Combined Human Umbilical Cord-Derived Stem Cell and Hyperbaric Oxygen Therapy Effectively Preserved the Neurological Function and Brain Architecture in Rat After Acute Intracerebral Hemorrhage

Background: This study tested the hypothesis that combined therapy with human umbilical cord-derived mesenchymal stem cells (HUCDMSCs) and hyperbaric oxygen (HBO) was superior to either one on preserving neurological function and reducing brain-hemorrhagic volume (BHV) in rat after acute intracerebral hemorrhage (ICH) induced by intracranial injection of collagenase. Methods and Results: Adult-male SD rats (n=30) were equally divided into group 1 (sham-operated control), group 2 (ICH), group 3 (ICH + HUCDMSCs/1.2x10 6 cells/intravenous injection at 3h and days 1/2 after ICH), group 4 (ICH + HBO/at 3h and days 1/2 after ICH) and group 5 (ICH + HUCDMSCs-HBO), and euthanized by day 28 after ICH. By day 1, the neurological function was signicantly impaired in groups 2 to 5 than in group 1 (p<0.001), but it did not differ among groups 2 to 5. By days 7/4/28, the integrity of neurological function was highest in group 1, lowest in group 2 and signicantly progressively improved from groups 3 to 5 (all p<0.001). By day 28, the BHV was lowest in group 1, highest in group 2 and signicantly lower in group 5 than in groups 3/4 (all p<0.0001). The protein expressions of inammation (HMGB1/TLR-2/TLR-4/MyD88/TRAF6/p-NF-κB/IFN-γ/IL-1ß/TNF-α), oxidative-stress/autophagy (NOX-1/NOX-2/oxidized protein/ratio of LC3BII/LC3BI) and apoptosis (cleaved-capspase3/PARP), and cellular expressions of inammation (CD14+, F4/80+) in brain tissues exhibited an identical pattern, whereas cellular levels of angiogenesis (CD31+/vWF+/small-vessel number) and number of neurons (NeuN+)

which are complicated could explain why the effective treatment of intracranial hemorrhage is still unmet need.
Hypoxic neurons performing anaerobic metabolism result in acidosis, lactate production, and an unsustainable reduction in cellular metabolic reserve [16]. As the hypoxic microenvironment persists, the neuronal cells lose their ability to maintain ionic homeostasis, follow by free oxygen radical accumulation and cell membrane degradation [16], resulting in an irreversible cell death [ [16]. This gives some basis to the assertion that therapy designed to increase oxygen availability in the early period following brain hemorrhage may improve long-term outcomes [17]. Interestingly, hyperbaric oxygen therapy (HBO) has been reported to be effective in improving the outcomes after traumatic brain injury [17]. Other study has reviewed that HBO could be a new look on treating ischemic stroke and traumatic brain injury [18]. HBO is the therapeutic administration of 100% oxygen at environmental pressures greater than 1 atmosphere absolute (ATA) in an airtight vessel. In this way, it is possible to deliver a greatly increased partial pressure of oxygen to the tissue, including the ischemic brain tissues.
Experimental studies have further demonstrated that MSCs therapy markedly preserved neurological function and reduced brain infarct volume in rodent after acute ischemic stroke attack [27,28]. The results of these previous reports [19,[21][22][23][24][25][26][27][28] raise the hypothesis that MSCs therapy may offer unexpected bene t for patients after ICH, especially for those who have large ICH zone non-candidate for surgical intervention. However, prior to apply MSC therapy for human being in a clinical setting of ICH, a preclinical study had to be rst performed by using human being derived MSC (i.e., xenogeneic MSC) to prove not only the safety and e cacy but also the immune privilege of the MSC regardless for what kind of biological species. Furthermore, due to the complexity of the pathophysiologic mechanisms involved in the hemorrhagic stroke, a single therapeutic strategy would be inadequate for disease improvement.
Therefore, we performed a combined therapy with human umbilical cord blood-derived stem cell (HUCDMSCs) and HBO for rodent after acute ICH. Procedure and protocol of animal model of intracerebral hemorrhage (ICH) induced by type IV collagenase proteolytic enzyme

