Central Irisin Administration Attenuates Hypothalamic TLR4/MyD88-Mediated Neuroinflammatory Signaling in Diet-Induced Obese Mice

Abstract

Hypothalamic inflammation represents a central mechanism linking obesity to metabolic dysfunction. This process involves glial activation and persistent innate immune signaling, with Toll-like receptor 4 (TLR4) emerging as a critical interface between inflammatory pathways and impaired central insulin signaling. Irisin, a myokine released in response to physical exercise, has been shown to exert metabolic and anti-inflammatory effects in peripheral tissues, as well as neuroprotective actions in the brain. However, whether irisin directly modulates obesity-associated hypothalamic inflammation, particularly through TLR4-dependent pathways, remains unknown. Here, we investigated the effects of short-term intracerebroventricular delivery of recombinant irisin on hypothalamic inflammatory signaling in diet-induced obese mice. Central irisin administration reduced glial reactivity, downregulated components of the TLR4/MyD88 pathway, and increased the expression of anti-inflammatory cytokines in the hypothalamus. In addition, irisin restored insulin-stimulated AKT phosphorylation and selectively reduced inguinal white adipose tissue mass without affecting overall body weight. Together, these findings indicate that central irisin administration attenuates obesity-related hypothalamic inflammation and modulates central insulin signaling, supporting a role for irisin as a regulator of neuroinflammation-linked metabolic dysfunction.

Graphical Abstract

Graphical abstract created with Biorender.com under publication license (MR296VUFQG).

Supplementary Information

The online version contains supplementary material available at 10.1007/s12035-026-05729-8.

Introduction

Obesity affects more than one billion individuals worldwide and represents one of the most pressing global public health challenges [1]. Characterized by excessive adipose tissue accumulation, this chronic condition is strongly associated with increased mortality, largely due to its close link with metabolic comorbidities such as type 2 diabetes mellitus, hypertension, and cardiovascular disease [1, 2]. The pathogenesis of obesity involves a sustained disruption of energy homeostasis, driven primarily by sedentary lifestyles and the widespread availability of ultra-processed, energy-dense foods [3, 4]. Over time, chronic nutritional overload exceeds the capacity of physiological mechanisms that regulate body weight, ultimately leading to persistent metabolic dysfunction [4].

A direct consequence of overnutrition is the establishment of a low-grade systemic inflammatory state that disrupts hypothalamic control of energy homeostasis [5]. High-fat diet (HFD) exposure triggers inflammatory responses in the hypothalamus within hours, preceding overt weight gain or the development of systemic adiposity [6, 7]. This early response is characterized by reactive gliosis, involving rapid activation of microglia and astrocytes, along with infiltration of peripheral immune cells [8]. The resulting inflammatory milieu promotes hypothalamic insulin resistance, thereby impairing neuronal signaling pathways that regulate satiety and energy expenditure [9]. Through sustained disruption of these core homeostatic processes, hypothalamic inflammation contributes to a self-reinforcing cycle that facilitates obesity onset and progression [5, 7, 8].

At the molecular level, this neuroinflammatory response engages innate immune signaling pathways, with Toll-like receptor 4 (TLR4) emerging as a key mediator [10]. TLR4 is predominantly expressed in glial cells and is activated by metabolic stressors, including saturated fatty acids [10, 11]. Upon activation, TLR4 signals through the MyD88/NF-κB axis, driving the sustained production of pro-inflammatory mediators such as IL-1β, IL-6, and TNF-α. These mediators directly impair hypothalamic insulin signaling, thereby compromising central metabolic regulation [12]. In this context, the TLR4 signaling cascade constitutes a critical mechanistic link between nutritional excess, hypothalamic inflammation, and metabolic dysfunction, representing a promising therapeutic target for obesity-related disorders.

In light of the pivotal role of hypothalamic inflammation in obesity pathophysiology, identifying strategies capable of attenuating neuroinflammatory signaling is of considerable interest. Irisin, a myokine released from skeletal muscle in response to exercise, has gained attention as a mediator of the systemic metabolic and anti-inflammatory benefits of physical activity [13]. Derived from proteolytic cleavage of fibronectin type III domain-containing protein 5 (FNDC5), irisin was initially described as a peripheral hormone that promotes white adipose tissue browning and thermogenesis [14]. Beyond skeletal muscle, FNDC5/irisin is expressed in adipose tissue, liver, pancreas, and the central nervous system (CNS), suggesting autocrine, paracrine, and central actions [15]. Importantly, irisin can cross the blood–brain barrier [16–18] and has been detected in cerebrospinal fluid, hippocampus, and hypothalamus [19–23]. Its receptor, integrin αVβ5 — originally identified in bone tissue [24] — is also expressed in microglia and astrocytes, the principal effectors of neuroinflammation [25, 26].

Most studies investigating the central effects of irisin have focused on brain regions vulnerable to neurodegeneration, particularly the hippocampus, where irisin enhances synaptic plasticity and memory, and confers neuroprotection in models of stroke, hypoxic-ischemic injury, and intracerebral hemorrhage [16, 26–28]. In contrast, research addressing hypothalamic actions of irisin has largely focused on its role in thermoregulation, energy balance, and cardiovascular control, frequently using central administration approaches [29–31]; however, whether irisin modulates obesity-associated hypothalamic inflammation via TLR4-dependent signaling remains unknown. To address this gap, the present study evaluated the effects of short-term intracerebroventricular (ICV) administration of recombinant irisin on hypothalamic neuroinflammation in a mouse model of diet-induced obesity, with a specific focus on the TLR4/MyD88 signaling axis.

