Scraping Therapy Improved Muscle Regeneration through Regulating GLUT4/Glycolytic and AMPK/mTOR/4EBP1 Pathways in Rats with Lumbar Multifidus Injury

Background High morbidity of nonspecific low back pain (NLBP) and large consumption of medical resources caused by it have become a heavy social burden. There are many factors inducing NLBP, among which the damage and atrophy of multifidus (MF) are most closely related to NLBP. Scraping therapy can have significant treatment effects on NLBP with fewer adverse reactions and less medical fund input than other modalities or medications. However, the mechanism of scraping therapy treating NLBP remains unclarified. Here, we wanted to investigate the effects of scraping therapy on promoting MF regeneration and the underlying mechanisms. Methods A total of 54 male rats (SD, 6-7 weeks old) were randomly divided into nine groups, namely, K, M6h, M1d, M2d, M3d, G6h, G1d, G2d, and G3d, with six rats in each group. They were injected with bupivacaine (BPVC) to intentionally induce MF injury. We then performed scraping therapy on the rats that had been randomly chosen and compared treatment effects at different time points. In vitro data including skin temperature and tactile allodynia threshold were collected and histological sections were analyzed. mRNA sequencing was applied to distinguish the genes or signaling pathways that had been altered due to scraping therapy, and the results were further verified through reverse transcription polymerase chain reaction and Western blot analysis. Results Transitory petechiae and ecchymosis both on and beneath the rats' skin raised by scraping therapy gradually faded in about 3 d. Cross-sectional area (CSA) of MF was significantly smaller 30 h, 2 d, and 4 d after modeling (P=0.007, P=0.001, and P=0.015, respectively, vs. the blank group) and was significantly larger in the scraping group 1 d after treatment (P=0.002 vs. the model 1d group). Skin temperature significantly increased immediately after scraping (P < 0.001) and hindlimb pain threshold increased on the 2nd day after scraping (P=0.046 and P=0.028, respectively). 391 differentially expressed genes and 8 signaling pathways were characterized 6 h after scraping; only 3 differentially expressed genes and 3 signaling pathways were screened out 2 d after treatment. The amounts of mRNAs or proteins for GLUT4, HK2, PFKM, PKM, LDHA (which belong to the GLUT4/glycolytic pathway), p-mTOR, p-4EBP1 (which belong to the AMPK/mTOR/4EBP1 pathway), and BDH1 were enhanced, and p-AMPKα was decreased after scraping therapy. Conclusions Scraping therapy has therapeutic effects on rats with multifidus injury by promoting muscle regeneration via regulating GLUT4/glycolytic and AMPK/mTOR/4EBP1 signaling pathways.


