Untargeted metabolomics analysis of plasma metabolic characteristics in patients with acne and insulin resistance

Acne vulgaris is a chronic inflammatory disease with high incidence, diverse clinical manifestations, poor clinical efficacy, and easy recurrence. Recent studies have found that the occurrence of acne is related to metabolic factors such as insulin resistance; however, the specific mechanism of action remains unclear. This study aimed to identify significantly different metabolites and related metabolic pathways in the serum of acne vulgaris patients with or without insulin resistance. LC–MS/MS was used to analyze serum samples from patients about acne with insulin resistance (n = 51) and acne without insulin resistance (n = 69) to identify significant metabolites and metabolic pathways. In this study, 18 significant differential metabolites were screened for the first time. In the positive-ion mode, the upregulated substances were creatine, sarcosine, D-proline, uracil, Phe–Phe, L-pipecolic acid, and DL-phenylalanine; the downregulated substances were tridecanoic acid (tridecylic acid), L-lysine, cyclohexylamine, sphingomyelin (d18:1/18:0), gamma-L-Glu-epsilon-L-Lys, and 1,2-dioleoyl-sn-glycero-3-phosphatidylcholine. In the negative-ion mode, the upregulated substance was cholesterol sulfate, and the downregulated substances were D(-)-beta-hydroxybutyric acid, myristic acid, D-galacturonic acid, and dihydrothymine. Cholesterol sulfate showed the most significant expression among all differential metabolites (VIP = 7.3411). Based on the KEGG database, necroptosis and ABC transporters were the most significantly enriched metabolic pathways in this experiment. The differential metabolites and pathways identified in this study may provide new possibilities for the clinical diagnosis and development of targeted drugs for acne patients with insulin resistance.


Introduction
Acne vulgaris is a chronic inflammatory skin disease involving multiple factors, mainly the pilosebaceous glands (Skroza et al. 2018), that often occurs on the face, chest, and back.Acne typically manifests as comedones in the early stage, which may then evolve into papules, pustules, nodules, cysts, and even scars (Bernardis et al. 2020).Traditionally, four distinct processes have been thought to play key roles: increased sebaceous secretion, hyperkeratosis of the hair follicle and sebaceous ducts, microbial colonization of the hair follicle, and immune responses to infection (Hazarika 2021).However, the pathogenesis of acne vulgaris is still not well understood.
Insulin resistance (IR) refers to a decrease in glucose uptake and utilization efficiency of body organs, such as the liver and muscle, under the action of various internal and external factors.To maintain the stability of blood glucose, excessive insulin secretion results in hyperinsulinemia.The homeostatic model assessment for IR (HOMA-IR) is often used in clinical practice and academic fields to evaluate IR (Gonzalez-Saldivar et al. 2017).
In acne vulgaris, sebaceous lesions are thought to occur after abnormal desquamation of keratinocytes within the sebaceous hair follicles (Kircik 2016).Nonunion erosion and accumulation of granulation tissue often occur in some acne cysts (Bernardis et al. 2020).Studies have shown that the physiological secretion of insulin/IGF-1 modulates keratinocyte differentiation, proliferation, and migration in healthy subjects (Zhang et al. 2016), and can also regulate the growth of fibroblasts and promote the proliferation of granulation tissue through the Grb2-ERK1/2 and PI3K-Akt signaling pathways (Takahashi et al. 1997).Therefore, hyperinsulinemia often causes a variety of skin manifestations, and acne is another important clinical feature of insulin resistance-hyperandrogenemia (Nagpal et al. 2016;Slayden et al. 2001).
Acne was originally thought to be a self-limiting disease of adolescence.More than 85% of teenagers suffer from acne, and this may be due to "physiological IR" during adolescence (Pektas et al. 2020).Some patients achieve remission or even a complete cure of acne in late adolescence, which may be due to a decrease in insulin/insulin-like growth factor (IGF-1) levels in the body (Del et al. 2012;Kumari and Thappa 2013).However, some people still have recurrent acne, even after the age of 25 years.This may be related to the high levels of androgen and pathological IR in adults (Del et al. 2012;Kumari and Thappa 2013).A study of the association between acne and IR in 100 adult men with acne found a significantly higher prevalence of IR than in controls (Nagpal et al. 2016).In women, especially those with polycystic ovary syndrome (PCOS), acne is more closely related to insulin levels (Nagpal et al. 2016).However, there are no reports on the mechanisms of action of acne vulgaris and IR.
Metabolomics is a tool used to study the types, quantities, and changes in metabolites caused by external stimuli, pathophysiological changes, and genetic mutations in organisms (Blazenovic et al. 2018).Untargeted metabolomics aims to obtain a comprehensive profile of all measurable small molecules in a given sample, including unknown analytes (Blazenovic et al. 2018).Metabolomics can be used in clinical applications, such as the etiology, pathogenesis, and clinical diagnosis of diseases.For example, LC-MS/ MS-targeted metabolomics has been used to analyze differences in metabolic characteristics between acne patients and healthy controls, which may provide new ideas for the treatment of acne (Yu et al. 2022).
Therefore, in this study, we performed the first non-target metabolomic analysis of serum samples from patients about acne with IR (n = 51) and acne without IR (n = 69).We aimed to explore the role of biological pathways involved in differential metabolites in patients about acne with IR, to better understand the biological mechanism of acne, and to provide new possibilities for clinical diagnosis and targeted drug development in patients with acne and IR.

