Combination of isoflurane and propofol as general anaesthesia during orthopaedic surgery of perioperative cerebral hypoperfusion rats to avoid cognitive impairment Anaesthesia during perioperative cerebral hypoperfusion

Background: Perioperative cerebral hypoperfusion (CH) is common, although the underlying mechanism of cognitive impairment that results due to perioperative cerebral hypoperfusion remains to be determined. Isoflurane anaesthesia induces neuronal injury via endoplasmic reticulum (ER) stress, whereas a sub-anaesthetic dose of propofol improves postoperative cognitive function. However, the effects of the combination of isoflurane plus propofol, which is a common aesthetic combination administered to patients, on ER stress and cognition remain unknown. Methods: We sought to determine the effects of isoflurane plus propofol on ER stress and cognitive function in rats insulted by cerebral hypoperfusion. Ligation of the bilateral common carotid arteries (CCA) was adopted to develop the cerebral hypoperfusion rat model. A second surgery, open reduction and internal fixation (ORIF), requiring general anaesthesia, was performed 30 days later so that the effects of anaesthetics on the cognitive function of CH rats could be assessed. Rats received isoflurane alone (1.9%), propofol alone (40 mg·kg-1·h-1) or a combination of isoflurane and propofol (1% and 20 mg·kg-1·h-1 or 1.4% and 10 mg·kg-1·h-1). Behavioural studies (fear conditioning FC test), histological analyses (Nissl staining) and biochemical analyses (western blotting of the harvested rat brain tissues) were employed. Results: The combination of 1% isoflurane plus 20 mg·kg-1·h-1 propofol did not aggravate cognitive impairment or ER stress in ageing rats with CH that were further subjected ORIF surgery. Conclusions: These data suggest that ER stress contributes to the underlying mechanism of cognitive impairment and that the combination of isoflurane and propofol did not aggravate cognitive impairment and ER stress in ageing rats with CH that were further subjected ORIF surgery.

most common configuration is a transmembrane pentamer composed of 2α 1 2β 2 γ 2 , accounting for 43% of all GABA A R configurations and representing the most abundant configuration in the hippocampus and the cerebral cortex [18]. The GABA A R α1 subunit, which is related to cognition, is the most widely distributed in the mammalian brain, [19,20]. Its main function is to maintain CNS arousal and the sensitivity of the receptor to sedative hypnotics (propofol, isoflurane, etc.). Mutation of the M2 domain Ser270 and the M3 domain Ala291 in the α1 subunit affects the potency of isoflurane and propofol on GABA A Rs [21]. Kelley et al. confirmed that cerebral ischaemia can induce miniature inhibitory postsynaptic current (mIPSC) reduction and GABA-activated current inhibition [22]. Further studies found that mIPSC frequency and kinetic parameters did not change, only amplitude decreased, while oxygen-glucose deprivation (OGD) inhibited neuronal GABA A R α1 subunit expression [22]. This finding suggests that the change in GABA A R activity is triggered by a decrease in the expression of its functional subunit α1.
Furthermore, our previous study showed that a sub-anaesthetic dose (20 mg·kg − 1 ·h − 1 ) of propofol exerts post-treatment brain protection by activating the KCC2-GABA A R pathway.
Propofol post-treatment can reverse the decrease in hippocampal IPSCs after OGD injury, promote KCC2 expression, and maintain the normal function of GABA A R. However, administration of KCC2 antagonists only partially reversed the effect of propofol on mIPSC [23]. IT remains unknown whether or not cerebral ischaemia triggers the expression change and structural regulation of GABA A R functional subunit protein. Is there any upstream mechanism other than KCC2 that regulates the GABA A R structure, thereby affecting its function? To answer these questions, we chose the GABA A R α1 subunit as a target of research in this study. GABA A R undergoes post-synthesis modification and folding in the ER. Prolonged ER stress has been well known to be related with neurodegenerative diseases [24,25]. The unfolded protein response (UPR) triggered by ER stress is an important quality control system for maintaining protein homeostasis (Proteostasis). Proteostasis refers to an equilibrium state of specific protein synthesis, folding and unfolding, modification and degradation in the intracellular proteome at a specific time point. The ER of the cell is a site for the folding and post-translational processing of secreted proteins and membrane proteins (approximately 1/3 of the human proteome). Binding immunoglobulin protein (BiP), also known as glucose-regulated protein 78 (GRP78), is an ER chaperone protein whose expression is part of the UPR and is required to alleviate ER stress [26]. Once ER stress occurs, BiP binds to unfolded proteins and activates downstream receptor proteins, increasing molecular chaperone expression, reducing global protein translation, and increasing unfolded/misfolded proteins. It degrades and reduces ER stress and protects cells through endoplasmic reticulum-associated degradation (ERAD).
The expression of C/EBP homologous protein (CHOP), a transcriptional factor, is also induced by ER stress but indirectly regulates apoptosis [27]. During stress, UPR attempts to increase protein-folding capacity and remove misfolded and unfolded proteins. If homeostasis is inadequately restored under chronic ER stress, terminal UPR will trigger apoptosis through abundant signalling mechanisms, mainly mediated by CHOP, c-Jun Nterminal kinase (JNK), and caspase-12, with CHOP as the most widely studied [28].
Thus, the expression levels of BiP, CHOP and the GABA A R α1 subunit were used to evaluate the cellular mechanisms accounting for the neural substrate conditions that allow normal cognitive functions in this study.
The objective of the current study was to explore general anaesthetics for rats with CH that are subjected to ORIF surgery to protect cognitive function. By using behavioural and biochemical analyses, we tested the hypothesis that a combination of isoflurane and propofol better protects cognitive function than isoflurane or propofol administered alone during ORIF surgery.

