSTRACT Tang-luo-ning (TLN) is a traditional Chinese herbal recipe that has been used to treat diabetic peripheral neuropathy (DPN); nevertheless, the underlying mechanism remains unclear. This study was aimed to investigate the microRNA (miRNA) expression profile in diabetic rats treated with TLN, and the target genes were predicted. Male Sprague-Dawley rats were randomly divided into control, diabetes, and TLN-treated diabetes groups. Diabetes was induced with streptozotocin, and TLN (5 g/kg/day) was orally given for eight weeks. Then, the sciatic nerves were harvested for miRNA microarray analyses. The differentially expressed miRNAs and their target genes were analyzed. Compared with the control rats, 24 miRNAs were significantly upregulated, and 59 were downregulated in the sciatic nerves of the diabetic rats by more than two folds (all P < 0.05). In TLN-treated diabetes rats, 26 miRNAs were upregulated, and 14 were downregulated compared with diabetic rats without TLN treatment (all P < 0.05). DPN-induced alterations of the miRNA profile were reversed by the TLN treatment. A total of 1402 target genes were screened. In GO analysis, genes in localization, cytoplasm, and protein binding processes were enriched, and the most significantly enriched pathways included the neurotrophin, Fc epsilon RI, and Wnt signaling pathways. Further analyses revealed that DVL1 and NTF3 genes were involved in these pathways. Our findings indicate that TLN may affect the Wnt and neurotrophin pathways by acting on DVL1 and NTF3 genes.


Introduction
Diabetic peripheral neuropathy (DPN) is a common chronic complication of diabetes and potentially affects the distal sensory, motor, and autonomic nerves. Clinically, DPN is characterized by pain, numbness, and dysfunction of the autonomic nervous system, which seriously impairs individual quality of life [1]. Tang-luo-ning (TLN) is a traditional Chinese herbal recipe that is comprised of Astragalus root, Fructuscorni, and Salvia miltiorrhiza. In our previous study, the clinical effective rate of TLN for treating DPN was 89.9%, and TLN was significantly superior to mecobalamin [2]. TLN could effectively improve DPN-related symptoms, including numbness, pain, muscle spasms, and impaired tendon reflexes [3]. In animal experiments, TLN could reduce the sorbitol levels in the sciatic nerves of diabetic rats [4]. Additionally, TLN increased the rat serum level of nerve growth factor (NGF) as well as the mRNA expression of NGF in the sciatic nerves [5]. Moreover, TLN also increased the serum insulin-like growth factor-1 (IGF-1) levels and the mRNA expression of IGF-1 in the liver of diabetic rats [6,7]. Further experiments revealed that TLN could activate the Nrg1/ErbB2 and Erk/bad signal pathways to promote Schwann cell survival [8]. In our previous studies, we found that TLN was effective in treating DNP in both clinical and animal experiments [2,[4][5][6][7]. These conclusions were supported by the experiments conducted by Yang et al. [9]. However, the mechanism underlying the therapeutic effects of TLN in DPN has not yet been clarified.
MicroRNAs (miRNAs) are a type of non-coding RNA of 21-25 nucleotides in length, which are involved in the post-transcriptional regulation of gene expression. miRNAs have been reported to mediate the therapeutic effects of natural herbs [10,11]. In the previous study, we performed a microarray bioinformatics analysis of genes that were differentially regulated by TLN, and cell survival genes (NRG1, MTOR, GAB1, and PI3KCB) identified from the microarray analysis were validated by qPCR. We found that TLN significantly upregulated the expression levels of NRG1, MTOR, and GAB1 [8].
The aim of this study was to investigate the microRNA (miRNA) expression profile in diabetic rats treated with TLN, and the target genes were predicted.