Introduction
Nonspecifc low back pain (NLBP) refers to acute and chronic pain in lumbosacral region localized between costal margin and inferior gluteal folds, with or without leg pain, excluding specifc pathogenic factors [1,2]. It accounts for 90% of cases in low back pain [1]. Pathological mechanism of NLBP is complex, which includes neurological, muscular, joint, and psychological factors. A study found that multifdus (MF) has an important role in maintaining spinal stability and performing normal physiological functions of the spine, and the damage and atrophy of MF were closely related to the development of NLBP [3]. MF is located deep in dorsal musculature, and it is attached to the spinous process, transverse process, and vertebral plates of the spine. Compared to other lumbar muscles, MF has larger crosssectional areas and shorter fbers, which allow MF to generate tremendous forces in a small range of motion [4,5].
Tere are many kinds of therapeutic applications towards NLBP including bed rest, pain medication, physical/ rehabilitation therapy, and cognitive behavioral therapy [1,2]. Nevertheless, each method has its own limitation or side efect. Bed rest is recommended as the primary treatment in some of the literature, but it is highly controversial, with some studies showing that bed rest is not only unbenefcial but may also have adverse efects compared to patients who maintain normal activities, especially when prolonged bed rest tends to cause atrophy of paravertebral muscles like MF, and induce complications including deep vein thrombosis [6][7][8]. Medications commonly used to treat NLBP include nonsteroidal anti-infammatory drugs (NSAIDS), opioids, and antidepressants, all of which relieve pain symptoms through their respective targets of action but accompanied by inevitable side efects, with NSAIDS prone to gastrointestinal adverse efects and opioids predisposed to cause dizziness, nausea, and constipation [9,10]. Te limitation and addiction of these medications may compromise ultimate therapeutic efects. Physiotherapy, including heat therapy, electrotherapy, microwave, and other remedies, is now attracting more extensive clinical interest, as it is supposed to be capable of improving vascular microcirculation, promoting anti-infammation, and relieving muscle spasm. However, its long-term efcacy on patients with low back pain is very limited [11]. Lumbar traction can relieve symptoms such as oozing, edema, and spasm of lumbar tissues and relax local muscles [12], but in a meta-analysis that analyzed 32 randomized controlled trials, researchers concluded that traction therapy was not signifcantly effective compared to the placebo group [13]. Exercise is currently one of the most recommended treatments for low back pain, and there are varieties of individualized projects aiming to relieve NLBP symptoms, restore daily function, and reduce recurrence rate by enhancing strength of patients' core spinal muscles, regulating physiological status, and improving psychological regulation [14,15]. Exercise requires patients' long-term adherence, and some studies have pointed out that this therapy is not efective in relieving pain in the short term [16,17].
Scraping therapy, also known as Gua Sha or coining therapy, is one of the characteristic external therapies in traditional Chinese medicine, which has been commonly practiced in Asia and partly distributed in Western countries. Tis modality involves scraping or rubbing lubricated area of the body repeatedly and unidirectionally with a smooth-edged instrument to intentionally raise transitory petechiae and ecchymosis which normally fade within 1 week [18]. Scraping therapy is empirically efective for acute or chronic pain and for other conditions such as respiratory diseases as well as musculoskeletal problems (from fbromyalgia to severe strain, spasm, or injury) [19]. However, most reports available mainly focused on therapeutic impacts as well as complications of scraping therapy [20], and there are only few studies that have limitedly investigated underlying mechanisms [21]. Implicit mechanism of therapeutic efects in scraping therapy makes its distribution confned to smaller regions compared with acupuncture, besides the noticeable abusive-like skin manifestations after treatment.
According to our long-term clinical observation on the efects of this modality, scraping therapy does relieve patients' muscular stifness and soreness and partly improves muscle strength. Skeletal muscle has a strong regenerative capacity, which is closely related to the function of satellite cells (SCs) [22]. A growing amount of evidence suggested that SC function is largely dependent on two metabolic states of cells: oxidation and glycolysis [23]. Glycolysis plays an important role in muscle regeneration [24]. Adenosine 5′monophosphate (AMP)-activated protein kinase (AMPK) is an important energy receptor in human body that is involved in a variety of signaling pathways and plays a regulatory role in glycometabolism, protein metabolism, lipid metabolism, and autophagy in the organism [25], thus also connected with the process of myocyte repair [26,27]. In the present study, a rat model of lumbar multifdus injury was adopted to simulate NLBP of humans. We aimed to screen changed epigenetic expressions through transcriptomics after single scraping therapy and verify its myocyte-growthpromoting function, as well as the underlying mechanisms involved in glycolysis and AMPK pathway.

Animals.
Male rats (Sprague-Dawley) were purchased from Charles River Laboratories (Beijing, China) at 6 to 7 weeks of age. Tey were kept under standardized pathogen-free conditions in the animal facility of 12 h light/ dark cycles, temperature (24 ± 2°C), and 40%-50% relative humidity. All animal experiments have been approved by the Ethical Committee of Medicine of Navy Medical University (Shanghai, China; No. 2021GW0309) and have been conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. A total of 54 rats were adaptively fed for 1 week and randomly divided into nine groups using a randomized digital table, namely, the blank group (K), the model 6h group (M6h), the model 1d group (M1d), the model 2d group (M2d), the model 3d group (M3d), the scraping 6h group (G6h), the scraping 1d group (G1d), the scraping 2d group (G2d), and the scraping 3d group (G3d), with 6 rats in each group, calculated by ANOVA using assumption that α � 0.05 and power � 0.80. All rats were euthanized by intraperitoneal overdose of 3% sodium pentobarbital (3 mL/kg) at appropriate time and the order of treatments and measurements were randomized. Apart from the conductor, no one was aware of the group allocation at the diferent stages of the experiment.

