Transcriptome analysis of brown adipose tissue in Brandt’s vole treated with tannic acid under cold exposure

Abstract

Background

Tannic acid (TA) is a hydrolysable plant secondary metabolite known to influence multiple physiological processes in animals; however, its role in regulating brown adipose tissue (BAT) thermogenesis remains poorly understood. Notably, the overwinter food caches of Brandt’s voles predominantly consist of Artemisia species, which are rich in TA. This study aimed to determine whether TA contributes to cold tolerance in Brandt’s voles by activating BAT thermogenesis. Adult male voles were administered TA, after which the masses of BAT and inguinal white adipose tissue (iWAT) were measured, and temperature changes in BAT, the body surface, and the rectum were recorded following exposure to − 20 °C. In addition, transcriptomic analyses of BAT were performed, and the expression and protein levels of key thermogenic markers were assessed.

Results

The results showed that TA reduced iWAT mass while exerting minimal effects on BAT mass. TA-treated voles exhibited significantly elevated temperatures in BAT, the body surface, and the rectum after cold exposure. Histological analyses revealed that TA treatment reduced adipocyte area in iWAT while increasing the number of nuclei in brown adipocytes in BAT. In BAT, differentially expressed genes (DEGs) in voles receiving a low TA dose were significantly enriched in pathways related to fat digestion and absorption and peroxisome proliferator-activated receptor (PPAR) signaling. In contrast, DEGs in voles administered a high TA dose were predominantly associated with brown adipocyte differentiation and the upregulation of cold-induced thermogenesis. Moreover, TA administration increased the expression of FFAR4 and UCP1, as well as the protein levels of PGC-1α, PPARγ, and UCP1 following cold exposure.

Conclusions

Collectively, these findings demonstrate that TA enhances cold tolerance in Brandt’s voles by promoting thermogenic gene expression and stimulating brown adipocyte differentiation in BAT, providing novel insights into the role of plant secondary metabolites in mammalian cold adaptation and herbivore–plant interactions.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12864-026-13187-8.

Introduction

Survival of endothermic animals in cold environments requires efficient thermoregulatory mechanisms [1]. Non-shivering thermogenesis (NST)—the process of generating heat with minimal shivering—is a crucial mechanism for many small mammals (e.g., rodents) when exposed to low temperatures, as NST rapidly produces heat during acute cold exposure to enhance thermal tolerance without a corresponding high resting metabolic rate (RMR) [2–6]. In mammals, NST is predominantly mediated (approximately 60%) by brown adipose tissue (BAT), which is specialized for energy expenditure due to its abundance of mitochondria containing uncoupling protein 1 (UCP1). UCP1, a transmembrane protein comprising six α-helical domains, disrupts the mitochondrial proton gradient, thereby producing heat in lieu of ATP synthesis [7–9].

Non-shivering thermogenesis in BAT is regulated by multiple signaling pathways [10]. In BAT, adaptive thermogenesis is initiated upon cold stimulation or adrenergic receptor activation via the PGC-1α–UCP1 axis. Peroxisome proliferator-activated receptor gamma coactivator-1 alpha (PGC-1α) serves as a master transcriptional regulator of mitochondrial remodeling and biogenesis [11, 12]. Peroxisome proliferator-activated receptor gamma (PPARγ) collaborates with the coactivator PGC-1α to coordinate distinct transcriptional programs through interactions with transcription factors that modulate thermogenic gene expression [13]. Moreover, PPARγ plays a critical role in the differentiation of both brown and white adipocytes, thereby influencing the development of both white adipose tissue (WAT) and BAT [14]. AMPK (adenosine monophosphate-activated protein kinase) is a pivotal regulator of various metabolic pathways. It is highly expressed in BAT, where it facilitates non-shivering thermogenesis by activating PGC-1α and plays a crucial role in maintaining BAT mitochondrial function [10].

It is reported that many plant secondary metabolites (PSMs), such as resveratrol [10], curcumin [15], epigallocatechin gallate [16], and berberine [17] can stimulate the AMPK signaling pathway to activate BAT thermogenesis. Tannins, a group of phenolic PSMs commonly found in plants, are hypothesized to act as chemical deterrents against insect and vertebrate herbivores. Their adverse effects include interference with digestion and nutrient absorption, decreased food palatability, impaired kidney or liver function, weight loss, depletion of endogenous nitrogen, and even mortality [18–21]. Tannic acid (TA) is a form of hydrolysable tannin [19]. Recent studies have revealed that TA also exhibits antibacterial [22, 23], anti-inflammatory [24, 25], and antioxidant properties [26]. Additionally, it demonstrates physiological activities such as lowering blood glucose [27] and decreasing lipid levels [28], and improve hippocampal function and spatial memory and ultimately modulate food-hoarding behavior in Siberian chipmunk (Tamias sibiricus) [29]. However, it remains unknown whether TA, a phenolic PSM, influences the activation of thermogenesis in BAT.

Brandt’s vole (Lasiopodomys brandtii), a small herbivorous mammal, is a representative rodent species of temperate grasslands. It exhibits gregarious behavior and primarily inhabits the grasslands of Inner Mongolia, China [30–32]. In its winter habitat, surface temperatures may range from − 25 °C to − 42 °C, while the temperature inside its burrows remains around − 20 °C [33]. The survival strategies and tolerance mechanisms exhibited by the Brandt’s vole under low-temperature conditions have attracted considerable attention from researchers. Previous research indicates that the overwinter food caches of Brandt’s voles predominantly consist of plants from the genus Artemisia. Notably, Artemisia frigida and A. scoparia account for over 90% of the stored forage and are characterized by high concentrations of TA [34–37]. Accordingly, we speculate that TA may contribute to the cold tolerance exhibited by Brandt’s voles.

In the present study, adult male Brandt’s voles were administered graded doses of TA via gastric gavage over an 18-day period. We then examined variations in the mass of BAT and inguinal white adipose tissue (iWAT), as well as changes in the temperatures of BAT, the body surface, and the rectum following exposure to − 20 °C. Furthermore, gene expression changes were evaluated using RNA sequencing and RT-qPCR, and the protein expression levels of genes associated with BAT thermogenesis were assessed. This investigation aimed to elucidate the effects of TA on cold tolerance and its influence on the gene expression profile in BAT in Brandt’s voles. Our findings provide critical insights into the role of TA in cold tolerance in voles and improve our understanding of how plant secondary metabolites influence herbivore physiology and adaptation.