Ethics
The procedure and protocol of ICH was based on previous report [29]. In details, the rats were anesthetized by inhalation of 2.0% iso urane and then were placed on a warming pad at 37 °C, followed by securing the head and shaving the scalp hair. Under sterile conditions, 1 cm long midline incision of the scalp was carefully created to expose the perpendicular intersection point of the coronal and sagittal suture (i.e., bregma). The Hamilton syringe (250 µl) was then mounted onto the injection pump and stereotaxically guided the needle (26 Gauge) over bregma. The stereotactic manipulator arms were adjusted the position of the needle 1.4 mm anterior and 3.2 mm lateral to the right. A small cranial burr hole was then created by using a 1 mm drill bit. After this procedure, 1.0 µl collagenase type IV (0.25 IU/ µl) was then carefully injected into corpus/dorsal striatum (5 mm below the skull) by Hamilton syringe 26 G at a rate of 0.2 µl/min. The syringe was removed slowly after the injection is completed and sterile bone wax is used to plug the hole quickly. The skin on the surface of head was then closed by using 4 − 0 prolene suture. Finally, the animals were cared for in a portable animal intensive care unit (ThermoCare®) with food and water for 24 hours.

Animal grouping
Pathogen-free, adult male Sprague-Dawley (SD) rats (n = 30) weighing 320-350 g (Charles River Technology, BioLASCO Taiwan Co. Ltd., Taiwan) were used in the present study. The animals were equally categorized into group 1 [sham-operated control (SC), i.e., only incision of the skin over scalp], group 2 [acute ICH induced by directly intracranial injection of collagenase (1.0 µl)], group 3 [ICH + HUCDMSCs (1.2 × 10 6 cells) by intravenous administration at 3 h and on days 1 and 2, respectively after ICH procedure, i.e., total amount was 3.6 × 10 6 cells], group 4 [ICH + hyperbaric oxygen therapy (HBO) (3 h duration for each time) at 3 h and at days 1 to 4 after ICH procedure] and group 5 (ICH + combined HUCDMSCs + HBO). Animals in each group were euthanized by day 28 after ICH induction and the brain specimen was harvested from each animal for individual study.

Hyperbaric oxygen therapy
The procedure and protocol of HBO therapy was based on a previous report [30]. Brie y, to induce tissuelevel hyperoxia, SD rats were subjected to HBO administration in an animal tabletop chamber (Piersol-Dive, model 4934) with the animals exposed to 100% oxygen at 2.4 atmospheres absolute (ATA) for 90 minutes (3 h/one time) at 3 h and days 1 to 4 after ICH induction.
Corner test for assessment of neurological function prior to and after ICH induction The sensorimotor functional test (corner test) was conducted for each rat of each group (i.e., n = 6 per group) at baseline and on days 1, 7, 14 and 28 after acute ICH induction as we previously described [28,30,31]. In detail, the rat could walk through a tunnel and then turn into a 60-degree corner. To exit the corner, the rat could turn either left or right. The results were recorded by a technician blinded to the study design. This test was repeated 10 to 15 times with at least 30 seconds between each trial. We recorded the number of right and left turns from 10 successful trials for each animal and used the results for statistical analysis.

Measurement of brain hemorrhagic area
To evaluate the impact of HUCDMSCs-HBO treatment on preserving the brain parenchyma, coronal sections of the brain were obtained from four extra animals in each group as 2 mm slices by day 14 after ICH induction. Each cross section of brain tissue was then stained with 2% 3,5-Triphenyl-2H-Tetrazolium Chloride (TTC) (Alfa Aesar) for BIA analysis. Brie y, all brain sections were placed on a tray with a scaled vertical bar to which a digital camera was attached. The sections were photographed from directly above at a xed height. The images obtained were then be analyzed using Image Tool 3 (IT3) image analysis software (University of Texas, Health Science Center, San Antonio, UTHSCSA; Image Tool for Windows, Version 3.0, USA).
The hemorrhagic area was observed as either whitish or pale reddish regions (i.e., dis-colored region). Intracerebral hemorrhagic region was further con rmed by microscopic examination. The percentages of hemorrhagic area were then obtained by dividing the area with total cross-sectional area of the brain. The rest of the brain tissue was then cut into pieces for speci c studies.
Brie y, equal amounts (50 µg) of protein extracts were loaded and separated by SDS-PAGE using acrylamide gradients. After electrophoresis, the separated proteins were transferred electrophoretically to a polyvinylidene di uoride membrane (GE, UK). Nonspeci c sites were blocked by incubation of the hour at room temperature. Horseradish peroxidase-conjugated anti-rabbit immunoglobulin IgG (1:2000, Cell Signaling, Danvers, MA, USA) was used as a secondary antibody for one-hour incubation at room temperature. The washing procedure was repeated eight times within one hour. Immunoreactive bands were visualized by enhanced chemiluminescence (ECL; Amersham Biosciences, Amersham, UK) and exposed to Biomax L lm (Kodak, Rochester, NY, USA). For the purpose of quanti cation, ECL signals were digitized using Labwork software (UVP, Waltham, MA, USA).

