Nonsteroidal anti-inflammatory drug-activated gene-1 (NAG-1) is identified as a member of the transforming growth factor-β (TGF-β) superfamily that is involved in regulating cellular functions, such as proliferation and differentiation [1, 2]. It is also known as growth differentiation factor-15 (GDF-15), which can be increased by cellular stress and disease [3]. Baek et al. [2] reported that NAG-1/GDF-15 is overexpressed in human colorectal cells by the treatment with nonsteroidal anti-inflammatory drugs (NSAIDs), resulting in the induction of apoptosis and anti-tumorigenesis. NAG-1 has also been investigated in diverse diseases including cancer, obesity, diabetes, and cardiovascular disease [3-6]. Transgenic (TG) mice expressing human NAG-1 showed the suppression of colorectal and lung carcinogenesis as well as the alleviation of obesity and diabetes [4-8]. Although it has been demonstrated that NAG-1/GDF-15 is expressed in the central nervous system and highly upregulated in regions adjacent to the injury site [9, 10], it is unclear whether NAG-1 plays a role in the nervous system on nociceptive signaling under the physiological and pathological conditions.

Inflammation or tissue injury induces pain that is related to not only the dysfunction of neurons but also the pathological activation of glial cells including astrocytes and microglial cells [11]. It has been suggested that TGF-β1 is a relevant mediator of nociception and has protective effects against the development of chronic pain via inhibition of the neuroimmune responses in neurons and glial cells [12]. Emerging evidence shows that NAG-1/GDF-15 is associated with low back pain-associated disability [13], suggesting that the pathological changes in the nervous system could be related to the function of NAG-1. Furthermore, the function of NAG-1 is involved in the reduction of inflammatory responses [14, 15], while there is limited understanding of the role of NAG-1 in inflammation-induced spinal pain processing.

In this regard, the purpose of this study was to determine the role of NAG-1 on nociceptive behaviors following mechanical, thermal, and chemical stimulation. To do this, we examined whether: (1) female and male NAG-1 TG mice have similar characteristics for basal nociceptive responses when compared with those of wild-type mice; (2) NAG-1 TG mice show nociceptive behaviors after formalin-induced chemical stimulation; and (3) the activation of spinal neurons and glial cells is differently regulated in NAG-1 TG and WT mice following intraplantar formalin administration.

Animals

Ten-week-old male and female NAG-1 transgenic (TG) mice and wild-type (WT) littermates were used in the present study. C57BL/6 mice overexpressing human NAG-1 were previously reported [4]. Animals were housed in standard laboratory animal cages with wooden chip flakes bedding. They were maintained under standard laboratory conditions (22±2°C, 12/12 h light/dark cycle) with free access to food and water. All mice were allowed an acclimatization period of at least 3 days before being used in experiments. Animal breeding and the experimental protocols for animal usage were reviewed and approved by the Institutional Animal Care and Use Committee at Seoul National University (SNU-170417-23-2 and SNU-200408-4). All the necessary efforts were undertaken to minimize the number of mice used in the experiments as well as to avoid any discomfort to them.

Genotyping

For NAG-1 genotyping, genomic DNA was taken from tail tips with DNeasy Blood & Tissue Kit (Qiagen, Hilden, Germany) in accordance with the provided protocol, followed by polymerase chain reaction (PCR). PCR products were amplified with GoTaq^® Green Master Mix (Promega, Madison, WI, USA) using the following primer pair: hNAG-1 (F: 5’-GTGCTGGTTATTGTGCTGTCTC-3’, R: 5’-AGTCTTCGGAGTGCAACTCTGAGG-3’). The thermal cycling condition was as follows: initial denaturation at 94°C for 2 min, followed by 40 cycles of 94°C for 30 s, 52°C for 30 s, and 72°C for 1 min, and final extension at 72°C for 5 min. The product was electrophoresed on a 1.4% agarose gel and photographed under UV light.