Results
Combination treatment with 1% isoflurane and 20 mg·kg -1 ·h -1 propofol protected cognitive function in ageing rats with CH and being subjected to an ORIF surgery.
To observe the effects of different dosages of isoflurane and propofol on cognitive function, a contextual FC test was performed on the first and seventh days after ORIF. The percentage of freezing time in Group C and Group IP 1 was not significantly different on the first day (C vs IP 1 : 44.23 ± 6.60 vs 42.86 ± 7.12, P = 1.00) or the seventh day (C vs IP 1 : 35.70 ± 5.21 vs 34.85 ± 5.02, P = 1.000) after ORIF (Fig. 1A). However, in Groups IP 2 , I and P, the percentage of freezing time was significantly reduced compared with Group C all P < 0.05) (Fig. 1A). The results suggest that the combination of 1% isoflurane and 20 mg·kg -1 ·h -1 propofol could protect cognitive function, while other dosages could not.
Treatments with isoflurane or propofol alone were not able to prevent CA1 neuronal death in ageing rats with CH that were subjected to ORIF surgery.
Hippocampal slices were stained with cresyl violet (Nissl staining) to investigate potential neuronal damage caused by anaesthetics on days 1 and 7 after ORIF. Compared with Group C, the number of surviving neurons decreased one day after ORIF only in Group I (C vs I: 193.13 ± 23.94 vs 150.88 ± 20.19, P = 0.039, Fig. 1 B-C). On the seventh day after ORIF, the number of surviving neurons in Groups I and P was significantly lower than that  (Fig. 1 B-C).
Combination treatment with 1% isoflurane and 20 mg·kg -1 ·h -1 propofol maintained the expression level of cell GABA A R α1 in the hippocampus.
As described above, GABA A R α1 is a key functional component of the neural substrate involved in cognitive functions. Therefore, western blotting was performed on the first and seventh days after ORIF to evaluate the expression of the GABA A R α1 subunit. There was no difference in the expression of GABA A R α1 between Group C and Group IP 1  100.00 ± 14.72 vs 56.23 ± 8.12, P < 0.001; C vs I: 100.00 ± 14.72 vs 27.92 ± 4.39, P < 0.001; C vs P: 100.00 ± 14.72 vs 24.71 ±4.01, P < 0.001) (Fig. 2).

BiP
The expression levels of BiP in Groups IP 1 , IP 2 , I and P were all upregulated compared with that in Group C on day 1 (C vs IP 1 : 100.00 ± 18.58 vs 442.86 ± 69.09, C vs IP 2 : 100.00 ± 18.58 vs 248.02 ± 35.15, C vs I: 100.00 ± 18.58 vs 165.13 ± 25.53, C vs P: 100.00 ± 18.58 vs 188.54 ± 27.90, all P < 0.05). The highest expression level was found in Group IP 1 , and the lowest expression level was found in Group I. On day 7, the expression of BiP fell in all four groups, and there was no significant difference between Group I and Group C (C vs I: 100.00 ± 13.91 vs 142.57 ±18.70, P = 0.053). However, the expression of BiP in Groups IP 1 , IP 2 and P was significantly higher than that in Group C (C vs IP 1 : 100.00 ± 13.91 vs 268.27 ± 46.51, C vs IP 2 : 100.00 ± 13.91 vs 199.47 ±31.66, C vs P: 100.00 ± 13.91 vs 154.64 ± 27.93, all P < 0.05, Fig. 4). The highest expression was observed in Group IP 1 (Fig. 4).