2
Pain Research and Management

Te Construction of Animal
Model. We adopted the method described in the previous research [28]. Te rats in the model and scraping groups were anesthetized by intraperitoneal injections of 3% sodium pentobarbital (1 mL/ kg) and then fxed in a prone position with hair shaved. Muscle injury was induced by bilateral injections of 0.5% bupivacaine (BPVC; Sigma, USA) into MF at vertebral level L4 and L5. Te injections of BPVC were made over 3 s with a 27-gauge needle, after which the needle was rotated out to facilitate the absorption of the drug. Te needle was advanced into the muscle just beside the spinous process until it contacted the bone of vertebral arches. Ten, the cannula was withdrawn 1 mm and the material deposited close to the bone to make sure that MF was injected. 100 μL of BPVC was injected at each entry point and a total of 400 μL were deposited. Te whole operation was kept sterile.

Te Manipulation of Scraping Terapy.
Te rats in scraping groups were given scraping therapy and the target range was from the frst thoracic vertebra down to the sacrum on the back (three scraping lines were located on the middle of the back, and about 0.8 cm next to the midline on each side; the scraping length of each was equal, Figure 1(a)). Te selected skin area was shaved and wiped with glycerin for lubrication. Te rats were then scraped 30-35 times for each scraping line with a smooth-edged instrument made of bull horn (Figure 1(c)) in a unidirectional manner. Te scraping force was optimized and standardized at 7-8 N; the frequency was 40 times/min, and the total time of therapy was about 3 min.

Skin Temperature Measurement.
Skin temperatures at the vertebral L4, L5, and L6 were measured at approximately 10 cm above the skin of the rat's back using an infrared thermal imager (FLIR, USA). Rats were tested for the baseline temperature before scraping therapy, and those in the scraping groups were additionally tested both at the time immediately after scraping and before harvested.

Te Assessment of Tactile
Allodynia. Te assessment of tactile allodynia was performed as previously described [29]. Testing was performed during the day portion of the circadian cycle only (06:00-18:00 h). Rats were placed in a von Frey testing cage with a wire mesh bottom which allowed full access to the paws. Behavioral accommodation was allowed for approximately 20 min, until cage exploration and major grooming activities ceased. Te area tested was the midplantar hind paws, in the sciatic nerve distribution, avoiding the less sensitive tori (footpads was recorded if the paw was sharply withdrawn or the rat finched immediately upon removal of the hair. Ambulation was considered an ambiguous response, and in such cases, the stimulus was repeated. Te 50% withdrawal threshold was determined using the up-down method through the formula: 50% g threshold � (10 [X f +kδ] )/10, 000 [30]. Rats were tested for the baseline pain thresholds, and those in the model and scraping groups were tested 24 h after BPVC injection and at the time before sacrifced, and each rat was tested in both hind paws.
2.6. Immunohistochemistry. Te skin and MF which had been harvested were fxed in 4% paraformaldehyde for 48 h and then washed with running water to remove residual paraformaldehyde and impurities. Te tissues were sealed and dehydrated in gradient concentrations of ethanol (50%, 70%, 85%, 95%, and anhydrous ethanol) for 2 h each. Te washed and dehydrated tissues were soaked in a mixture of anhydrous ethanol and xylene for 2 h followed by xylene immersion for 2 h. Te process was repeated once to make the tissues transparent. Te tissues were frst impregnated in a mixture of melted parafn and xylene in equal amounts for 2 h, and then placed in parafn for 2 h twice. Te transparent tissue blocks were placed in the melted parafn wax and kept warm in the molten wax chamber. Te blocks were sliced at 5 μm using a microtome (RM2235, Leica, Germany). Te tissue slices were put into an incubator (Huitai Company, China) at 65°C for baking for 30 min. Te slices were then put into xylene for 15 min twice, and dehydrated in gradient concentrations of ethanol (100%, 95%, 90%, 80%, and 70% ethanol to water), each for 15 min. Te tissue sections were put into hematoxylin solution, stained for 5 min, rinsed under running water for 3 min; 1% hydrochloric acid alcohol fractionation for 2-3 s, rinsed under running water for 15 min, and then put into 95% ethanol for 30 s of dehydration; the tissue sections were then put into alcohol eosin solution, stained for 2 min, and put into 95% ethanol for 3 min twice, and anhydrous ethanol for 3 min twice. Te slices were transparentized with xylene for 3 min twice, blocked with neutral resin, dried at 65°C, and were observed using a microscope (DMi8, Leica, Germany). Crosssectional area (CSA) of MF fber (primary outcome) was analyzed by randomly selecting 30 fbers using ImageJ analysis software (National Institutes of Health, https://rsb. info.nih.gov/ij/).