Methods

Analysis of Single-Cell RNA Sequencing Data

Publicly available single-cell RNA sequencing (scRNA-seq) datasets from the adult mouse hypothalamus were obtained from three independent studies [32–34]. Each dataset was subjected to quality control and normalized individually using the SCTransform method implemented in the Seurat package (v3.1.2) within the R environment. To integrate the datasets and minimize batch effects, a weighted anchor-based integration approach was applied to a latent space generated by canonical correlation analysis (CCA). This procedure enhanced biological signals shared across datasets while reducing study-specific technical variability. Following integration, a total of 51,623 cells were retained for downstream analyses. Cell-type annotations were transferred from a reference hypothalamic atlas using a CCA-based transfer learning method within Seurat.

For visualizing cellular heterogeneity and gene expression, we utilized diffusion-based Manifold Approximation and Projection (dbMAP), an unsupervised dimensionality reduction algorithm. This method preserves the intrinsic structure of the high-dimensional transcriptomic data in two- or three-dimensional space, grouping transcriptionally similar cells and separating distinct populations [35]. Using this approach, we examined the distribution and variability of Fndc5, Tlr4, Itgav, and Itgb5 expression across hypothalamic cell types. Gene expression patterns on dbMAP embeddings, as well as quantitative expression distributions across annotated hypothalamic cell clusters, were visualized using the Cerebro (v1.2.2) scRNA-seq analysis platform.

Animals and Ethics Statement

Four-week-old male C57BL/6 J mice were obtained from the Multidisciplinary Center for Biological Research in the Area of Laboratory Animal Science (CEMIB). The mice were housed under controlled conditions, including an average temperature of 22 °C (±  2 °C), a 12-h light–dark cycle, and ad libitum access to food and water. During acclimation and the pre-surgical phase, mice were group-housed (four per cage); following stereotaxic surgery, they were singly housed to protect the cannula and prevent social interference. All experimental procedures were approved by the Animal Use Ethics Committee (CEUA) of the State University of Campinas (protocol number 5926–1/2021).

Experimental Design Overview

Mice were fed a chow/control diet (CD; 10% of calories from fat, Nuvilab CR-1, Nuvital®, SP, Brazil) or a high-fat diet (HFD; 45% of calories from fat, 228 BAN, Pragsoluções Biociências®, SP, Brazil) throughout the experimental period (diet compositions are available in Supplementary Tables S1–S2). Body weight and food intake were monitored weekly to track DIO progression. In the 12th week of the diet, six animals from each group (CD and HFD) were randomly selected for a glucose tolerance test (detailed in the Supplementary Methods), complementing metabolic phenotyping.

In the 13th week, all animals underwent stereotaxic surgery for the implantation of a cannula into the right lateral ventricle of the brain. Cannula functionality was assessed via ICV infusion of angiotensin II, followed by monitoring water intake. After confirming cannula viability, treatments were administered each morning for 7 consecutive days. A subset of HFD-fed animals received ICV treatment with irisin (HFD-Irisin), while the remaining HFD-fed animals received sterile saline as a control (HFD-Vehicle). CD-fed animals also received saline (CD-Vehicle). On the 8th day, approximately 24 h after the last ICV injection, and following a 12-h overnight fast, mice were euthanized under deep anesthesia and decapitated for brain extraction. The hypothalamus was dissected and stored at −80 °C until further analysis. Inguinal (iWAT), epididymal (eWAT), and retroperitoneal (rWAT) white adipose tissues; interscapular brown adipose tissue (iBAT); and the gastrocnemius, soleus, tibialis anterior, and extensor digitorum longus (EDL) muscles were collected and weighed.

Independent cohorts of animals were used for immunofluorescence, RT-qPCR, and western blot analyses (inflammation and hypothalamic insulin sensitivity) to ensure full hypothalamic availability and to avoid interference from prior procedures. An overview of the experimental protocol is shown in Fig. 1.

Fig. 1.

Experimental design overview. Four-week-old C57BL/6 J mice were fed a chow/control diet (CD) or a high-fat diet (HFD) throughout the experimental period. In the 13.^th week, the animals underwent stereotaxic surgery for implantation of a cannula into the right lateral ventricle. Cannula functionality was assessed by the intracerebroventricular (ICV) infusion of angiotensin II. A subset of HFD-fed animals received ICV treatment with irisin (HFD-Irisin), while the remaining HFD-fed animals received sterile saline (HFD-Vehicle). All CD-fed animals also received saline as a control (CD-Vehicle). Approximately 24 h after the last day of treatment, the mice were euthanized, and their tissues were dissected, weighed, and stored. Figure created with BioRender.com under publication license (DH28KXJWCV)

Stereotaxic Surgery and Angiotensin II Test

The animals were anesthetized with a combination of Ketamine® (100 mg/kg) and Xylazine® (10 mg/kg), administered via intraperitoneal [36]. Subsequently, pre-surgical analgesia was ensured via subcutaneous injection of tramadol (5 mg/kg) [37]. The animals were then positioned in a stereotaxic frame (Stoelting, IL, USA), equipped with mouse-specific nose and ear bars, and a cannula was implanted into the right lateral ventricle of the brain using the following coordinates: anteroposterior −0.34 mm; mediolateral −1.0 mm; dorsoventral −2.2 mm, relative to bregma [38]. The cannula was fixed to the skull with resin cement (Dual RelyX™ ARC, 3 M, SP, Brazil) and autopolymerizing acrylic (JET, Clássico®, SP, Brazil), with additional stabilization provided by a metal bone screw, as illustrated in Supplementary Fig. S1. Finally, the mice were housed individually in their cages, and clinical signs, including body weight and behavioral conditions, were monitored daily. After recovery, cannula functionality was assessed by ICV infusion of 2 μL angiotensin II (10⁻⁴ M) (#A9525, Sigma-Aldrich™, MO, USA), followed by monitoring of water intake responsiveness [36].