Immunoblotting

Cortex, brainstem, and spinal cord from NAG-1 transgenic and wild-type mice were sampled and lysed by RIPA buffer supplemented with protease and phosphatase inhibitors as previously described [16]. After determining concentration with Pierce BCA Protein Assay Kit (Thermo Fisher Scientific, Waltham, MA, USA), tissue lysates were loaded on a 12% sodium dodecyl sulfate-polyacrylamide gel, followed by transfer to a nitrocellulose membrane (GVS North America, Sanford, ME, USA). After blocking with 5% non-fat milk in Tris-buffered Saline with 0.05% Tween-20 (TBST buffer) for 1 h at room temperature, the membrane was incubated for overnight at 4°C with a primary antibody specific for GDF15 (G-5, cat# sc-377195, Santa Cruz Biotechnology) or β-actin (C4, cat# sc-47778, Santa Cruz Biotechnology) followed by incubation with HRP conjugated secondary antibody for 1 h at room temperature. The membrane was washed three times with TBST buffer for 10 min each between each step. The membrane was incubated with Pierce™ ECL Western Blotting Substrate (Thermo Fisher Scientific), and the blot was visualized with Alliance Q9 mini (UVITEC, Cambridge, UK).

Behavioral assessments

Mechanical sensitivity was assessed using von Frey filaments (North Coast Medical, Morgan Hill, CA) and the up-and-down method following the procedure as previously described [17, 18]. The results were expressed as the 50% withdrawal threshold (g), which represented force of the von Frey hair to which an animal reacts in 50%.

Thermal sensitivity was examined using a plantar analgesia meter (Model 390, IITC Life Science Inc., Woodland Hills, CA) as previously described by Hargreaves et al. with minor modification [18, 19]. Mice were placed into a plastic chamber on the glass floor and a radiant heat source was positioned under the floor beneath the hind paw. The test was duplicated in the hind paw of each mouse, and the mean paw withdrawal latency (sec) was calculated. Cutoff time in the absence of a response was set at 20 s to prevent tissue damage.

Hot plate test was performed using a hot-plate apparatus (Model-35100, Ugo Basile, Comerio, Italy) as described in a previous study [20]. The temperature of the plate was maintained at 50, 52 or 55°C. Mice were placed into an acrylic cylinder (20 cm in diameter, 25 cm height) on the heated surface, and the time (in seconds) between placement and shaking, licking, or lifting of their hind paws or jumping (whichever occurred first), was recorded as the response latency (sec). The test was duplicated in each mouse, and the mean response latency was calculated.

Acetone test was performed by applying acetone (30 μl) to the plantar surface of the hind paw. The score of responses to acetone-induced evaporative cooling was represented to 0, no response; 1, brief withdrawal of the paw; 2, prolonged and repetitive withdrawal along with licking and/or shaking. The test was repeated 5 times to the hind paw of each mouse with a 5-min interval between each application, and the mean score was calculated. All behavioral analyses were performed blindly.

Formalin test

Mice were acclimated to an acrylic observation chamber for at least 30 min before formalin administration. Formalin (2.5% formalin solution in 20 μl physiological saline) was injected subcutaneously into the plantar surface of the left hind paw using a 100 μl Hamilton syringe connected to a 30-gauge needle. Following injection, animals were immediately placed in an observation chamber and nociceptive behaviors including licking or biting the injected area were recorded for a 60-min period. The cumulative response time (s) of these behaviors was measured and analyzed during the first phase of pain (phase I; 0~10 min post-administration) and the second phase of pain (phase II; 10~60 min post-administration) [21].

