Ruthenium red (11103-72-3; FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) [20], A967079 (1170613-55-4; TOCRIS Bioscience, Bristol, UK) [21] and ionomycin (56092-81-0; FUJIFILM Wako Pure Chemical Corporation) were used as ion channel inhibitors and inducers. Ruthenium red was dissolved in distilled water, and A967079 and ionomycin were dissolved in dimethyl sulfoxide. The final concentrations of ruthenium red, A967079, and ionomycin were 10, 100, and 200 mM, respectively. Rabbit anti-UCP1 polyclonal antibody (bs-1925R; Thermo Fisher Scientific, Waltham, MA, USA), rabbit anti-TRPA1 polyclonal antibody (NB110-40763; Novus Biologicals LLC, Centennial, CO, USA), and rabbit anti-TRPM8 polyclonal antibody (NBP1-97311; Novus Biologicals LLC) were used as primary antibodies for immunostaining. Cy3-conjugated anti-rabbit IgG goat antibody (AP132C; Merck, Darmstadt, Germany) was used as the secondary antibody.

3T3-L24 mouse white adipocytes is a cell line obtained by inducing 3T3-L1 mouse progenitor cells (ECACC strain cells, EC91031101-F0) into white adipocytes followed by isolation [22]. HPA primary human preadipocytes derived from subcutaneous adipocytes were obtained from ScienCell Research Laboratories (7220; Carlsbad, CA, USA) and induced to HA human white adipocytes using preadipocyte differentiation medium (ScienCell Research Laboratories; 7221). 3T3-L24 and HA adipocytes underwent 2-weeks passaging culture at 37˚C before use.

Dulbecco’s Minimum Essential Medium and Minimum Essential Medium Eagle (modified) with Earle’s salts (MP Biomedicals Inc., IIIkirch, France) adjusted to pH 7.0 were respectively mixed with fetal bovine serum (A5256701; Thermo Fisher Scientific, Waltham, MA, USA) at a ratio of 9:1, and penicillium-streptomycin solution (168-23191; FUJIFILM Wako Pure Chemical Corporation) was added at a final concentration of 25 μg/mL. These media are hereafter referred to as DMEM and MEM, respectively.

An RNeasy Lipid Mini Kit (74804; Qiagen, Hilden, Germany) was used for mRNA purification, and cDNA was synthesized using a QuantiTect Reverse Transcription Kit (205311; Qiagen). qPCR was performed with the obtained cDNA using a Rotor-Gene device (Qiagen). Commercially available QuantiTect Primer Assays (Qiagen) were used as DNA primers for qPCR. The reaction solution comprised a Rotor-Gene SYBR Green PCR Kit (204443; Qiagen). Reactions were performed for at least 50 cycles of 95˚C for 5 s and 65˚C for 30 s according to the manufacturer’s instructions. β-actin was used as a standard for Ct values, and ΔΔCt values were calculated using the ΔCt values of control cells. The DNA primers used in this study are shown in Table 1.

Table 1. Primers used in this study.

3T3-L24 adipocytes (1 × 10⁵ cells/flask) and HA human adipocytes (4 × 10⁴ cells/well) were seeded into 25 cm²-culture flasks containing 5 mL DMEM and 6-well plates containing 4 mL MEM, respectively, and cultured at 37˚C for 2 days. The respective medium was quickly changed to medium at 37˚C and 15˚C, and incubated for 1 h in an incubator set at 37˚C and 15˚C, respectively. After incubation, the cells were washed once with phosphate-buffered saline (PBS), detached using 0.25% trypsin, and collected by centrifugation (800 × g, 3 min). Thereafter, mRNA was immediately extracted to examine gene expressions by qPCR. All solutions used for cell collection were preheated at 37˚C, and the operation was performed on aluminum plates (WakenBtech, Kyoto, Japan) at 37˚C and room temperature (30˚C) within 15 min.

The amount of NADH per cell, which corresponds to mitochondrial activity, was measured using a Cell Counting Kit 8 (341-07761; FUJIFILM Wako Pure Chemical Corporation), according to the manufacturer’s instructions. Briefly, 10,000 cells were inoculated into a 96-well plate containing 0.1 mL of DMEM and attached for 16 h through incubation. Subsequently, the medium was replaced with 0.1 mL DMEM containing 10 μL CCK-8 solution at 15˚C, and the plate was incubated at 15˚C. The plates were sequentially removed from the incubator at 0, 45, 90, and 120 min, and absorbance was measured at 450 nm. The total cell number was determined using a cell counter (WC2-100; WakenBtech) after the cells were detached with 0.25% trypsin and collected using centrifugation. The increase in the percentage absorbance per cell at 45, 90, and 120 min relative to that of the control cells (0 min) was calculated.