Diabetic rat model and TLN treatment
Male Sprague-Dawley (SD) rats (200-250 g) were purchased from the Vital River Laboratories (Beijing, China). The rats were housed in the Animal Center at the Beijing University of Chinese Medicine, according to the National Standards (Gb14925-2001) [9]. The ambient environment was maintained at a humidity of 60 ± 10% and a temperature of 25 ± 1 °C. Ten rats were randomly chosen as the normal group, and the remaining 20 rats were subjected to the DPN model. The DPN model was induced by intraperitoneal injection of 60 mg/kg streptozotocin (STZ) (Sigma-Aldrich; St. Louis, MO, USA). After 72 hours of injection, the blood glucose was measured; a value of >16.7 mmol/L was defined as diabetes. A total of 20 diabetic rats were randomly divided into groups with (10 rats) or without (10 rats) TLN treatment. TLN was given by gavage at a dose of 5 g/kg/d for eight consecutive weeks. The dosage of TLN was determined based on our experience in previous studies. An equivalent volume of distilled water was used as control. After the last administration, all rats were fasted for 12 hours. Then, all rats were anesthetized using xylazine (10 mg/kg)/ketamine (75 mg/kg), and sciatic nerves were harvested. This study was approved by the Ethics Committee of Beijing University of Chinese Medicine (No. BUCM201738). These animal experiments were conducted according to the NIH guidelines for the care and use of laboratory animals [8].

RNA extraction and miRNA array
Total RNAs were extracted from rat sciatic nerves and quantified by NanoDrop (ND-2100, Thermo Fisher Scientific, Waltham, MA, USA). RNA integrity was assessed using Agilent 2100 (Agilent Technologies, Santa Clara, CA, USA). The sample labeling, microarray hybridization, and washing were performed based on the manufacturer's standard protocols. Briefly, total RNAs were tailed with Poly A and then labeled with Biotin. The labeled RNAs were then hybridized onto the microarray (Affymetrix miRNA 4.0, Santa Clara, CA, USA). After washing and staining the slides, the arrays were scanned by the Affymetrix Scanner 3000 (Affymetrix, Santa Clara, CA, USA).

Data analysis
Affymetrix GeneChip Command Console software (Version 4.0, Affymetrix, Santa Clara, CA, USA) was used to analyze the array images to obtain raw data. Next, R software was used for the following data analysis. Differentially expressed miRNAs were determined by fold changes as well as P values, which was calculated using the t-test. A P value < 0.05 was considered to indicate statistical significance. Target genes of differentially expressed miRNAs were predicted using three databases (miRanda, miRDB, miRTarbase). GO analysis and KEGG analysis were performed to predict the functions of target genes. Hierarchical clustering was performed to show the distinguishable miRNAs expression patterns among the samples.

Alterations of the miRNA expression profile after TLN treatment
Compared with the miRNA expression profile in the sciatic nerves of the control rats, 24 miRNAs were significantly upregulated, and 59 miRNAs were downregulated in diabetes rats by more than two folds (all P < 0.05, Table 1). Clustering analysis shows that these 83 genes could be clearly clustered into two categories ( Figure 1). In diabetes rats treated with TLN, 26 miRNAs were upregulated, and 14 were downregulated compared with diabetic rats without TLN treatment (all P < 0.05, Table 2). These 40 genes were also clustered into two categories ( Figure 2). Among the differentially expressed genes in the above two analyses, 19 overlapped genes were observed (Table 3). In these 19 genes, the upregulation and downregulation in diabetic rats compared with control rats were reversed by the TLN treatment (Table 3).

Prediction of target genes
Target genes of the differentially expressed miRNAs were predicted using the miRanda, miRDB, and miRTarbase databases, with at least two databases included as the screening criteria. A total of 1402 target genes were screened (Table S1).
GO analysis of the target genes was performed, which were designated to the categories of biological process, cellular component, and molecular function. In the biological process category, the most significantly changed items between diabetes rats treated with or without TLN were localization, transport, and establishment of localization ( Figure 3). In the cellular component category, the most significantly altered items were cell part, cytoplasm, and cell ( Figure 3). The most significant entries in the molecular function were binding and protein binding ( Figure 3).
We performed a pathway analysis of the miRNAtargeted genes using the KEGG database. The number of target genes in each pathway was counted, and the significance of the enrichment of the target genes was analyzed. The top 10 significantly enriched pathways were presented in Table 4. After the TLN treatment, the most significantly enriched pathways were metabolic pathway, endocytosis, and cytokinecytokine receptor interaction (Table 4).