Immunohistochemistry & Image analysis

The immunohistochemistry was performed as described in previous studies [22, 23]. Under general anesthesia with isoflurane, different groups of mice were euthanized 2 h after the formalin injection. Mice were transcardially perfused with 4% paraformaldehyde solution. The spinal cords were post-fixed in the identical fixative for 2 h at room temperature and then placed in 30% sucrose in PBS (pH 7.4) at 4°C. L4-5 spinal cords were sectioned (40 μm) using a cryostat (Leica CM1520, Leica Biosystems, Germany). Transverse spinal cord sections were incubated in blocking solution for 1 h at room temperature and then incubated for 2 days at 4°C with a primary antibody specific for c-Fos (cat# ab190289, Abcam plc.), glial fibrillary acidic protein (GFAP; cat# MAB360, Millipore Co.) or ionized calcium binding adaptor molecule-1 (Iba-1; cat# 019-19741, Wako Pure Chemical Industries, Ltd.). Tissue sections were incubated with the secondary antibody, Alexa Fluor^® 488-conjugated anti-rabbit or anti-mouse antibody (1:400, Life Technologies) for 90 min at room temperature. Fluorescent images were acquired by a Nikon Eclipse TE2000-E confocal laser scanning microscope with the aid of the EZ-C1 Gold version 3.80 software (Nikon Instech Co., Ltd.).

To analyze c-Fos-, GFAP- and Iba-1-immunofluorescence, tissue sections were randomly selected and analyzed using a computer-assisted image analysis system (Metamorph version 7.7.2.0; Molecular Devices Corporation, PA, USA). The following three dorsal horn regions were analyzed: 1) the superficial dorsal horn (SDH, laminae I and II); 2) the nucleus proprius (NP, laminae III and IV); and 3) the neck region (NECK, laminae V and VI). The number of c-Fos-immunoreactive cells was counted in both the ipsilateral and contralateral side of the spinal cord. The positive pixel area of GFAP- or Iba-1-immunofluorescence was counted on % threshold area [(positive pixel area/pixel area in each region)×100]. All analytical procedures were performed blindly without knowledge of the experimental conditions.

Statistical analysis

Statistical analyses were performed using Prism 5.0 (Graph Pad Software, San Diego, USA). One-way ANOVA was used to determine differences in the data obtained from time course changes in formalin-induced nociceptive behavioral experiments and this was followed by a Bonferroni’s multiple comparison test for post-hoc analysis. Comparisons between 2 groups were analyzed by two-tailed Student’s t-test. All data are expressed as the mean±SEM and p values less than 0.05 were considered statistically significant.

Genetic overexpression of NAG-1 in the central nervous system

NAG-1 transgenic mice used in the present study overexpress the human NAG-1 DNA throughout the body (Fig. 1A). The expression of NAG-1 in the central nervous system was detected in cortex, brainstem, and spinal cord using immunoblotting. The results show that NAG-1 protein is expressed in the cortex, brainstem, and spinal cord in NAG-1 TG mice (Fig. 1B). Female NAG-1 TG mice display significantly lower body weights relative to female WT mice (Fig. 1C; *p<0.05 vs. WT; t (10)=2.566, p=0.0281). Male NAG-1 TG mice showed a tendency to reduce body weights as compared with male WT mice, but this reduction did not reach statistical significant (Fig. 1C; t (10)=1.533, p=0.1563).

Nociceptive response to mechanical, thermal, or cold stimulation in NAG-1 TG and WT mice

To measure basal sensitivity to mechanical stimuli, von Frey filament test was performed in the hind paw of NAG-1 TG and WT mice (Fig. 2A). The 50% withdrawal thresholds to mechanical stimuli were measured to 0.4 g in both female and male mice. Thermal sensitivity was measured in the hind paw of NAG-1 TG and WT mice using Hargreaves test and hot plate test. There was no significant difference between NAG-1 TG and WT mice in paw withdrawal latency to thermal stimuli (Fig. 2B; Female: t (10)=1.653, p=0.1294; Male: t (10)=0.1085, p=0.9157). Latency to respond to the hot plate maintained at 50, 52, or 55°C was examined in NAG-1 TG and WT mice (Fig. 2C), and there was no difference between NAG-1 TG and WT female (50°C: t (10)=1.123, p=0.2877; 52°C: t (10)=1.351, p=0.2065; 55°C: t (10)=0.7508, p=0.4701) and male (50°C: t (10)=0.1885, p=0.8543; 52°C: t (10)=2.150, p=0.0571; 55°C: t (10)=0.3502, p=0.7335) mice. In addition, acetone drop test was performed to measure cold sensitivity in NAG-1 TG and WT mice. There was no significant difference in acetone test score between NAG-1 TG and WT mice (Fig. 2D; Female: t (10)=0.9583, p=0.3605; Male: t (10)=0.8984, p=0.3901).

