All animal procedures were approved by the Institutional Animal Care and Use Committee of Nanjing Medical University (IACUC: 2103051) under animal use permit number SYXK(Su)2020-0022. A total of 60 adult male Sprague-Dawley rats (Jiangsu Laboratory Animal Center, Nanjing, Jiangsu, China), 8 weeks old and weighing 280–320 g, were housed under a 12/12 h light/dark cycle at 22–24 °C with 50%–60% humidity. They had unlimited access to standard rodent chow and water.

The comprehensive experimental design is outlined in Fig. 1. A total of 60 Sprague-Dawley (SD) rats (Jiangsu Laboratory Animal Center, Nanjing, Jiangsu, China; 8 weeks, 280–320 g) were divided into four groups: Sham operation group (n = 15), CLP group (n = 15), CLP+MFGE8 group (n = 15), and CLP+MFGE8+CGT group (n = 15). Sepsis rat models were induced by CLP surgery [18]. Extrinsic MFGE8 (LMAI Bio, Shanghai, China; ICA286Hu01, 3.3 µg) was injected intracerebroventricularly (ICV) 24 h post CLP surgery [19], while 10 mg/kg CGT (MedChemExpress, Monmouth Junction, NJ, USA; HY-16141) was injected intraperitoneally (IP) 1 h post-CLP surgery to the CLP+MFGE8+CGT group [20]. Behavioral experiments using MWM and OFT tests were performed 3 days after CLP. The MWM test involved four consecutive days of training sessions, followed by a probe test session on the seventh day (n = 8 per group). After behavioral experiments, the animals underwent fMRI scans to detect brain changes in structure and function on day 7 post-CLP surgery (n = 8 per group). The rats were then euthanatized by deep isoflurane anesthesia. The hippocampus was removed to determine the level of the inflammatory cytokines interleukin (IL)-1$\beta$, IL-6 and C-reactive protein (CRP) by enzyme-linked immunosorbent assay (ELISA). Efferocytosis-related protein expression (Rac1, MFGE8 and LC3) was assayed by western blots, and pathological changes in the brain were evaluated by hematoxylin-eosin (HE) staining and immunofluorescence.

Fig. 1.

Schematic illustration of the experimental design. All rats underwent either CLP or sham surgery. In the CLP+MFGE8+CGT group, CGT was administered via intraperitoneal injection (IP) 1 hour after CLP surgery, while MFGE8 was administered via ICV 24 hours post-CLP surgery. Cognitive function was assessed 3 days after CLP using open field task (OFT) and the 5-day Morris water maze (MWM) tests. T2 MRI scans and fMRI scans were conducted 7 days post-CLP, after which the animals were euthanatized for histological and biochemical assessments. CLP, cecal ligation and puncture; CGT, cilengitide; MFGE8, Milk fat globule EGF factor 8 protein; ICV, intracerebroventricular injection; IP, intraperitoneal injection; OFT, open field task; MWM, Morris water maze; fMRI, functional magnetic resonance imaging; MRI, magnetic resonance imaging; ELISA, Enzyme-linked immunosorbent assay.

The sepsis rat models were established by CLP surgery in male SD rats. After fasting for 6 h, rats were anesthetized with an intraperitoneal injection of 35 mg/kg pentobarbital sodium (Sigma, 11715, Darmstadt, Germany) following the protocol by Rittirsch et al. [21]. A 2 cm incision was made on the lower quadrant of the abdominal wall under aseptic conditions. The mesentery and cecum were then dissociated. The mesentery and cecum were exposed, and the cecum was ligated 3 cm distal to the ileocecal valve using 4-0 non-absorbable suture. Perforation was induced by puncturing the cecum twice with a 22-gauge needle, leading to the extrusion of a small amount of feces. Immediately after the CLP surgery, rats were subcutaneously injected with saline solution (5 mL/100 g) for fluid resuscitation. Rats were then allowed free access to food and water post-surgery. In the sham group, all surgical procedures were performed except for cecum ligation and perforation. After successfully establishing the sepsis model, a total of 20 rats experienced acute respiratory failure or renal failure and died post CLP surgery. No antibiotics were administered to the experimental rats at any time during the study. The survival rate was monitored every 24 hours for 7 days, and a survival curve was constructed.