Intracerebroventricular Treatment

Recombinant irisin (#100–65, PeproTech®, NJ, USA) was reconstituted in sterile water and diluted in sterile 0.9% saline (0.1% BSA) according to the manufacturer's instructions. Mice received daily ICV injections of 2 μL irisin (7.5 μg/kg) [39, 40] or vehicle (0.9% saline) for 7 consecutive days, administered between 9:00 and 10:00 a.m. The animals were gently immobilized, without the influence of sedatives, and the substances were slowly administered into the cannula using a 22-gauge gingival needle attached to a Hamilton® syringe. The needle was kept in place for 5 s after injection to prevent reflux of the substances.

Food Intake and Body Weight Assessment

During the 7-day treatment period, mice were housed individually in their cages and received a known amount of diet each morning. Daily food consumption was measured by weighing the remaining diet the following morning. Body weight was monitored daily, and naso-anal length was recorded immediately before euthanasia to calculate the Lee index [41]. Body weight change, tissue mass, and caloric intake were calculated from these records, as detailed in Supplementary Table S3.

Hypothalamic Insulin Signaling Assay

On the last day of treatment, following a 12-h overnight fast, the mice received a single ICV injection of 2 μL insulin (10 mU/mouse, Humulin®, Lilly Laboratory, SP, Brazil) [42, 43] through the cannula previously implanted, without the use of sedatives. Twenty minutes post-injection, animals were euthanized by overdose of 9% isoflurane (Isoforine®, Cristália, SP, Brazil). The hypothalamus was dissected and stored at −80 °C for subsequent western blot analysis.

Heart Perfusion and Immunofluorescence Assay

Mice were anesthetized with Ketamine® and Xylazine®, as previously described, and perfused through the left cardiac ventricle with 0.9% saline followed by 4% paraformaldehyde (PFA). [44]. Brains were post-fixed for 24 h in 4% PFA and cryoprotected in 30% sucrose at 4 °C for 36 h. Serial coronal  sections (20 µm) were cut using a cryostat (Leica CM1860) with Tissue-Tek O.C.T. embedding medium and stored at −20 °C in antifreeze solution (30% sucrose, 60% ethylene glycol, prepared in 1 × phosphate-buffered saline [PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄, pH 7.4]) until immunostaining.

For free-floating immunostaining, slices were first incubated for 10 min in a 0.15% hydrogen peroxide solution to neutralize endogenous peroxidase activity, followed by PBS washes to remove any residue. Slices were then blocked with 0.2% Triton X-100 and 5% donkey serum for 2 h at room temperature and subsequently incubated overnight at 4 °C in a fresh blocking solution containing primary antibodies: rabbit anti-Iba1 (1:250, #019–19741, FujiFilm Wako, Thermo Fisher Scientific Inc., MA, USA) or Cy3-conjugated rabbit anti-GFAP (1:200, #Bs-0199R-Cy3, Bioss, MA, USA). Since the anti-GFAP antibody was directly conjugated to Cy3, no secondary antibody incubation was required for these sections. After PBS washes, slices incubated with anti-Iba1 were further incubated in a blocking solution containing the secondary antibody donkey anti-rabbit FITC (1:500, #ab6798, Abcam, MA, USA) for 2 h at room temperature. During the final 10 min of antibody incubation, 4′,6-diamidino-2-phenylindole (DAPI) (1:10000, #D9542, Sigma-Aldrich™, MO, USA) was added. Sections were then washed, mounted with Vectashield mounting medium (#H-1200, Vector Laboratories™, CA, USA), and coverslipped.

Images from at least three to five brain sections per animal were captured from the hypothalamic region using a confocal microscope (Upright LSM780-NLO Zeiss), with identical acquisition parameters (exposure, gain, laser power) across all samples to minimize technical variability. Z-stacks comprising 34 sequential layers were acquired and projected at maximum intensity using a 40 × objective. Quantitative analysis was performed in ImageJ/Fiji (Bethesda, USA) by measuring the percentage of area occupied by Iba1- and GFAP-positive immunoreactivity, using identical threshold settings for all images. The detailed list of antibodies used for immunofluorescence is available in Supplementary Table S4.

