All protocols in this study were approved by the Affiliated Wuxi No. 2 People’s Hospital of Nanjing Medical University. Male Sprague–Dawley (SD) rats (250–300 g) were bought from the Animal Center of Chinese Academy of Sciences (Shanghai, China). The rats were maintained on a standard diet and housed in temperature- and humidity-controlled animal quarters under a 12-h light/dark cycle.

The “two-hemorrhage” SAH model of rat was performed as previously described in our other paper (Zhao et al., 2016a). Rats were anesthetized with pentobarbital intraperitoneal injection (40 mg/kg), and the spontaneous respiration was maintained. A small suboccipital incision was made; the occipital bone and the atlanto-occipital membrane were exposed with the aid of a surgical microscope. A 27-gauge needle was used to puncture the atlanto-occipital membrane carefully into the cisterna magna. Fresh autologous nonheparinized blood (0.2 mL) withdrawn from the femoral artery of the same rat was injected into the cisterna magna within 2 min. The bone wax was used to seal the hole to prevent fistula. Then, the incision was immediately sutured after the blood injection. A 30° head-down prone position was maintained for 30 min to ensure adequate blood distribution around the basal intracranial arteries. After 48 h (day 2), the same procedure was repeated. After the operations, the animals were allowed to recover from the effects of anesthesia and returned to their cages. During the period of unconsciousness, the rats were given 2 mL of water every 3 h using a feeding tube.

We divided male SD rats randomly into three groups: Sham + vehicle group (Sham + V, n = 12), SAH + vehicle group (SAH + V, n = 12), and SAH + sulforaphane group (SAH + S, n = 12). Sulforaphane (SFN, Sigma-Aldrich, St. Louis, MO, United States) was dissolved in corn oil (Sigma-Aldrich, St. Louis, MO). A dose of 50 mg/kg in 250–280 μL corn oil was intraperitoneally (IP) injected every 24 h starting from 30 min after the first blood injection. All rats were then followed up with the daily injection of corn oil for 4 days (Zhao et al., 2016a). In the Sham + V group, the same procedures were applied without blood injection, but the same volume of corn oil was injected intraperitoneally. Three days after the second SAH induction (day 5), the animals were euthanized by cervical dislocation.

Two independent observers, who were unaware of the study’s objectives, recorded the appetite scores after the first SAH induction. The modified appetite-scoring table, as previously described (Zhou et al., 2007), was used to record the scores daily (Table 1).

The rats were euthanized on day 5 for tissue assays. A 3-cm segment of the mid-ileum was taken and flushed with ice-cold saline. One-half of the segment was stored in liquid nitrogen immediately for enzyme-linked immunosorbent assay (ELISA), RNA reverse transcriptase–polymerase chain reaction (RT-PCR), and Western blot. The other half was immersed in 10% buffered formalin for histopathological studies, including hematoxylin–eosin (HE) staining and ultrastructural observations. Another 3-cm segment of the mid-ileum was taken for the assessment of the intestinal wet/dry weight ratio.

Ileum tissues from each group were fixed in 4% formaldehyde and embedded in paraffin. Deparaffinized sections were sectioned in 5–6 slices with a thickness of approximately 5 μm. These slices underwent routine dewaxing and dehydration processes before HE staining (Abcam, United States) was performed according to established protocols.

The wet/dry weight ratio is a reliable index to assess tissue microvascular permeability, and it represents the percentage of water in tissue. The intestinal wet/dry weight ratio was evaluated as described in a previous study (Zhu et al., 2009). Excess fluid was blotted from specimens after the gut tissue samples were taken, and wet weights were measured. Then, the specimens were dried at 80°C for 72 h and re-weighed to measure the weight of the dry content. The gut wet/dry weight ratio was then calculated.

