Male Sprague-Dawley (SD) rats (Academia Sinica, Taipei, Taiwan) weighing 280-330 g were maintained at a density of two animals per cage for at least 5 days after they arrived at our laboratory. The rats were given ad libitum access to food and water while housed within a room with a 12 : 12 h light-dark cycle. All rats were treated as per the guidelines stipulated by the Experimental Animal Laboratory and approved by the Animal Care and Use Committee of Chang Gung University (IACUC Approval No.: CGU14-052).

The surgical procedure used in the lateral hemisection of the spinal cord in rats can be found in our previous studies (Lin et al., 2011, 2012). Briefly, isoflurane was delivered to the rats as general anesthesia, before the animals were fastened within a stereotaxic apparatus (David Kopf Instruments, Tujunga, CA). An adjustable wire knife was used to create a spinal cord lesion on the left side. Control animals were subjected to a sham operation (laminectomy without hemisection). Posterior decompression was achieved via laminectomy at the eleventh thoracic vertebrae using a fine diamond drill. The wire knife guide was positioned vertically adjacent to the lateral surface at the spinal cord’s lower thoracic level. This permitted analysis of the healthy cephalad and caudal segments of the spinal cord. The knife was turned medially and then extended 1.5 mm, at which point the guide was lifted by 4.0 mm to hemi-transect the spinal cord. Iridectomy scissors were then used to cut to ascertain the thoroughness of the hemisection. The wound was closed in layers.

The rats were divided into three groups: (1) Vehicle-treated sham-operated control (sham; laminectomy only) (n = 6); (2) Vehicle-treated SCI group (SCI) (n = 6); (3) SCI rats post-treated with CSC treatment (SCI + CSC) (n = 6).

The cryogen is 1, 1, 1, 2-tetrafluoroethane, an FDA-approved replacement for Freon, biodegradable, non-toxic, and incombustible. Following spinal cord hemisection, cryogen spurts were immediately sprayed onto the injured spinal cord through an electrically powered nozzle from a distance of 3 cm above the dural surface. Coolant was delivered continuously in spurts of 20 milliseconds at intervals of 5 spurts per minute for 10 minutes. CSC treatment lowered the surface temperature in the treatment region to -10 ${}^{\circ }$ C. The surface temperature returned to normal within roughly 500 milliseconds after CSC treatment (Anvari et al., 1996; Smithies et al., 1998). Note that spraying was precisely regulated using a programmable digital delay generator. A schematic diagram showing the SCI model and CSC treatment are presented in Fig. 1A.

Fig. 1.

CSC reduced injury-induced inflammation following spinal cord hemisection in rats. (A) Schematic illustration of the CSC technique applied to tissue adjacent to the spinal cord hemisection under the following conditions: (i) Sham controls; (ii) SCI rats undergoing SCI; (iii) SCI rats undergoing SCI followed by CSC (SCI + CSC); mRNA expression levels of (B) NF $\kappa$ B P105, (C) TNF- $\alpha$ , (D) iNOS, and (E) RANTES. Vertical bars indicate the mean ( $\pm$ SEM) mRNA expression (n = 3 in each group). *, P 0.05 and ***, P 0.001 indicate the level of statistical significance.

At 24 h post-treatment, all of the animals in all groups underwent deep anesthesia via isoflurane. The animals were then sacrificed, and the spinal cord tissue was removed to investigate mRNA expression (n = 3 in each group) and perform immunohistochemical analysis (n = 3 in each group).

Total RNA was prepared through the direct lysis of animal tissue using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). Then, reverse transcribing the mRNA into cDNA using oligo-dT, and SuperScript II reverse transcriptase (Invitrogen, Carlsbad, CA, USA). Quantitative RT-PCR was performed using the ABI StepOne sequence detector system (Applied Biosystems, Foster City, CA, USA) in accordance with the SYBR Green methodology, normalized to the expression of a housekeeping gene (GAPDH), as previously described (Lee et al., 2015). The primer sets and product size of each cDNA of interest were as follows: rat CISD2 (Gene ID: 295457, accession number: NM_001191608, 5’- CCCTTCCTTGGTGTACTTGCA -3’ and 5’- TTTCGTTCACCACTTTGGGATT -3’, 129 bp); NF $\kappa$ B P105 (Gene ID: 81736, accession number: NM_001276711, 5’- GGGCCTGCAAAGGTTATCGT -3’ and 5’- TGCCGTCCTCACAGTGCTT -3’, 128 bp); TNF- $\alpha$ (Gene ID: 24835, accession number: NM_012675, 5’- GACGTGGAACTGGCAGAAGAG -3’ and 5’- AGGAATGAGAAGAGGCTGAGACA -3’, 147 bp); iNOS (Gene ID: 24599, accession number: NM_012611, 5’- GTCTTTGACGCTCGGAACTGT -3’ and 5’- CGACCTGATGTTGCCATTGT -3’, 101 bp); RANTES (Regulated on Activation, Normal T Cell Expressed and Secreted) (Gene ID: 81780, accession number: NM_031116, 5’- AACCCCCTACTCCCACTCATTC -3’ and 5’- GGCCATAGGAGAGGACACAGTT -3’, 106 bp); and GAPDH (Gene ID: 24383, accession number: NM_017008, 5’- CGTGTTCCTACCCCCAATGT -3’ and 5’- CCTGCTTCACCACCTTCTTGAT -3’, 103 bp).