Quantitative Real-Time Polymerase Chain Reaction (RT-qPCR)

Total RNA was extracted from the hypothalamus using TRIzol® reagent (Thermo Fisher Scientific, MA, USA), following the manufacturer’s protocol. RNA concentration and purity were assessed by spectrophotometry (SpectraMax® i3, Molecular Devices, CA, EUA). Next, 1 µg of total RNA was reverse transcribed into cDNA using the High-Capacity cDNA Reverse Transcription Kit (#4,368,814, Applied Biosystems™, Thermo Fisher Scientific Inc., MA, USA). RT-qPCR was performed on the 7500 Fast Real-Time PCR system (Applied Biosystems™, MA, USA) using SYBR® Green Master Mix (#PB20.14–05, PCR Biosystems™, PA, USA). Relative mRNA expression levels were calculated using the 2^−ΔΔCT method [45], normalized to Gapdh, and expressed relative to the control group (CD). Primer sequences, validated for specificity and efficiency, are listed in Supplementary Table S5.

Proteins Extraction and Western Blot

Hypothalamic samples were homogenized in lysis buffer (7 M urea, 2 M thiourea, 10 mM EDTA, 100 mM sodium fluoride, 100 mM sodium pyrophosphate, 10 mM sodium orthovanadate, 2 mM PMSF, 1 µg/mL aprotinin, 1% Triton X-100) and incubated for 40 min at 4 °C. Homogenates were centrifuged at 11,000 × g for 20 min at 4 °C, and the supernatants were collected. Protein concentration was determined using the Bradford reagent (BioAgency Biotecnologia, SP, Brazil), with bovine serum albumin (BSA) as the standard. After quantification, proteins were denatured in 20% Laemmli buffer 5x (0.5 M Tris–HCl pH 6.8, 30% glycerol, 10% SDS, 0.6 M DTT, 0.012 M bromophenol blue) for 5 min at 95 °C.

Subsequently, 30 μg of protein per sample was loaded onto a 10% SDS-PAGE gel and size-separated using running buffer (1.8 mM EDTA, 50 mM Tris base, 380 mM glycine, 0.1% SDS). Proteins were then transferred to nitrocellulose membranes (BioRad®) at 120 V for 2 h in transfer buffer (25.15 mM Tris base, 191.75 mM glycine, and 0.02% SDS) containing 20% methanol. Transfer efficiency was confirmed by visualizing protein bands after staining with Ponceau for 5 min.

For immunoblotting, membranes were blocked with 5% BSA for 2 h at room temperature and subsequently incubated overnight at 4 °C with the following primary antibodies: mouse anti-TLR4 (1:250, #sc-293070, Santa Cruz, CA, USA), rabbit anti-MD2 (1:1000, #ab-24182, Abcam, MA, USA), mouse anti-MyD88 (1:500, #ab-2064, Santa Cruz, CA, USA), rabbit anti-phospho-AKT 1/2/3 Ser473 (1:1000, sc-7985, Santa Cruz, CA, USA), rabbit anti-AKT 1/2/3 (1:1000, sc-8312, Santa Cruz, CA, USA). After washing with TBS-T (150 mM sodium chloride, 100 mM Tris base, pH 7.5, 0.02% Tween 20), membranes were incubated with horseradish peroxidase-conjugated secondary antibodies: goat anti-mouse IgG (1:10,000, #31,430, Invitrogen, Thermo Fisher Scientific Inc., MA, USA) or goat anti-rabbit IgG (1:10,000, #31,460, Invitrogen, Thermo Fisher Scientific Inc., MA, USA) for 2 h at room temperature.

Protein bands were detected by chemiluminescence using the Amersham™ Imager 600 (Life Sciences, UK, USA) and quantified using ImageJ/Fiji software (Bethesda, USA). Densitometry data, expressed as a percentage of the control group (100%), were obtained in arbitrary optical density units. Alpha-tubulin (1:5000, #T9026, Sigma-Aldrich™, MO, USA) was used as the loading control. A detailed list of antibodies used for Western Blot is available in Supplementary Table S6.

Statistical Analysis

Data normality was verified using the Shapiro–Wilk test. Comparisons among multiple groups with normally distributed data were performed using a one-way analysis of variance (ANOVA) followed by Tukey’s post-hoc test, while non-normally distributed data were analyzed using the Kruskal–Wallis followed by Dunn’s post-hoc test. Pairwise comparisons of normally distributed data, available in the supplementary data, were conducted using the unpaired Student’s t-test. Statistical analyses were performed using GraphPad Prism Software (version 9.5.1, GraphPad, USA), with significance set at p < 0.05. Results are presented as mean ± standard error of the mean (SEM).

Results

Transcriptomic Distribution of Tlr4, Fndc5, and αVβ5 Integrin Across Hypothalamic Cell Types

To map the transcriptomic landscape of irisin-related signaling in the hypothalamus, we analyzed publicly available scRNA-seq datasets [32–34] using dbMAP embeddings to visualize cell-type heterogeneity (Fig. 2A). Fndc5 transcripts were primarily detected in neuronal clusters (Fig. 2B; Supplementary Fig. S2A), whereas the receptor subunits Itgav and Itgb5 showed enriched expression in microglia, with lower levels in astrocyte-tanycyte populations (Fig. 2C–E; Supplementary Fig. S2B–D). In parallel, Tlr4 was detected in both microglial and astrocyte–tanycyte clusters, with relatively higher expression in microglia (Fig. 2F; Supplementary Fig. S2E). These expression profiles support the rationale for exploring the potential interaction between irisin and hypothalamic TLR4 signaling in obesity.

Fig. 2.

