Peripheral Transplantation of Mesenchymal Stem Cells at Sepsis Convalescence Improves Cognitive Function of Sepsis Surviving Mice

Objective To investigate the effects of peripheral transplantation of mesenchymal stem cells (MSCs) at sepsis convalescence on post-sepsis cognitive function and underlying mechanisms in mice. Methods Sepsis was induced by cecal ligation and puncture (CLP) in mice. Bone marrow-derived MSCs from mice were cultured and injected via tail vein on the 8th day after CLP. Cognitive function was detected in open field, novel object recognition task, and delayed matching-to-place water maze task during 10-26 days after CLP. Neuroinflammation, neurogenesis, and peripheral inflammation were detected on the 12th and 31th days after CLP. MSCs tracing was detected during 8-10 days after CLP. Results Transplanted MSCs were located at peripheral organs (lung, spleen, liver) and had no obvious effects on survival and weight of sepsis mice. Transplanted MSCs mitigated cognitive impairments and hippocampal microglial activation, improved hippocampal neurogenesis of sepsis surviving mice, and had no obvious effect on the leukocyte amount, the neutrophil percentage, and the inflammatory factors of peripheral blood, and the hippocampal inflammatory factors. Conclusions Our data indicated that MSCs transplantation via peripheral vein at later phase of sepsis can improve post-sepsis cognitive impairment and hippocampal neurogenesis of sepsis surviving mice.


Introduction
Sepsis and its associated complications are one of the foremost causations of death in intensive care units worldwide [1,2]. More than half of septic patients have sepsisassociated encephalopathy (SAE), which remains badly understood and is deemed to be reversible [2,3]. However, septic survivors who had been diagnosed with SAE may have permanent neurocognitive dysfunction, which could result in some serious socioeconomic burden [3][4][5]. Currently, SAE is mainly considered to be a peripheral inflammationinduced brain dysfunction [3]. However, numerous efforts focusing on targeting the inflammatory cytokines network have proved unprofitable for the treatments of sepsis and SAE [6]. The search for novel potential therapeutic targets still needs to be continued.
Stem cell therapy is proved to be beneficial for central nervous system diseases, such as Alzheimer's diseases, Parkinson's diseases, and traumatic brain injury, due to the self-renewal potentialities, multiple differentiation activities, neurotrophic properties, and immuno-regulatory effects of the stem cells [7,8]. A recent study has demonstrated that bone marrow-derived mesenchymal stem cell (BM-MSCs) transplantation in the first 6 hours after sepsis improves the cognitive and behavioral impairments of septic surviving mice [9]. However, if the MSCs treatment will be applied to the sepsis patients at the early phase to prevent cognitive and behavioral alterations, some problems should be taken into account. First, symptomatic support treatment and antiinflammatory treatment are preferred at the early phase of sepsis patients. Second, complicated inflammatory environment is a main feature of sepsis at the early phase [10], which can affect the efficacy of MSCs treatment [11]. In contrast, it may be easy to implement and achieve stable therapeutic effects, if the stem cell therapy is given during sepsis convalescence.
The present study investigated whether peripheral administration of MSCs at the sepsis convalescence could mitigates post-sepsis cognitive function in the wellestablished cecal ligation and puncture (CLP) model mice of sepsis. Toward this end, we further explored the underlying mechanisms of MSCs treatment.

Materials and Methods
2.1. Animals. Adult old C57BL/6 J male mice weighing 25-30 g were obtained from the experimental animal center of Central South University, China. The protocol [LLSC(LA)2017-061] was approved by the ethics committee of the 3rd Xiangya Hospital of Central South University. All mice were housed under a 12 h light-dark cycle, an adequate temperature of 23°C, and a relative humidity of 50%-60% with free access to water and food. Animals were utilized with age-and weight-match and also were treated to minimize their suffering in experiments.

Sepsis
Model. Mice were randomly divided into Sham +NS, Sham+MSC, CLP + NS, and CLP + MSC groups. The last two groups received CLP surgery, while the former two groups did not. CLP surgery was performed as previously described [12,13]. Briefly, animals were anesthetized with the 2% sevoflurane (Maruishi Pharmaceutical Co., Ltd., Japan) in a well-ventilated room. 40% of the cecum was ligated and punctured once with a 20-gauge needle. The Sham+NS and Sham+MSC groups underwent only identical laparotomy but without CLP. Bupivacaine (3 mg/kg) and Buprenorphine (0.1 mg/kg) were injected subcutaneously once to avoid postoperative pain. All animals were given 1 mL of saline subcutaneously every 8 hours for 5 days.
2.3. MSC Culture and Delivery. Bone marrow-derived MSCs (Cyagen Biosciences Inc, Guangzhou, China) from the C57BL/6 J mice were cultured in an incubator (Thermo, USA) under an appropriately maintained temperature of 37°C and an atmosphere of 5% CO 2 . Culture medium (MUBMX-90011, Cyagen) was subsequently renewed every 2 days after removing nonadherent cells on day 2. When the cells population density reached 80%-90%, it was digested by 0.25% trypsin and subcultured according to the density of 2.5~4:0 × 10 4 /cm 2 . The mice in CLP + MSC and Sham+MSC groups were injected with MSCs (1 × 10 6 diluted in 200 μl of physiological saline) within twelve passages via tail vein on the 8 th day after CLP sur-gery, while the other groups were admitted equal volume of normal saline.