Participants
Patients with acne vulgaris were selected from December 2019 to October 2020 in the dermatology outpatient department of the Affiliated Hospital of Southwest Medical University of Sichuan Province.The inclusion criteria included: (1) meeting the diagnostic criteria of acne vulgaris in Guideline for diagnosis and treatment of acne (Ju 2019); (2) patients or their guardians can sign and fully understand the content of the informed consent form; (3) between 14 and 41 years old, regardless of gender and disease duration; (4) no factors included in the exclusion criteria.Exclusion criteria included: (1) other types of acne or other skin and accessory diseases, such as androgenetic alopecia, acanthosis nigricans, and psoriasis; (2) oral contraceptives, isotretinoin, metformin, and other drugs that affect insulin or androgen metabolism are in use or have been used in the past 3 months; (3) pregnant and lactating patients or patients with PCOS or adrenal hyperplasia; (4) patients with mental disorders, liver and kidney dysfunction, or malignant tumors; and (5) patients with diabetes or other endocrine diseases.IR was determined as follows: HOMA-IR ≥ 2.14 [HOMA-IR = FBG(mU/L) * FINS(mmol/L) / 22.5] (Vora et al. 2008).Enrolled patients with acne were divided into two groups: the control group-those with IR (n = 51) and the test group-those without IR (n = 69).The study was approved by the Ethics Committee of the Affiliated Hospital of Southwest Medical University (No. KY2019040).

Collection of peripheral venous blood
Two tubes of fasting venous blood, 7-8 mL each, were drawn from patients after fasting for 8 h.After one tube of blood was centrifuged at 3000 RPM for 10 min at 4 °C, the supernatant was collected in EP tubes, numbered, and stored at -80 °C for LC-MS/MS analysis.Another blood tube was used to detect fasting glucose (FBG) and fasting insulin (FINS) levels using an automatic blood biochemical analyzer.