Discussion
In our study, ageing (16-18 month) rats were chosen as test subjects. All rats received ligation of the bilateral CCA to mimic the pathological process of CAS. Thirty days after ligation surgery, ORIF surgery was performed under different anaesthesia regimens according to the group. After ORIF surgery, behavioural experiments (FC test) were carried out to evaluate the cognitive function of the rats. Histological analyses (Nissl staining) were performed to explore neuronal damage, and biochemical analyses (western blotting) of harvested rat brain tissues were performed to detect molecular changes.
The first consideration that must discussed is the selection and intervention of the test subject. The incidence of PND in orthopaedic patients varies from 16% to 45%, although it can be as high as 72% [29], and it has been proven that ageing is a risk factor [30].
Therefore, we chose ageing rats as the test subjects. CH has been reported to be a key factor in the development of cognitive impairment [31]. The underlying mechanism could be hypoxia-induced white matter damage, microvascular inflammation and neuro-gliovascular dysfunction [32]. We deem that ageing patients with perioperative CH require more attention to be paid to the selection of surgery and anaesthesia. Moreover, CAS has been detected in 75% of men and 62% of women older than 65 years, with a prevalence of ≥50% stenosis of 7% in men and 5% in women [33]. Taking incidence into account, we therefore used ligation of the CCA to induce CH in ageing rats in this study. As it is difficult to separate clinical anaesthesia and surgery, and our main purpose was to explore the combined effects of the two factors, no separate anaesthesia group was used, which is The FC test is a very sensitive and effort-independent test of learning and memory [37].
To eliminate effects on motor ability caused by tibial fracture, the FC test was chosen to inspect cognitive function after ORIF surgery. Isoflurane has been reported to suppress learning in a dose-dependent fashion. Hence, we trained animals before surgery and anaesthesia to remove the influence of the acquisition phase on the assessment of memory postoperatively [38]. After ORIF surgery and anaesthesia, the rats were placed in the same chamber that was used during the FC training phase. No tone was delivered while the rats were in the chamber. In this circumstance, freezing behaviours rely on hippocampal memory. On the other hand, the freezing behaviour of rats in a different chamber and exposed to the tone stimulus relies on hippocampal-independent memory.
We found that the freezing time of rats was significantly shorter in Groups I, P and IP 2 than in Group C, while there was no obvious difference between Groups C and IP 1 . The only difference in the intervention among Groups I, P, IP 2 and IP 1 was the anaesthesia method. Our results suggest that hippocampal-dependent memory impairment was not exhibited by the rats anaesthetized with 1% isoflurane and 20 mg·kg -1 ·h -1 propofol, in contrast to all other groups of rats. Such an obvious difference aroused our interest in evaluating the state of related anatomic structures.
The hippocampal CA1 area is crucial for context-specific memory retrieval and spatial memory. After CA1 lesions, both recent and remote memory are impaired [39]. compared with Group C, and on day 7, the number of neurons in Groups I and P were markedly decreased compared with Group C. The difference between Group C and the combined anaesthesia groups was not significant. Thus, we can draw the conclusion that, compared with the combination groups, the high dose of isoflurane or propofol alone can cause irreversible damage to the nervous system.
The GABA A R α1 subunit has also been linked to brain cognitive functions [41]. More recently, the expression level of GABA A R α1 in the hippocampal CA1 region was found to be significantly downregulated in rats with chronic ischaemic encephalopathy [41]. In our study, the expression of GABA A R containing the α1 subunit decreased in all but one group (the 1% isoflurane and 20 mg·kg -1 ·h -1 propofol group). This trend was coincident with the change in freezing time, indicating that improper usage of anaesthetics could aggravate cognitive impairment even though neurons are alive.
A previous study confirmed that isoflurane affects cognitive function by ER stress [42]. Based on previous studies and our experiments, we can make the following inferences.
Anaesthetics stimulate neurons and cause accumulation of unfolded or misfolded proteins in the ER, thereby inducing ER stress. BiP dissociates with transducers of UPR to transport unfolded/misfolded proteins to the cytoplasm and trigger ERAD [43]. Meanwhile, the transcription and translation of proteins within neurons, except for UPR downstream biomarkers, are inhibited. The expression of BiP and CHOP increases due to UPR. If the increased expression of BiP is sufficient to bind unfolded/misfolded proteins, then ER stress could be alleviated and the neuron could adapt, as observed for Group IP 1 in this study. Otherwise, prolonged or excess expression of CHOP will lead to apoptosis [44], as observed for Groups I, P and IP 2 . Different anaesthetic schedules alter various neural substrate components (GABA A R α1 subunit) that play roles in cognitive functions. Although the stimulation may not be severe enough in some cases to cause apoptosis, cognitive impairment may still be aggravated (or induced) or prevented (Fig. 5).