Single-cell RNA sequencing analysis of the adult mouse hypothalamus from public datasets. (A) dbMAP embedding of 51,623 hypothalamic cells, colored by annotated clusters (neurons, oligodendrocytes, astrocyte–tanycytes, microglia, vascular and mural cells). Each point represents a single cell, with spatial proximity reflecting transcriptional similarity and defining distinct cell-type populations. (B–F) Normalized expression feature plots of Fndc5 (B), Itgav (C), Itgb5 (D), Itgav/Itgb5 co-expression (E), and Tlr4 (F)

Central Irisin Administration Selectively Reduces iWAT Mass Without Affecting Body Weight or Food Intake

To assess the central effects of irisin in obesity, male mice were fed a HFD for 12 weeks and then received ICV irisin or vehicle for 7 consecutive days. The obese phenotype was confirmed prior to ICV surgery by increased body weight, higher caloric intake, elevated fasting glucose levels, and impaired glucose tolerance in HFD-fed mice versus CD-fed controls (Supplementary Fig. S3A–I).

After the surgical procedure and treatment period, HFD-Veh animals maintained a higher body weight than CD-Veh animals (Fig. 3A–C). However, no significant differences in body weight or food intake were observed between HFD-Irisin and HFD-Veh groups (Fig. 3A–E). Throughout the treatment period, both HFD groups exhibited a gradual reduction in body mass, which was more pronounced than that observed in CD-fed controls (Fig. 3A). This differential response likely reflects the increased sensitivity of obese animals to surgical stress and ICV procedures, rather than a specific effect of irisin.

Fig. 3.

Central irisin infusion for 7 days promotes depot-specific iWAT reduction without altering body weight or food intake in DIO mice. (A) Body weight during treatment (n = 12). (B) Body weight measured at euthanasia (n = 12). (C) Body weight change (%) during treatment (n = 12). (D-E) Cumulative caloric intake (kcal) and area under the curve (AUC) of food intake (n = 12). (F) Lee index (g/mm) (n = 12). (G) Masses of inguinal white adipose tissue (iWAT), epididymal white adipose tissue (eWAT), retroperitoneal white adipose tissue (rWAT), and interscapular brown adipose tissue (iBAT) (n = 9–12). (H) Masses of the gastrocnemius (Gastro), soleus, tibialis anterior, and extensor digitorum longus (EDL) muscles (n = 6–12). Body weight and food intake were monitored across all experimental cohorts, and the data shown are representative of the consistent phenotype observed throughout the study. All statistics were performed using one-way ANOVA followed by Tukey's post-hoc test and are presented as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; ^##p < 0.01 and ^####p < 0.0001 (CD-Veh vs. HFD-Veh and HFD-Irisin)

Tissue weight measurements confirmed the expected increase in adiposity and a concomitant reduction in muscle mass (gastrocnemius and tibialis anterior) in HFD-Veh mice compared to controls, along with an increased Lee index (Fig. 3F–H). Interestingly, even in the absence of differences in body weight and food intake between the HFD-Irisin and HFD-Veh groups (Fig. 3A–E), irisin-treated animals exhibited a selective reduction in iWAT mass (Fig. 3G), suggesting that central irisin administration may exert depot-specific effects on peripheral fat accumulation.

Central Irisin Administration Attenuates Hypothalamic Glial Reactivity in Obese Mice

Given the established association between obesity and sustained low-grade neuroinflammation, characterized by increased hypothalamic glial reactivity [46], we investigated whether central irisin infusion modulates this response. Immunofluorescence staining revealed that HFD-Veh mice showed an increased percentage of Iba1 + (microglia/macrophages) and GFAP + (astrocytes) positive area in the hypothalamus compared with CD-Veh controls. This obesity-induced glial activation was attenuated by ICV irisin infusion, restoring staining density to levels comparable to those of lean mice (Fig. 4A–D).

Fig. 4.

Central irisin infusion for 7 days reduces hypothalamic glial reactivity markers (Iba1 + and GFAP +) in DIO mice. (A–B) Representative images immunofluorescence (green) and quantification of the percentage of Iba1 + area in the hypothalamus (n = 4). (C–D) Representative images immunofluorescence and quantification of the percentage of GFAP + area (red) in the hypothalamus (n = 4). Cell nuclei were stained with DAPI (blue). The hypothalamic region shown in the images corresponds to the arcuate nucleus. Scale bar = 50 μM; 3V = third ventricle. Independent cohort of animals were used for these analyses. All statistics were performed using one-way ANOVA followed by Tukey's post-hoc test and are presented as mean ± SEM. *p < 0.05, ^**p < 0.01

Central Irisin Administration Attenuates TLR4/MyD88 Pathway Markers and Increases Anti-inflammatory Cytokine mRNA Levels in the Hypothalamus of Obese Mice

Hypothalamic TLR4 signaling is a primary driver of obesity-associated neuroinflammation [46]. Following the observed reduction in glial activation, we next examined the effect of irisin on this pathway. TLR4 activation requires receptor dimerization and recruitment of MyD88 to initiate the downstream inflammatory signaling cascade [47] (Fig. 5A). HFD-Veh mice displayed an upregulation of Tlr4, Md2, and Myd88 mRNA compared to CD-Veh controls, with similar upward trends in protein levels (TLR4: p = 0.07; MyD88: p = 0.06) (Fig. 5B–E). In contrast, irisin treatment significantly blunted TLR4 and MyD88 proteins levels, suppressed Tlr4 and Md2 gene expression, and reduced NFκB p65 levels compared to HFD-Veh (Fig. 5B–G).