LC-MS/MS analysis
Analyses were performed using a UHPLC (1290 Infinity LC, Agilent Technologies) coupled to a quadrupole timeof-flight (AB Sciex TripleTOF 6600) at Shanghai Applied Protein Technology Co., Ltd.(Shanghai, China).For HILIC separation, samples were analyzed using a 2.1 mm × 100 mm ACQUITY UPLC BEH 1.7 µm column (Waters, Ireland).In both ESI positive and negative modes, the mobile phase contained 25 mM ammonium acetate and 25 mM ammonium hydroxide in water (B = acetonitrile).The gradient was 85% B for 1 min and was linearly reduced to 65% in 11 min, reduced to 40% in 0.1 min and kept for 4 min, and then increased to 85% in 0.1 min, with a 5 min re-equilibration period.For RPLC separation, a 2.1 mm × 100 mm ACQUITY UPLC HSS T3 1.8 µm column (waters) was used.In the ESI positive mode, the mobile phase contained A = water with 0.1% formic acid and B = acetonitrile with 0.1% formic acid, whereas in the ESI negative mode, the mobile phase contained A = 0.5 mM ammonium fluoride in water and acetonitrile (B).The gradient was 1% B for 1.5 min and was linearly increased to 99% in 11.5 min and kept for 3.5 min.Then, it was reduced to 1% in 0.1 min, and a 3.4 min of re-equilibration period was employed.The gradients were set at a flow rate of 0.3 mL/min, and the column temperatures were kept constant at 25 °C.A 2 µL aliquot of each sample was injected.The ESI source conditions were set as follows: ion source Gas1 (Gas1) as 60, ion source Gas2 (Gas2) as 60, curtain gas (CUR) as 30, source temperature at 600 °C, and ion spray voltage floating (ISVF) ± 5500 V.In MS acquisition, the instrument was set to acquire over the m/z range 60-1000 Da, and the accumulation time for the TOF MS scan was set at 0.20 s/spectra.In the auto MS/ MS acquisition, the instrument was set to acquire over the m/z range of 25-1000 Da, and the accumulation time for the product ion scan was set at 0.05 s/spectra.The product ion scan was acquired using information-dependent acquisition (IDA) with a high-sensitivity mode.The parameters were set as follows: the collision energy (CE) was fixed at 35 V with ± 15 eV; declustering potential (DP), 60 V ( +), and − 60 V ( −), excluding isotopes within 4 Da; candidate ions to monitor per cycle: 10.

Metabolomics data analysis
The raw data were converted to mzXML files using Proteo Wizard MSconvent and processed using XCMS for feature detection, retention time correction and alignment.The metabolites were identified by accuracy mass (< 25 ppm) and secondary mass spectrometry data which were matched with standard database (Shanghai Applied Protein Technology Co., Ltd.).In the extracted ion features, only the variables having more than 50% of the nonzero measurement values in at least one group were kept.For the multivariate statistical analysis, the MetaboAnalyst (www.metab oanal yst.ca) web-based system was used.After the Pareto scaling, principal component analysis (PCA) and orthogonal partial least-squares-discriminant analysis (OPLS-DA) were performed.The leave-one-out cross-validation and response permutation testing were used to evaluate the robustness of the model.The significant different metabolites were determined based on the combination of a statistically significant threshold of variable influence on projection (VIP) values obtained from OPLS-DA model and two-tailed Student's t test (p value) on the raw data, and the metabolites with VIP values > 1.0 and p values < 0.1 were considered as significant.The Kyoto Encyclopedia of Genes and Genomes (KEGG; http:// www.genome.jp/ kegg/) database was applied to pathway analysis.

Statistical analysis
Quantitative demographic and clinical data with normal distribution were expressed as mean ± standard deviation (SD) and analyzed using a t test.SPSS software was used for analyses.A two-tailed p value of < 0.05 indicates a statistically significant difference.

Comparison of baseline data between the two groups
In this study, there were 51 patients (25 men and 26 women) with acne combined with IR, with a prevalence of 42.5%, and there were 69 patients without IR, including 40 women and 29 men.The average age, FBG, FINS, and HOMA-IR were statistically different between the two groups, but there was no significant difference in the sex composition ratio (p > 0.05) (Table 1).