mRNA-seq and Data Analysis.
Total RNAs from MF tissues were extracted using TRIzol reagent (Life Technologies, USA) according to the manufacturer's instructions.
RNA purity was measured using NanoDrop ™ One/OneC (TermoFisher Scientifc, USA) and was quantifed with Qubit ™ RNA HS Assay Kit (Qiagen, German). RNA integrity (RIN) was analyzed using Agilent 4200 TapeStation System. Te mRNA with polyA structure was captured using magnetic beads with Oligo dT. Ten, the fragmented mRNA was used as template and random oligonucleotides used as primers. Te frst strand of cDNA was synthesized in the M-MuLV reverse transcriptase system, and dNTPs degraded Pain Research and Management from RNA with RNaseH were used in the DNA polymerase I system to synthesize the second strand of cDNA [31]. After purifed and repaired, cDNA of about 200 bp was screened with AMPure XP beads and amplifed through PCR to construct the library, which was sequenced on an Illumina HiSeq PE150 sequencer [32]. Sequenced fragments were mapped to reference genome with HISAT2 software [33]. Gene expression analysis was performed using HTSeq and value conversion was reported in fragments per kilobase per million (FPKM).

Pilot Study.
A total of 9 rats were adaptively fed for 1 week and randomly divided into 3 groups, namely, the blank group, the model 1d group (M1d), and the scraping 1d group (G1d), with 3 rats in each group. Te results of crosssectional area for each group are as follows in Table 1. 2.11. Statistics. All statistical analyses were performed using SPSS 23.0 software (all the raw data were included in the supplemental fle, named S3 raw data). Quantitative data were reported as the mean ± standard error. One-way analysis of variance (ANOVA) followed by Fisher's least signifcant difference (LSD) post hoc test was performed for multiple comparisons of cross-sectional area, and expressions of mRNA and proteins, and paired t-test was performed for data of skin temperature and allodynia. Shapiro-Wilk test was used for the normality assessment. Statistics of tactile allodynia were not normally distributed, and Wilcoxon signed-rank test was used.
Other data were all normally distributed. For the comparison of skin temperature and allodynia, the independent variables were time intervals after scraping therapy in the scraping groups and time intervals without scraping therapy in the blank group and the modeling groups; the dependent variables were skin temperature and tactile allodynia, respectively. For the comparison of cross-sectional area and expressions of mRNA and proteins, the independent variable was whether given scraping therapy or not; the dependent variables were cross-sectional area and expressions of mRNA and proteins, respectively. Te mathematical formula in use for the sample size calculation was n � [((μ α + μ β )σ/δ)] 2 + (1/4)μ 2 α , and we used the program called "Tests for Two Means" of PASS 11 for calculating. σ is the standard deviation of the change in cross-sectional area of lumbar multifdus of rats after scraping therapy, δ is the difference value in the population mean of the change in crosssectional area of lumbar multifdus of rats, and μ α and μ β are values of μ corresponding to α and β, respectively, which had all come from our preliminary experiments. For all analyses, a p value of <0.05 was considered to be statistically signifcant.