Nociceptive response to chemical stimulation induced by intraplantar formalin administration in NAG-1 TG and WT mice

In order to examine the effect of NAG-1 on inflammatory pain, we compared the nociceptive responses evoked by a formalin injection in the hind paw of NAG-1 TG mice and WT littermates. Intraplantar administration of formalin (2.5%) elicited a biphasic period of nociceptive behaviors including flinching, licking, and biting of the hind paw [21]. No significant differences were observed between the nociceptive behaviors of the NAG-1 TG and WT mice during the first phase of pain (0~10 min). During the second phase of pain (10~60 min), nociceptive behaviors were significantly increased in WT female mice at 20 min after formalin injection compared with those of NAG-1 TG female mice (Fig. 3A; *p<0.05 vs. NAG-1 TG mice; Group: F (1,130)=1.150, p=0.2854; Time: F (12,130)=11.49, p<0.0001; Interaction: F (12,130)=2.873, p=0.0014). By contrast, NAG-1 TG female mice showed an increase in formalin-induced nociceptive behaviors at 40 min after formalin injection compared with those of WT female mice (Fig. 3A; **p<0.01 vs. WT mice). There was no significant difference in the cumulative response time of nociceptive behaviors during the first or second phase (phase I or II, respectively) of the formalin test between NAG-1 TG and WT female mice (Fig. 3B; Phase I: t (10)=0.2266, p=0.8253; Phase II: t (10)=0.5558, p=0.5905). In addition, there was no difference in formalin-induced nociceptive behaviors between NAG-1 TG and WT male mice (Fig. 3C; Group: F (1,130)=0.09721, p=0.7557; Time: F (12,130)=10.77, p<0.0001; Interaction: F (12,130)=0.4094, p=0.9578). The cumulative response time spent licking or biting injected paw was similar during phase I or II between these two groups (Fig. 3D; Phase I: t (10)=1.179, p=0.2658; Phase II: t (10)=0.4469, p=0.6645).

c-Fos immunoreactivity in the spinal cord dorsal horn of NAG-1 TG and WT mice

Intraplantar administration of formalin (2.5%) increased c-Fos immunoreactivity in the superficial dorsal horn (SDH, laminae I~II), nucleus proprius (NP, laminae III~IV), and neck region (NECK; laminae V~VI) of the ipsilateral lumbar spinal cord in NAG-1 TG and WT female mice compared with that in the contralateral side of the spinal cord (Fig. 4A; ***p<0.001 vs. contralateral side; SDH: F (3,20)=23.50, p<0.0001; NP: F (3,20)=42.86, p<0.0001; NECK: F (3,20)=23.87, p<0.0001). No significant difference was observed in c-Fos immunoreactivity between NAG-1 TG and WT female mice (Fig. 4). Representative images of c-Fos immunostaining in the ipsilateral and contralateral lumbar spinal cord dorsal horns are shown in Fig. 4B.

GFAP and Iba-1 immunoreactivity in the spinal cord dorsal horn of NAG-1 TG and WT mice

There was a tendency of increase in GFAP immunoreactivity in the ipsilateral side of the spinal cord of NAG-1 TG and WT female mice compared with that in the contralateral side of the spinal cord, but not significant (Fig. 5A; SDH: F (3,20)=2.107, p=0.1313; NP: F (3,20)=1.440, p=0.2610; NECK: F (3,20)=0.2696, p=0.8465). Iba-1 immunoreactivity did not change in the ipsilateral spinal cord dorsal horn after the intraplantar injection of formalin (Fig. 5B; SDH: F (3,20)=0.06614, p=0.9772; NP: F (3,20)=0.3488, p=0.7904; NECK: F (3,20)=0.4669, p=0.7086). Representative images of GFAP and Iba-1 immunostaining in the ipsilateral and contralateral lumbar spinal cord dorsal horns are shown in Fig. 5C and D, respectively.

