A total of 172 C57BL/6 male mice (12–14 months; 24–36 g) were purchased from the Animal Center of Nanjing Medical University (Nanjing, China). The mice were housed in pairs in a colony room maintained at 24±1°C and 40–50% relative humidity with a 12-h light-dark cycle (lights on at 07:00 a.m.). Mouse chow and water were available ad libitum. All studies were approved by the Institutional Animal Care and Use Committee of Jinling Clinical Medical College of Nanjing Medical University, and met the Guide for the Care and Use of Laboratory Animals from the National Institutes of Health (NIH; Bethesda, MD, USA) guidelines for the Use of Experimental Animals in research.

Mice were randomly assigned to the following groups: Control (n=22); surgery (n=30); surgery + repeated LPS (−72 h) (n=30); surgery + LPS (−24 h) (n=30); surgery + LPS (−6 h) (n=30); and surgery + LPS (0 h) groups (n=30). Exploratory laparotomy was performed as previously described (16,17). Mice were allowed to acclimate for 2 weeks prior to the experiments. The surgery was performed under 1.5% isoflurane anesthesia to mimic exploratory abdominal surgery in humans. A median abdominal incision (~1-cmlong vertical incision) was made to allow for penetration of the peritoneal cavity. Thereafter, the investigator inserted blunt forceps into the opening and explored the viscera, intestines and musculature. Next, sterile 4-0 chromic gut sutures were used to close the peritoneal lining and skin. The wound was dressed with polysporin (Pfizer, Inc., New York, NY, USA). The animals were maintained under isoflurane anesthesia during the 10 min of the surgical procedure. Sham controls received neither anesthesia nor surgery. In the present study, all efforts were made to minimize animal suffering and reduce the number of animals used.

All injections were performed intraperitoneally (i.p.) at a volume of 5 ml/kg body weight. LPS (from Escherichia coli 0111:B4, L-2630; Sigma-Aldrich; Merck KGaA, Darmstadt, Germany) was dissolved in 0.9% NaCl. The dose (0.2 mg/kg) of LPS was based on previous studies, in which it was demonstrated that LPS preconditioning abolished the exacerbated inflammation and protected against cognitive impairment in other animal models (13,14). Animals in the control group received an injection of 0.9% NaCl with the same regimen. All mice were injected with LPS or saline between 8.00 and 9.00 a.m. Mice were sacrificed and tissue was collected at 24, 72 and 168 h following preconditioning. Tissue was additionally collected from six untreated mice to include as a baseline control group. In the present study, no animal succumbed during the observation period in all groups. The flow chart of the experimental protocol is presented in Fig. 1.

An open field test was performed to evaluate the exploratory and anxiety behaviors. Mice were placed individually in the center of a clear Plexiglas box (50×50×40 cm) and left free to explore the environment for 5 min. Total movements and time spent in the center of the open field were recorded during a 5-min exploration time period. The behavior of the mice was recorded using a video camera (Software Technology Co., Ltd., Ningbo, China). The apparatus was cleaned with 70% ethanol prior to testing each mouse to avoid the presence of olfactory cues.

Fear conditioning tests were performed to evaluate fear learning, as previously described (16,17). Mice were placed into a conditioning chamber (32×25×25 cm), with a stainless-steel shock grid floor. Mice learn to associate an environment (context) with a conditional stimulus (CS), for example, a tone, and an unpleasant stimulus [foot shock; unconditional stimulus (US)]. The training paradigm was as follows: Tone at 75 dB for 30 sec; shock at 0.75 mA for 2 sec. Contextual memory was tested 24 h subsequent to the training. The animals were placed back in the original training chamber to monitor freezing behavior, which was defined as an absence of any movement for >3 sec. The cued fear memory was tested 2 h subsequently in a novel context with a continuous 3-min training tone presentation to monitor freezing behavior.

The concentrations of TNF-α (cat. no. 22351), IL-1β (cat. no. 23107), IL-6 (cat. no. 21724), and IL-10 (cat. no. 26271) in the hippocampus were determined using ELISA kits (North China Institute of Science and Technology, Beijing, China), according to the manufacturer's instructions, as previously described (16,17). Mice were sacrificed with an i.p. injection of 2% sodium pentobarbital (60 mg/kg) and the hippocampus was collected, separated, and placed in a homogenizer. Homogenates were centrifuged at 10,000 × g for 10 min at 4°C. The quantity of TNF-α, IL-1β, IL-6 and IL-10 in each sample was standardized to its protein content.

Under deep isoflurane anesthesia, mice were perfused transcardially with normal saline, followed by 4% paraformaldehyde (PFA) in PBS for 15 min. Brains were harvested and post-fixed in 4% PFA overnight at room temperature and then with 30% sucrose for 24 h at 4°C. Brains were freeze-mounted in optimal cutting temperature (−20°C) embedding medium, cut into 10-µm-thick sections using a cryostat, and mounted on slides. Slices were blocked with 3% bovine serum albumin (Thermo Fisher Scientific, Inc., Waltham, MA, USA) for 1 h at room temperature. The sections were incubated with a rat anti-allograft inflammatory factor 1 [IBA1 (a marker of microglial activation), 1:200, cat. no. ab48004; Abcam, Cambridge, UK] antibody overnight at 4°C, followed by a 1 h incubation with the secondary antibodies [Cy3-conjugated donkey anti-rat immunoglobulin G (1:300, cat. no. sc53682; Santa Cruz Biotechnology, Inc., Dallas, TX, USA] at room temperature. Following three washes for 5 min at room temperature using PBS, sections were counterstained with DAPI for 15 min at room temperature, mounted on glass slides and coverslipped with fluorescence mounting medium. A total of three independent microscopic fields in each section were randomly acquired in the hippocampal CA1, CA3 and dentate gyrus regions, and three sections per mouse were imaged. Images were processed, and the area of the microglia was quantified using ImageJ software (version 1.50i; National Institutes of Health).