Immuno uorescent (IF) staining of brain specimens
The procedure and protocol of IF staining were based on our previous reports [28,[30][31][32][33]. In detail, frozen sections (4 µm thick) was obtained from the brain hemorrhagic area/at risk area of each animal, permeated with 0.5% Triton X-100, and incubated with antibodies against NueN (1:100, Merck), CD31 Fluor594-conjugated goat anti-mouse or rabbit IgG was used to localize signals. Sections were nally counterstained with DAPI and observed with a uorescent microscope equipped with epi uorescence (Olympus IX-40). Three brain sections were analyzed for each rat. For quanti cation, three randomly selected high-power elds (HPFs; 400x for IF study) were analyzed in each section. The mean number of positively-stained cells per HPF for each animal was then be determined by summation of all numbers divided by 9.
Vessel density in brain infarct zone The procedure and protocol for identifying number of small vessels in the brain hemorrhagic area/at risk area were based on our previous reports [31][32][33]. In detail, staining of small blood vessels was performed with alpha smooth muscle actin (α-SMA) (1:400) as primary antibody at room temperature for 1 hour, followed by washing with PBS thrice. Ten minutes after the addition of anti-mouse-HRP conjugated secondary antibody, the tissue sections were washed with PBS thrice. Then 3,3' diaminobenzidine (DAB) (0.7 gm/tablet) (Sigma) will be added, followed by washing with PBS thrice after one minute. Finally, hematoxylin was added as a counter-stain for nuclei, followed by washing twice with PBS after one minute. Three brain sections were analyzed in each rat. For quanti cation, three randomly selected HPFs (100x) were analyzed in each section. The mean number per HPF for each animal was then be determined by summation of all numbers divided by 9.
Procedure and protocol of brain magnetic resonance imaging (MRI) for determining the brain hemorrhagic volume (BHV) The procedure and protocol for brain magnetic resonance imaging (MRI) study were based on our previous report [28]. The MRI was performed at day 28 after IS induction. Brie y, during MRI measurements, mice were anesthetized by 2% inhalational iso urane with room air and placed in an MRIcompatible holder (Biospec 94/20, Bruker, Ettingen, Germany). Rectal temperature and respiration were monitored throughout the procedure to ensure normal physiological conditions were maintained. MRI data were collected using a Varian 9.4T animal scanner (Biospec 94/20, Bruker, Ettingen, Germany) with a rat surface array. The MRI protocol consisted of 40 T2-weighted images. Forty continuous slice locations were imaged with a eld-of-view of 30 mm x 30 mm, an acquisition matrix dimension of 256 × 256 and slice thickness of 0.5 mm. The repetition time (TR) and echo time (TE) for each fast spin-echo volume were 4200 ms and 30 ms, respectively. Custom software, ImageJ (1.43i, NIH, USA), was used to process the region of interest (ROI). Planimetric measurements of images from MRI T2 were performed to calculate the stroke volumes of cortex. Collectively, the BHV was calculated by summation of total coronal sections and then divided the numbers of coronal sections to obtain the means of infarct areas. Additionally, the height of the infarct zone was calculated by summation of the thickness of each coronal sections. Finally, the BHV was obtained by mean of infarct area x height.

Statistical analysis
Quantitative data were expressed as means ± SD. Statistical analysis was adequately performed by ANOVA, followed by Bonferroni multiple-comparison post hoc test. SAS statistical software for Windows version 8.2 (SAS institute, Cary, NC) was utilized. A P value of less than 0.05 was considered statistically signi cant.