Calcium ion (Ca²⁺) uptake in response to direct cold stimulation was examined using a Calcium Kit-Fluo4 (348-91281; FUJIFILM Wako Pure Chemical Corporation). Briefly, 3T3-L24 cells (2 × 10⁴ cells) were attached by 16 h of incubation at 37˚C in a glass bottom dish (D141400; Matsunami Glass, Osaka, Japan) containing 2 mL DMEM and precultured in loading buffer (20 mM HEPES, 115 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl₂, 1.8 mM CaCl₂, 13.8 mM glucose, pH 7.6, 1.25 mM probenecid, 0.04% of 5% pluronic F-127, and 50 μL Flo4-AM solution) for 1 h at 37˚C. After the sample was placed in the insulation device set on an EVOS^® FL Auto Cell Imaging System (Thermo Fisher Scientific, Massachusetts, USA), the cells were washed with 2 mL of recording buffer (loading buffer without Flo4-AM) at 37˚C. Thereafter, the buffer was replaced with recording buffer at 37˚C, recording buffer containing 5 μM ionomycin at 37˚C, recording buffer at 15˚C, and recording buffer without Ca²⁺ at 15˚C, and the fluorescence intensity was measured every 10 s. For the inhibition experiments, the cells were precultured with loading buffers containing 10 μM ruthenium red or A97079 at 37˚C for 1 h, and recording medium containing 10 μM ruthenium red or A967079 at 15˚C was used for fluorescence intensity analysis.

Next, immunostaining of transport receptor potential ankyrin 1 (TRPA1) and transport receptor potential melastatin 8 (TRPM8) was conducted. Briefly, the cells cultured in glass bottom dishes were washed twice with 2 mL PBS solution, fixed in phosphate buffer containing 4% formaldehyde for 10 min, washed thrice with PBS solution, incubated with PBS solution containing 0.2% TritonX-100 for 10 min, and washed thrice with PBS solution. Subsequently, the sample was blocked with PBS containing 1% BSA for 1 h and washed thrice with PBS solution. Next, rabbit anti-TRPA1 and anti-TRPM8 polyclonal antibody solutions diluted to 1/200 volume in PBS containing 0.1% albumin were added to the pretreated cells, which were then incubated for 1 h at 30˚C and washed four times with PBS. Thereafter, secondary antibody (Cy3-conjugated anti-rabbit IgG goat polyclonal antibody) solution diluted to 1/500 volume in PBS solution containing 0.1% albumin was incubated at 30˚C for 1 h and washed four times with PBS solution. The cells were then observed for fluorescence intensity using an EVOS^® FL Auto Cell Imaging System.

Next, the effect of Ca^2⁺ uptake on gene expression was assessed. Briefly, 3T3-L24 cells (1 × 10⁵ cells) were cultured in 24-well plates at 37˚C for 16 h, after which the cells were replaced with recording medium and Ca²⁺-free recording medium (recording medium without Ca²⁺) and incubated for 30 s at 15˚C; these were replaced with 5 mL DMEM at 37˚C and incubated at 37˚C for 1 h. Subsequently, the resulting cells were rapidly collected, mRNA was extracted, and gene expression was analyzed using qPCR.

Preliminary experiments showed that gene expression after short-term cold stimulation at 15˚C for 1 hour was maintained for more than a day, so this was repeated for long-term cold stimulation. 3T3-L24 adipocytes (5 × 10⁴ cells/flask) cultured in 25-cm² culture flasks were replaced with 5 mL DMEM at 15˚C, maintained for 1 h in an incubator, and then replaced with 5 mL DMEM at 37˚C for 23 h. This cold stimulation procedure was repeated for 5 days, and on the fifth day, the cold-stimulated cells were collected, and mRNA was extracted from a portion of the cells in the same manner as that for short-term direct cold stimulation. The rest of the cells (5 × 10⁴ cells) were inoculated into 25-cm flasks containing 6 mL DMEM and cultured at 37˚C for 2 days. Long-term direct cold stimulation with one cycle of 7 days was continued for 4 weeks.