Materials and methods

Animals and treatments

The Brandt’s voles used in this experiment were captured on the Xilinhot Grassland in Inner Mongolia and subsequently housed and bred at the animal facility of the College of Bioscience and Technology, Yangzhou University. The voles were maintained in plastic cages measuring 22 cm × 18 cm × 15 cm, with wood-chip bedding lining the floor. They had ad libitum access to distilled water and were fed a standard rodent chow. The facility conditions were controlled with a relative humidity of 45–55%, a temperature of 20–24 °C, and a light-dark cycle of 16 h of light and 8 h of dark. At 21 days of age, the pups were weaned from their parents and housed separately. Ninety-day-old male Brandt’s voles, with body weights of 50 ± 5 g, were selected as experimental subjects. All procedures were carried out in accordance with the guidelines approved by the Animal Care and Use Committee of the Faculty of Veterinary Medicine at Yangzhou University (No. NSFC2020-SKXY-6).

At 10:00 AM daily, the animals underwent intragastric administration for 18 consecutive days. The experimental voles were randomly assigned to three groups (n = 8 per group): a control group, a low-dose group, and a high-dose group. Animals in these groups received distilled water, 600 mg·kg⁻¹ tannic acid (TA), and 1200 mg·kg⁻¹ TA by gavage, respectively. Under natural field conditions, the maximum daily tannin intake of pubertal Brandt’s voles has been estimated to be approximately 75 mg [38]. Based on the average body mass of the experimental animals (approximately 50 g), the low- and high-dose treatments corresponded to daily TA intakes of approximately 30 mg and 60 mg, respectively. Moreover, because adult voles generally have greater body mass and higher absolute food intake than pubertal voles, their potential daily tannin intake under field conditions may be even higher. Therefore, the TA doses used in this study are within a biologically relevant range of natural tannin exposure. During the gavage period, the animals had free access to distilled water and were provided with 15 g of food daily. This amount exceeded the normal daily food intake of adult Brandt’s voles under laboratory conditions and ensured that food was available ad libitum throughout the experiment. Body weight and food intake were recorded each day. Food intake was calculated by subtracting the remaining food from the amount provided the previous day. Following the 18-day gavage period, the voles were fasted for 24 h, weighed at 10:00 AM the following day, and then decapitated for blood collection and tissue sampling. Interscapular BAT and iWAT and were carefully excised, weighed, and recorded. Adipose tissue from the left side was fixed in 4% paraformaldehyde and stored at 4 °C. The adipose tissue index was calculated using the following formula: Adipose tissue index = (adipose tissue weight (g)) / (vole body weight (g)).

In a separate cohort of 24 Brandt’s voles (n = 8 per group) receiving the same treatment regimen, animals were fasted for 24 h after 18 days of gavage treatment and were then subjected to cold exposure at 10:00 a.m. on the following day. Rectal temperature was measured and recorded with a thermometer (UT321, Uni-Trend Technology(China)Co., Ltd, Dongguan, Guangdong, China) prior to cold exposure. The animals were subsequently placed individually in ventilated plastic boxes with small air holes and exposed to − 20 °C in a freezer for 1 h. Food was not provided during the cold-exposure period to ensure consistent experimental conditions among animals. Following cold exposure, rectal temperature was measured again, and infrared thermal images were captured. Immediately after imaging, the animals were euthanized by decapitation, and the iWAT and interscapular BAT were excised, rapidly frozen in liquid nitrogen, and stored at − 70 °C. Infrared thermal imaging was employed to measure the body surface temperature and BAT temperature of Brandt’s voles. The animals were placed in plastic boxes, and images were acquired with the camera positioned approximately 45 cm away. Thermal images captured when the animals were in a fully extended posture were analyzed using FLIR Tools software. Body surface temperature was calculated as the average of three body sites, whereas the highest temperature in the interscapular dorsal region was defined as the BAT temperature. Body surface temperature and BAT temperature were measured only after cold exposure. Baseline measurements of these parameters were not obtained prior to the cold challenge. All animals survived throughout the TA treatment and cold-exposure procedures, and no animals were excluded due to mortality.

H and E staining

The fixed tissues of interscapular BAT and iWAT were dehydrated in ethanol, embedded in paraffin, and sectioned at a thickness of 4 μm. Sections were stained with hematoxylin and eosin according to the manufacturer’s instructions. The adipocyte area in iWAT and the number of nuclei in BAT adipocytes from the H&E-stained images were quantified using ImageJ. For each sample, five white adipocytes in iWAT were randomly selected for measurement, and five random fields in interscapular BAT were analyzed to quantify the number of nuclei per field. Samples with damaged or unsuitable histological sections were excluded from the morphometric analysis.

Quantitative real-time polymerase chain reaction (qRT-PCR) analysis

The cDNA sequences of target genes obtained via RNA-seq were analyzed using the NCBI Primer–BLAST tool to design qRT-PCR primers (Table 1). Primer specificity was verified by melt curve analysis in qRT-PCR to confirm the absence of primer dimers and non-specific amplification products (Table 1). qRT-PCR reactions were conducted in a real-time PCR system (Applied Biosystems, Grand Island, NY, USA) using TB Green Premix Ex Taq II (Vazyme Biotech Co., Ltd., Nanjing, China). From each sample, 4 µL of diluted (250 ng µL^{− 1}) RNA was reverse-transcribed into cDNA using the PrimeScript™ 1st Strand cDNA Synthesis Kit (R323, Vazyme Biotech, Nanjing, China). Thermal cycling conditions for qRT-PCR were as follows: 95 °C for 30 s followed by 40 cycles of 95 °C for 10 s, 60 °C for 30 s, and 95 °C for 15 s, then 60 °C for 1 min, 95 °C for 30 s, and 60 °C for 15 s. Standard curves were constructed for each gene via serial 5-fold dilutions of cDNA. Amplification efficiency ranged between 0.9 and 1.0, confirming the validity of the comparative quantification method. No amplification was detected in the absence of template or non-RT control. The fold change of gene expression was calculated using the 2^−ΔΔCt method [39], using β-actin [40] as housekeeping gene.