Scraping Terapy Raised Transitory Petechiae and Ecchymosis Which Would Fade in about Tree Days and Promoted the Recovery of Myocytes.
Transitory petechiae and ecchymosis both on and beneath rats' skin raised by scraping therapy gradually faded in about 3 days (Figures 2(a) and 2(b)). 3 days after scraping, skin surface of rats basically restored to the baseline level, with no visible ecchymosis, and subfascial stasis was basically absorbed, with a few scattered light red petechiae in sight. Tere were visibly less cell debris, fewer infammatory cells, and more neonatal myocytes in the G groups compared with the M groups at the same time (Figure 2(c)). CSA of MF was signifcantly smaller 30 h, 2 d, and 4 d after modeling (P � 0.007, P � 0.001, and P � 0.015, respectively) and was signifcantly larger in the G1d group than in the M1d group 1 d after scraping (P � 0.002) (Figure 2(d) and Table 2). However, there were no signifcant diferences in CSA between the M group and the G group at other time points.

Scraping Terapy Increased Rats' Skin Temperature
Transiently and Partly Enhanced Withdrawal Treshold of Rats' Hindlimb. Local skin temperature signifcantly increased immediately after scraping (P < 0.001), while no signifcant diferences were found at other time points (Figure 3(a) and Table 3). Pain threshold (also called withdrawal threshold) increased in the 2nd day after scraping (P � 0.046 and P � 0.028, respectively) with no prominent diferences recorded at other time points (Figure 3(b), Tables 4 and 5). Our previous fndings indicated that single scraping treatment may have limited efect of pain relieving, which had little to do with temperature increase of local tissues.

Te Changes of mRNA Expressions Induced by Scraping
Terapy Were Analyzed through mRNA Sequencing. Tere were 391 diferentially expressed genes in the G6h group compared with the M6h group, including 322 upregulated genes and 69 downregulated genes. Only 3 diferentially expressed genes were identifed between the G2d group and the M2d group, with 2 upregulated and 1 downregulated genes (Figures 4(a) and 4(b)). 2 upregulated genes were Hmgb2l1 and Bdh1, and 1 downregulated gene was B3galt2.
Te results of gene ontology (GO) enrichment analysis of diferentially expressed genes between the G6h group and the M6h group included the following: (i) biological processes: mainly enriched in skeletal muscle development, carbohydrate catabolism, and intracellular metal ion homeostasis; (ii) cellular composition: mainly enriched in muscle nodes, contractile fber fraction, and myogenic fbers; and (iii) molecular functions: no relevant functions were enriched. Te results between the G2d group and the M2d group included the following: (i) biological processes: mainly enriched in sphingolipid metabolism, oligosaccharide metabolism, and response to growth hormone; (ii) molecular functions: mainly enriched in DNA binding, galactosyltransferase activity, and enhancer binding; and (iii) cellular composition: no relevant functions were enriched (Figure 4(c)).
Te main pathways with a signifcant concentration of diferentially expressed genes between the G6h group and the M6h group were analyzed through Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment, and eight signaling pathways, including glycolysis/glycogen isomerization, starch and sucrose metabolism, carbon metabolism, and fructose and mannose metabolism, were obtained. Te main diferent pathways between the G2d group and the M2d group were three signaling pathways, namely, ketone body synthesis and degradation, methyl butyrate metabolism, and sphingolipid biosynthesis (Figure 4(d)).

Scraping Terapy Altered the Expression of Transcription Factors Functioning in Glycolysis and Ketone Body
Metabolism. We next sought to verify the results of mRNAseq and identify molecular mechanisms for the increase in CSA of MF with scraping therapy by analyzing relevant transcription factors of anabolic and catabolic pathways. As       uncovered, and we wanted to conduct a preliminary discussion in the next section. Te statistics for mRNA expressions are as follows in Table 6.