Results
Identi cation of hemorrhagic area by day 14 after ICH induction and time courses of neurological function assessed by corner test (Figure 1) First, to assess the early stage of intracerebral hemorrhagic zone, the whole brain cross section was stained by TTC by day 14 after ICH induction. The result demonstrated that the intracranial hemorrhagic area was lowest in group 1 (SC), highest in group 2 (ICH only) and signi cantly lower in group 5 (ICH + HUCDMSC + HBO) than in group 3 (ICH + HUCDMSC) and 4 (ICH + HBO), but it showed no different between the groups 3 and 4 ( Fig. 1-C to 1-H).
Next, to elucidate the therapeutic impact of HUCDMSC-HBO on preservation of the neurological function (i.e., by corner test), the time courses of corner test ( Fig. 1-I) were performed for the animals in each group. The results showed that by day 0 prior to ICH induction, the neurological function did not differ among the ve groups ( Fig. 1-J). However, by day 1, the neurological function was signi cantly impaired in groups 2 to 5 than in group 1, but it showed no different among the groups 2 to 5 ( Fig. 1-K). However, by days 7, 14 and 28, the neurological function was signi cantly impaired in group 2 than in group 1, while signi cantly progressively improved in groups 3 and 4 and further improved in group 5 ( Fig. 1-L to 1-N).
Brain hemorrhagic volume and expression of NeuN+ cells in brain hemorrhagic zone by day 28 after ICH induction ( Figure 2) To further evaluate the anatomical integrity of brain parenchyma, we utilized the brain MRI instrument. As we expected, by day 28 the brain MRI demonstrated that the BHV (Fig. 2-A to 2-E) was lowest in group 1, highest in group 2 and signi cantly lower in group 5 than in groups 3 and 4 ( Fig. 2-F). However, this parameter did not differ between groups 3 and 4 ( Fig. 2-F).
To realize the ultrastructure of the brain architecture in these animals, the high manifested IF microscopic analysis was done in the present study. The result revealed that number of NeuN+ cells (Fig. 2-G to 2-K), an indicator of neurons, was highest in group 1, lowest in group 2 and signi cantly higher in group 5 than in groups 3 and 4 but it showed no difference between groups 3 and 4 ( Fig. 2-L).
Impact of HUCDMSCs and HBO therapies on upregulating the angiogenesis capacity in brain parenchyma by day 28 after IHC induction (Figures 3, 4) By using the IF microscopic examination, we further determined the angiogenesis ability of HUCDMSC-HBO therapy in ischemic zone. The result showed that the expression of CXCR4+ cells (Fig. 3-A to 3-E), an index of endothelial progenitor cells that play a crucial role for angiogenesis, was notably progressively increased from groups 1 to 5, implicating an intrinsic response to ischemic stimulation that could be more augmented by HUCDMSC-HBO therapy (Fig. 3-F). Additionally, the α-SMA stain demonstrated that the number of small vessels (i.e., de ned as the diameter ≤ 25 μM) (Fig. 3-G to 3-K), an indicator of angiogenesis/neovascularization, was highest in group 1, lowest in group 2 and signi cantly higher in group 5 than in groups 3 and 4 but it did not differ between groups 3 and 4 ( Fig. 3-L).
On the next stop, we also utilized the IF microscopic examination to identify the expressions of endothelial-cell surface biomarkers of CD31 (Fig. 4-A to 4-E) and vWF (Fig. 4-G to 4-K), an integrity of endothelial cells. The result showed that these cellular expressions exhibited an identical pattern of number of small vessels among the ve groups ( Fig. 4-F, Fig. 4-L).
Cellular and molecular level of in ammatory downstream signalingby day 28 after IHC induction ( Figures  5, 6, 7) It is well recognized that tissue ischemia and necrosis would always induce in ammatory reactions. Thus, we evaluated the protein and cellular level of in ammatory responses. As expected, the cellular expressions of CD14 (Fig. 5-A to 5-E) and F4/80 (Fig. 5-G to 5-K) in brain tissue, two indicators of in ammation, were highest in group 1, lowest in group 1, signi cantly lower in group 5 than in groups 3 and 4 and signi cantly lower in group 3 than in group 4.
We then returned to see the molecular level of in ammation. Consistently, the Western blot analysis showed that the protein expressions of HMGB1, TLR-2, TLR-4, MyD88 and TRAF6 (Fig. 6-A to 6-E), ve indices of upstream in ammatory signaling, displayed a similar pattern of CD14+ cells among the ve groups. Additionally, the protein expressions of p-NF-κB, IFN-γ, IL-1ß and TNF-α (Fig. 7-B to 7-E), four indicators of downstream in ammatory signaling, also exhibited a similar pattern of CD14+ cell, whereas the protein expression of IKB-ß (Fig. 7-A), another indicator of downstream signaling, displayed an opposite pattern of CD14+ cells among the ve groups.
Protein level of oxidative stress and autophagic biomarkers by day 28 after IHC induction (Figure 8) It is also well recognized that tissue ischemia and necrosis always elicit oxidative stress which, in turn, induces apoptosis and upregulated the autophagy. Based on this concept, we performed Western blot analysis. The result demonstrated that the protein expressions of NOX-1, NOX-2 and oxidized protein (Fig.  3-A to 3-C), three indices of oxidative stress, were lowest in group 1, highest in group 2 and signi cantly lower in group 5 than in groups 3 and 4, but they showed no difference between groups 3 and 4. Additionally, the protein expressions of cleaved caspase 3 ( Fig. 3-D), and indicator of apoptosis, was lowest in group 1, highest in group 2, signi cantly lower in group 5 than in groups 3 and 4 and signi cantly lower in group 3 than in group 4. On the other hand, the cleaved PARP (Fig. 3-E), another indicator of apoptosis and the ratio of LC3B-II/LC3B-I (Fig. 3-F), an indicator of autophagic biomarker, exhibited an identical pattern of oxidative stress among the ve groups.