Metabolite identification and attribution classification
By matching the metabolite structure identification with the database, 346 metabolites were identified in the two modes, of which 221 were identified in the positive-ion mode and 125 were identified in the negative-ion mode, all were classified and counted according to their chemical classification information.The results showed that there were 60 lipids and lipid-like molecules, accounting for 17.34% of the total; 58 organic acids and their derivatives, accounting for 16.76%; 27 organic heterocyclic compounds, accounting for 7.8%; 16 nucleosides, nucleotides, and analogs, accounting for 4.62%; 16 organic oxygen compounds, 4.62%; 14 benzene ring compounds, 4%; 9 organic nitrogen compounds, 2.6%; 6 carbohydrate polyketo compounds, 1.73%; 1 organic sulfur compound, 0.29%; and 139 unclassified compounds, accounting for 40.17% (Fig. 1).

Analysis of differences between groups
To retain the original information to the greatest extent possible, we used univariate statistical analysis and multidimensional statistical analysis to explore more potential information in the data, so as to more accurately screen out significant differential metabolites between groups.
Based on the univariate analysis, FC analysis was used to analyze the differences between all metabolites detected in the positive-and negative-ion modes.The differential metabolites with an FC > 1.5, FC < 0.67, and p value < 0.05 were visualized in the form of a volcano plot (Fig. 2A, B).Next, we performed PCA on the two sets of data, and the results showed that sample overlap was high (Fig. 3A,  B).To maximize the difference between groups and better screen for differential metabolites, OPLS-DA analysis was performed, which can better distinguish the two groups of samples (Fig. 4A, B).After seven cycles of cross-validation, Q 2 = 0.144 in the positive-ion mode and Q 2 = -0.0923 in the negative-ion mode were obtained.We then performed a permutation test on the model again, and the permutation test chart showed that the original model did not overfit (Fig. 4C, D).

Bioinformatics analysis of metabolic differences
The VIP values of each metabolite were obtained by the OPLS-DA model analysis of the two groups of samples.According to the VIP value (VIP > 1 and p value < 0.05), 18 metabolites with biological significance were screened out.In the positive-ion mode, the upregulated substances were creatine, sarcosine, D-proline, uracil, Phe-Phe, L-pipecolic acid, and DL-phenylalanine; the downregulated substances were tridecanoic acid (tridecylic acid), L-lysine, cyclohexylamine, sphingomyelin (d18:1/18:0), epsilon-(gamma-Lglutamyl)-L-lysine, and 1,2-dioleoyl-sn-glycero-3-phosphatidylcholine.In the negative-ion mode, the upregulated substance was cholesterol sulfate, and the downregulated substances were D(-)-beta-hydroxybutyric acid, myristic acid, d-galacturonic acid, and dihydrothymine.Cholesterol sulfate showed the most significant expression among all differential metabolites (VIP = 7.3411).A bar chart was used to analyze the significant differential metabolites and fold changes between the groups.Compared to the control group, the expression of six substances, such as tridecanoic acid and L-lysine, was downregulated in the positive-ion mode.In the negative-ion mode, only cholesterol sulfate expression showed an upward trend (Table 2, Fig. 5A, B).
To explore the functional relevance or positive/negative correlation between the significantly different metabolites, correlation analysis was performed, and the results showed that there was a significant positive correlation between creatine and sarcosine in the positive-ion mode.L-lysine was positively correlated with epsilon-(gamma-L-glutamyl)-L-lysine, and sphingomyelin (d18:1/18:0) with 1,2-dioleoyl-sn-glycero-3-phosphatidylcholine.The remaining significant metabolites were negatively or not correlated with each other the negative-ion mode, and d-galacturonic acid was positively correlated with dihydrothymine.Myristic acid was positively correlated with d-galacturonic acid and dihydrothymine levels.Cholesteryl sulfate was negatively correlated with myristic acid, d-galacturonic acid, and dihydrothymine.Other significant metabolic differences had less correlation (Fig. 6A, B).