MSC Labeling and
Tracing. The MSCs were directly labeled via incubating in 1 μM Lipophilic Tracers DiR (Invitrogen, USA) for 30 minutes at 37°C [14] . MSCs then were washed twice with new culture medium for 10 minutes each time to remove residual DiR solution. Mice were injected with 3 × 10 6 of DiR labeled MSCs (1 × 10 6 diluted in 200 μl of normal saline) via tail vein on the 8 th day after surgery. These mice were sacrificed at 6 h, 24 h, or 72 h after injection, and the organs (brain, lung, kidney, spleen, liver) were dissected for fluorescent signal by IVIS Lumina II (PerkinElmer, USA). The fluorescence intensity was quantified with ImageJ in a blind manner.
2.5. Behavior Test. Mice were subjected to open field, novel object recognition, and delayed matching-to-place tasks on postoperative day 10-26 as shown in Figure 1(a). Before the test began, mice were placed in a sound-isolated behavioral testing room for half an hour. All tests were carried out by the same person who was blind to the animal's group.

Open Field Test.
Open field test was used to assess locomotive activities. Each mouse was gently placed in the center of apparatus (width:50 cm, length:50 cm, height:38 cm) and was left alone to explore the arena for 5 min. At the end, the mouse was immediately taken back to its home cage. Using 75% alcohol solution to clean the apparatus after the fecal and urine were removed left by the previous mouse. The total ambulatory distances and time in central zone were analyzed.

Novel Object Recognition (NOR) Test.
Object recognition experiment was conducted to test hippocampusrelated learning and memory described in previously studies [15]. It was performed in a 30 cm ×30 cm ×38 cm open-field apparatus. The test included training and testing phases. In the training phase, two identical objects were parallelized equidistant from the center of arena, and equidistant from the arena walls. The mice were gently placed in the arena keeping their heads away from the two identical objects. Then, the mice were permitted to freely explore the objects for 10 min. 24 hours later, one of the familiar objects was replaced with a novel object, and animals were also allowed to freely explore for 10 min. The apparatus and objects were cleaned with 75% alcohol solution between trials to remove the remaining scent. The behaviors were recorded using a digital camera. The preference index (defined as novel object investigation time/(novel object investigation time + familiar object investigation time)) was used to assess the learning and memory.

Delayed
Matching-to-Place (DMP) Task. The DMP water maze task is used to evaluate working memory [16][17][18]. Following the reported protocol [16][17][18], mice were pretrained using four trials per day for 5 days. After pre-training, the mice were given the testing tasks using memory intervals (intertrial intervals between trials 1 and 2) of 5 sec, 20 min, and 2 h. Each testing phase period was 3 days. Testing phase 2 Oxidative Medicine and Cellular Longevity  3 Oxidative Medicine and Cellular Longevity performance was recorded by an overhead video camera (Logitech, China) connected to video recorder and a computer running custom-written Smart 3.0 software, and was calculated by subtracting the trial 2 time/path-length for each mouse from its trial 1 time/path-length. Greater time/pathlength differences indicated better performance.
The photographs were acquired under a microscope (Nikon, Tokyo, Japan) and LSM800 confocal microscope (Carl Zeiss, Jane, Germany). Based on the Iba-1 staining (n = 12 slices from three mice for each group), the number of Iba-1 positive (Iba-1 + ) microglia and the percent of activated microglia in the CA1 and dentate gyrus (DG) were counted and analyzed as described by our previous study [19]. According to the report, resting microglia was defined as when the cell body was small and round and the branches were thin, highly ramified, and equally distributed around the cell body. In contrast, activated microglia was defined as when the cell body was bigger, pleomorphic bi-or tripolar, or spindle/rod-shaped, and the branches were shortened, twisted or displayed no ramification [19]. For the  2.10. Statistical Analysis. All data were analyzed with Prism 9 (Graphpad Software Inc., La Jolla, CA, USA), and were presented as mean ± standard error (SEM). Log-rank (Mantel-Cox) test was used to analyze the factors of the survival time for any significant differences. The body weight changes and DMP task performance were analyzed with two-way ANOVA followed by multiple comparison tests. The oneway ANOVA followed by multiple comparison tests was used for other results. A significant level was set at p < 0:05.