Table 1.

Primers used in the qRT-PCR study

Primer	Forward(5’-3’)	Reverse(5’-3’)
β-actin	TTTGCGTGACATCAAAGAG	ATGCCAGAAGATTCCATACC
PDRG-1	GACACCAAGAGGAATCAGAA	CCCAAAGCAAACCATCAC
UCP1	ACAGTACCCGAGTGTACCCA	ATGACATTCCAGGACGCCAG
FFAR4	GTCACTGGTGTACCCTGCTT	TCCTTAATTTGCTGGCCCTG

Western blot analysis

Western blot analysis was applied to measure the protein level of UCP1, PGC-1α, PPARγ, p-AMPK and AMPK in the BAT, according to the method of Ye et al. [41], with minor modifications. Proteins extracted from the BAT were homogenized by Ultrasonicator (VCX 130 PB, Sonics and Materials Inc., Connecticut, USA) in a chilled radio immunoprecipitation assay lysis buffer with a protease inhibitor (Phenylmethanesulfonyl fluoride) (Beyotime Institute of Biotechnology, Shanghai, P.R. China). The protein content in the samples was determined using a BCA Protein Assay Kit (E112; Vazyme Biotech Co., Ltd., Nanjing, P.R. China). Then, 15 µL samples containing 50 µg of boiled protein (10 min at 95 °C) were separated in a 10% sodium dodecyl sulfate (SDS) – polyacrylamide gel by gel electrophoresis (Bio-Rad, Laboratories, Inc., Hercules, California, USA) and transferred to polyvinylidene difluoride transfer membranes (ImmobilonVR-P, catalogue No. IPVH 00010; Merk Millipore Ltd., Ireland). The molecular masses of the bands were evaluated by comparison with a color molecular mass marker (P0080, Beyotime Institute of Biotechnology, Shanghai, P.R. China) that was loaded and transferred simultaneously. The membranes containing UCP1(34 kDa: 23673-1-AP; Wuhan Sanying, Wuhan, China), PGC-1α (100 kDa: 66369-1-Ig; Wuhan Sanying, Wuhan, China), PPARγ (58 kDa: AF6284; Affinity bioscience LTD, Changzhou, China), p-AMPK (62 kDa: AF3423; Affinity bioscience LTD, Changzhou, China), AMPK(62 kDa: 10929-2-AP; Wuhan Sanying, Wuhan, China) and β-actin (42 kDa: bs-0061R; Bioss Biotechnology, Beijing, China) were cut separately, referring to the position of the marker. After blocking with 5% skim milk for 1 h, the membranes were respectively incubated in 1:10000 rabbit anti-AMPK, 1: 5000 rabbit anti-p-AMPK, 1:8000 rabbit anti-UCP1, 1:5000 rabbit anti-PPARγ, 1:50000 mouse anti-PGC-1α, and 1:5000 anti-β-actin overnight at 4 °C, according to the manufacturer’s protocols. After three washes with Tris-buffered saline with 0.1% Tween-20, the membranes were exposed to 1:5000 goat anti-rabbit horseradish peroxidase conjugated IgG (YFSA02; Yi Fei Xue Biotechnology, Nanjing, China) for UCP1, PPARγ, p-AMPK, AMPK and β-actin and goat anti-mouse horseradish peroxidase (HRP) conjugated IgG (YFSA01; Yi Fei Xue Biotechnology, Nanjing, China) for PGC-1α for 1 h at room temperature. The bands were visualized using the enhanced chemiluminescence detection method by exposing them for 2 min at room temperature in a 1:1 mixture of ECL reagents A and B (E411; Vazyme Biotech Co., Ltd., Nanjing, China). The chemiluminescence images were captured by GelCap software (Tanon, Shanghai, China) with 1–2 s exposure. The optical densities of the individual bands were measured with Tanon Gis image software version 4.2 (Tanon, Shanghai, China) in the Gel Imaging Analysis System (5200Multi; Tanon, Shanghai, China), referring to Fan et al. [42]. The expression of UCP1, PGC-1α, and PPARγ were all normalized to β-actin. p-AMPK was normalized to AMPK.

RNA-seq analysis

For RNA-seq analysis, five animals were randomly selected from each experimental group (n = 5 per group) for transcriptome sequencing. Total RNA (1 µg per sample) was used for library construction with the NEBNext^® Ultra™ RNA Library Prep Kit for Illumina^® (NEB, USA) according to the manufacturer’s protocol. Briefly, mRNA was enriched using poly-T oligo-attached magnetic beads, fragmented, and reverse-transcribed into cDNA. Following end repair, A-tailing, and adaptor ligation, fragments of approximately 240 bp were selected using the AMPure XP system (Beckman Coulter, USA). The libraries were then PCR-amplified and assessed for quality and insert size distribution using an Agilent 2100 Bioanalyzer. Library clustering was performed on a cBot Cluster Generation System, and sequencing was carried out on an Illumina platform to generate paired-end reads. Raw reads were processed to obtain clean reads by removing adaptor sequences, poly-N reads, and low-quality reads. Sequencing quality metrics, including Q20, Q30, GC content, and duplication levels, were subsequently calculated. De novo transcriptome assembly was conducted using Trinity [43] with a minimum k-mer coverage of 2 based on pooled left.fq and right.fq reads. Functional annotation of assembled transcripts was performed by comparison against the NR, Pfam, KOG/COG/eggNOG, Swiss-Prot, KEGG, and GO databases. Gene expression levels for each sample were quantified using RSEM [44]. The RNA sequencing data generated in this study have been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA1444008, with individual run accessions SRR37832804–SRR37832818 (https://www.ncbi.nlm.nih.gov/bioproject/PRJNA1444008). Differential expression analysis between groups was carried out using the DESeq R package (v1.10.1), which models count data based on the negative binomial distribution. Genes with a fold change ≥ 1.5 and a P-value < 0.05 were initially identified as differentially expressed. P-values were further adjusted using the Benjamini–Hochberg method to control the false discovery rate (FDR), and genes with an adjusted P-value < 0.05 and a fold change ≥ 1.5 were considered significantly differentially expressed. GO enrichment analysis of differentially expressed genes (DEGs) was performed using the topGO R package based on the Kolmogorov–Smirnov test, while KEGG pathway enrichment analysis was conducted using KOBAS software [45].