Conclusion
In conclusion, 1% isoflurane and 20 mg·kg -1 ·h -1 propofol is a more favourable aesthetic combination to avoid further damage to cognitive function of ageing rats with CH during orthopaedic surgery. The underlying mechanism of this phenomenon is related to alleviation of ER stress.

Methods
In our study, a ligation of bilateral CCA surgery [44] was adopted to prepare rats as CH administered intraperitoneally [46]. Surgery was carried out via aseptic techniques. The left hind paw of the rat was shaved and disinfected with iodine tincture. After the skin was incised, a 0.38 mm pin was inserted into the intramedullary canal. Once the tibia was internally fixated, the bone was fractured at the middiaphysis (tibial, midshaft) using surgical pliers. The skin was sutured with 8/0 Prolene sutures. In Group C, only the skin was incised and sutured. During surgery, a heating lamp was used to help maintain the body temperature of the anaesthetized rats at 37 ± 0.5°C. Postintervention rats were moved to heated pads for recovery and then returned to their home cage supplied with sufficient food and water. For post-procedural pain relief, the rats were administered buprenorphine (0.05 mg/kg, subcutaneous) twice daily for 3 days [48].

Fear conditioning test
The FC test was utilized to evaluate cognitive function [38]. The FC test consisted of a training phase at 24 h prior to ORIF surgery and an evaluation phase on days 1 and 7 after ORIF, when hippocampal-dependent memory was assessed.
During the training phase, rats were placed in a chamber (Ugo Basile, Italy) and allowed to adapt to the environment for 120 s. After adaption, a 20 s 70-dB tone (conditional stimulus) was delivered, followed by an interval of 25 s. After the interval, an 0.70 mA electrical foot shock was delivered to the rat for 2 s (unconditional stimulus). After six pairs of conditional-unconditional stimuli, the rats learned the association and had established long-term memory. The pairs of conditional-unconditional stimuli were separated by 60 s inter-training intervals. Each training chamber was cleaned with 95% ethyl alcohol before placement of the next rat and was illuminated only with a 10 W bulb in a dark experimental room.
During the evaluation phase, rats were placed again in the training chamber for 5 min without tone and foot shock. Each animal's freezing behaviour (without any movements) was analysed by using the ANY-Maze Video Tracking System (Stoelting, Illinois, USA). The percentage of time spent exhibiting freezing behaviour was calculated using the formula of 100*f/5 min, where f was the total of freezing time within 5 min. Freezing time measured during exposure to the known context or after a conditional stimulus in the known context reflects hippocampal-dependent memory, whereas assessment during delivery of the conditional stimulus (tone) assesses hippocampal-independent memory [38]. Thus, the results in this experiment were used to assess hippocampus-dependent memory.

Nissl staining
On days 1 and 7 after ORIF, rats (n=8/group) were first anaesthetized with 10% thiobutabarbital (100 ml/kg, i.p.). Rats were perfused with saline before the heart stopped, followed by perfusion with 4% paraformaldehyde solution. Then, the brain was taken out and fixed in 4% paraformaldehyde for 24 h. Coronal slices (3.0-mm thick) from each brain containing the dorsal hippocampus and the medial dorsal prefrontal cortex were dehydrated and embedded in paraffin. A series of 10-μm-thick coronal sections was obtained from each slice, and the sections were stained with cresyl violet [49]. For each brain, five sections at the dorsal hippocampus located at coordinates -3.14 from the bregma to -4.52 from bregma were analysed for Ammon's horn pyramidal cell counts

Statistical analysis
The data were analysed using SPSS 20.0 software (IBM Corp., Armonk, NY, USA). Data are presented as the mean ± standard deviation (SD). Behavioural data were tested using a two-way analysis of variance (ANOVA) with repeated measures. Other data were analysed using a one-way ANOVA with Tukey post hoc comparisons. P < 0.05 was the criterion for statistical significance.

46.
Hu N, Guo D, Wang H, Xie K, Wang C, Li Y, Wang C, Wang C, Yu Y, Wang G: Involvement of the blood-brain barrier opening in cognitive decline in aged rats following orthopedic surgery and high concentration of sevoflurane inhalation. Brain research 2014, 1551:13-24.

47.
Bayne K: Revised Guide for the Care and Use of Laboratory Animals available.

Supplementary Files
This is a list of supplementary files associated with the primary manuscript. Click to download.