Transplanted MSCs Were Located at Peripheral Organs and Had no Obvious Side Effects.
To determine the safety of MSCs transplantation at later phase of sepsis, we detected the survival and weight of sepsis mice (Figure 1(a). Compared to the Sham+NS and Sham+MSC groups, the survival rate and weight of CLP + NS group decrease significantly (Figures 1(b) and 1(c)). There were no significant difference of survival rate and weight between CLP + NS and CLP + MSC groups (Figures 1(b) and 1(c)). These suggested no obvious side effects of MSCs transplantation at later phase of sepsis. To trace the transplanted MSCs with DiR marker, we detected the fluorescent signal of the organs (brain, lung, kidney, spleen, liver) at 6 h, 24 h, and 72 h after tail vein injection. Fluorescent signal was mainly detected in lung, spleen, and liver. (Figure 1(d), suggesting that MSCs were mainly located at peripheral organs.

MSCs Transplantation Mitigated
Cognitive Impairments of Sepsis Surviving Mice. We next investigated whether MSCs transplantation at later phase of sepsis could improve cognitive function of sepsis surviving mice via open field test, novel object recognition test and delayed matching-toplace (DMP) water maze task (Figure 1(a). There was no significant difference between four groups in open field test (Figure 2(a). In the novel object recognition test, the preference index of the CLP + NS group was significantly less than that of Sham+NS, Sham+MSC, and CLP + MSC groups (Figure 2(b). DMP task was used to evaluate working memory which began at 13 days after CLP surgery. In this task, mice were pre-trained for 5 days and then, tested using four sessions with intertrial intervals of 5 sec, 20 min, or 2 h between trials 1 and 2. Greater latency and path-length saving between trial 1 and trial 2 indicated better performance. The latency and path length saving of CLP + NS group at 5 s and 20 min intertrial intervals (ITI) were significantly less than that of Sham+NS, Sham+MSC, and CLP + MSC groups (Figure 2(c). This above information supported an improvement of cognitive function of sepsis surviving mice with MSCs transplantation at later phase of sepsis.

MSCs Transplantation Improved Hippocampal
Neurogenesis of Sepsis Surviving Mice. Neuroplasticity is   of open field test on the 10 th day after CLP surgery (One-way ANOVA followed by multiple comparison tests, N = 10 for each group); (b) preference index for novel object in novel object recognition task on the 11 th and 12 th days after CLP surgery (One-way ANOVA followed by multiple comparison tests, N = 10 for each group); and (c) the performances of delayed matching-to-place (DMP) task in a water maze from the 13 th to 26 th days after CLP surgery. Greater latency and path length saving between trial 1 and trial 2 with intertrial intervals (ITI) of 5 sec, 20 min, or 2 h indicated better performance (two-way ANOVA followed by multiple comparison tests, N = 10 for each group). The data were expressed as mean ± SEM; * P < 0 · 05, * * P < 0 · 01, * * * P < 0 · 001.Sham+NS: mice received laparotomy and tail vein injection of normal saline. Sham+MSC: mice received laparotomy and tail vein injection of MSCs; CLP + NS: mice received CLP surgery and tail vein injection of normal saline; CLP + MSC: mice received CLP surgery and tail vein injection of MSCs. 6 Oxidative Medicine and Cellular Longevity   Figure S1). In contrast, no significant changes of hippocampal PSD-95 and synaptophysin were detected between Sham+NS, Sham+MSC, CLP+ NS, and CLP + MSC groups (Figure 3(c)).

MSCs Transplantation Alleviated Microglial
Activation of the Hippocampus of Sepsis Surviving Mice. Neuroinflammation plays an important role in pathogenesis of postsepsis cognitive dysfunction [11]. Thus, we detected microglial activation and inflammatory factor levels (IL-1, IL-6 and TNF-α) in the hippocampus. The percentages of activated microglial cells in CLP + NS group on the 12 th and 31 th days after CLP surgery significantly increased, relative to the Sham+NS, Sham+MSC, and CLP + MSC groups (Figures 4(a) and 4(b)). However, we did not observe differences in the mRNA levels of IL-1, IL-6, and TNF-α expressed in the the hippocampus among each group on the 12 th day after CLP surgery (Figure 4(c).