Intersection analysis of GO analysis and pathway analysis
We performed an intersection analysis of GO and pathway analyses. Between the normal and diabetes groups, there were 469 intersection genes. Between the diabetic rats treated with or without TLN, there were 235 intersection genes. There were 124 overlapped intersection genes between the above two analyses. We further analyzed the top 10 pathways and identified two genes, DVL1 and NTF3, which were involved in the Wnt ( Figure 4) and neurotrophin ( Figure 5) signaling pathways.

Discussion
DPN is a chronic microvascular complication of diabetes. According to global epidemiological surveys, at least 50% of patients with type 1 or type 2 diabetes suffer from DPN, which is one of the major causes of amputation in diabetic patients [12]. Clinical manifestations in patients with DPN include decreased motor and sensory nerve conduction velocity, microvascular and microcirculatory dysfunction, the increased threshold of vibration and thermal pain perception, progressive sensory loss, and peripheral nerve degeneration. Most patients with DPN usually have self-described symptoms such as paresthesia, allodynia, hyperalgesia, and spontaneous pain [13][14][15][16][17][18].
In the differential expression analysis of the miRNA expression profile, we found that 19 miRNAs were overlapped in the comparisons of control vs. diabetic rats and diabetic rats treated with vehicle vs. diabetic rats treated with TLN. In addition, diabetes-induced changes of these 19 miRNAs were reversed by the TLN treatment, suggesting that the therapeutic effects of TLN on DPN may be associated with an altered Table 1. Differential analysis of miRNA expression profiles between control and diabetic rats. Compared with the miRNA expression profile in the sciatic nerves of the control rats, 24 miRNAs were significantly upregulated, and 59 miRNAs were downregulated in diabetic rats by more than 2-fold (all P < 0.05).    Through further pathway and gene analysis, DVL1 from the Wnt signaling pathway and neurotrophin-3 (NTF3) from the neurotrophin signaling pathway were identified. A previous study found that the Wnt signaling pathway may mediate oxidized lowdensity lipoprotein (ox-LDL)-induced endothelial cell damage through oxidative stress and that endothelial cell damage can be ameliorated by suppressing oxidative stress and inhibiting the Wnt signaling pathway [19,20]. The protein encoded by the DVL1 gene is a cytoplasmic phosphoprotein that plays an important role in the Wnt signaling pathway [20]. Figure 4 shows the relationship between DVL1 and other components in the Wnt pathway. These results suggest that treatment with TLN may influence the DVL1 expression and then affect the Wnt signaling pathway, resulting in reduced oxidative stress and improved symptoms of DPN. According to previous studies, NTF3 belongs to the nerve growth factor family and is widely expressed in the central nervous system and surrounding tissues [21]. NTF3 can induce the growth of neurites and promote the Table 3. The common miRNAs in both differential analyses. In these 19 genes, the upregulation and downregulation in diabetic rats compared with control rats were reversed by treatment with TLN.    expression of key enzymes involved in neurotransmitter synthesis. NTF3 plays an important role in the development, differentiation, growth, regeneration, and functional characteristics of the central and peripheral neurons [22][23][24][25]. A lack of NTF3 is associated with a variety of neurological disorders [25][26][27][28]. In addition, NTF3 plays a critical role in the survival and function of sensory neurons and sympathetic neurons. Studies have also shown that NTF3 has therapeutic effects on peripheral neuropathy [29][30][31].
Notably, NTF3 protects sensory neurons in diabetes rats and promotes axonal regeneration in diabetic lesions [32,33]. Taken together, NTF3 may mediate the therapeutic effects of TLN on DPN. In this study, we predicted the target genes but did not quantify the protein expression levels. In the future, we will conduct validation and functional experiments for specific genes.

Conclusion
Our findings indicate that TLN may affect the Wnt and neurotrophin pathways by acting on DVL1 and NTF3 genes. Moreover, TLN may improve the symptoms of DPN by affecting the localization, cytoplasm, and protein binding processes.