Statistical analysis was performed using SPSS 16.0 for Windows (SPSS, Inc., Chicago, IL, USA). Data are presented as the mean ± standard error of the mean. Multiple comparisons were analyzed using one-way analysis of variance followed by the post hoc Tukey test. P<0.05 was considered to indicate a statistically significant difference.

To first exclude possible impairments in locomotor activity and exploratory behavior following surgery, mice were evaluated in an open field arena prior to the contextual assessment. As presented in Fig. 2A and B, no alterations were observed in the total distance and time spent in the center of the arena between any group (total distance: F_(5,66) = 0.546, P=0.740; time spent in the center: F_(5,66) = 1.871, P=0.111).

To assess whether surgery induced cognitive impairments, cognitive function was assessed by fear conditioning tests. As revealed in Fig. 2C and D, the contextual fear response was significantly decreased in mice subjected to surgery compared with the control group. Notably, a single LPS preconditioning treatment at 24 h prior to surgery restored freezing behavior, an indicator of memory retention in rodents when performed at day 7 following surgery (F_(5,66) = 5.274, P<0.001; Fig. 2C). Repeated low-dose LPS administration did not reverse surgery-induced cognitive impairment when compared with the control group (P=0.099). However, no significant difference was observed in post-tone freezing time in the auditory-cued fear test between groups (F_(5,66) = 1.460, P=0.215; Fig. 2D), suggesting that surgery impaired hippocampal-dependent memory (5).

In general, repeated injections of LPS at an increasing dose may provide greater suppression of pro-inflammatory responses. Therefore, LPS was first administered at 72 h prior to surgery for 3 consecutive days to investigate whether this regime may decrease the signs of neuroinflammation and consequently improve cognitive impairment following surgery. As presented in Fig. 3, repeated LPS administration significantly decreased the hippocampal expression levels of pro-inflammatory mediators, including: TNF-α, IL-1β and IL-6, at 24 h post-surgery (TNF-α: F_(4,25) = 4.730, P=0.001; IL-1β: F_(4,25) = 5.435, P<0.001; IL-6: F_(4,25) = 3.534, P=0.008). However, repeated LPS administration did not significantly alter the hippocampal levels of the anti-inflammatory mediator IL-10 at 24 h post-surgery (IL-10: F_(4,25) = 0.785, P=0.587). In addition, it was observed that there was no difference in the hippocampal levels of TNF-α, IL-1β, IL-6 and IL-10 among the groups at 72 h post-surgery, returning to the baseline levels as in the control mice. Based on these results, and to minimize the number of animals used, cytokine levels were only measured at 24 and 72 h post-surgery in the subsequent experiments.

Although repeated LPS administration downregulated the signs of neuroinflammation, it was observed that this regime did not confer beneficial cognitive outcomes. Subsequently, the present study tested the hypothesis that a single low-dose LPS administration at 24 h prior to surgery may decrease the levels of pro-inflammatory mediators, in addition to reversing cognitive impairment. As expected, one single low-dose treatment with LPS at 24 h prior to surgery significantly reduced the hippocampal levels of TNF-α, IL-1β and IL-6 at 24 h post-surgery (TNF-α: F_(4,25) = 9.254, P<0.001; IL-1β: F_(4,25) = 6.598, P=0.001; IL-6: F_(4,25) = 4.410, P=0.008; Fig. 4A-C). Similarly, a single low dose of LPS at 24 h prior to surgery did not significantly alter the hippocampal level of the anti-inflammatory mediator IL-10 following surgery (IL-10: F_(4,25) = 0.890, P=0.484; Fig. 4D).

To further investigate whether a shorter interval between LPS administration and surgery may produce similar effects to the 24 h regime, LPS was administered 6 h prior to surgery. A single low-dose LPS administration at 6 h prior to surgery did not alter the cytokine levels (TNF-α: F_(4,25) = 10.713, P<0.001; IL-1β: F_(4,5) = 9.9655, P<0.001; IL-6: F_(4,25) = 5.058, P=0.004; IL-10: F_(4,25) = 1.434, P=0.252; Fig. 5).

To evaluate whether LPS administration can aggravate the inflammatory response to surgical trauma, low-dose LPS was administered immediately prior to surgery. Unexpectedly, low-dose LPS administration did not further increase the expression levels of the assessed cytokines in the hippocampus, measured at 24 and 72 h post-surgery, when compared with the surgery group (TNF-α: F_(4,25) = 7.596, P<0.001; IL-1β: F_(4,25) = 11.135, P<0.001; IL-6: F_(4,25) = 7.028, P=0.001; IL-10: F_(4,25) = 0.896, P=0.481; Fig. 6).

Since microglial cells are among the principal immune cells of the brain involved in the amplified neuroinflammatory response during the development of POCD, the present study evaluated whether low-dose LPS preconditioning was able to modulate the activity of microglia. Similar to the results of the analysis of cytokine expression, surgery induced a significant increase in the number of hippocampal IBA1-positive cells in the hippocampus, which was reversed by repeated LPS administration at 72 h or a single LPS administration at 24 h prior to surgery. Consistent with the above results, one single low-dose LPS administration at 6 h or immediately prior to surgery did not further increase the number of IBA1-positive cells in the hippocampus (Fig. 7).