Protein–protein interaction (PPI) networks were predicted using the STRING database (http://string-db.org/) and visualized in Cytoscape (v3.10.2). Gene Set Enrichment Analysis (GSEA) was performed with gene sets derived from KEGG pathways as well as the Biological Process (BP), Cellular Component (CC), and Molecular Function (MF) categories of GO.

Statistical analysis

Differences in initial body mass, percentage changes in body mass, food intake, iWAT and interscapular BAT indices, number of nuclei in interscapular BAT adipocytes, area of iWAT cells, mRNA expression levels of PDRG-1, FFAR4, and UCP1, as well as protein levels of UCP1, PGC-1α, PPARγ, and the p-AMPK/AMPK ratio in brown adipose tissue were analyzed using one-way analysis of variance (ANOVA), followed by either LSD or Dunnett’s T3 post hoc tests depending on the assumptions of normality and homogeneity of variance. Absolute masses of iWAT and interscapular BAT were initially analyzed using one-way ANOVA and were subsequently reanalyzed using analysis of covariance (ANCOVA) with final body mass as a covariate to account for potential body mass effects. Surface temperature, rectal temperature, rectal temperature reduction ratio, and interscapular BAT temperature were also analyzed using ANCOVA with final body mass as a covariate, followed by LSD post hoc tests. Statistical significance was defined as a P-value less than 0.05. All ANOVA and ANCOVA analyses were conducted using IBM SPSS Statistics version 26 (IBM, Armonk, NY, USA).

Result

Effect of TA on body mass and food intake in Brandt’s vole

Initial body mass did not differ significantly among the control, low-dose, and high-dose groups (F₂, ₂₁ = 1.190, P = 0.324) (Fig. 1A). TA significantly affected percentage changes in body mass in Brandt’s voles (F₂, ₂₁ = 101.705, P < 0.001) (Fig. 1B). Compared with the control group, both the low- and high-dose groups exhibited lower percentage changes in body mass (P < 0.001 for both). TA also exerted a significant effect on food intake (F₂, ₂₁ = 9.602, P = 0.001), with food intake in the control group exceeding that of the low- and high-dose groups (P = 0.004 and P < 0.001, respectively) (Fig. 1C).

Fig. 1.

Effect of tannic acid (TA) on body mass and food intake in Brandt’s vole. A Initial body mass; B Percentage change in body mass; C Food intake. Data are presented as mean ± SEM (n = 8 per group). ** and *** indicate P < 0.01 and P < 0.001 vs. control group, respectively

Effect of TA on mass and morphology of iWAT and interscapular BAT in Brandt’s vole

TA significantly influenced iWAT mass and the iWAT mass index (F₂, ₂₁ = 27.831, P < 0.001; F₂, ₂₁ = 18.089, P < 0.001, respectively). Both parameters were lower in the low- and high-dose groups than in the control group (iWAT mass: P < 0.001 for both; iWAT index: P = 0.033 and P < 0.001, respectively) (Fig. 2A–B). To account for differences in body mass, ANCOVA was performed using final body mass as a covariate. After adjustment, the effect of TA on iWAT mass remained significant (F₂, ₂₀ = 6.160, P = 0.008), with higher iWAT mass in the control group than in the high-dose group (P = 0.008). In contrast, interscapular BAT mass was not significantly affected by TA (F₂, ₂₁ = 0.214, P = 0.809) (Fig. 2C). However, the BAT index differed significantly among groups (F₂, ₂₁ = 23.935, P < 0.001), with higher values observed in the low- and high-dose groups compared with the control group (P < 0.001 for both) (Fig. 2D). Furthermore, after adjustment for final body mass using ANCOVA, TA treatment significantly affected interscapular BAT mass (F₂, ₂₀ = 14.013, P < 0.001), with higher BAT mass in the low- and high-dose groups than in the control group (P < 0.001 for both). Histological analysis revealed that TA significantly reduced the area of BAT adipocytes (F₂, ₁₉ = 18.151, P < 0.001), with smaller adipocyte areas in the low- and high-dose groups relative to the control group (P = 0.032 and P = 0.010, respectively) (Fig. 3A). Conversely, TA significantly increased the number of nuclei in BAT adipocytes (F₂, ₁₇ = 7.269, P = 0.005), with higher nuclei numbers in the low- and high-dose groups compared with the control group (P = 0.010 and P = 0.002, respectively) (Fig. 3B).

Fig. 2.

Effect of tannic acid (TA) on inguinal white adipose tissue (iWAT) and interscapular brown adipose tissue (BAT) mass in Brandt’s vole. A iWAT mass; B iWAT mass index; C BAT mass; D BAT mass index. Data are presented as mean ± SEM (n = 8 per group). * and *** indicate P < 0.05 and P < 0.001 vs. control group, respectively

Fig. 3.

Effect of tannic acid (TA) on cellular morphology of inguinal white adipose tissue (iWAT) and interscapular brown adipose tissue (BAT) in Brandt’s vole. A iWAT; B BAT. Data are presented as mean ± SEM. * and ** indicate P < 0.05 and P < 0.01 vs. control group, respectively. For panel A, n = 7, 7, and 8; for panel B, n = 6, 7, and 7 (control, low-dose, and high-dose groups, respectively)

Effect of TA on rectal, body surface and BAT temperatures after cold exposure

According to ANCOVA, rectal temperature prior to cold exposure did not differ among groups (F₂, ₂₀ = 0.457, P = 0.640) (Fig. 4A). However, rectal temperature after cold exposure was significantly influenced by TA (F₂, ₂₀ = 4.171, P = 0.031), with higher temperatures observed in the low- and high-dose groups compared with the control group (P = 0.009 and P = 0.038, respectively) (Fig. 4B). In contrast, the rectal temperature reduction ratio did not differ significantly among groups (F₂, ₂₀ = 2.714, P = 0.091) (Fig. 4C). TA significantly affected body surface temperature after cold exposure (F₂, ₂₀ = 5.579, P = 0.012). Body surface temperatures in the low- and high-dose groups were significantly higher than those in the control group (P = 0.010 and P = 0.004, respectively) (Fig. 4D). TA also significantly increased BAT temperature after cold exposure (F₂, ₂₀ = 5.427, P = 0.013), with higher BAT temperatures in the low- and high-dose groups compared with the control group (P = 0.037 and P = 0.004, respectively) (Fig. 4E–F).