Scraping Terapy Advanced the Expression or Phosphorylation of Proteins Involved in Glucose Transport, Glycolysis, and AMPK/mTOR/4EBP1 Pathways Six Hours after
Treatment. Since diferentially expressed genes had been recorded, we then further analyzed the proteins involved in the pathways mentioned previously from the extractions of MF 6 hours after scraping through Western blot analysis. As indicated in Figures 6(a)-6(d), compared with the M group, scraping therapy resulted in an increased level of GLUT4 (P < 0.001), and it also led to increased expressions of HK2 and PFKM (P < 0.001 and P � 0.006, respectively), and a trend for increased PKM expression (P � 0.066) (which are three rate-limiting enzymes in glycolysis), as well as increased LDHA expression (P � 0.001) (which is the enzyme for the fnal step of glycolysis). In addition, this manipulation inhibited the phosphorylation of AMPKα (P � 0.012) and enhanced the phosphorylation of mTOR and 4EBP1 (P � 0.021 and P � 0.035, respectively). Taken together, scraping therapy can rapidly enhance glycolysis and regulate the AMPK/mTOR/4EBP1 pathway. Te statistics for protein expressions are as follows in Table 7.

Discussion
Scraping therapy, termed Gua Sha, which is an efective treatment modality of traditional Chinese medicine, has a wide range of applications and can treat more than 30   kinds of diseases associated with the nervous system, motor system, respiratory system, digestive system, cardiovascular system, and endocrine system. In this study, we observed the efect of scraping therapy on pathological changes of MF cells. Within 3 days after treatment, compared with the model groups, the disruption of MF cells in the scraping groups was alleviated; the infltration of infammatory factors was decreased, and the CSA of MF had an increase tendency, indicating that repair process of myocytes was accelerated. Skin temperature of rats after scraping increased  Proteins/Actin Figure 6: Te efects of scraping therapy on targeted protein expression or phosphorylation in MF 6 hours after manipulation. (a, b) Representative immunoblots showed the levels of targeted protein expression or phosphorylation in MF. (c, d) Quantitative analysis of Western blots for the levels of GLUT4, HK2, PFKM, PKM, LDHA, p-AMPKα, p-mTOR, and p-4EBP1 proteins in the three groups. Te expression level of each targeted protein was normalized with a Western blot antibody to actin. * P < 0.05, * * P < 0.01, and * * * P < 0.001 versus K; △P < 0.05, △△P < 0.01, and △△△P < 0.001 versus M. N � 3 per group.
immediately compared with the baseline, from which we can presumed that scraping therapy can temporarily increase skin surface temperature, promote local tissue microcirculation and energy metabolism, and regulate the microenvironment of the body in a short period, which is consistent with the fndings of Xu et al. [37]. Withdrawal threshold of rats' hind paws in the scraping group was elevated on the 2nd day after intervention compared with the model group, indirectly demonstrating that scraping therapy could alleviate symptoms of low back pain to some extent, and the efect may be related to the enhancement of local microcirculation and other unelaborated mechanisms of action [19].
Skeletal muscle has a strong regenerative capacity and can rapidly restore muscle strength even in the presence of extensive muscle fber necrosis, which is closely related to the function of satellite cells (SCs), a class of skeletal muscle stem cells with the ability to promote muscle growth and tissue repair, located in the niche between the myofbril and basement membrane on the surface of the muscle fber [22,38]. Under normal conditions, SCs are in mitotic quiescence (G0 phase), and once the microenvironment in which SCs are located changes, such as when muscle tissue injury occurs, they exit G0 phase and instead begin to proliferate rapidly and migrate to the targeted site, fusing with damaged myogenic fbers to regenerate and restore their function [38][39][40]. In addition to transcriptional regulation, SC metabolic regulation is another key regulatory mechanism. A growing amount of evidence suggested that SC function is largely dependent on two metabolic states of the cells: oxidation and glycolysis [23]. Glycolysis