Inflammatory scores were determined using a scoring system, as follows: epithelium (E): 0, normal morphology; 1, loss of goblet cells; 2, extensive loss of goblet cells in large areas; 3, loss of crypts; and 4, significant loss of crypts in large areas. Infiltration (I): 0, no infiltrate present; 1, infiltration observed around the bases of crypts; 2, infiltration extending to the lamina muscularis mucosa layer; 3, extensive infiltration reaching the muscularis mucosa accompanied by abundant edema; and 4, infiltration into the submucosal layer. The histological score was defined as the sum of these two parameters (total score = E + I).

The frozen ileum tissues were homogenized using a glass homogenizer in 1 mL of the buffer. The buffer was composed of 1 mg/L of pepstatin A, 1 mmol/L of PMSF, 1 mg/L of leupeptin, and 1 mg/L of aprotinin in PBS solution (pH 7.2). Then, the tissue lysates were centrifuged at 12,000 g for 20 min at 4°C. The intestinal levels of inflammatory mediators were quantified using specific ELISA kits for rats, according to the manufacturers’ instructions (TNF-α kit Diaclone Research, France; IL-1β and IL-6 kits, Biosource Europe SA, Belgium). The cytokine levels in the ileum tissue were expressed as the content of cytokine per gram of protein.

The protein levels were quantified as previously described (Zhao et al., 2016a). In brief, protein lysates were prepared from the ileum tissues of rats and rat intestinal epithelial (IEC-6) cells using the precooled RIPA lysis buffer containing the protease inhibitor PMSF; then, they were separated using 12% SDS-PAGE and transferred to nitrocellulose membranes. The membrane was blocked with 5% skimmed milk for 2 h at room temperature and then incubated with primary antibodies at 4°C; the primary antibodies used were as follows: anti-Keap1 (1:500, Santa Cruz Biotechnology, CA), anti-Nrf-2 (1:500, Santa Cruz Biotechnology, CA), anti-HO-1 (1:1,000; Santa Cruz Biotechnology, CA), anti-NQO-1 (1:1,000; Santa Cruz Biotechnology, CA), anti-LC3 (1:2000; Abcam, United States), anti-P62 (1:5,000; Abcam, United States), anti-ATG4D (1:800; Proteintech Group, United States), anti-β-actin (1:1,000, Santa Cruz Biotechnology), and anti-GAPDH (1:1,000, Santa Cruz Biotechnology). Optical densities were obtained using Glyko BandScan software (Glyko, Novato, CA). All experiments were repeated at least thrice.

Total RNA was extracted using Trizol (Invitrogen, CA, United States), and cDNA was synthesized from 2 μg of total RNA using the BU-Script RT-Kit (Biouniquer, Jiangsu, China) and stored at −20°C. Reverse transcription was conducted using GoTaq Green Master Mix (Promega, WI, United States). The primers and PCR parameters are shown in Table 2. PCR products were detected through agarose gel electrophoresis. The intensity of the bands was analyzed using the ImageJ program. GAPDH was used as a housekeeping gene.

IEC-6 cells were fixed with 4% of paraformaldehyde and permeabilized with 0.5% Triton X-100 in ice-cold PBS. The pretreated cells were blocked with 1% bovine serum albumin (BSA) and incubated with primary anti-LC3B (1:200; Abcam, United States). The nuclei were stained with 4′, 6-diamidino-2-phenylindole dihydrochloride (DAPI). Images of the stained cells were acquired using a fluorescence microscope (BX41, Olympus, Japan).

IEC-6 cells were first fixed with fresh 4% paraformaldehyde at 4°C for 45 min and washed using PBS (pH 7.2); then, 0.4% Triton X-100 was used for percolation for 30 min at room temperature. Finally, Hoechst 33258 solution (10 mg/L) was added and placed at room temperature for 5 min in the dark. A fluorescence microscope was used to observe cell morphology. The normal nucleus showed diffused and uniform low-intensity fluorescence, and the nucleus of apoptotic cells showed granular fluorescence.