Fig. 4.

Effect of tannic acid (TA) on body surface, rectal, and brown adipose tissue (BAT) temperatures in Brandt’s vole. A Rectal temperature before cold exposure; B Rectal temperature after cold exposure; C Rectal temperature reduction ratio; D Body surface temperature after cold exposure; E BAT temperature after cold exposure; F Representative thermal infrared images of Brandt’s voles. Data are presented as mean ± SEM (n = 8 per group). * and ** indicate P < 0.05 and P < 0.01 vs. control group, respectively

Effect of TA on gene expression in BAT after cold exposure

RNA sequencing was performed on 15 BAT samples, generating 94 Gb of high-quality data, with Q30 scores exceeding 93.38% for all samples. Clean reads were mapped to the de novo assembled transcriptome, yielding alignment rates of 89.37–92.36% (Table 2), and a total of 37,015 unigenes were identified. Differential expression analysis after cold exposure identified 162 DEGs between the control and low-dose groups and 377 DEGs between the control and high-dose groups, resulting in 487 unique DEGs in total (Fig. 5A–B). Fifty-two DEGs were shared by both TA-treated groups relative to the control group (Fig. 5B). Compared with the control group, the low-dose group exhibited 59 upregulated and 103 downregulated genes, whereas the high-dose group showed 169 upregulated and 208 downregulated genes (Fig. 6A–B). The top 10 DEGs between the control and low-dose groups were PAK1IP1, LOC101392181, TC2N, PI16, CABLES1, APOA4, SLC22A17, PAL_GLEAN10011656, FD755_025127, and HES1 (Fig. 6A). The top 10 DEGs between the control and high-dose groups were WNT5A, IKBKB, COL12A1, ADH6, CKMT1, CLINT1, MT-ND4, SP140, FADS2, and SPON1 (Fig. 6B). Among the shared DEGs, PDRG-1, FFAR4, and UCP1 were selected for further validation.

Table 2.

Summary of clean reads mapped to the de novo assembled transcriptome of Brandt’s vole (Lasiopodomys brandtii) (n = 5 per group)

Group	Samples	Total Reads	Mapped Reads	GC Content	%≥Q30
Control	a1	21,853,034	19,530,950(89.37%)	47.46%	94.11%
	a2	19,198,347	17,573,329(91.54%)	46.88%	94.24%
	a3	20,713,153	18,862,340(91.06%)	46.32%	93.72%
	a4	21,192,090	19,037,177(89.83%)	46.09%	93.68%
	a5	19,926,840	18,043,547(90.55%)	45.98%	94.21%
Low dose	b1	21,890,329	20,001,011(91.37%)	45.83%	94.28%
	b2	20,888,694	18,893,990(90.45%)	46.18%	94.05%
	b3	19,695,850	17,956,200(91.17%)	46.20%	94.29%
	b4	23,669,498	21,493,799(90.81%)	47.22%	94.38%
	b5	20,614,006	18,920,848(91.79%)	45.97%	94.23%
High dose	d1	22,754,193	20,424,554(89.76%)	45.69%	93.38%
	d2	21,093,271	19,201,951(91.03%)	46.55%	94.02%
	d3	20,885,668	18,956,819(90.76%)	46.26%	94.34%
	d4	20,656,558	18,710,102(90.58%)	45.81%	93.99%
	d5	21,946,958	20,269,663(92.36%)	46.59%	94.34%

Fig. 5.

Differentially expressed genes (DEGs) in brown adipose tissue (BAT) of Brandt’s vole after cold exposure (n = 5 per group). A Number of DEGs; B Venn diagram showing total and shared DEGs among groups

Fig. 6.

Volcano plots of differentially expressed genes in brown adipose tissue (BAT) of Brandt’s vole after cold exposure (n = 5 per group). A Control vs. low-dose group; B Control vs. high-dose group

qRT-PCR analysis confirmed that TA significantly affected the expression of FFAR4 (F₂, ₁₂ = 11.635, P < 0.001), PDRG-1 (F₂, ₁₂ = 5.995, P = 0.016), and UCP1 (F₂, ₁₂ = 12.980, P < 0.001) in BAT (Fig. 7A–C). Specifically, FFAR4 and UCP1 expression levels were significantly higher in the low- and high-dose groups than in the control group, whereas PDRG-1 expression was significantly lower in both TA-treated groups. These expression patterns were consistent with the RNA-seq results (Fig. 6A–B). At the protein level, TA significantly increased UCP1 (F₂, ₂₁ = 21.661, P < 0.001), PGC-1α (F₂, ₂₁ = 51.721, P < 0.001), and p-AMPK/AMPK (F₂, ₂₁ = 24.167, P < 0.001) levels in BAT after cold exposure, with higher levels observed in both TA-treated groups compared with the control group (P < 0.001 for all) (Fig. 8A–C). TA also significantly affected PPARγ protein expression (F₂, ₂₁ = 4.824, P = 0.019), with higher expression detected in the high-dose group relative to the control group (P = 0.006) (Fig. 8D).

Fig. 7.

Validation of differentially expressed genes in brown adipose tissue (BAT) of Brandt’s vole after cold exposure by qRT-PCR. A FFAR4; B PDRG-1; C UCP1. Data are presented as mean ± SEM (n = 5 per group). *, **, and *** indicate P < 0.05, P < 0.01, and P < 0.001 vs. control group, respectively

Fig. 8.