Long-term cold-stimulated cells were collected weekly, mRNA was extracted, and gene expression was analyzed by qPCR, according to the previously described method. Cells subjected over 2 weeks of long-term direct cold stimulation and 3T3-L24 white adipocytes were cultured for 16 h at 37˚C in 35-mm glass bottom dishes containing 4 mL DMEM, and cell color, mitochondrial content, and UCP1 expression were evaluated. Cell color was observed under an optical microscope (EVOS^® FL Auto Cell Imaging System). For UCP1 immunostaining, rabbit anti-UCP1 polyclonal antibody was used, and immunostaining was conducted similarly to that of TRPA1. For fluorescent mitochondrial staining, cells in 35-mm glass bottom dishes were fixed in 3.7% formamide in PBS for 30 min, washed with PBS, incubated in Molt View Green solution (100 nM Molt View Green in PBS) for 30 min, washed with PBS, and Molt View Green-stained cells were observed for fluorescence intensity using the EVOS^® FL Auto Cell Imaging System.

NADH amounts in beige and white adipocytes were measured using a Cell Counting Kit 8 as a measure of mitochondrial activity. Briefly, 5,000 cells were seeded in 96-well plates, and 3 h later, the medium was replaced with 100 μL DMEM containing 10 μL CCK-8, and after 2 h of incubation at 37˚C, absorbance at 450 nm was measured. Glucose uptake rate was measured using a Glucose Uptake Cell-Based Kit (600470; Cayman Chemical, Ann Arbor, MI, USA). Briefly, cells (3 × 10⁴) were inoculated in 96-well plates and cultured in 100 μL DMEM for 16 h, after which the medium was replaced with 100 μL DMEM containing 2 μg/mL insulin and 1/100 volume μL of Cell Base Assay NDB Glucose (Cayman Chemical), and incubated for 1 h at 37˚C. After washing twice with 100 μL Cell Base Assay Buffer (Cayman Chemical), fluorescence (485/538 nm) was measured.

More than three independent experiments were performed to confirm the reproducibility of the results (the total independent experiments number is indicated as n). For statistical evaluation, p-value based on a two-tailed Student’s t-test were calculated using KaleidaGraph software (HULINKS Inc., Tokyo, Japan). Results with a p-value greater than 0.05 are not shown in the figures. To determine relative fluorescence intensity, several fluorescence micrographs were obtained under the same imaging conditions as those of the controls, and the fluorescence intensities of approximately 60 - 100 cells were calculated and compared using ImageJ software (NIH, Bethesda, MD, USA).

The effects of direct short-term cold stimulation on the gene expression were examined.Figure 1(a) andFigure 1(b) show the ratio of gene expression in 3T3-L24 white adipocytes exposed to 1 h of direct cold stimulation at 15˚C relative to that in 3T3-L24 cells cultured at 37˚C (control cells). Specific beige adipocyte genes (Ucp1and Prdm16), genes associated with the amounts of mitochondria (PGC1a, Cider, Cdk5, and Cox8b), and browning induction genes (Cebpα, Pparg, Fabp4, and Zfb516) were used as marker genes [23][24]. The average expression values of specific beige adipocyte genes and genes associated with the amount of mitochondria in cells exposed to direct cold stimulation were approximately 5–10-fold higher than those in control cells (Figure 1(a)). Moreover, in the cells exposed to direct cold stimulation, Pparg and Zfp516 genes [25][26], which are the most important genes for initial induction to beige adipocyte, were upregulated relative to those in control cells (Figure 1(b)). Figure 1(c)shows the increased NADH per cell when cells were exposed to direct cold stimulation at 15˚C, and the value corresponds to mitochondrial activity. A 15% increase occurred during 2 h of direct cold stimulation. These results suggest that direct cold stimulation of white adipocytes may activate a signaling pathway that induces beige adipocytes.

Figure 1. Effects of direct cold stimulation for 1 h at 15˚C on gene expression related to differentiation into beige adipocytes. Ratio of gene expression in 3T3-L24 adipocytes incubated at 15˚C for 1 h to that of those incubated at 37˚C for 1 h (mean ± SD, n = 6 - 9) of beige adipocyte marker genes (a) and initial induction and maintenance factors of beige adipocytes (b). (c) Increased ratio of the amount of NADH in 3T3-L24 adipocytes incubated at 15˚C to those incubated at 37˚C (mean ± SD, n = 5). (d) Ratio of gene expression of HA adipocytes incubated at 15˚C for 1 h to those incubated at 37˚C for 1 h (mean ± SD, n = 5 - 6).

Whether this pathway is present in human cells was examined. Figure 1(d) shows the ratio of gene expression in human HA white adipocytes exposed to 1 h of direct cold stimulation at 15˚C to that in HA cells cultured at 37˚C (control cells). In the case of human HA white adipocytes, the expression of genes related to direct cold stimulation at 15˚C varied but did not significantly increase (p > 0.05). Thus, the response mechanism to short-term direct cold stimulation differs between human and mouse adipocytes.