Magnetic resonance imaging (MRI) scans were performed in a 7-Tesla horizontal bore magnet (Bruker Biospec 7 T/20 USR; Bruker, Karlsruhe, Germany). Using a rat head volume coil, we acquired blood oxygen level-dependent (BOLD) fMRI responses [17]. Eight rats in each group were included in imaging examinations. Fast spin-echo T2 weighted MRI was obtained to visualize the structure of the rats’ brains. The rats were anesthetized by 2% isoflurane (792632, Sigma, Shanghai, China), and then the head was fixed in a capsule with a tooth-bar to limit movement. Inhaling 2% isoflurane and 2 L/min oxygen, the rats breathed at a rate of 60 times per minute. Fast spin-echo T2-weighted MRI was obtained with the following parameters: slice = 25, repetition time (TR) = 3000 ms, echo time (TE) = 11 ms, scan time = 4 min 48 s, Flip angle = 90°, SI = 1.00/1.00 mm, field of view (FOV) = 3.17/2.50 cm, Matrix = 256. The scan parameters of BOLD were slice = 18, TR = 2000 ms, TE = 25 ms, scan time = 5 min 0 s, Flip angle = 90°, slice thickness = 1 mm, FOV = 2.88/1.8 cm, Matrix = 128. We used SPM8 (https://www.fil.ion.ucl.ac.uk/spm/) and DPABI (http://rfmri.org/dpabi) for pre-processing, including slice timing, correction, realignment, motion correction, co-registration of T2 images to a rat brain atlas and spatial normalization of functional data. Images in which the head moved by more than 0.2 mm and 2 degrees were excluded. After the images were smoothed with a Gaussian kernel of 0.4 mm $\times$ 0.4 mm $\times$ 0.4 mm full-width at half-maximum to improve the signal-to-noise ratio, ALFF was abstracted from the BOLD images. In order to make inferences about group data, we analyzed the areas by T-contrast images between every two groups, which showed significant activation (p 0.005), with a cluster size more than 10 pixels.

The total hippocampus is composed of the following subfields: cornu ammonis (CA1, CA2, and CA3), dentate gyrus (DG) and the subiculum, which was based on T2-MRI scan. The segmentation of total hippocampus and its subfields was implemented using voxel-based morphometry (VBM) with statistical parametric mapping (SPM) 12 (https://www.fil.ion.ucl.ac.uk/spm/). The relevant hippocampal areas were determined according to the functional Waxholm space atlas of the rat brain. Hippocampal volume was calculated and analyzed by MATLAB 2013b (MathWorks Inc., Torrance, CA, USA) and SPSS 22.0 (IBM, Armonk, NY, USA).

Motor function and anxiety were assessed by OFT. The open field apparatus was a black square (100 $\times$ 100 cm; height of 50 cm), and the arena was divided into the center area (60 cm $\times$ 60 cm square) and the edge area (20 cm away from the walls). Light levels were kept low (~300 lx) to minimize anxiety. The test sessions were recorded by a video camera placed 160 cm above the arena and analyzed with the ANY-maze Video Tracking System (Stoelting Co., Wood Dale, IL, USA). The experiment was performed on the third day after surgery. The total distance traveled by the animals and the duration spent in the central area were measured in a 10-minute session, after which the arena was cleaned with 75% alcohol to remove odors.

For the MWM, cognitive functions related to spatial learning and memory were evaluated according to the methods of Wang [22]. The apparatus consisted of a circular pool (diameter 180 cm; height 45 cm) filled with opaque water (22–24 °C) and a submerged platform (diameter 4 cm, smooth-faced) located at a fixed position in the target quadrant. The pool was divided into four quadrants, and the tests were recorded by a camera above the pool and analyzed using the ANY-maze Video Tracking System. The MWM test comprised navigation and probe tests. Navigation tests were conducted from the third to sixth day post-CLP, with each rat trained four times from every quadrant each day to find the hidden platform. On the seventh day, the platform was removed for the probe test, during which the time spent in the target quadrant and the number of times crossing the platform location were recorded.

The rats were injected intraperitoneally with pentobarbital sodium (11715, Sigma, Shanghai, China) 50 mg/kg on day 8 post-CLP surgery, and the brain tissues were fixed, cut into slices, dehydrated, cleared, and embedded in paraffin. Sections were then subjected to immunohistochemistry with primary antibodies against ionized calcium-binding adapter molecule 1 (Iba-1) (1:100; Abcam, ab178846, Cambridge, UK), and then incubated with goat anti-rabbit goat anti-rabbit immunoglobulin G heavy and light chains (IgG H&L) (Cy3, 1:100, Abcam, ab6939, Cambridge, UK) and secondary antibodies. Images were captured and analyzed using a fluorescence microscope and ImageJ software (ImageJ 1.0, NIH, Bethesda, MD, USA).