KEGG differential metabolic pathway analysis
The significant differential metabolites screened under the positive-and negative-ion models were merged, and the above metabolites were annotated using the KEGG database (Kyoto Encyclopedia of Genes and Genomes; http:// www.kegg.jp/), with a total of 24 annotated pathways.Fisher's exact test was performed on the above metabolic pathways and their Rich factor levels were evaluated.There were eight pathways with high enrichment, among which the p values of arginine and proline metabolism were the lowest (p < 0.05).The necroptosis pathway had the largest enrichment factor (0.1000) and contained the most significant differential metabolites.To screen out the metabolic pathway of the overall change by differences in abundance and chart analysis, in article 8 metabolic pathways, the experimental group was compared to the control group: glycine, serine, and threonine metabolism, and arginine and proline metabolism showed a trend of increase; sphingolipid signaling pathway and necrotizing apoptosis expression showed a trend of cut; the expression of ABC transporter cut trend was not obvious; the expression of protein digestion and absorption, lysine degradation, and aminoacyl-tRNA biosynthesis showed no up-or down-trend (Fig. 7A, B).

Discussion
In this study, we included 120 samples, including acne with IR (n = 51) and acne without IR (n = 69).By LC-MS/ MS analysis, 18 significantly differential metabolites were identified.In the positive-ion mode, the upregulated substances were creatine, sarcosine, D-proline, uracil, Phe-Phe, L-pipecolic acid, and DL-phenylalanine.The downregulated substances were tridecanoic acid  the correlation coefficient, that is, the higher the degree of positive or negative correlation, the darker the color.The size of the dot is related to the significance of the correlation; the more significant, the smaller the p value and the larger the dot Fig. 7 KEGG differential pathway analysis.A KEGG enrichment pathway map (bubble plot).Each bubble in the figure represents a metabolic pathway (the top 20 with the highest significance were selected according to the P value).The horizontal axis of the bubble and the size of the bubble represent the size of the influencing factor of the pathway in the topological analysis, and the larger the size, the larger the influencing factor.The vertical coordinate of the bubble location and the color of the bubble represent the P value of enrichment analysis (taking the negative common logarithm, that is, -log10 p value).The darker the color, the smaller the P value, the more significant the enrichment degree.B Differential abundance score plots for all differential metabolic pathways.The Y-axis represents the name of the differential pathway, and the coordinates on the X-axis represent the DA score.DA score is the overall total change of all metabolites in the metabolic pathway.A score of 1 indicates that all identified metabolites in this pathway tend to be upregulated and -1 all identified metabolites in this pathway tend to be downregulated.The length of the line segment indicates the absolute value of DA score, the size of the dot at the end of the line segment indicates the number of metabolites in the pathway, and the larger the dot indicates the more metabolites.The color of the line segments and dots is proportional to the DA score value.The darker the red is, the more inclined the overall expression of the pathway is to be upregulated, and the darker the blue is, the more inclined the overall expression of the pathway is to be downregulated (tridecylic acid), L-lysine, cyclohexylamine, sphingomyelin (d18:1/18:0), epsilon-(gamma-L-glutamyl)-L-lysine, and 1,2-dioleoyl-sn-glycero-3-phosphatidylcholine.In the negative-ion mode, the upregulated substance was cholesterol sulfate, and the downregulated substances were D(-)-beta-hydroxybutyric acid, myristic acid, D-galacturonic acid, and dihydrothymine.Cholesterol sulfate showed the most significant expression among all differential metabolites (VIP = 7.3411).Based on the KEGG database, necroptosis and ABC transporters were the most significantly enriched metabolic pathways in this experiment.
The pathogenesis of acne and IR is not fully understood, and current studies have found that insulin/IGF-1 may lead to the occurrence and development of acne by directly or indirectly regulating androgen levels (Melnik and Schmitz 2009).When the body becomes tolerant to insulin and the blood insulin level increases, insulin/IGF-1 can activate the PI3K/Akt or AMPK pathways, thereby increasing the androgen content in the body (Fan et al. 2007).High levels of androgen binding to the androgen receptor (AR) activate Wnt/β-catenin and mammalian target of rapamycin 1 (mTORC1) signaling pathways.It induces the differentiation of cortical cells and promotes adipogenesis and sebaceous duct obstruction, thereby inducing acne (Melnik 2014).In addition, IGF-1 can promote the production of inflammatory factors by activating the NF-κB pathway, which causes skin inflammation (Cong et al. 2019).In addition to insulin/IGF-1, AR play a key role in the development of acne.AR can bind to a variety of regulatory factors through its domains or ligands to exert its biological functions and is an important hub in signaling pathways (van de Wijngaart et al. 2012).When AR function is decreased, the effect of androgens on the skin and accessory organs will also be inhibited (van de Wijngaart et al. 2012).