MSCs Transplantation Did Not Affect Peripheral
Inflammation Level of Sepsis Surviving Mice. Compared to the Sham+NS, Sham+MSC groups, the leukocyte amount and the neutrophil percentage in the peripheral blood of CLP + NS group significantly increased on the 12 th day after surgery, but the HMGB1 level did not ( Figure 5(a) and 5(b)). There were no significant differences between CLP + NS group and CLP + MSC group in the leukocyte amount, the neutrophil percentage and the HMGB1 level of the peripheral blood on the 12 th day after surgery (Figure 5(a) and (b)).

Discussion
Post-sepsis cognitive impairment as a common post-sepsis sequela is influentially associated with discount life quality and decreased life independence in sepsis survivors [4,5].
Our study tried to explore whether MSCs transplantation via tail vein at the later phase of sepsis could improve the post-sepsis cognitive function in sepsis surviving mice. We found that sepsis mice with MSCs transplantation at the later phase of sepsis showed better long-term memory and working memory in novel object recognition test and delayed matching-to-place (DMP) water maze task, respectively. No obvious side effects marked by weight and survival rate were detected. This was in line with previous study about MSCs transplantation at the early phase of sepsis [9]. They found that MSCs transplantation at the first 6 hours after sepsis was also protective for the cognitive function of septic surviving mice. These data may suggest that MSCs transplantation via peripheral vein is a simple, safe, and effective intervention for the prevention and treatment of post-sepsis cognitive impairment, although its impact on long-term outcomes or survival warrants further study. Neuroinflammation plays an important role in pathogenesis of post-sepsis cognitive dysfunction [10,22,23]. Limiting neuroinflammation and reverting microcirculatory dysfunction improved post-sepsis cognitive function by the statins in experimental models, although they had been unprofitable in a randomized controlled clinic trial [24,25]. In the study reported by Silva et al. [9], MSCs transplantation at the first 6 hours after sepsis improved post-sepsis cognitive function, corresponding to the inhibition of The HMGB1 protein level in peripheral blood on the 12 th day after CLP surgery (one-way ANOVA followed by multiple comparison tests, N = 6 per group). The data were expressed as the mean ± SEM; * P < 0 · 05, * * P < 0 · 01, * * * P < 0 · 001. Sham+NS: mice received laparotomy and tail vein injection of normal saline; Sham+MSC: mice received laparotomy and tail vein injection of MSCs; CLP + NS: mice received CLP surgery and tail vein injection of normal saline; CLP + MSC: mice received CLP surgery and tail vein injection of MSCs. Oxidative Medicine and Cellular Longevity neuroinflammation and peripheral inflammation (marked by the levels of IL-1β, IL-6, and MCP1 in plasma) in mice. Thus, in this study, we also evaluated the peripheral inflammation with leukocyte number, neutrophil percentages, and HMGb1 level (a critical molecule in pathogenesis of sepsis [26]) in peripheral blood, and the neuroinflammation with microglial activation and inflammatory factors (IL-1,IL-6, and TNF-α) in the hippocampus. However, we found, MSCs transplantation at the later phase of sepsis had no significant effects on the peripheral inflammation and the level of hippocampal inflammatory factors, suggesting that antiinflammation was not the underlying protective mechanism of MSCs transplantation at the later phase of sepsis. Interestingly, we found that MSCs transplantation at the later phase of sepsis significantly improved the neurogenesis impairment of septic surviving mice. These suggest that the protective mechanism of transplanted MSCs to post-sepsis cognitive function is sepsis-phase dependent. In addition, we did not trace transplanted MSCs in brain, but traced them in peripheral organs. Previous studies had shown that MSCs transplanted in the hippocampus could increase hippocampal neurogenesis in immunodeficient mice [27,28]. Transplanted MSCs could promote tissue recovery and regeneration via secreting lots of specific mediators, conferring immunomodulatory, anti-inflammatory, antimicrobial, angiogenic, antifibrotic, and structural reparative properties [9,18]. Notably, Islam et al. revealed that MSCs protected against acute lung injury via transferring its mitochondria to pulmonary alveoli [29]. This provides a further explanation for the protective mechanism of peripherally transplanted MSCs to central nervous system connected cognitive function in our study. However, the intermediators need further investigation.

Conclusion
In conclusion, our results demonstrated that MSCs transplantation via peripheral vein at later phase of sepsis could be considered as a 'proof of concept' for the prevention and treatment of post-sepsis cognitive impairment ( Figure 6). The effectiveness of this therapy and the underlying mechanism needed to be evaluated further in the future trial(s) and model animals.

Data Availability
The data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
The authors declare that they have no competing interests.  Figure 6: The graphical abstract of this study. MSCs transplantation can improve post-sepsis cognitive impairment and hippocampal neurogenesis of sepsis surviving mice via peripheral vein on the eighth day after CLP surgery.