Fig. 5.

Central irisin infusion for 7 days attenuates hypothalamic TLR4/MyD88 signaling and upregulates anti-inflammatory cytokine transcripts in DIO mice. (A) Schematic representation of the simplified TLR4-MyD88 signaling pathway. (B) Gene expression of pro-inflammatory markers Tlr4, Md2, Myd88, Il1b, and Il6 in the hypothalamus (n = 5–7). (C–F) Quantification of hypothalamic TLR4 (C), MD2 (D), MyD88 (E), and NFκB p65 (F) protein levels detected by Western blot, using α-tubulin as a loading control (n = 6–7). (G) Representative Western blot bands for the analyzed proteins. Black lines indicate non-adjacent lanes or proteins from independent blots processed under identical experimental conditions and analyzed in parallel. (H) Gene expression of Bax and Bcl2 in the hypothalamus, pro-apoptotic and anti-apoptotic markers, respectively (n = 5–7). (I) Gene expression of anti-inflammatory markers Il10, Il4, and Bdnf in the hypothalamus (n = 5–7). Gene expression levels were normalized to Gapdh. Separate cohorts of animals were used for mRNA and protein quantification. Statistical analysis was performed using one-way ANOVA with Tukey's post-hoc test, or Kruskal–Wallis followed by Dunn's post-hoc test. Data are presented as mean ± SEM. ^✱p < 0.05, ^✱✱p < 0.01, and ^✱✱✱✱p < 0.0001. The illustration in Panel A was created with BioRender.com under publication license (RF28MRV318)

As NFκB triggers the transcription of pro-inflammatory cytokines [47], we also evaluated Il1b and Il6 expression. Both cytokines were elevated in the HFD-Veh group; irisin treatment significantly reduced Il6 mRNA levels, whereas Il1b showed a visual reduction that did not reach statistical significance, likely due to high variability in the non-parametric analysis (Fig. 5B). Considering that sustained hypothalamic inflammation is a known stimulus for cellular stress [48], we next evaluated key regulators of the apoptotic machinery. The pro-apoptotic gene Bax was upregulated in the HFD-Veh group and remained elevated following the 7-day irisin treatment, with no concomitant changes in the anti-apoptotic gene Bcl2 across groups (Fig. 5H).

Beyond TLR4 inhibition, irisin treatment stimulated compensatory anti-inflammatory responses. HFD-Veh mice exhibited decreased Il10 expression and a trend toward reduced Il4 levels (p = 0.08), whereas irisin treatment significantly upregulated both cytokines, returning their expression to control levels (Fig. 5I). Although BDNF is often linked to irisin's anti-inflammatory actions, its transcript levels remained unchanged (Fig. 5I). Notably, irisin treatment led to a significant reduction in hypothalamic Fndc5 expression in HFD-fed animals (Supplementary Fig. S4), suggesting a potential negative feedback mechanism.

Central Irisin Administration Restores Hypothalamic Insulin-stimulated Akt Phosphorylation in Obese Mice

Chronic hypothalamic TLR4 activation impairs central insulin signaling in obesity [7] (Fig. 6A). Based on this, we investigated whether reducing hypothalamic inflammation through central irisin treatment could improve insulin sensitivity. To this end, mice received an ICV injection of insulin (10 mU per animal), and the hypothalamus was collected 20 min later to assess insulin-stimulated AKT phosphorylation (Fig. 6B). Compared to CD-Veh mice, the HFD-Veh group showed reduced insulin-induced AKT phosphorylation (p = 0.05), indicating an impaired response to the hormone (Fig. 6C–D). Irisin-treated HFD animals displayed an increase in pAKT levels following the insulin challenge, reaching levels comparable to those observed in CD-Veh mice (Fig. 6C–D). These findings indicate that central irisin treatment restores insulin-induced AKT phosphorylation in the hypothalamus of diet-induced obese mice.

Fig. 6.

Central irisin infusion for 7 days restores hypothalamic insulin-stimulated AKT phosphorylation in DIO mice. (A) Schematic overview of the crosstalk between the TLR4–MyD88 and insulin signaling pathways. (B) Experimental protocol for ICV insulin administration and hypothalamic tissue collection. (C–D) Representative Western blot bands and quantification of phosphorylated AKT and total AKT protein levels in the hypothalamus, using α-tubulin as a loading control (n = 5). Black lines indicate non-adjacent lanes or proteins from independent blots processed under identical experimental conditions and analyzed in parallel. Independent cohort of animals were used for these analyses. All statistics were performed using one-way ANOVA followed by Tukey's post-hoc test and are presented as mean ± SEM. ^✱p < 0.05. The illustration in Panel A was created with BioRender.com under publication license (RF28MRV318)

Discussion

Our findings provide new evidence that central irisin administration mitigates hypothalamic inflammation in a DIO mouse model. Specifically, irisin treatment attenuated glial reactivity, downregulated key components of the TLR4/MyD88 axis, and restored insulin-induced AKT phosphorylation in obese mice. These central anti-inflammatory effects coincided with a significant reduction in iWAT mass, although the mechanisms underlying this depot-specific decrease were not directly addressed here.