is the process of conversion from one molecule of glucose to two molecules of pyruvate along with two molecules of adenosine triphosphate (ATP) in the presence of multiple enzymes. While in an oxygen-rich environment, cells can use oxidative phosphorylation (OXPHOS) to produce ATP more efciently by oxidizing pyruvate to acetyl-coenzyme A (acetyl-CoA) followed by productions of carbon dioxide and water through the tricarboxylic acid cycle (TCA cycle) that occurs in the mitochondria, producing an average of 34 molecules of ATP per molecule of glucose. In contrast, glycolysis seems to be an inefcient energy-producing metabolic pathway, but it can play an important role in rapid provision of ATP and the synthesis of key cellular macromolecules [41][42][43]. Ryall et al. [24] suggested that skeletal muscle stem cells shift their metabolic processes from original fatty acid oxidation to a glycolytic-based metabolic pattern during the transition from quiescence to proliferative phase. Chen et al. [44] used a YY1knockdown mouse model to investigate how SC uses glycolytic metabolism to activate its own proliferative and diferentiation functions. Te researchers found that muscle regeneration was severely reduced after acute injury in mice, and SC showed defective autonomic activation and proliferative capacity. Transcriptome analysis revealed that YY1 could inhibit mitochondrial gene expression and activate Hif1α-mediated glycolytic gene expression, whereas YY1 gene deletion leads to upregulation of mitochondrial gene expression and inhibition of glycolysis, resulting in defective SC activation and proliferative capacity, demonstrating that Glucose is transported into cells in two main ways: one is by energy-consuming counter-concentration gradient cotransport with Na + , which occurs mainly in renal tubular epithelial cells, and small intestinal mucosal cells; the other way is through glucose transporter proteins on cell membrane, which is a non-energy-consumingcis-concentration gradient transport process [45]. Tere are fve isoforms of glucose transporter proteins, of which facilitated glucose transporter member 4 (GLUT4), encoded by solute carrier family 2 member 4 gene (Slc2a4), is mainly found in adipose and muscle tissues and encodes glucose transport process, which is a rate-limiting step in glucose metabolism by skeletal muscle cells using glucose [46]. Hexokinase (HK) is the rate-limiting enzyme of glycolysis, which catalyzes the production of glucose 6-phosphate (G-6-P) from glucose and this reaction is irreversible. HK has 4 subtypes, among which hexokinase 2 (HK2) is mainly present in skeletal muscle cells and is a key enzyme promoting glycolysis of skeletal muscle cells [47]. Fructokinase 6-phosphofructokinase 1 (PFK1) catalyzes the production of fructose 1,6-diphosphate from fructose 6-phosphate, which is the second rate-limiting enzyme in the glycolytic pathway.
Tere are three isozymes, and we analyzed PFKM (phosphofructokinase, muscle) in our study because the specimens used were MF tissues. Te third rate-limiting enzyme in the glycolytic pathway, pyruvate kinase (PK), catalyzes the phosphoenolated pyruvate to generate ATP and consists of M-type and L-type isoenzyme. PKM was the one we tested as it is mainly distributed in skeletal and cardiac muscles. In the current experiment, gene expression levels of Hk2, Pfkm, and Slc2a4 increased within 2 days except for temporary decline of the latter two gene expression on the 1st day after scraping therapy, indicating that glycolysis was generally enhanced within 2 days after treatment, especially 6 hours after manipulation. Tis is consistent with the signifcant increase of protein expression levels of GLUT4, HK2, PFKM, PKM, and LDHA which are all involved in glycolysis. AMPK is composed of the catalytic α subunit and the regulatory β and c subunits, and the α subunit has two isoforms, α1 and α2. AMPKα1 is widely present in diferent tissues and is encoded by the Prkaa1 gene [48]. Fu   found that AMPKα1 has an important role in muscle regenerative capacity, which can activate SC aerobic glycolysis through nonclassical sonic hedgehog (Shh) signaling pathway, thus promoting SC activation and proliferation. Mammalian target of rapamycin (mTOR) likewise plays