ELISA was performed to determine the levels of IL-1$\beta$, IL-6, and CRP in the hippocampus. The hippocampus was lysed, homogenized, and centrifuged, and the levels of the cytokines were measured using ELISA kits (Solarbio, Beijing, China) according to the manufacturer’s instructions. Absorbance was measured with a microplate reader (Bio-Rad 680, iMark, Hercules, CA, USA) at 450 nm.

The hippocampus on both sides was lysed and immediately homogenized in radio immunoprecipitation assay (RIPA) lysis buffer containing phenylmethylsulfonyl fluoride (n = 5). After protein concentration was assayed with a bicinchoninic acid (BCA) kit (Solarbio, Beijing, China), total protein samples were separated by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride filters. Membranes were incubated overnight at 4 °C with anti-rat monoclonal antibodies raised against MFGE8 (1:2000, Abcam, ab200649, Cambridge, UK), Rac1 (1:2000, Abcam, ab155938) and LC3 (1:2000, Abcam, 192890). On the second day, the membranes were washed three times using phosphate buffer saline (PBS) with 0.1% Tween-20, then reacted with horseradish peroxidase-conjugated rabbit anti-rat IgG (H+L) (1:5000, Abcam, ab6734) for 2 h at 22–25 °C. The integral optical density of each band was measured using ImageJ software.

Data are presented as means $\pm$ standard deviation and were analyzed with SPSS 22.0 (SPSS, Inc., Chicago, IL, USA) and Graphpad Prism 8 (Graphpad Software, San Diego, CA, USA). Multiple group comparisons were performed by one-way analysis of variance (ANOVA) followed by Tukey’s test for multiple comparisons. For results from behavioral tests, western blots and ELISA, p-values less than 0.05 (p 0.05) were considered statistically significant. The differences in ALFF values were analyzed by two-sample t-tests with REST software (http://www.restfmri.net/) between every two groups. The correction thresholds were limited by the false discovery rate (FDR) at p = 0.005. Cluster connectivity criteria: mm = 5, cluster size = 10, p = 0.001.

CLP surgery induced rat deaths from sepsis with serious systemic inflammatory response and multiple organ failure. Totally 20 rats suffered acute organs failure and died post CLP surgery. The post-CLP 7-day survival rates indicated the severity of sepsis in the different groups, which were 100% in the sham group, 50.79% in the CLP group, 63.03% in the CLP+MFGE8 group, and 38.96% in the CLP+MFGE8+CGT group (Fig. 2A). We ensured that each group will get at least 8 rats during 7 day post CLP surgery.

Fig. 2.

T2-MRI tracking of structural damage in the hippocampus of rats from the four groups. (A) Survival rates after CLP surgery. The 7-day survival proportions in the sham group, CLP group, CLP+MFGE8 group and CLP+MFGE8+CGT group were 100%, 50.79%, 63.03% and 38.96%, respectively. (B) The bilateral hippocampal areas were extracted according to the functional Waxholm space atlas of the rat brain. (C) T2-MRI scans showed a regional signal increased significantly in the hippocampus of rats after CLP surgery. However, this abnormal signal was reduced in the CLP+MFGE8 group. In the CLP+MFGE8+CGT group, the hyperintense focus in the region of the hippocampus, and the extent of damaged areas was milder than in the CLP group. (D) Volume of the left hippocampus (F [3,28] = 78, ANOVA, n = 8). (E) Volume of the right hippocampus (F [3,28] = 25, ANOVA, n = 8). *p 0.05, compared with the sham group; #p 0.05, compared with the CLP group; &p 0.05, compared with the CLP+MFGE8 group, one-way ANOVA, n = 8. ANOVA, analysis of variance.

T2-MRI detected structural changes in rat brains after CLP surgery. We extracted bilateral hippocampal areas based on the Waxholm space atlas of the rat brain, and compared the volume of hippocampus in the four groups (Fig. 2B). Compared with the sham group, signals in the hippocampus and external capsule were increased in the CLP group (Fig. 2C). The hyperintense focus in the region of the hippocampus indicated the brain cells were experiencing edema post-CLP surgery. In the CLP+MFGE8 group, the hyperintense region was significantly decreased in the hippocampus. The volumes of the bilateral hippocampus were decreased in the CLP+MFGE8 group compared with CLP group, suggesting that treatment with MFGE8 could reduce brain damage in the SAE rats (Fig. 2D,E). The serious damage to the hippocampus was observed in the CLP+MFGE8+CGT group. CGT inhibited the therapeutic function of MFGE8, and increased the high signal regions in the left hippocampus (Fig. 2D).