On the second day after SCI, the animals underwent deep anesthesia via isoflurane before being sacrificed. They were then thoroughly perfused with phosphate-buffered saline (PBS), followed by cold 4% paraformaldehyde in 0.15 M sodium phosphate buffer. The spinal cords were removed immediately, fixed at 4 ${}^{\circ }$ C for 8 h, and cryoprotected in 15% and 30% sucrose for 3 days. Tissue from the spinal cords was frozen (-80 ${}^{\circ }$ C) before analysis. Sections (5 $\mu$ m in thickness) were cut from the center of the spinal cord hemisection on a frozen stage using a sliding microtome. The sections were then washed in PBS and incubated in 3% normal goat serum with 0.3% Triton X-100 in PBS for 1 h. The free-floating sections were incubated at 4 ${}^{\circ }$ C with the following antibodies: Anti-GFAP (Catalog: 16825-1-AP, Proteintech, Chicago, USA); Anti-caspase-3 (Catalog: 19677-1-AP, Proteintech, Chicago, IL, USA); Anti-NeuN (clone A60, Catalog: MAB377, Merk, Taipei City, Taiwan); Anti-CISD2 (Catalog: 13318-1-AP, Proteintech, Chicago, IL, USA). Immunoreactivity was visualized using the Vectastain Elite ABC Peroxidase method (Vector Laboratories, Burlingame, CA, USA) with diaminobenzidine (DAB) as the chromagen. The morphological and quantitative analysis focused on injured grey and white matter adjacent to the spinal cord hemisection (epicenter), which is indicated as the region of interest (ROI) within the dashed box in Fig. 3A. Cell counting was performed at a magnification of 400 $\times$ on every sixth section stained using the antibodies mentioned above. Only cells presenting visible indications of staining were counted. All data are presented as mean $\pm$ SEM of three consecutive measurements.

Independent two-sample t-tests were used to compare data between the experimental groups. One-way analysis of variance (ANOVA) was used to determine whether there were significant differences between the means of three or more independent (unrelated) groups. In cases where the differences were apparent, multiple comparisons were performed using the Newman-Keuls method. Data are presented as mean $\pm$ SEM. All statistical assessments were two-sided and evaluated at a 0.05 level of significance. Statistical analysis was performed using GraphPad Prism software version 5.0 (GraphPad Software, Inc; La Jolla, CA).

In the current spinal cord hemisection model, rats were subjected to CSC treatment involving short spurts (20 milliseconds) at intervals of 5 spurts per min for 10 min, or untreated with CSC (Fig. 1A). qRT-PCR was used to examine the mRNA expression of proinflammatory mediators at the site of SCI. Compared to the sham control group, animals in the SCI only group presented higher expression levels of NF $\kappa$ B (P105), TNF- $\alpha$ , iNOS, and RANTES (*P 0.05, Fig. 1B; *P 0.05, Fig. 1C; ***P 0.001, Fig. 1D; and *P 0.05, Fig. 1E, respectively). Compared to the SCI only group, animals in the SCI + CSC group presented significantly lower mRNA expression levels of NF $\kappa$ B (P105), TNF- $\alpha$ , iNOS, and RANTES (*P 0.05, Fig. 1B; *P 0.05, Fig. 1C; *P 0.05, Fig. 1D; and *P 0.05, Fig. 1E, respectively). Overall, these findings indicate that CSC treatment had an anti-inflammatory effect on rats previously subjected to SCI.

The qRT-PCR analysis revealed that CISD2 expression levels were significantly lower in the SCI only group than in the sham control group (P 0.05, Fig. 2). CISD2 expression levels were considerably higher in the SCI only group than in the SCI + CSC group (P 0.05). In a previous paper, we reported on the anti-inflammatory effects of CISD2 (Lin et al., 2017). In the present work, CSC treatment was shown to attenuate the injury-induced pro-inflammatory response (i.e., SCI-induced reduction in CISD2 expression).

Fig. 2.

Spinal cord hemisection produced a corresponding decrease in CISD2 expression. CSC mediated the injury-induced decrease in CISD2 expression. Vertical bars indicate the mean ( $\pm$ SEM) mRNA expression (n = 3 in each group). *, P 0.05 and ***, P 0.001 indicate the level of statistical significance.