Effect of tannic acid (TA) on protein expression levels of thermogenesis-related markers in brown adipose tissue (BAT) of Brandt’s vole after cold exposure. A UCP1; B PGC-1α; C AMPK; D PPARγ. Data are presented as mean ± SEM (n = 8 per group). ** and *** indicate P < 0.01 and P < 0.001 vs. control group, respectively. Full-length, uncropped Western blot images are provided in Supplementary Figures S1–S8

GO and KEGG functional enrichment analysis of DEGs

Functional enrichment analysis was performed to evaluate the biological functions of differentially expressed mRNAs via GO and KEGG enrichment analyses. Notably, in the low-dose group relative to the control, DEGs were primarily associated with negative regulation of cytokine production involved in immune response (GO:0002719) in the biological process category, chylomicron (GO:0042627) in the cellular component category, and intronic transcription regulatory region sequence-specific DNA binding (GO:0001161) in the molecular function category (Fig. 9A–C). In contrast, the high-dose group DEGs, when compared to the control, were mainly enriched in responses to electrical stimulus involved in the regulation of muscle adaptation (GO:0014878), brown fat cell differentiation (GO:0050873), and positive regulation of cold-induced thermogenesis (GO:0120162) in the biological process category; in the cellular component category, they were mainly linked to the hemoglobin complex (GO:0005833); and in the molecular function category, to oxygen carrier activity (GO:0005344) (Fig. 9D–F).

Fig. 9.

Gene Ontology (GO) functional enrichment analysis of differentially expressed genes in brown adipose tissue (BAT) of Brandt’s vole after cold exposure (n = 5 per group). A–C Control vs. low-dose group; D–F Control vs. high-dose group

Furthermore, in comparison to the control, DEGs in the low-dose group showed significant enrichment in the fat digestion and absorption (ko04975) and PPAR signaling (ko03320) pathways. Conversely, the low-dose group exhibited downregulation of DEGs in pathways associated with tuberculosis (ko05152), complement and coagulation cascades (ko04610), pertussis (ko05133), Staphylococcus aureus infection (ko05150), D-glutamine and D-glutamate metabolism (ko00471), mineral absorption (ko04978), systemic lupus erythematosus (ko05322), and the phagosome (ko04145). Notably, the pathways for vitamin digestion and absorption (ko04977), cholesterol metabolism (ko04979), and transcriptional misregulation in cancer (ko05202) demonstrated both upregulated and downregulated gene profiles (Fig. 10A). KEGG analysis for the high-dose group, relative to the control, revealed that DEGs were primarily associated with human papillomavirus infection (ko05165), breast cancer (ko05224), malaria (ko05144), African trypanosomiasis (ko05143), systemic lupus erythematosus (ko05322), Staphylococcus aureus infection (ko05150), steroid hormone biosynthesis (ko00140), steroid biosynthesis (ko00100), rheumatoid arthritis (ko05323), and pathways in cancer (ko05200). Among these pathways, only the malaria pathway was exclusively downregulated, while the others exhibited both upregulation and downregulation of gene expression (Fig. 10B). Additionally, pathway analyses using Gene Set Enrichment Analysis (GSEA) based on the KEGG database for the comparisons between the control and both the low-dose and high-dose groups further corroborated these findings (Figures S9–S10).

Fig. 10.

KEGG pathway enrichment analysis of differentially expressed genes in brown adipose tissue (BAT) of Brandt’s vole after cold exposure (n = 5 per group). A Control vs. low-dose group; B Control vs. high-dose group

Gene network analysis

Following PPI analysis using STRING and subsequent visualization via Cytoscape, a comprehensive protein interaction network of DEGs was constructed (Figure S11; PPI enrichment P-value < 5.13 × 10⁻¹⁴). Among the 487 DEGs identified, 358 formed a large and intricate interaction network. To elucidate closely interacting DEGs, we applied further clustering using the Markov Cluster Algorithm (MCL) with an inflation parameter of 1.8, which resulted in the identification of 43 clusters. The top five clusters comprised 55, 9, 7, 7, and 6 genes, respectively (Fig. 11A–E). Notably, Cluster 1, containing 55 genes, remains functionally uncharacterized, whereas Cluster 2, which includes 9 genes (Lep, Ifnar2, UCP1, Gk, Mpzl2, Adrb1, Arrdc3, FFAR4, Lifr), is associated with brown fat cell differentiation. Furthermore, Cluster 3 is linked to extracellular matrix (ECM)–receptor interactions, Cluster 4 to collagen biosynthesis and modifying enzymes, and Cluster 5, containing 6 genes (Cyp51, Cyp7a1, Fads2, Idi1, Dhcr7, Msmo1), to cholesterol biosynthesis.

Fig. 11.

Top five clusters identified by protein–protein interaction (PPI) network analysis using the Markov Cluster Algorithm (MCL) based on total differentially expressed genes. A Cluster 1; B Cluster 2; C Cluster 3; D Cluster 4; E Cluster 5

Discussion

In the present study, we demonstrated for the first time that TA influences cold tolerance in Brandt’s voles. Given the similar initial body mass among groups, the lower percentage change in body mass observed in the TA-treated groups suggests that TA treatment reduced body mass gain in Brandt’s voles. TA treatment significantly reduced both the mass and index of iWAT, and this effect remained significant after adjustment for final body mass, suggesting that TA effectively decreases WAT. Consistent with this finding, TA potently inhibited fatty acid synthase activity in a concentration-dependent manner, suppressing its expression during adipocyte differentiation and downregulating the mRNA levels of fatty acid synthase and PPARγ [46]. In addition, TA prevented the differentiation of 3T3-L1 preadipocytes, thereby reducing intracellular lipid accumulation [46]. Notably, Pin1—an isomerase that regulates the differentiation of mouse embryonic fibroblasts into adipocytes—was significantly inhibited by TA [28]. Intragastric administration of TA also decreased triglyceride and phospholipid levels in plateau pikas (Ochotona curzoniae) [47]. The reductions in iWAT mass and food intake may together contribute to the overall decrease in body mass observed following TA treatment.

All of the mass, its index, and the area of iWAT cell were sharply reduced by TA treatment. By contrast, although the initial analysis indicated that interscapular BAT mass remained relatively unchanged, BAT index increased significantly following TA treatment. Furthermore, after adjustment for final body mass using ANCOVA, interscapular BAT mass was significantly greater in TA-treated voles than in controls. These findings suggest that TA exerts a more pronounced reductive effect on WAT than on BAT and may promote BAT recruitment or maintenance. Similar outcomes were observed with grape seed proanthocyanidin extract; eleven weeks of administration reduced both iWAT mass and its index in Wistar rats, while the BAT index remained largely unaffected [48]. Moreover, in diet-induced obese Wistar rats, administration of grape seed proanthocyanidin extract did not reduce BAT weight but increased the BAT index due to reduced body weight gain [49]. Likewise, Berberine treatment significantly decreased body weight and iWAT mass without affecting food intake in diet-induced obese mice [17]. In the present study, both BAT index and body-mass-adjusted BAT mass were elevated in TA-treated voles, suggesting that TA may preserve or promote BAT development while reducing WAT deposition. Enhancements of the nuclei number of BAT adipocytes indicate that TA treatment might stimulate the proliferation or differentiation of brown adipocytes in BAT. Similarly, Berberine treatment increased the nuclei number of brown adipocytes in HFD-fed male C57BL/6J mice [17]. Together with the increased BAT index and body-mass-adjusted BAT mass observed in the present study, these findings suggest that TA may promote BAT remodeling and recruitment.