Discussion
This study which investigated the therapeutic impact of HUCDMSC-HBO on ICH rodent yielded several preclinical implications. First, HUCDMSC or HBO therapy substantially reduced the brain hemorrhagic area (i.e., estimated by TCC stain at early stage of ICH) and brain hemorrhagic volume (i.e., estimated by brain MRI at day 28 after ICH) and remarkably preserved neurological function (i.e., estimated by corner test). Second, the extensive woks were done in the present study and found that in ammatory signaling and oxidative stress could be the mainly underlying mechanism for brain damage after ICH induction.
Third, therapeutic impact of HBO was comparable with HUCDMSC for protecting the brain architecture against hemorrhagic damage. Finally, the combined HUCDMSC-HBO therapy was superior to either one alone on protecting the brain volume and neurological function in setting of ICH.
The effective treatment of severe ICH is still lacking. This could explain for why this disease entity is still a high cause of death and disability in adults worldwide [1][2][3]. Interestingly, our recent studies [27,28] have shown that exogenous MSCs treatment for rodent IS or brain hemorrhage was safe and without immune rejection. Of importance was that these MSCs therapy effectively protected the brain from ischemic related/hemorrhagic relayed injury [27,28]. Additionally, our other recent studies have shown that HBO therapy effectively preserved the brain infarct volume and neurological function in rat after IS [30,34]. The most important nding in the present study was that as compared with ICH group, the brain hemorrhagic volume and brain hemorrhagic zone were signi cantly preserved in IHC animals treated by HUCDMSCs and HBO. Of distinctive importance was that combined HUCDMSCs and HBO offer additional bene t than single one therapy alone on preserving the brain architecture and neurological integrity after ICH attack. Our ndings, in addition to strengthening the ndings of previous studies [27,28,30,34], could, at least in part, explain for why the neurological function of ICH animals was greatly improved with HUCDMSC-HBO therapy.
It is well documented that the mechanical damage is always immediately elicited after ICH which commonly wrecks the brain tissue as it enlarges [2]. This, is frequently associated with the pressure created enough by blood and surrounding brain edema, resulting in a life-threatening situation [4]. The secondary damage is caused by cytotoxicity of blood [5,6], impaired calcium homeostasis [7], excitotoxicity from excitatory neurotransmitters/glial cells [8,9] and upregulation of oxidative stress and in ammation [6,[9][10][11][12][13][14]. Additionally, an association between the upregulation of in ammation and oxidative stress and organ damage in setting of ischemia or ischemia-reperfusion injury have been extensively investigated by previous studies [19][20][21][22][23][24]. An essential nding in the present study was that as compared with SC group, the brain hemorrhagic volume (i.e., caused mechanical damage) and the in ammatory downstream signaling and oxidative stress (i.e., caused by cellular and molecular perturbations) were remarkably upregulated in ICH animals (refer to Fig. 9). Our ndings, in addition to being consistent with the ndings of the previous studies [2, 4-7, 9-14, 19-24], could explain for why much higher apoptosis autophagic phenomenon and the unfavorable outcomes developed in the IHC animals, and suggest that in ammatory signaling/oxidative stress plaid a ultimately crucial role on damage (i.e., second damage) of the brain architecture and neuron integrity after ICH.
The mechanisms for why MSCs and HBO therapy effectively protect the cells/tissues and organs from any ischemia-related damage always elicit high interest from scientists [35,36]. Currently, abundant data have shown that anti-in ammation, immunomodulation and tissue regeneration as well as cytokine/paracrine effects could be the main mechanisms for MSCs therapy on safeguarding the cells/tissues and organs far away from in ammatory and ischemic related damage [27,28,36]. Our ndings demonstrated even xenogenic HUCDMSCs (i.e., from human derived MSC to rodent) was still safe without any immune reaction and offered additional bene t to IHC rats (refer to Fig. 9). In this way, our ndings corroborated with the ndings of previous studies [27,28,36]. On the other hand, the mechanistic basis of HBO therapy on protecting the ischemic cells/tissues and organs from ischemic damage has been emphasized mainly due to upregulation of circulating levels of endothelial progenitor cells/MSCs, an increase in vascular permeability, high content of oxygen in ischemic tissues and antiapoptotic and anti-in ammatory effects [27,28,36]. These ndings [27,28,36] could explain for why a synergic effect of combined HUCDMSC + HBO therapy on reducing the brain hemorrhagic volume and improving the neurological function to be identi ed in the present study.
We proposed the therapeutic regimen of the multiple dosages of HUCDMSCs (i.e., like the regimen of antibiotics to be utilized for treatment the bacterial-induced sepsis in our daily clinical practice) could be the key to success for remarkable preservation the brain architecture and neurological function. Interestingly, our previous [36] and recent [37] studies have shown that early and multiple dosages of stem cell administration was the crucial role on reducing the post-heart transplant acute rejection [37] and preservation of the neurological function after acute ischemic stroke [38] in rodent. In this way, the results of the present study, in addition to being comparable with the ndings of our previous [36] and recent [37] studies, highlight that this regimen could be seriously considered in our future stem cell therapy for the patients.