Despite the relatively well-characterized peripheral actions of irisin [49–51], the relevance of its circulating and tissue-specific levels in obesity remains uncertain. Emerging evidence suggests that circulating irisin levels decline in obesity and type 2 diabetes [52–55], a trend supported by reports of decreased skeletal muscle Fndc5 expression — the primary source of circulating irisin [56]. While some conflicting findings exist [57, 58], these discrepancies likely stem from cohort heterogeneity and technical challenges in reliable irisin quantification [59]. In the hypothalamus, the Fndc5 status under metabolic stress remains scarcely investigated; our data indicate that obesity per se does not alter hypothalamic Fndc5 mRNA levels, in line with prior reports in HFD-fed mice [22]. Conversely, we observed that central irisin administration significantly downregulated local Fndc5 expression (Supplementary Fig. S4), pointing toward a potential negative feedback mechanism. This finding not only reinforces the biological activity of the recombinant protein but also underscores the value of our pharmacological approach in assessing the central actions of irisin independently of endogenous expression levels.

Previous studies have demonstrated irisin's signaling via αVβ5 receptors on astrocytes [17, 25] and microglia [26, 60], supporting the notion that these glial populations are responsive to irisin. In line with this evidence, our scRNA-seq analyses confirmed integrin subunit (Itgav and Itgb5) expression in hypothalamic glial cells and showed that both Tlr4 and the integrin subunits are enriched in microglia, suggesting that these cells may integrate inflammatory signaling and irisin responsiveness at the molecular level. Although mRNA expression does not necessarily reflect functional protein abundance, our histological findings supported this possibility, as central irisin treatment attenuated the obesity-induced elevation in Iba1⁺ and GFAP⁺ immunoreactivity. This reduction in glial activation markers indicates a shift in the hypothalamic inflammatory landscape, even though direct receptor engagement remains to be functionally validated.

Glial reactivity changes do not inherently indicate a specific functional phenotype [61]. Recent studies have reported paradoxical metabolic benefits of hypothalamic microgliosis in experimental obesity models, including improved glucose tolerance via TNFα-dependent activation of POMC neurons [62] and enhanced insulin sensitivity through IL-2–mediated neuroimmune signaling [63]. Nevertheless, in chronic DIO models, gliosis is more consistently associated with sustained pro-inflammatory signaling and insulin resistance [7, 64]. Aligning with this pathological framework, we observed a significant upregulation of TLR4 pathway components in obese mice, an effect effectively countered by a 7-day ICV administration of recombinant irisin. By suppressing the TLR4/MyD88/NF-κB axis at the transcriptional and/or protein levels, our findings extend irisin’s known immunomodulatory capacity [25, 65–68] specifically to the obese hypothalamus. Furthermore, the increased hypothalamic Il10 and Il4 expression observed in our study suggests a shift toward an anti-inflammatory milieu, as reported in other models of neuroinflammation [26].

Regarding the magnitude of these molecular changes, it is noteworthy that while Tlr4, Myd88, and Md2 mRNA levels were significantly elevated in obese mice, the corresponding protein levels exhibited only a non-significant upward trend. This discrepancy likely reflects the higher sensitivity of RT–qPCR compared with western blotting, particularly in small hypothalamic samples where localized or cell-specific protein changes may be diluted in whole-tissue lysates. Additionally, chronic metabolic stress often involves compensatory protein turnover mechanisms, and the ICV cannulation protocol combined with daily handling may have induced a degree of basal inflammation in both groups, potentially narrowing protein-level differences between HFD-fed and control animals. Despite these technical considerations, the impact of recombinant irisin administration in suppressing these pro-inflammatory markers remains evident.

The coordinated reduction in glial reactivity and in the expression of TLR4-associated inflammatory components may be relevant for hypothalamic function, as a pro-inflammatory glial phenotype is known to impair insulin signaling in neighboring neurons [43]. In obesity, inflammatory pathways such as IKK/NF-κB and JNK impair insulin receptor substrate (IRS1) signaling through inhibitory serine phosphorylation, often via SOCS3 or PTP1B induction [69–72]. To assess the functional impact of irisin on this axis, we measured AKT protein, a key mediator of hypothalamic insulin signaling and energy homeostasis [73, 74]. Obese mice exhibited a reduction in insulin-induced AKT phosphorylation compared to controls, suggesting an impairment in the activation of this pathway; nevertheless, central irisin treatment was able to restore these levels. This aligns with prior evidence indicating that irisin modulates insulin-related signaling pathways and promotes cellular resilience under metabolic stress [65, 75, 76]. While further studies are needed to determine whether this recovery involves a direct reduction in IRS1 inhibitory phosphorylation or other specific signaling intermediates, these data suggest that irisin may rescue hypothalamic insulin responsiveness by attenuating local inflammatory tone.

Bidirectional communication between the hypothalamus and peripheral energy stores via the autonomic nervous system is a fundamental axis for metabolic control [77, 78]. This axis is often dysregulated in obesity, characterized by altered sympathetic outflow and impaired responsiveness of adipose tissue to central signals [79]. In the present study, ICV irisin selectively reduced iWAT mass without affecting food intake or total adiposity. This observation points to a potential influence of central irisin infusion on the hypothalamic–adipose tissue axis, possibly involving sympathetic modulation. Previous reports have shown that acute ICV administration of irisin can activate hypothalamic neurons [31], increase energy expenditure [29], and elevate norepinephrine levels [30]. The reduction in iWAT may reflect an early window of adipose remodeling that precedes broader systemic changes in body weight. Thus, these hypothesis-generating observations provide a basis for future research to explore whether central irisin drives adipose tissue remodeling independently of weight loss, requiring further validation through sympathetic denervation and histological analysis.

