This study was approved by the Animal Care Committee of the Health Sciences Center, Federal University of Rio de Janeiro (CEUA-190/13), and registered with the Brazilian National Council for Animal Experimentation Control. All animals received humane care in compliance with the “Principles of Laboratory Animal Care” formulated by the National Society for Medical Research and the U.S. National Academy of Sciences Guide for the Care and Use of Laboratory Animals.

Male 129/SVxC57/BL6 mice, 10–12 weeks of age, weighting 22–25 g, were kept under a 12/12 h light/darkness cycle, in a controlled-temperature environment (23°C).

For fasting, in three independent experiments, 30 mice were divided into two groups (n = 15/group): one without access to food for 48 h (fasting group) and another group with food available ad libitum (control group). Both groups had free access to water. Twenty animals (n = 10/group) were used to evaluate mRNA expression and hormone levels. Ten animals (n = 5/group) were used to evaluate thyroid monocarboxilate transporter 8 by immunohistochemistry.

For sepsis, three independent experiments were performed. Sepsis was induced by cecal ligation and puncture surgery in 17 animals, whereas, 10 mice were sham-operated and used as control. Briefly, animals were anesthetized with intraperitoneal (i.p.) injection of ketamine and xylazine (0.25 and 0.025 mg/kg, respectively) and a midline laparotomy (2-cm incision) was performed. The cecum was carefully isolated to prevent damage to blood vessels and ligated with a 3-0 cotton suture. The ligature was placed below the ileocecal valve to prevent bowel obstruction. Finally, the cecum was punctured once with an 18-gauge needle and the animals left to recover from anesthesia (de Araujo et al., 2012). In Sham group, the abdominal cavity was opened and the cecum was isolated without ligation and puncture. Mice were fasted for 16 h before surgery in both groups. Postoperative care was similar in both groups, and consisted of a subcutaneous injection of tramadol hydrochloride (20 mg/g body weight) in 1 mL warm (37°C) normal saline (NaCl 0.9%). At 24 h, seven animals from CLP group died.

Mice were killed under anesthesia with isoflurane (Cristalia, Brazil) after 48 h of fasting or 24 h of sepsis induction, at 10:00 am in order to avoid hormonal and gene expression circadian variation. After killing, blood, thyroid, liver, hypothalamus, and pituitary were collected for molecular biology analysis (mRNA and protein expression by qPCR and Western Blot, respectively). Serum was obtained from the blood and stored at −20°C and all tissues were immediately frozen in liquid nitrogen and stored at −70°C. Bone marrow, peritoneal, and pleural fluids were collected for leukocyte analysis.

Serum total T₃ and T₄ concentrations were measured by radioimmunoassay (RIA), using a commercial kit (Total T₄ MAb RIA Kit and Total T₃ RIA Kit, MP Biomedicals, CA, USA) and following the manufacturer's recommendations. This kit is based on solid phase method and the limit of detection for T₄ was 1 μg/dL and for T₃ was 25 ng/dL.

Total RNA from liver, hypothalamus, pituitary, and thyroid samples was extracted using the TRIzol method, according to manufacturer's protocol (TRIzol Reagent; Life Technologies, CA, USA). In the fasting experiment, total RNA from PVN and ARC nuclei cryosections was extracted using the RNA easy microkit (Qiagen, Hilden, Germany) following manufacturer's recommendations. The cDNA synthesis was performed using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, CA, USA) with 1 or 0.5 μg of total RNA, for tissue or nuclei cryosections respectively. All RNA samples were reverse transcribed in a single reaction. After the cDNA synthesis, mRNA expressions were evaluated by qPCR using the Maxima SYBR Green/ROX qPCR Master Mix 2X (Thermo Fisher Scientific, MA, USA) and the Master Cycler Realplex system (Eppendorf, Germany). Primer pair's sequences are shown in Table 1. The efficiency range accepted for each assay was 95–105%. qPCR quality and genomic DNA contamination was checked using intron-spanning primers, reverse transcriptase-negative samples from cDNA synthesis and melting curve analysis obtained from each reaction. Quantification of the samples mRNA expression was calculated from the quantification cycle (Cq) by the 2^−ΔΔCq method (Livak and Schmittgen, 2001) and corrected using 36β4 (Rplp0) expression as the reference gene. The results are expressed relative to the values of the control group (fasting) or Sham (CLP-induced sepsis), which were considered to be equal to 1.

Hypothalamic PVN and ARC nuclei were obtained from snap frozen mouse brains (n = 10/group), previously collected with the other tissues, as described in “animal experiment” section. The nuclei were collected in cryosections by the punch technique, previously described (Franco et al., 2012, 2015), with particularities for the mouse brain, and following the mouse brain stereotaxic atlas coordinates (Paxinos and Franklin, 2004). The PVN and ARC nucleus were obtained through subsequently 640 and 1,500 μM slices. The slices were collected, respectively, at 0.58 mm and 1.22 mm posterior to the bregma.