Study limitations
This study has limitations. Despite a lot of works had been done in the present study, the exact mechanisms of combined HUCDMSCs and HBO on preserving the brain architecture and neurological function were still not fully investigated. Based on the results of the present study, Fig. 9 illustrated the proposed mechanism for how the HUCDMSC + HBO therapy on improving the outcomes after IHC in rodent. Second, although the short-term outcomes (i.e., at a study period of 28 days) were attractive and promising, the long-term impact of such kind of therapy remains uncertain.

Conclusions
In conclusion, the results of the present study demonstrated that combined HUCDMSC and HBO therapy offered a synergic effect on protecting the macro-and micro-ultrastructure of the brain and preserving neurological integrity in rat after ICH.    Analytical result of number of vWF+ cells, * vs. other groups with different symbols ( †, ‡, §), p<0.0001. Scale bar in right lower corner represent 20µm. All statistical analyses were performed by one-way ANOVA, followed by Bonferroni multiple comparison post hoc test (n=6 for each group). Symbols (*, †, ‡, §) indicate signi cance (at 0.05 level). SC = sham-operated control; ICH = intracerebral hemorrhage; HUCDMSC = human umbilical cord-derived mesenchymal stem cell; HBO = hyperbaric oxygen.

Figure 5
In ammatory cell in ltration in brain parenchyma by day 28 after ICH induction A to E) Illustrating the immuno uorescent (IF) microscopic nding (400x) for cellular expression of CD14 (green color