Astrocytes can be stimulated to release various pro-inflammatory mediators in response to SCI, as indicated by GFAP over-expression. Immunostaining for GFAP (arrow in Fig. 3C) was more robust in the SCI only group than in the sham control group (Fig. 3B) (***P 0.001, Fig. 3E). Immunostaining for GFAP was lower in the SCI + CSC treatment (indicated by an arrow in Fig. 3D) than in the SCI only group (**P 0.01, Fig. 3E). These results indicate that CSC therapy mediated the activation of astrocytes and astrocyte-mediated proinflammatory responses following SCI.

Fig. 3.

Representative photographs of GFAP-stained sections of tissue samples were obtained from the spinal cords of rats in the sham control group (B), SCI (C), and SCI + CSC (D) after SCI. (A) Schematic depiction of ROI (labeled as the dashed box) in injured spinal cords for all groups. Staining of GFAP-positive cells was weaker in the sham group (B) than in the SCI group (C). CSC treatment decreased GFAP immunofluorescence. (A-C: magnification 400 $\times$ ). (E) Vertical bars indicate the mean ( $\pm$ SEM) number of GFAP-stained cells per tissue section in each group (n = 3). *, P 0.05 and ***, P 0.001 indicate the level of statistical significance.

The blue arrows in Fig. 4A-C indicates the immunoreactivity of caspase-3 as a proxy of cellular apoptosis. Following spinal cord hemisection, immunostaining for caspase-3 was significantly higher in the SCI only group than in the sham control group (***P 0.001, Fig. 4D). Immunostaining for caspase-3 was significantly lower in the SCI + CSC group than in the SCI only group (**P 0.01, Fig. 4D), indicating that CSC attenuated cellular apoptosis.

Fig. 4.

Representative photographs of caspase-3-stained sections of tissue samples obtained from the spinal cords of rats in the (A) sham control group, (B) SCI only group, and (C) SCI + CSC group (A-C: 400 $\mathrm{\times }$ ). The blue arrows in A-C indicate caspase-3-positive cells. (D) Vertical bars indicate the mean ( $\pm$ SEM) number of caspase-3-stained cells per tissue section in each group (n = 3). **, P 0.01 and ***, P 0.001 indicate the level of statistical significance.

The blue arrows in Fig. 5A present the immunohistochemistry indicative of the neuron-specific marker NeuN in the sham group. As indicated by the blue arrows in Fig. 5B, the morphological integrity was far worse in the SCI only group than in the sham control group, and neuronal loss was also evident (***P 0.001, Fig. 5D). The blue arrows in Fig. 5C indicate that NeuN staining was more preserved in the SCI + CSC group than in the SCI only group, and neuronal loss was less evident (*P 0.05, Fig. 5D). Overall, the anti-inflammatory and anti-apoptotic effects of CSC could potentially help to prevent neuronal deficits.

Fig. 5.

Representative photographs of NeuN-stained sections of tissue samples obtained from the spinal cords of rats in the (A) sham control group, (B) SCI only group, and (C) SCI + CSC group (A-C: 400 $\mathrm{\times }$ ). Compared to sham controls, the arrows in (A-C) indicate NeuN immunoreactivity (exhibited neuronal loss in SCI group). CSC treatment preserved neurons after SCI. (D) Vertical bars indicate the mean ( $\pm$ SEM) number of NeuN-stained cells per tissue section in each group (n = 3). *, P 0.05 and ***, P 0.001 indicate the level of statistical significance.

As in our previous research, we determined in the current study that spinal injury led to a decline in CISD2 expression, as evidenced by a reduction in CISD2 immunoreactivity in the SCI only group (***P 0.001, Fig. 6D). Far more CISD2-positive cells were observed in the sham group (blue arrows in Fig. 6A) than in injured tissue (showing signs of vacuolation) at the site of the injury in the SCI only group (dashed circle in Fig. 6B). The number of CISD2-positive cells in the same injured tissue in the SCI + CSC group (dashed circles in Fig. 6B and 6C) was far higher than in the SCI only group (*P 0.05, Fig. 6D). Our results demonstrate the CISD2-preserving beneficial effects of CSC. The eventual anti-inflammatory property could contribute to the prevention of astrocytic activation, astrocyte-mediated neuroinflammation, apoptosis, and neuron loss.

Fig. 6.

SCI-induced a decrease in CISD2 expression in the injured spinal cord, the effect of which was attenuated by CSC. Representative photographs of CISD2-stained sections of tissue samples obtained from the spinal cords of rats in the (A) sham control group, (B) SCI only group, and (C) SCI + CSC group (A-C: 400 $\times$ ). (A) The sham group presented a high degree of CISD2 immunoreactivity. (B) CISD2 expression was lower in tissue samples with follicles from the hemisected spinal cord (indicated by a dashed circle). At the injury site of a similar situation in (C) (dashed circle), a densely packed pattern of CISD2 immunoreactivities proved the CISD2-elevating effect of CSC treatment in SCI rats, compared to (B). (D) Vertical bars indicate the mean ( $\pm$ SEM) number of CISD2-stained cells per tissue section in each group (n = 3). *, P 0.05 and ***, P 0.001 indicate the level of statistical significance.