Next, the mechanism of the signaling response to short-term direct cold stimulation in mouse adipocytes was examined. Studies have reported an increase in Ca²⁺ when adipocytes or progenitor adipocytes change to brown adipocytes [27], and TRPA1 and TRPM8, which are activated at temperatures below around 20 and 10˚C, respectively, are present in vagus nerve cells and take up Ca²⁺ for activation [28]. Based on these findings, we surmised that a small number of TRP channels may be present in mouse adipocytes, and these channels take up Ca²⁺ in response to direct cold stimulation.

First, the relationship between direct cold stimulation and Ca²⁺ uptake by TRP channels was examined. Figure 2(a) shows the amount of Ca²⁺ influx when 3T3-L24 adipocytes were incubated with recording buffer solutions containing ionomycin, ruthenium red, and A967079 at 15˚C for 10 s; an increase in Ca²⁺ results in green fluorescence. Figure 2(b) shows the average fluorescence intensity calculated from the experiment shown in Figure 2(a) using ImageJ software. At 37˚C, a small amount of Ca²⁺ influx was observed even after 4 min (sample A), but at 15˚C, much Ca²⁺ influx occurred within 10 s (sample D), and the amount

Figure 2. Involvement of Ca²⁺ in signaling in short-term direct cold stimulation. (a) Fluorescence micrographs of cells uptaking Ca²⁺ at 10 s. (b) Relative fluorescence intensity calculated from several images (mean ± SD, n = 80–100). (A) Recording medium at 37˚C (negative control); (B) Recording medium + ionomycin at 37˚C (positive control); (C) Recording medium without Ca²⁺ at 15˚C; (D) Recording medium with Ca²⁺ at 15˚C; (E) Recording medium with Ca²⁺ + ruthenium red at 15˚C; (F) Recording medium with Ca²⁺+ A967079 at 15˚C. (c) Immunostaining of 3T3-L24 adipocytes using anti-TRPA1 and anti-TRPM8 antibodies. (d) Ratio of the gene expression of 3T3-L24 adipocytes incubated in a solution without Ca²⁺ at 15˚C for 30 s to those incubated in a solution with Ca²⁺ at 15˚C for 30 s (mean ± SD, n = 5 - 7).

reached levels approximately similar to those of the positive control (sample B). In contrast, a small amount of influx was observed even in the case of direct cold stimulation at 15˚C with recording solution without Ca²⁺ (sample C), suggesting that the increase in Ca²⁺at 15˚C was not only due to uptake from outside the cells but also due to efflux from the intracellular smooth endoplasmic reticulum. Moreover, Ca²⁺ influx was strongly suppressed when ruthenium red, a TRP channel inhibitor, was added and incubated at 15˚C for 10 s (sample E), and when A967079, a specific TRPA1 channel inhibitor, was added, Ca²⁺ influx was suppressed to approximately 20% of that without A967079 addition (sample F). These results suggest that Ca²⁺ is most likely primarily taken up into the cell by several TRP channels, including TRPA1 channels. Moreover, whether TRPA1 and TRPM8 were present on the cell surface was examined. Figure 2(c) shows the results of immunostaining of mouse white adipocytes using anti-TRPA1 and anti-TRPM8 antibodies. While TRPA1 and TRPM8 expression was low, TRPA1 was shown to be slightly expressed.

Next, we examined whether Ca²⁺ influx into the cells caused increased gene expression. 3T3-L24 adipocytes were subjected to direct cold stimulation at 15˚C for 30 s with Ca²⁺-free buffer and Ca²⁺-containing buffer, which were replaced with DMEM for 1 h at 37˚C, and gene expression levels were compared. Figure 2(d) shows the ratio of gene expression in cold-stimulated cells with the buffer without Ca²⁺ at 15˚C to that in cold-stimulated cells with the buffer containing Ca²⁺ at 15˚C. Pparg and Zfp516 gene expression in cells treated with Ca²⁺-free buffer was significantly downregulated, whereas a reduction in the expression of the other genes was observed but was not significant. Direct cold stimulation at 15˚C may not have resulted in a predominant difference in gene expression ratios because a small amount of Ca²⁺ is secreted from intracellular organelles. In summary, although the detailed mechanism is still unclear, short-term direct cold stimulation likely causes Ca²⁺ uptake via TRP channels, which in turn induces genes such as Pparg and Zfp516.