The REST software was used to analyze the ALFF with BOLD signals among the four groups using one-way ANOVA analysis. The results indicated that the significant difference among the four groups was only in the molecular layer of right cerebellum (montreal neurological institute (MNI) coordinates: x = –1, y = –86.05, z = 21.2; Voxel size 10 voxel, voxel significance p = 0.005). There were no significant findings in the cerebral cortex. Therefore, a two-sample t test was used to compare each two groups. The results revealed that, in comparison to the sham group, the ALFF value increased in the left hypothalamic region (HR) and right cornu ammonis 3 (CA3) of the hippocampus in rats from the CLP group (p 0.005, FDR corrected, Fig. 3A, n = 8). Several brain regions exhibited lower ALFF signals in the CLP+MFGE8 group compared to the CLP group, including the right CA1, CA3 and DG (p 0.005, FDR corrected, Fig. 3B, n = 8). In contrast, the ALFF in the right CA1 of hippocampus increased, while it decreased in the left molecular layer of the cerebellum (p 0.005, FDR corrected, Fig. 3C, n = 8). The number of voxels and MNI coordinates of the cluster in the brain are listed in Table 1.

Fig. 3.

fMRI detected brain functional changes in the four groups. (A) Compared with the sham group, the ALFF values increased in the left hypothalamic region and right CA3 of the hippocampus (yellow areas) in the CLP group (p 0.005, FDR corrected, t-test, n = 8). (B) In the CLP+MFGE8 group, several brain regions including the right CA1, CA3 and DG (green areas) emerged with lower ALFF signals than the CLP group (p 0.005, FDR corrected, t-test, n = 8). (C) Compared with the CLP group, CGT decreased the ALFF value in the left molecular layer of the cerebellum (blue area) and increased the ALFF in the right CA1 of the hippocampus (yellow area) (p 0.005, FDR corrected, t-test, n = 8). (D) The bilateral hippocampal areas were extracted according to the functional Waxholm space atlas of the rat brain. (E) The average zALFF value of left hippocampus (F [3,28] = 2.7, p = 0.0598, ANOVA, n = 8). (F) The average zALFF value of right hippocampus (F [3,28] = 124, p = 0.0004, p 0.05, ANOVA, n = 8). CA3, cornu ammonis 3; CA1, cornu ammonis 1; DG, dentate gyrus. *p 0.05, compared with the sham group; #p 0.05, compared with the CLP group; &p 0.05, compared with the CLP+MFGE8 group, one-way ANOVA, n = 8. ALFF, Amplitude of low-frequency fluctuations; FDR, false discovery rate; zALFF, z-score transformation amplitude of low-frequency fluctuations.

In this study, the average ALFF value in the bilateral hippocampal region of rats among the four groups was extracted for further analysis (Fig. 3D). The results indicated that there was no significant difference in the functional activities in the left hippocampus among four groups (Fig. 3E, p $>$ 0.05). However, the z-score transformation amplitude of low-frequency fluctuations (zALFF) value in the right hippocampus of rats in the CLP+MFGE8 group decreased compared to the CLP group, suggesting that MFGE8 decreased the neural activity in the right hippocampal region of SAE rats (Fig. 3F, p 0.05). Additionally, when compared with the CLP+MFGE8 group, the zALFF value in the right hippocampus of the rats in the CLP+MFGE8+CGT group was higher, and the spontaneous neural activity in the hippocampus was significantly enhanced (Table 2).

To test the motor function and anxiety level of rats in the four groups, we recorded the total running distance and duration of time spent in the central area by OFT (Fig. 4A). To evaluate cognitive dysfunction among rats in the four groups, we tested the rats by MWM (Fig. 4B). The results showed no significant difference in the total running distance among the four groups (Fig. 4C; p $>$ 0.05). The duration of running in the central area by the CLP group was less than the sham group (Fig. 4D; p 0.05). In the MWM test, MFGE8 shortened the training time (Fig. 4E; p 0.05) and increased the time spent in the target quadrant and the number of times the platform was crossed in the space exploration section (Fig. 4F,G; p 0.05). These results showed that MFGE8 improved the exploratory behavior of rats. In contrast, the MFGE8 antagonist CGT caused significant impairment of spatial memory in the CLP+MFGE8+CGT group (p 0.05).

Fig. 4.