The NAG-1 TG mice used in the present study were initially generated and characterized in 2006 by Baek et al. [4]. These mice exhibit high expression of human NAG-1 in various tissues, including the skin, kidney, colon, and brain. In the present study, we showed that NAG-1 TG mice expressed not only the genomic DNA but also the protein of NAG-1 in the cortex, brainstem, and spinal cord, which was not observed in WT littermates. This result suggests the potential involvement of NAG-1 expressed in the central nervous system in the functioning of the nervous system. We next examined the role of NAG-1 on basal nociceptive responses to mechanical and thermal stimulation in both female and male mice. Both female and male NAG-1 TG mice showed no difference from WT mice in their sensitivity to mechanical, noxious heat, and cold stimulation. These results indicate that NAG-1 overexpression in the nervous system does not alter the basal nociceptive sensitivity to mechanical and thermal stimulation in transgenic mice. Regarding the levels of NAG-1/GDF-15 increased in the central nervous system of rodents by cold-induced lesion of the cerebral cortex or kainate-induced excitotoxicity [9, 10], we need to investigate the major function of NAG-1 under the pathological conditions.

The formalin test is considered as a reliable way of generating continuous pain by injured tissue and quantifying analgesia [24, 25]. The subcutaneous injection of diluted formalin into the hind paw causes tissue injuries and generates biphasic behavioral responses, with an early and short-lasting first phase followed by a prolonged second phase [25, 26]. The early response may be due to the direct effects on peripheral sensory receptors, while the late response is related to inflammation and changes in the central nervous system function induced by neural activity generated during the early phase [24, 27]. In the present study, the nociceptive behaviors evoked by a formalin injection in the hind paw were tested in female and male NAG-1 TG and WT mice. During the first or second phase of pain the cumulative time spent licking or biting of the injected hind paw was similar in the four groups, but the peak time in the second phase was delayed in female NAG-1 TG mice. These results suggest the possibility that NAG-1 in female mice may have an inhibitory effect on the second phase of inflammatory pain in part, while it may be not enough to block nociceptive responses to subcutaneous formalin. It has been demonstrated that NSAIDs and steroids produce analgesia in the second phase of pain, but have little or no effect on the first phase of pain produced by formalin injection [27]. Since NSAIDs are able to induce the expression of NAG-1, which contributes to mediating cyclooxygenase-independent effects [28], NAG-1 may at least partially be involved in the analgesic effect of NSAIDs on formalin-induced pain. We plan to investigate the possible mechanisms underlying the relationship between NAG-1 and NSAID-induced analgesic effects in future studies.

Several studies have suggested the possible molecular mechanisms underlying the protective role of NAG-1 against cancer, obesity, hyperglycemia, and diabetes [6, 7, 29]. Yang et al. [29] showed that NAG-1 interacts with glial cell-derived neurotrophic factor family receptor α-like (GFRAL) leading to the anorexigenic effect in mice with obesity. Since GFRAL expression was found in hindbrain neurons, but not in peripheral tissues, NAG-1/GFRAL pathway acts as a central mechanism of the regulation of food intake [29]. However, there is another research showing that anti-obesity effect of NAG-1 is mediated by GFRAL-independent pathway, which is related to the increased expression of thermogenic and lipolytic genes in adipose tissues [6]. Chen et al. [7] showed that NAG-1 has an anti-diabetic effect both in mouse and cellular model of diabetic nephropathy, while it is not clear which receptors are responsible for this action. In addition, NAG-1 downregulated the genes involved in the AGE/RAGE signaling pathway that provokes oxidative stress and chronic inflammation, and reduced the expression of TLR4, MyD88, and NF-κB p65 phosphorylation [7]. Regarding that activation of AGE/RAGE pathway can contribute to the spinal central sensitization in persistent pain and diabetic peripheral neuropathy [30, 31], NAG-1 may play a protective role in chronic disease states by decreasing the inflammatory responses. Although it is still largely unclear how NAG-1 plays a role in pain perception and which molecular mechanisms are related to the actions of NAG-1, we plan to further investigate not only the role of NAG-1/GDF-15 on peripheral and central sensitization in chronic neuropathic pain but also the detailed mechanisms underlying the protective effect of NAG-1 against pathological changes in the nervous system.