Although rectal temperatures were similar between the TA treatment and control groups at room temperature, following cold exposure the TA-treated voles consistently maintained higher surface, core, and BAT temperatures. These differences remained significant after adjustment for final body mass, indicating that the improved thermal status of TA-treated voles could not be explained solely by body weight differences. Consistently, a similar pattern was observed for the rectal temperature reduction ratio; however, the group difference was no longer statistically significant after adjustment for body mass. Collectively, these findings suggest that TA may enhance BAT thermogenic responses and thereby improve cold tolerance in Brandt’s voles. Interestingly, tannins from grains have been proposed to play a role in chilling tolerance regulation in plants [50], and intra-oral injection of flavangenol has been shown to elevate BAT temperature [51]. Similarly, although the basal rectal temperature was unchanged in both groups at room temperature, berberine treatment notably increased the rectal temperature of lean mice housed at 4 °C [17]. Similar rectal temperature after TA treatment implies that reduction of body mass in normal Brandt’s vole induced by TA treatment is not due to the enhancement of energy expenditure.

In the BAT RNA-seq analysis, UCP1 was identified as a differentially expressed gene with consistently higher expression in both low- and high-dose groups compared to the control group, a finding confirmed at the protein level. Uncoupling proteins (UCPs) are mitochondrial transporter family members localized in the inner mitochondrial membrane, with UCP1—also known as thermogenin—being a well-characterized protein expressed exclusively in BAT [52, 53]. Similarly, kaki-tannin has been shown to enhance UCP1 gene expression at both the transcriptional and protein levels in the BAT of type 2 diabetic NSY/Hos mice [54]. Additionally, Nomura et al. [55] reported that tea catechins upregulate UCP1 mRNA in rat BAT, while repeated cold exposure elevated UCP1 protein levels, without significantly affecting mRNA levels, in the BAT of striped hamsters (Cricetulus barabensis). These observations suggest that in Brandt’s voles treated with TA, UCP1 expression is markedly enhanced following cold exposure, resulting in increased surface, core, and BAT temperatures.

KEGG pathway analysis revealed that the differentially expressed genes were significantly enriched in fat digestion and absorption (ko04975). Since BAT utilizes free fatty acids as substrates for heat production [56], enhanced lipid digestion and absorption may supply additional substrates to support thermogenesis [57]. Furthermore, some DEGs were associated with the regulation of cholesterol metabolism (ko04979) in the low-dose group relative to the control, a finding corroborated by PPI analysis, which identified a cluster related to cholesterol biosynthesis with downregulated Cyp51, Fads2, Idi1, and Dhcr7 in high-dose group compared to the control. These results align with previous reports demonstrating that cold acclimation-induced BAT activation improves cholesterol metabolism in hypercholesterolemic patients [57] and that cholesterol biosynthesis is enriched in the BAT of UCP1 knockout mice following cold exposure [58]. In BAT, AMPK activation increases non-shivering thermogenesis with inhibition of fat and cholesterol synthesis in brown and beige adipocytes [10]. Therefore, we speculate that TA may enhance the metabolism of WAT, thereby reducing its mass, while simultaneously supplying triglycerides to BAT. BAT, endowed with improved lipid digestion and absorption capabilities, subsequently utilizes these triglycerides to fuel thermogenesis in response to cold exposure.

Moreover, enrichment of the PPAR signaling pathways (ko03320) was observed. Peroxisome PPARγ is a key regulator of adipocyte differentiation [59, 60], and during brown adipocyte differentiation, adipogenic signals stimulate UCP1 gene transcription via PPARγ and its coactivator, PGC-1α [53]. PPARγ binds to the coactivator PGC-1α, coordinating distinct transcriptional programs by interacting with transcription factors that regulate thermogenic gene expression [13]. The observed enrichment of the PPAR signaling pathways may therefore indicate an increase in brown adipocyte differentiation and thermogenic gene expression in the BAT of TA-treated Brandt’s voles in response to cold exposure.

In addition, Gene Ontology analysis revealed that several DEGs were associated with brown fat cell differentiation (GO:0050873) and the positive regulation of cold-induced thermogenesis (GO:0120162) in the high-dose group compared to the control group. This finding is consistent with the KEGG and PPI analyses, which also identified clusters related to brown fat cell differentiation. Although mRNA levels of PPARγ and PGC-1α did not change significantly in the RNA-seq data, their protein levels were substantially elevated in both the low- and high-dose groups after cold exposure. FFAR4, previously known as G-protein-coupled receptor 120 (Gpr120), promotes the browning of white fat and regulates thermogenesis in beige and brown mouse adipocytes [61, 62]. Consistently, FFAR4 within the cluster related to brown fat cell differentiation significantly enhanced expression in both low- and high-dose groups. Similarly, it is reported that Omija fruit extract can enhance endurance and energy metabolism by upregulating the expression of PGC-1α in skeletal muscle [63]. Therefore, TA treatment appears to enhance cold tolerance in Brandt’s voles by promoting thermogenic gene expression and stimulating brown adipocyte differentiation in BAT. Moreover, the preservation of BAT mass following TA administration may further contribute to the improved cold tolerance observed in these animals. Previous studies have demonstrated that, in Leymus chinensis, A. frigida, and Cleistogenes squarrosa grasslands, the overwinter food caches of Brandt’s voles are predominantly composed of plants in the genus Artemisia, with A. frigida and A. scoparia accounting for over 90% of the stored forage [34, 35, 37]. Notably, A. frigida contains higher levels of tannic acid compared to Stipa krylovii and C. squarrosa [36]. Accordingly, we hypothesize that the capacity of tannic acid to enhance cold tolerance may drive Brandt’s voles to preferentially hoard Artemisia plants during winter. A similar phenomenon has been observed in plateau zokors (Eospalax baileyi), which cache plants containing high levels of secondary metabolites, such as phenols and tannins, in contrast to plants that are less frequently cached during winter [64]. This behavior is likely attributable to the cold tolerance benefits associated with tannic acid. Furthermore, the role of tannic acid in enhancing cold tolerance may lend support to the “high tannin hypothesis” advanced by PD Smallwood and WD Peters [65] and Fleck and Woolfenden [66]. Specifically, Food hoarding animals prefer to cache high-tannin food over those containing a low level of tannins, thereby capitalizing on tannic acid’s protective effects during winter.