MFGE8 improved the performance of neurobehavioral tests in SAE rats. (A) The tracks of rats in OFT test among four groups. (B) The tracks of rats in probe tests of MWM. (C) The total distance rats traveled in 10 min of OFT test (F [3,28] = 1.122, p $>$ 0.05, one-way ANOVA, n = 8). (D) Duration in the central area the rats traveled in 10 min of OFT (F [3,28] = 13.08, one-way ANOVA, n = 8). (E) The latency difference in the MWM navigation test during four days (repeated measurements ANOVA, n = 8). (F) The number of crossings of the platform location (F [3,28] = 14.42, ANOVA, n = 8). (G) The total time in the target quadrant (F [3,28] = 10.23, n = 8). *p 0.05, compared with sham group; #p 0.05, compared with CLP group; &p 0.05, compared with CLP+MFGE8 group.

We examined the expression of IL-1$\beta$ (Fig. 5A), IL-6 (Fig. 5B) and CRP (Fig. 5C) in the hippocampus of SAE rats (n = 5) by ELISA on day 8 post-CLP surgery. It was observed that the levels of IL-1$\beta$, IL-6 and CRP in the CLP group were significantly higher than those in the sham group (p 0.05). In contrast, the CLP+MFGE8 group exhibited significantly lower levels of IL-1$\beta$, IL-6, and CRP compared to the CLP group (p 0.05). However, the expression of IL-1$\beta$, IL-6 and CRP increased in the CLP+MFGE8+CGT group when compared to the CLP+MFGE8 group (p 0.05). The differences in IL-1$\beta$, IL-6 and CRP levels between CLP group and CLP+MFGE8+CGT group were not significant (p $>$ 0.05).

Fig. 5.

Inflammatory factors and efferocytosis-related proteins in rat hippocampus. (A) Expression of IL-1$\beta$, (B) IL-6 and (C) CRP in rat hippocampus. (D) Expression of efferocytosis-related proteins in the hippocampus 8 days after CLP surgery by western blot. Expression of (E) MFGE8, (F) Rac1 and (G) LC3. *p 0.05, compared with the sham group; #p 0.05, compared with the CLP group; &p 0.05, compared with the CLP+MFGE8 group. One-way ANOVA, n = 5. IL, interleukin; CRP, C-reactive protein; Rac1, ras-related C3 botulinum toxin substrate 1; LC3, light chain 3.

We examined the expression of efferocytosis-related proteins in the hippocampus of the four groups (Fig. 5D, n = 5) on day 8 post-CLP using western blot. The concentration of MFGE8 in the CLP+MFGE8 and CLP+MFGE8+CGT groups was increased (Fig. 5E; p 0.05). The injection of MFGE8 increased the expression of Rac1 in the hippocampus after CLP surgery, while addition of CGT, the $\alpha$V$\beta$3 and $\alpha$V$\beta$5 integrin inhibitor, down-regulated the expression of Rac1, which plays a vital role in the migration of microglia (Fig. 5F; p 0.05). LC3 is a key molecule in intracellular phagocytosis during efferocytosis. Rats in the CLP+MFGE8 group had increased LC3 in the hippocampus, while CGT suppressed expression of LC3 and disrupted microglial phagocytosis (Fig. 5G; p 0.05).

To investigate effect of MFGE8 on microglial efferocytosis in the hippocampus, we performed immunofluorescence staining of hippocampal sections (Fig. 6A). The results indicated that CLP surgery significantly increased Iba-1 expression in the hippocampal region. However, treatment with MFGE8 dramatically promote the expression of Iba-1 in microglial cells compared with CLP-treated rats (Fig. 6B). Additionally, there were significantly more Iba1-positive cells in the hippocampus of rats in the CLP+MFGE8 group. In contrast, the Iba-1 level in the hippocampus was significantly decreased in the CLP+MFGE8+CGT group, and CGT also reduced the density of microglial cell in the hippocampal region (Fig. 6C).

Fig. 6.

MFGE8 treatment activated microglial aggregation in the hippocampus. Hippocampal sections were stained with Iba-1 as the microglial marker (red), with nuclear staining by DAPI (blue). (A) Hippocampal sections stained with Iba-1 were visualized by immunofluorescence to localize and assess microglial activation. (B) The increase in Iba-1-stained areas induced by MFGE8 was reduced after CGT treatment. (C) The number of activated microglia was increased in CLP+MFGE8 group. The number of activated microglia was reduced by CGT treatment. *p 0.05, compared with the sham group; #p 0.05, compared with the CLP group; &p 0.05, compared with the CLP+MFGE8 group (one-way ANOVA, n = 3). Iba-1, ionized calcium-binding adapter molecule 1; DAPI, 4${}^{\prime }$,6-diamidino-2-phenylindole.