Finally, we investigated whether prolonged direct cold stimulation induces a change to beige adipose tissue-like cells. Since cell browning and UCP1 protein expression could not observed following only one cycle of direct cold stimulation, repeated long-term direct cold stimulation (1 h of direct cold stimulation at 15˚C once a day for 5 days, followed by an interval of 2 days of incubation at 37˚C after the cells were transplanted) was conducted. Figure 3(a) shows the ratio of Cited1 and Tnfrfs9 (Cd137) gene expression in 3T3-L24 adipocytes exposed to long-term direct cold stimulation to that of 3T3-L24 cells cultured at 37˚C (control). Cited1and Tnfrfs9 are marker genes specifically expressed in beige adipocytes [23][24]. Figure 3(b) shows the average expression levels of these genes in long-term cold-stimulated cells from week 2 to week 4. The expression values of Cited1 and Tnfrfs9 genes gradually increased with long-term direct cold stimulation, almost reaching their maximum after 2 weeks of culture. The average expression values (2 - 4 weeks) ofCited1 and Tnfrfs9 in cells exposed to long-term direct stimulation were 1.2 × 10⁵- and 7.4-fold higher than those in the control. The strain of cells obtained following long-term cold stimulation for more than 2 weeks was designated 3T3-Beige (Figure 3(c)). The color of 3T3-Beige cells was more brownish than that of 3T3-L24 adipocytes (photos A and B) and dark brown spots were observed after more than 3 weeks of culture (photo C).

Figure 3. Effects of long-term direct cold stimulation on a change of 3T3-L24 adipocytes. (a) Ratio of gene expression of 3T3-L24 adipocytes incubated at 15˚C for 1 h to those incubated at 37˚C during long-term direct cold stimulation (n = 3). (b) Gene expression ratio averaged from week 2 to week 4 in Figure 3(a) (mean ± SD, n = 3). (c) Micrographs of 3T3-L24 adipocytes (A), 3T3-Beige cells (B), and brown spots observed in cells cultured by long-term direct cold stimulation (C).

Moreover, the characteristics of 3T3-Beige cells and 3T3-L24 white adipocytes were compared.Figure 4(a) shows the results of UCP1 protein immunostaining and fluorescent staining of mitochondria with anti-UCP1 antibody and Molt View Green; UCP1 and mitochondria emitted red and green fluorescence, respectively. Figure 4(b) shows their fluorescence levels calculated from several images of the experiments shown in Figure 4(a) using ImageJ software. The amounts of UCP1 and mitochondria in 3T3-Beige cells were approximately 1.4- and 1.3-fold higher than those in 3T3-L24 cells, respectively.

The mitochondrial activity and glucose uptake rates were also compared (Figure 4(c) and Figure 4(d)). The amount of NADH, corresponding to mitochondrial activity, in 3T3-Beige cells increased by approximately 1.8-fold, and glucose uptake also increased by approximately 1.7-fold. In addition, we investigated whether 3T3-Beige cells produced the four typical batokines,i.e., fibroblast growth factor 21 (fgf21), insulin-like growth factor 1 (Igf1), follistatin (Fst), and interleukin-6 (IL-6) [10][11]. Figure 4(e) shows the ratio of the gene expression of batokines in 3T3-Beige cells to that in 3T3-L24 adipocytes. SignificantFgf21 and Igf1 upregulation was observed, approximately 6- and 42-fold on average,

Figure 4. Characteristics of 3T3-Beige cells. (a) Comparison of UCP1 and mitochondria amounts in 3T3-L24 adipocytes and 3T3-Beige cells. Fluorescence staining of UCP1 and mitochondria with anti-UCP1 antibody and MoltVeiw Green. (b) Relative fluorescence values calculated from images of the experiments shown in Figure 4(a) using ImageJ software. (c) Amount of NADH in 3T3-L24 adipocytes and 3T3-Beige cells measured by using a Cell Counting Kit 8 (mean ± SD, n = 9). (d) Glucose uptake in 3T3-L24 adipocytes and 3T3-Beige cells measured using a Glucose Uptake Cell-Based Kit (mean ± SD, n = 7). (e) Ratio of batokine gene expression in 3T3-Beige cells to those in 3T3-L24 adipocytes (mean ± SD, n = 5 - 7).

whereas no increase in gene expression was observed inFst1 and Il-6, which are induced by adrenergic receptors. These results indicate that 3T3-Beige cells possess properties very similar to beige adipocytes, and that prolonged direct cold stimulation can induce a change from white adipocytes to beige adipose tissue-like cells.