In contrast to the results obtained from female mice, there was no difference in the peak time in the second phase of pain between male NAG-1 TG and WT mice, suggesting the possibility of sex differences between female and male NAG-1 transgenic mice. There are two possible mechanisms by which NAG-1 overexpression may lead to sex-differences in formalin-induced pain behaviors. First, several studies support the concept that NAG-1/GDF-15 expression may be modulated by sex hormones [32, 33]. Liu et al. [34] showed that NAG-1 secretion in prostate epithelial cell lines was downregulated in response to testosterone. Furthermore, the balance of sex hormone levels decreased NAG-1 secretion through androgen and estrogen receptors-mediated pathways resulting in inhibition of the effects of NAG-1 [33]. In the present study, the weight lowering effect of NAG-1 was observed in both female and male transgenic mice, while statistical significance was found only in female mice. Although we did not examine the levels of NAG-1 that is functionally activated, it could be possible that the effect of NAG-1 overexpression may be greater in female mice than in male mice, at least in our experimental conditions. Secondly, emerging evidence suggests that gonadal hormones modulate neuroinflammation, which may cause sex differences in pain [35]. As reviewed by Rosen et al. [35], testosterone can activate astrocytes and microglial cells, which then may produce tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), and brain-derived neurotrophic factor (BDNF). However, estrogen can inhibit these glial cells, leading to the alleviation of pain, hyperalgesia, and allodynia. In addition, sex-specific regulation of neurosteroids causes sex differences in their levels under physiological and pathological conditions such as traumatic brain injury, spinal cord injury, and peripheral neuropathy [36]. Regarding that glial cell-mediated neuroinflammation and neurosteroidogenesis play an important role in spinal pain processing during the induction phase of neuropathic pain [37, 38], it could be possible that sex-specific pathophysiological changes occur in the nervous system of NAG-1 TG female and male mice, which may bring the difference in nociceptive responses induced by inflammation.

Tissue damage or noxious stimuli activate nociceptors in the periphery leading to the activation of neurons and glial cells in the spinal cord, which mediates this process of nociception [39]. It has been demonstrated that c-Fos is an immediate early gene, which is induced in the nuclei of neurons in response to tissue injury or noxious stimulation [40, 41]. c-Fos, the protein product of c-Fos, can be used as a marker for neuronal activation in the central nervous system [41]. In the present study, c-Fos immunoreactivity was increased in the spinal cord dorsal horns of TG and WT female mice at 2 h after formalin administration, while there was no difference between TG and WT groups. These results indicate that neuronal activation induced by formalin injection was occurred in the spinal cord dorsal horn of NAG-1 TG female mice as well as WT female mice. In addition, the activation of glial cells in the spinal cord plays an important role in the development and maintenance of inflammatory pain [42-44]. In the present study, GFAP and Iba-1 were used as the molecular markers of astrocytic and microglial activation, respectively, but neither GFAP nor Iba-1 was increased in the spinal cord of TG and WT female mice. While the morphological changes of glial cells were not detected in the present study, neuro-glial interaction might be occurred in the spinal nociceptive processing induced by subcutaneous formalin in both NAG-1 TG and WT female mice.

In conclusion, the present study demonstrates that NAG-1 TG mice show similar characteristics in mechanical, heat, and cold nociception compared with those of WT mice. These basal nociceptive responses were not different in both female and male mice. Female NAG-1 TG mice showed delayed peak time in the second phase of formalin-induced inflammatory nociceptive behaviors compared with that of female WT mice, while formalin-induced neuronal activation in the spinal cord was not different in female NAG-1 TG and WT mice. These results suggest that NAG-1 delays the peak time of formalin-induced nociceptive behaviors in female mice, while it is not enough to block spinal neuronal activation and nociceptive behaviors induced by peripheral formalin injection.

This research was supported by the National Research Foundation of Korea (NRF) grants funded by the Korean Government (No. 2020R1A2C1102540) and BK21 Four Future Veterinary Medicine Leading Education & Research Center at Seoul National University, Republic of Korea.