Liver samples were homogenized using the Tissuelyser LT (Qiagen, Hilden, Germany) with lyses buffer (pH 6.4; 50 mM HEPES, 1 mM MgCl2, 10 nM EDTA, and 1% Triton X) and a protease inhibitor cocktail (Roche/DSM Nutritional). Total protein extract (20 μg), previously denatured, was submitted to SDS-PAGE electrophoresis and transferred to a polyvinylidene difluoride membrane (PVDF Hybond-P; Amersham Biosciences, Buckinghamshire, UK). After that, the membrane was blocked for 2 h with 5% non-fat dry milk (Molico; Nestle, São Paulo, Brazil) and incubated overnight with anti-MCT8 antibody (1:750, Santa Cruz Biotechnology) and anti-cyclophilin antibody (1:250.000; Affinity Bioreagents, Golden, CO, USA), the loading control. Both primary antibodies were diluted in Tween-TBS (TBS + 0.1% Tween 20) with 2% non-fat dry milk. Membranes were then incubated for 3 h with secondary antibodies: HRP anti-goat IgG (1:7,000; Santa Cruz Biotechnology) and HRP anti-rabbit igG (1.2:10,000; Amersham Biosciences, Buckinghamshire, UK). After seven wash steps, the membranes were then incubated for chemiluminescence visualization with ECL (ECL Western Blotting System, Amershan Biosciences) and densitometry analysis was performed using the Image Quant Las4000 software (GE HealthCare Life Sciences, Buckinghamshire, UK). The results are expressed relative to the values of the control group (fasting) or Sham (CLP-induced sepsis), which were considered to be equal to 1.

The thyroid glands were fixed in paraformaldehyde 4%. Cryopreservation was performed with sucrose gradient and posterior OCT embedding. A 5 μm cut was obtained with an ultramicrotome and the sections were then incubated with primary anti-MCT8 antibody (sc-47125, Santa Cruz Biotechnollogy), followed by the appropriate biotinylated secondary antibody (DCMT999, Spring), streptavidin-peroxidase (DHRR999, Spring) and finally 3,3-diamino-benzidine (DAB -ACB060, SCYTEK). The sections were then stained with Mayer's hematoxylin and mounted with glycerol. To quantify the MCT8 protein expression levels in the control and fasting groups, 30 random fields per animal were captured using an AXIOSCOPE microscope with an AXIOCAM HRc camera (CarlZeiss AG, Jena, Germany) at a magnification of 100× and analyzed with Image J analysis software.

In CLP induced sepsis animals, after 24 h, total and differential leukocyte counts were evaluated in bone marrow, peritoneal lavage fluid, and pleura. Peritoneal and pleural fluids were collected by flushing these cavities twice with 1.0 ml of 37°C sterile, PBS containing 10 mM EDTA using a micropippete. Mice femoral epiphysis were collected, centrifuged (98 g, 10 min, 4°C) and cell pellets were re-suspended in PBS for further quantification of leukocytes in the hemocytometer, as well as peritoneal and pleural washes, after dilution in Türk solution (1:40). After total leukocyte counts, samples from bone marrow and peritoneal/pleural washes were centrifuged (221 g, 10 min, 5°C). From the cell pellets obtained, differential cell counts were performed on May–Grunwald–Giemsa-stained cytospin preparations under an oil immersion objective to determine the percentage of mononuclear cells, eosinophils, and neutrophils in 100 leukocytes counted.

Sample size calculation was based on previous experimental studies of fasting (de Vries et al., 2015) and sepsis (Bloise et al., 2016). A sample size of eight animals per group (allowing for one animal as dropout) would provide the appropriate power (1-β = 0.8) to identify significant (α = 0.05) differences in variable analyzed between injured (fasting or sepsis) and Control or Sham groups, taking into account an effect size d = 1.8, a two-sided test, and a sample size ratio = 1 (G^*Power 3.1.9.2, University of Düsseldorf, Germany).

The Kolmogorov-Smirnov test with Lilliefors' corrections was used to test for normality of data, while the Levene median test was used to evaluate the homogeneity of variances. Rout test was used to remove outliers. The data are expressed as the mean ± the standard error of the mean (SEM). The Student's t-test was used for comparisons between two groups when data was normal otherwise Mann Whitney test was used. (mRNA expression, western blotting, RIA, and leukocyte counts). Statistical analyses were performed using the Graphpad prism 6 software (GraphPad Software, Inc., San Diego, CA, USA). Differences were considered to be significant at P < 0.05.

In fasting group, serum T₄ levels were undetectable and serum T₃ levels were lower (38%) in fasting mice compared to control (Table 2, P < 0.001). Fasting group also showed a 22% reduction in total body weight after a 48-h period of fasting (20.1 ± 1.4 g), compared to control (25.9 ± 0.9 g, P < 0.01).

In CLP-induced sepsis group, serum T₄ levels decreased 40% compared to Sham (Table 2, P < 0.01). No significant changes were observed in serum T₃ levels between fasting and CLP groups.