an important role in life activities of organism, participating in various metabolic processes such as protein translation, lipid synthesis, inhibition of autophagy, and cell cycle regulation [50], which can be inhibited by AMPK. Eukaryotic translation initiation factor 4E (eIF4E)-binding protein 1 (4EBP1) is one of the important target proteins downstream of mTOR, which is phosphorylated and dissociated from eIF4E. Te free eIF4E exposes active binding sites and binds to other translation initiation factors, thus promoting the translation of mRNA and increasing protein synthesis, which is an indispensable and critical component of cell proliferation [51]. In our study, Prkaa1 gene expression decreased within 3 days compared with the model groups, except for the temporary increase on the 1st day after scraping therapy, implying that the AMPK activity was relatively inhibited, especially 6 hours after treatment.
Western blot results showed that 6 hours after scraping, compared with the model group, the AMPK phosphorylation level decreased and mTOR and 4EBP1 phosphorylation levels increased in the scraping group. Te regulation of this signaling pathway was benefcial to protein synthesis and skeletal muscle regeneration. Taken together, the AMPK/ mTOR/4EBP1 pathway was regulated by scraping therapy in a short period. While AMPK can activate glycolytic pathway, when AMPK activity is inhibited, myocytes are prevented from appearing to overproliferate similar to growing mode of cancer cells due to excessive elevation of the glycolytic activity.
Ketone bodies are small molecules synthesized in the liver primarily from fat in the presence of fasting, prolonged exercise, or inadequate carbohydrate intake [52]. Tey circulate in the blood, with the ability to cross blood-brain barrier, and are taken up by energy-demanding peripheral tissues, where they are oxidized in mitochondria to produce acetyl-CoA, which drives the regeneration of ATP [53]. Te two major endogenous ketone bodies are acetoacetate (AcAc) and β-hydroxybutyrate (BHB), respectively, with BHB more abundant and stable in blood. In the presence of catalytic BDH1 (D-β-hydroxybutyrate dehydrogenase, mitochondrial), AcAc is converted to BHB in the liver and the reaction is reversed in peripheral tissues, accounting for the fnal step in ketone synthesis in the liver and the frst step in ketone catabolism in peripheral tissues, respectively [54]. Molecular functions of ketone bodies include providing energy, reducing infammation and oxidative stress, promoting muscle regeneration, and participating in lipid metabolism. AcAc has multiple signaling activities distinct from BHB, including binding to GPR43 to regulate lipid metabolism [55] and the activation of MEK1-ERK1/2 cell cycle protein D1 signaling pathway to accelerate muscle cell proliferation and muscle regeneration [56]. In skeletal muscle tissue, a rise in Bdh1 expression is associated with an increase in the conversion of BHB to AcAc and an increase in ketone body catabolic energy supply. Bdh1 gene expression was signifcantly enhanced within 2 days after scraping and decreased on the 3rd day, indicating that the modality promoted energy supply produced by ketone metabolism, which however need to be further verifed.

Conclusions
In summary, scraping therapy showed a clear therapeutic efect on injured multifdus and could alleviate pain symptoms to some extent in rats. Te potential mechanisms of its efcacy that we demonstrated were as follows: (i) regulating the GLUT4/glycolytic pathway and activating myosatellite cell proliferation and (ii) regulating the AMPK/ mTOR/4EBP1 signaling pathway and promoting protein synthesis (Figure 7). At a later stage, in vivo experiments including gene knockdown and overexpression techniques need to be used to further validate the role of GLUT4/ glycolytic and AMPK/mTOR/4EBP1 signaling pathways in promoting myocytes repair and regeneration, as well as regulatory mode of scraping therapy.

Data Availability
All the data generated or analyzed during this study are available upon reasonable request from the corresponding author.

Disclosure
Bin Zou, Juan Du, and Qiwen Xuan are co-frst authors.

Conflicts of Interest
Te authors declare that they have no conficts of interest.