Statistical analysis of the data was performed using SPSS 12.0. All data were presented as mean ± SD. The Mann–Whitney t-test was used to measure the appetite score from day 0 to day 5. Other measurements were analyzed using one-way analysis of variance and Turkey’s post hoc test. We used blinding throughout the data collection and analysis process and corrected for multiple comparisons. The overall significance level was set at 5%.

The rats exhibited a marked loss of appetite after SAH treatment, with the appetite score in the SAH + S group showing significant improvement compared to that in the SAH + V group on days 2, 3, 4, and 5, but still lower than that in the Sham + V group (Figure 1A). The wet/dry weight ratio is a hall marker to determine tissue microvascular permeability. SAH significantly increased the intestinal wet/dry weight ratio at day 5 following SAH, which was markedly decreased by SFN (Figure 1B).

Then, special indices such as villous height, crypt depth, villous diameter, and villous surface area were used to evaluate intestinal mucosal damages. As shown in Table 3, compared with the Sham + V group, all these indices were significantly decreased in the SAH + V group. Notably, SFN remarkably increased these indices. Furthermore, ultrastructural observation was performed to assess the effect of SFN on ileum tissue injury in rats after SAH. The structure of microvilli was all arranged in the Sham + V group (Figures 2A,B). However, some ultrastructural alterations, such as ruptured, distorted, and sparse microvilli, were detected in the SAH + V group (Figure 2C). Moreover, the mitochondria reduced, and their cristae were disrupted (Figure 2D). Following SFN administration, the ultrastructural alterations were dramatically attenuated in the SAH + S group (Figures 2E,F).

HE staining was performed to detect histopathological changes in ileum tissues. Compared with the Sham + V group, the SAH + V group showed significant inflammatory cell infiltration, mucosal interstitial edema, and other pathological phenomena, but SFN treatment significantly improved this effect (Figure 3A), which was further confirmed by the inflammatory score (Figure 3B). In addition, the control group (Sham + V) exhibited relatively low levels of IL-1β, TNF-α, and IL-6 (3.582 ± 0.492, 2.947 ± 0.291, and 5.903 ± 0.487 ng/g protein, respectively) in the ileum tissues, and SAH robustly increased the levels of these cytokines (Figure 3C). Interestingly, SFN administration significantly decreased the upregulated inflammatory cytokines induced by SAH in ileum tissues of rats (Figure 3C).

Considering that SFN exhibits the protective effect on oxidative stress and inflammation via the activation of Nrf-2 (Wagner et al., 2010; Hao et al., 2015), here, we tested the Keap1/Nrf-2/ARE signaling pathway using RT-PCR and Western blot. Compared with the Sham + V group, the expressions of Keap1, Nrf-2, HO-1, and NQO-1 were all significantly upregulated in the SAH + V group, and SFN further enhanced the effect of SAH on the Keap1/Nrf-2/ARE signaling pathway, both in mRNA (Figure 4A) and protein levels (Figures 4B,C). Western blot analysis also showed that autophagy-related proteins (LC3-II, ATG4D, and P62) were significantly upregulated in the SAH + V group, and SFN further upregulated the enhanced autophagy in rats after SAH (Figures 5A–C).

In vitro experiments were conducted to evaluate the role of autophagy in rats after SAH. TNF-α was used in IEC-6 cells to induce cell damage. LC3 immunofluorescence staining showed that TNF-α promoted the formation of autophagosomes, and SFN further upregulated the enhanced autophagosomes in IEC-6 cells after TNF-α treatment (Figure 6A). 3-MA (an inhibitor of autophagy) greatly counteracted the increased autophagosomes induced by SFN (Figure 6A). Moreover, SFN significantly reversed TNF-α-induced apoptosis in IEC-6 cells, and 3-MA dramatically rescued the decreased cell apoptosis induced by SFN (Figure 6B). In addition, SFN further enhanced the upregulated protein levels of Keap1, Nrf-2, HO-1, and NQO-1 induced by SAH in IEC-6 cells, which was similar to the results in vivo, whereas 3-MA had no effect on the Keap1/Nrf-2/ARE signaling pathway (Figure 7).