Cholesterol sulfate (CS) was the most significant differential metabolite in the negative-ion mode found in this study, and it has endogenous steroidogenic potential (Fritsch et al. 2001).Steroidogenesis in human skin can begin with cholesterol (Slominski et al. 2004;Thiboutot et al. 2003).Cholesterol sulfur transferase forms cholesterol sulfate in the basal and spinous layers of the epidermis and is converted to cholesterol by steroid sulfatase in the stratum corneum (Elias et al. 1984;Epstein et al. 1984).The most important physiological roles of CS are keratinocyte differentiation and epidermal barrier development (Jetten et al. 1989).Increased CS levels lead to retention hyperkeratosis, which is a factor in the development of acne (Elias et al. 1984;Williams and Elias 1981).In addition, cholesterol-sulfur transferase is stimulated by peroxisome proliferator-activated receptor (PPAR) activators in human keratinocytes (Jiang et al. 2005), which is known to increase sebum production (Trivedi et al. 2006).Moreover, PPAR ligands enhance androgen activation and lipid synthesis in human sebocytes (Makrantonaki and Zouboulis 2007).It has also been shown that CS can alleviate diabetes in STZ-induced diabetic mice by increasing pancreatic β-cell number and function, which provides a physiological pathway to preserve islet viability and prevent diabetes development (Zhang et al. 2022).In this study, cholesterol sulfate was significantly higher in acne with IR group than in the acne alone group, suggesting the possibility of cholesterol sulfate as a biomarker of insulin resistance in acne patients.
Creatine and sarcosine were the most significantly differentially expressed metabolites in the positive-ion mode, and creatine supplementation has been shown to improve glycemic control and IR in both healthy and diabetic patients (Solis et al. 2021).Normally, insulin is secreted by pancreatic beta cells in response to energy substrates (e.g., glucose, fatty acids, and amino acids), hormones, and changes in energy demand (e.g., fasting-feeding cycles, exercise, and stress) to maintain euglycemic levels (Defronzo et al. 2015).However, IR may lead to β-cell failure, resulting in a progressive decrease in insulin secretion (Abdul-Ghani et al. 2006;Defronzo and Tripathy 2009).IR is often associated with inhibition of the PI3K pathway, with increased serine phosphorylation and inhibition of tyrosine phosphorylation of IRS proteins (Copps and White 2012).IRS protein degradation also appears to occur under certain conditions (Bouzakri et al. 2006).Higher plasma creatine concentrations were associated with an increased incidence of T2DM in a prospective cohort study including more than 4700 participants (Post et al. 2021).The authors suggested that higher extracellular creatine concentrations and lower intracellular phosphorylated creatine/creatine content may be associated with impaired intracellular energy status, suggesting mitochondrial dysfunction, a hypothesized mechanism involved in the pathophysiology of T2DM (Post et al. 2021).In this study, creatine and sarcosine levels in acne patients with IR were significantly higher than those in the acne group, which was consistent with the positive correlation between creatine concentration and IR.
The necroptosis pathway was the most significantly enriched metabolic pathway in this study, and necrosis is the most typical form of programmed necrosis and is characterized by apoptosis and necrosis (Tang et al. 2019).Necrostatin-1 (NEC-1) is a necrosis inhibitor.NEC-1 improves cardiac function in prediabetic rats by directly inhibiting necrosis and myocardial mitochondrial dysfunction and increasing mitochondrial fusion independent of peripheral metabolic function (Apaijai et al. 2021).However, NEC-1 treatment directly improves cognitive function in prediabetic patients by inhibiting necrosis and reducing brain inflammation and mitochondrial dysfunction but does not improve insulin sensitivity (Jinawong et al. 2020).Therefore, although necrotic metabolic pathways can influence diabetes, the relationship between them and IR is not fully understood.In our analysis, the necrotic metabolic pathway was the most significantly enriched, which may be related to the fact that the number of unclassified metabolites was large, accounting for 40.17%, which may be limited by the number of metabolites in the database or the inability of the current technology to identify relevant metabolites.The specific relationship between necrotic metabolic pathway and acne with insulin resistance may be our future research direction.
The ABC transporter metabolic pathway was another significant metabolic pathway enriched in this experiment, and is characterized by the use of ATP-hydrolyzed energy to transport specific compounds across cell membranes, including lipids, amino acids, carbohydrates, vitamins, ions, glucuronate conjugates, and xenobiotics (Dean et al. 2001).It has been shown that the metabolic pathways most closely related to patients with moderate-to-severe acne include ATP-binding cassette transporter (ABC) and sphingolipid signaling pathways (Yu et al. 2022).Meanwhile, other study has shown that ABCA1 is important for controlling monocytosis, secretion of adipose and hepatic macrophages, and insulin sensitivity when mice become obese and require substantial lipid buffering, raising the possibility of enhancing ABCA1 in hematopoietic cells as a potential target for the treatment of IR and atherosclerosis (Tang et al. 2016).In summary, the ABC transporter metabolic pathway can affect both acne and insulin resistance, and it is very promising to be a potential biomarker for acne patients with insulin resistance.