Ultimately, despite evidence that irisin increases BDNF and reduces apoptosis in other brain regions [39], we detected no changes in hypothalamic Bdnf or Bax mRNA in obese mice following irisin treatment. This likely reflects low basal Bdnf expression in the hypothalamus compared to hippocampus [80, 81], where peripheral FNDC5/exercise robustly upregulates it [16]. Therefore, the central actions of irisin in the hypothalamus likely occur through region-specific, BDNF-independent pathways.

Limitations and Future Directions

This study provides new insights into the central role of irisin in hypothalamic neuroinflammation; however, some limitations must be acknowledged. Although ICV infusion was used to investigate the central actions of irisin, this route is not clinically translatable. Nevertheless, these results establish a proof-of-concept that increasing central levels of irisin attenuates key markers of hypothalamic inflammation in a DIO model. Future research should explore less invasive delivery strategies, such as BBB-permeable analogs, intranasal administration, or nanoparticle-based carriers. Furthermore, despite the observed reduction in glial markers, the specific involvement of the αVβ5 integrin receptor requires experimental validation. Genetic strategies targeting αVβ5 signaling in a cell-specific manner will be critical to confirm its role in mediating the effects of irisin and to determine if this modulation occurs directly on glia or through indirect neuronal pathways. Regarding neuroprotection, as Bax mRNA levels may not capture post-translational dynamics — such as the BAX/BCL-2 protein ratio or caspase activation — additional analyses of protein levels (via western blot or TUNEL staining) are necessary to fully define the anti-apoptotic potential of irisin in this model. Finally, as this study was conducted exclusively in male mice, investigating sex-specific responses is essential to ensure the broad applicability of these findings.

Conclusions

In conclusion, our study demonstrated that central irisin administration modulates the hypothalamic inflammatory landscape in chronic obesity. By attenuating TLR4/MyD88 signaling and glial activation markers, irisin treatment creates a microenvironment more permissive to metabolic regulation, as reflected by restored insulin-induced AKT phosphorylation in the hypothalamus of obese mice. The selective reduction in iWAT mass further supports a possible association with central irisin–dependent pathways in brain–periphery metabolic communication. Together, these findings position irisin as a central regulatory axis possibly linking hypothalamic inflammation to metabolic control, providing a foundation for future studies on underlying mechanisms and translational potential.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AKT

Protein Kinase B

BMI

Body Mass Index

CCA

Canonical correlation analysis

CD

Chow/Control Diet

CNS

Central Nervous System

DAPI

4′,6-Diamidino-2-phenylindole

dbMAP

Diffusion-Based Manifold Approximation and Projection

DIO

Diet-Induced Obesity

EDL

Extensor Digitorum Longus

eWAT

Epididymal White Adipose Tissue

FNDC5

Fibronectin type III domain-containing protein 5

GFAP

Glial Fibrillary Acidic Protein

GAPDH

Glyceraldehyde-3-Phosphate Dehydrogenase

HFD

High-Fat Diet

Iba1

Ionized Calcium-Binding Adapter Molecule 1

ICV

Intracerebroventricular

IL1β

Interleukin 1 Beta

IL6

Interleukin 6

iBAT

Interscapular Brown Adipose Tissue

iWAT

Inguinal White Adipose Tissue

IR

Insulin Receptor

IRS

Insulin Receptor Substrate

JNK

C-Jun N-terminal Kinase

MAPK

Mitogen-Activated Protein Kinase

MD2

Myeloid Differentiation Factor 2

MyD88

Myeloid Differentiation Primary Response 88

PFA

Paraformaldehyde

PI3K

Phosphoinositide 3-Kinase

PIP3

Phosphatidylinositol (3,4,5)-trisphosphate

POMC

Proopiomelanocortin

PTP1B

Protein Tyrosine Phosphatase 1B

rWAT

Retroperitoneal White Adipose Tissue

RT-qPCR

Reverse Transcription Quantitative Polymerase Chain Reaction

SOCS3

Suppressor of Cytokine Signaling 3

TLR4

Toll-Like Receptor 4

TNFα

Tumor Necrosis Factor Alpha

Authors Contribution

**K.C.P. Bem:** Conceptualization, Methodology, Investigation, Formal analysis, Data curation, Visualization, Writing – Original Draft, Writing – Review & Editing. **H.C.L. Barbosa:** Supervision, Conceptualization, Funding acquisition, Writing – Review & Editing. **T.C. Talpo, G.A.S. Nogueira, A.M. Zanesco, J.A.S. Junior:** Investigation, Resources, Writing – Review & Editing. **D.S. Oliveira:** Software, Formal analysis. **L.A. Velloso, A.C. Boschero:** Writing – Review & Editing, Funding acquisition.

Funding

The Article Processing Charge (APC) for the publication of this research was funded by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) (ROR identifier: 00x0ma614). This study is part of the master's research of Kelly Cristina Pereira Bem and was funded by the São Paulo Research Foundation (FAPESP; grant numbers 2021/01448–1, 2021/11189–3, 2021/04664–7, and 2013/07607–8).

Data Availability

The datasets generated and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Declarations

Clinical Trial Number

Not applicable.

Competing Interest

The authors declare no competing interests.