Total leukocyte counts (Figure 1) in bone marrow decreased by 37% in CLP-induced sepsis group (P < 0.05) and increased by 2.5- and 1.5-fold, respectively, in the peritoneal and pleural lavages of the CLP group (P < 0.01 and P < 0.05, respectively). Differential leukocyte counts (Figure 1) showed a 67% decrease in bone marrow neutrophils of the CLP-induced sepsis group (P < 0.05). It was also observed an important increase in neutrophils of the peritoneal and pleural lavages in the CLP mice, compared to Sham group (P < 0.05).

The influence of fasting (Figure 2A) or CLP-induced sepsis (Figure 2B) on liver thyroid hormone metabolism was evaluated through the analysis of deiodinases mRNA expression. Dio1 expression was decreased by 97% (P < 0.001) and Dio3 expression had a 3-fold increase in fasting mice, compared to control group (P < 0.05). Since genomic actions of thyroid hormones depend on their binding to nuclear receptors (THRs), we evaluated the expression of Thra and Thrb, which encode to the transcription factors and T₃ nuclear receptors Thrα and Thrβ, respectively. Thra expression decreased by 55% in fasting group compared to control (Figure 2A, P < 0.001), but Thrb expression did not change. To analyze whether the reduced Thra mRNA expression could reflect reduced T₃ action, we investigated the mRNA expression of Thrsp in the liver, a T₃ positive-regulated gene (Jump et al., 1984). Fasted mice showed a 98% reduction in Thrsp mRNA expression compared to control group (Figure 2A, P < 0.01).

In CLP-induced sepsis group, Dio1 expression was decreased by 49% (Figure 2B, P < 0.05) and Dio3 expression did not change between CLP-induced sepsis and Sham groups. CLP-induced sepsis mice showed a decrease in both Thra and Thrb mRNA expressions by 45% (P < 0.01) and 49% (P < 0.001), respectively, compared to Sham group. CLP-induced sepsis also promoted 87% decrease Thrsp mRNA expression in the liver (P < 0.05).

We then investigated the expression of the monocarboxylate transporters MCT8 and MCT10. Fasting promoted a 4-fold increase in Slc16a10 mRNA expression (Figure 2A, P < 0.001) and, in contrast, decreased by 77% Slc16a2 gene expression (Figure 3A, P < 0.001). Furthermore, fasted mice showed a 30% reduction in MCT8 protein expression in the liver compared to the control group (Figures 3B,C, P < 0.05). In contrast, CLP-induced sepsis promoted a 53% decrease in liver Slc16a10 mRNA expression (Figure 2B, P < 0.01) and a 49% decrease in liver Slc16a2 mRNA expression (Figure 3D, P < 0.05). No differences were seen on MCT8 protein expression in CLP-induced sepsis (Figures 3E,F).

We analyzed the expression of monocarboxilate transporters (MCT8 and MCT10) in the thyroid gland to investigate a possible role of this gland, during fasting and CLP-induced sepsis, causing changes in serum thyroid hormone levels. mRNA expression of Slc16a2 and Slc16a10 was decreased by 50% in fasting mice, compared to control group (Figure 4A, P < 0.05 and P < 0.01, respectively). In CLP-induced sepsis, Slc16a2 mRNA expression decreased 48% compared to control (Figure 4B, P < 0.05). Slc16a10 mRNA expression did not differ between Sham and CLP-induced sepsis (Figure 4B). Furthermore, immunohistochemistry analysis showed that fasting decreased MCT8 staining in the thyrocytes (Figure 4C, P < 0.01).

Since NTIS changes the set point of HPT axis through different mechanisms in fasting or CLP-induced sepsis, we set out to evaluate the expression of genes involved in thyroid hormone metabolism in the pituitary and the hypothalamus.

The expression of Slc16a2, Dio1, Dio2, and Thrb in the pituitary did no change in fasting mice (Figure 5A), although fasting promoted a 94% reduction in Tshb mRNA expression (Figure 5A, P < 0.0001), which suggest a decrease in the response of the pituitary to the low serum levels of T₄.

Slc16a2, Dio1, and Dio2 mRNA expressions did not differ between Sham and CLP-induced sepsis groups (Figure 5B). However, both Thrb and Tshb mRNA expressions reduced 27% (Figure 5B, P < 0.05) and 76% (Figure 5B, P < 0.001), respectively, confirming that in this model of NTIS, HPT axis is also less responsive to low T₄ serum levels.

In the total hypothalamus, the expression of Slc16a2, Slc16a10, Dio3, and Thrb did not change in these two models fasting and CLP-induced sepsis (Figures 6A,B). However, we observed a 2-fold increase in the expression of Dio2 mRNA in fasting mice, compared to control group (Figure 6A, P < 0.05). Dio2 expression did not change in the total hypothalamus of CLP-induced sepsis mice, compared to control group (Figure 6B). Thus, we decided to perform a more thorough analysis of Dio2 expression in PVN and ARC nuclei only on the fasting model.

We then investigated the Dio2 expression in the PVN and ARC nucleus. In the PVN, Dio2 mRNA expression presented a 2.3-fold increase in fasting group (Figure 7, P < 0.01). In the ARC, fasting promoted a 1.6-fold increase in Dio2 mRNA expression (Figure 7, P < 0.01).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.