Limitations
(1) Both the test and control groups were classified on the basis of acne patients.There were highly correlated metabolites and metabolic pathways in the body itself, so few species with significant metabolic differences were obtained; (2) among the metabolites screened in the positive-and negative-ion modes of this experiment, the number of unclassified metabolites was large, accounting for 40.17%, which may be limited by the number of metabolites in the database or the inability of the current technology to identify relevant metabolites; (3) further experiments were not performed to verify the findings of the study, such as detecting the concentrations of differential metabolites in the two groups of patients by biochemistry assay or ELISA, analyzing their correlation with HOMA, and conducting ROC analysis.

Conclusion
In conclusion, the differential metabolites screened in this study, such as cholesterol sulfate and creatine, have a certain impact on the pathogenesis of acne with IR and may be potential biomarkers for clinical symptoms of acne with IR, which provides new ideas and methods for subsequent targeted therapy.Necroptosis and ABC transporters were the most significantly enriched metabolic pathways in this study, which may be important pathways involved in the pathogenesis of acne with IR.This may provide new possibilities for the clinical diagnosis of acne in patients with IR and the development of targeted drugs.

Fig. 1
Fig. 1 Proportion of identified metabolites in each chemical category

Fig. 3
Fig. 3 PCA map.A PCA map of positive-ion differential metabolites; B PCA map of negative-ion differential metabolites.In the figure, t[1] represents principal component 1, t[2] represents principal component 2, and the ellipse represents the 95% confidence interval.Dots

Fig. 4
Fig. 4 OPLS-DA map and permutation test A OPLS-DA map of positive-ion differential metabolites; B OPLS-DA map of negativeion differential metabolites.In the figure, t[1] represents principal component 1, to[1] represents principal component 2, and the ellipse represents the 95% confidence interval.Dots of the same color represent biological replicates within a group, and the distribution of dots reflects the degree of difference between and within groups.C OPLS-DA permutation test of positive-ion differential metabolites; D

Fig. 5 Fig. 6
Fig. 5 Fold change analysis of significant differential metabolites expression A Fold change analysis of significant differential metabolites expression of positive-ion differential metabolites; B Fold change analysis of significant differential metabolites expression of negative-ion differential metabolites.In the figure, the abscissa rep-

Table 1
General characteristics of the two groups IR Insulin resistance; FBG fasting glucose; FINS fasting insulin; HOMA-IR Homeostatic Model Assessment for Insulin Resistance