Several limitations of the present study should be acknowledged. First, whole-body energy expenditure was not directly measured. Although physiological, transcriptomic, and protein-level evidence consistently supported enhanced thermogenic potential in BAT following TA treatment, indirect calorimetry measurements (e.g., oxygen consumption and energy expenditure) will be required in future studies to determine whether TA directly increases whole-body thermogenesis. Second, animals were maintained under a long-photoperiod (16 L:8D) laboratory regime rather than a winter-like short photoperiod. Because photoperiod is a major regulator of seasonal physiology in Brandt’s voles, future studies should investigate whether the effects of TA on adipose tissue remodeling and thermogenic responses are modified under short-day conditions. Such studies would provide a more ecologically relevant framework for understanding whether tannin consumption contributes to winter survival and cold adaptation in natural populations of Brandt’s voles. Third, although FFAR4 expression was significantly upregulated and previous studies have implicated FFAR4 in thermogenic regulation, its expression was assessed only at the transcript level in the present study. Therefore, the potential involvement of FFAR4 in BAT thermogenesis should be interpreted cautiously, and further protein-level and functional studies are needed to clarify its role in the thermogenic responses associated with TA supplementation. In addition, baseline body surface temperature and BAT temperature were not measured prior to cold exposure. Consequently, the observed differences reflect post-cold-exposure thermal responses rather than changes from individual baseline values. Finally, animals were fasted before cold exposure to minimize variation in feeding status and recent food intake. Although the fasting protocol was applied uniformly across all groups, fasting itself has been reported to influence thermogenic activity [67] and may interact with cold-induced BAT responses. Future studies incorporating both fasting and non-fasting conditions will help distinguish the independent and interactive effects of TA treatment, nutritional status, and cold exposure on thermogenic regulation.

Conclusions

This study shows that dietary TA, a widespread plant secondary metabolite, is associated with coordinated changes in adipose tissue traits and thermoregulatory responses in Brandt’s voles exposed to cold conditions. TA supplementation was accompanied by reduced iWAT mass and an increased BAT index, while total BAT mass remained largely unchanged, suggesting a shift in adipose tissue organization rather than overall fat accumulation. Consistently, TA-treated voles maintained higher surface, BAT, and rectal temperatures following cold exposure, indicating an association between TA intake and improved thermal performance. At the molecular level, TA intake was associated with dose-dependent transcriptional patterns in BAT. In the low-dose group, differentially expressed genes were mainly enriched in pathways related to lipid digestion, absorption, and PPAR signaling, whereas in the high-dose group, DEGs were predominantly linked to brown adipocyte differentiation and the positive regulation of cold-induced thermogenesis. Elevated expression of FFAR4 and UCP1, together with higher protein levels of PGC-1α, PPARγ, and UCP1, is consistent with an increased thermogenic potential in BAT, although direct causal relationships cannot be inferred from the present data. From an ecological and evolutionary physiology perspective, these physiological and molecular patterns may reflect an adaptive response that facilitates winter survival in cold steppe environments. The association between TA intake and enhanced thermogenic capacity provides a plausible mechanistic context for the winter hoarding behavior of Brandt’s voles, which preferentially store Artemisia species with relatively high tannin content. Collectively, our findings suggest that the selective use of tannin-rich plants may contribute to seasonal physiological adjustment, highlighting the potential role of plant secondary metabolites in shaping herbivore foraging strategies and thermal adaptation under natural environmental constraints.

Supplementary Information

Supplementary Material 1. ^{(2.1MB, zip)}

Abbreviations

TA

Tannic acid

WAT

White adipose tissue

BAT

Brown adipose tissue

iWAT

Inguinal white adipose tissue

DEGs

Differentially expressed genes

PPARγ

Peroxisome proliferator-activated receptorγ  

NST

Non-shivering thermogenesis

RMR

Resting metabolic rate

UCP1

Uncoupling protein 1

PGC-1α

Peroxisome proliferator-activated receptor gamma coactivator 1α

AMPK

Adenosine monophosphate-activated protein kinase

PSMs

Plant secondary metabolites

qPCR

Polymerase chain reaction

SDS

Sodium dodecyl sulfate

HRP

Horseradish peroxidase

PPI

Protein–protein interaction

GSEA

Gene set enrichment analysis

BP

Biological process

CC

Cellular component

MF

Molecular function

ANOVA

One-way analysis of variance

ECM

Extracellular matrix

MCL

Markov cluster algorithm

Gpr120

G-protein-coupled receptor 120

GO

Gene ontology

KEGG

Kyoto Encyclopedia of Genes and Genomes

Authors’ contributions

Data curation: QS, KC, XM. Formal analysis: QS, XD. Funding acquisition: XD, WW, SY. Investigation: QS, KC, XM. Methodology: QS, KC, XM. Project administration: XD. Resources: SY. Software: QS. Supervision: XD, WW, SY, FY. Writing-original draft: QS. Writing-review&editing: XD.

Funding

This work was supported by the National Natural Science Foundation of China (Grant number: 32370438, 31971418, and 31770422).

Data availability

The RNA-seq dataset has been deposited in the NCBI Sequence Read Archive (SRA) under BioProject accession number PRJNA1444008, and the corresponding accession information and web link have been provided in the manuscript.

Declarations

Ethics approval and consent to participate

All protocol and procedures were approved by the Animal Care and Use Committee of the Faculty of Veterinary Medicine at the Yangzhou University (No. NSFC2020-SKXY-6).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.