Formic Acid Induces Hypertension-Related Hemorrhage in hSSAO TG in Mice and Stroke Patients

Hypertension is a conrmed risk factor for cerebral hemorrhage in humans. Which endogenous factor directly induces hypertension-related stroke is unclear. In this study, 42 hemorrhagic stroke patients with hypertension and hyperlipidemia and 42 age-matched healthy controls were recruited. The contents of serum semicarbazide-sensitive amine oxidase (SSAO) and formic acid (FC, a nal product of SSAO metabolism) were examined in the patients after stroke. Hemorrhagic areas were quantied by computer tomography. In the animal study, hemorrhagic degree was assessed by hemotoxylin & eosin or tissue hemoglobin kits. The relationship between FC and blood pressure/hemorrhagic degree was explored in wild-type mice and hSSAO TG mice fed with high-fat diets or high-fat and -salt diets. The results showed that the levels of serum FC were positively correlated with blood pressure and hemorrhagic areas in stroke patients. Transfection of microRNA-134 could enhance SSAO expression in human vascular smooth muscle cells. Consistently, after treatment with high-fat and -salt diets, hSSAO TG mice exhibited higher levels of miR134 and FC, higher blood pressure, and more severe hemorrhage than wild-type mice. Interestingly, folic acid reduced hypertension and hemorrhage in hSSAO TG mice fed with high-fat diets. These ndings suggest that FC is a crucial endogenous factor for hypertension and hemorrhage. TATAGAAGAGCAGCGTGGG GGGGCACGATTGTCAAAGAT. SSAO AGCAGCAGC SSAO CCAGGCCACCAAATCCTTTC. GAPDH GGTCGGAGTCAA CGGATTTG. GAPDH GGAAGATGGTGATGGGATTTC. The cDNA was synthesized from total RNA with a Microcute-enhanced microRNA First-chain Synthesis Kit (kr211-02, TIANGEN, USA). Quantitative real-time PCR (qPCR) analysis for microRNA-134 was performed with a Microcute Enhanced MicroRNA Fluorescence Quantitative Detection Kit ( # kr411-02, TIANGEN, USA) by Applied Biosystems ( # QuantStudio 5, Thermo Fisher Scientic, USA). The expression in the tested samples was normalized to the expression of the housekeeping gene U6 35 .


Introduction
The most common causes of intracerebral hemorrhage are hypertension and amyloid angiopathy 1 .
Epidemiological investigations have revealed that high-fat and -salt diets increase the risk of hemorrhagic stroke 2 . Further studies have shown that high-fat diets can induce hyperlipidemia 3 , and that high-salt diets directly cause hypertension 4 , and hemorrhagic stroke 5 . Unsurprisingly, hyperlipidemia and hypertension increase the risk of hemorrhagic stroke by several folds 6 . However, the molecular mechanisms of hemorrhagic stroke remain unclear.
The activity and expression of serum semicarbazide-sensitive amine oxidase (SSAO) have been found to be rapidly elevated in ischemia or hemorrhagic stroke patients 7,8 . MicroRNA-134 (miR134) has been proposed to upregulation induced by high-fat diets 9 , and it can enhance SSAO expression (also named VAP-1) in mice 10 . Unexpectedly, overexpression of this gene in the blood vessels causes an abnormal elevation in blood pressure 11 . Notably, formic acid (FC) is the nal product of SSAO metabolism [12][13][14] . FC administration can induce tissue hemorrhage, blood pressure elevation 15 , and dilatation of cerebral vessels 16 . These data strongly suggest that SSAO-derived FC may be a direct trigger of hypertension cerebral hemorrhage. Folic acid, a promoting agent of FC metabolism 17 , can prevent FC toxicity 18 . Thus, it may be used to intervene this disease.
In the present study, we investigated whether high-fat diet increases SSAO expression via miR134/NF-kB p65 pathway, and test whether endogenous FC (a nal product of SSAO metabolism) directly initiates hypertension and hemorrhage in wild-type mice, hSSAO TG mice, and acute stroke patients. Our ndings suggest that SSAO-derived FC endogenously causes hypertension cerebral hemorrhage, and high-salt is a critical exogenous factor for stimulating vascular rupture by enhancing FC toxicity.

Methods
Data collection from human participants.
In this cross-sectional survey, hemorrhagic stroke patients with hypertension and hyperlipidemia (n = 42; age range: 51 ~ 53 years) and age-matched controls (n = 42; age range: 54 ~ 56 years) were recruited from Beijing Boai Hospital in 2018 and 2019 (Chinese Clinical Trial Registry, Clinical Trials: ChiCTR1900026397, China, http://www.chictr.org.cn/index.aspx) and. People in northern China preferred high-salt foods. The baseline and clinical characteristics upon admission were recorded, and touchscreen questionnaires were administered, as described in Supplementary Table S1 and Supplementary   Table S2 in the Supplement. Informed consent was obtained from each participant either directly or from a guardian before participation. All patients provided written informed consent for study participation. All research was performed in accordance with local policies and guidelines and in accordance with the Declaration of Helsinki. Ethical approval for the clinical investigations was obtained from the Clinical Ethics Committee at the Capital Medical University, China.
HASMC Human aortic smooth muscle cells (HASMC) were cultured as described previously 30 .
HAEC Primary human aortic endothelial cells (HAEC, # Clonetics, Allschwil, Switzerland) were cultured and passaged in EBM-2 medium ( # EGM-2 bulletkit, Clonetics, Allcell, USA), as described previously 31 . SSAO expression in HASMC detected by immuno uorescence HASMC incubated with the treatments for 6 and 12 h were xed for quantifying SSAO expression 32  In vitro cellular toxicity of formic acid and NaCL The HASMC were exposed to the treatments for 6 and 12 h, respectively. The reagents were PBS; sodium chloride (NaCL): 145 mM 33 ; FC: 10 mM 34 ; and a mixed solution of NaCL and FC. Cell viability was detected at 6 and 12 h with a Cell Counting Kit-8 ( # CK04-500T, DOJINDO, Japan). The changes in cellular morphology after FC and NaCL treatment for 6 and 12 h were imaged under a microscope ( # IX73, OLYMPUS, USA).
Examination of miR134, NF-kBp65, SSAO by Q-PCR Total RNA was extracted from samples with TRI Reagent solution ( # R1030, Applygen, USA). The quantity of total RNA was determined according to OD260 measurements. Human miR134 mimic (sequence Hemorrhagic areas of patients analyzed by CT. The ABC/2 method of computer tomography (CT) allows for rapid estimation of the hemorrhage volume of stroke patients, as described previously 36 .
Data collection from animal participants.

Generation of hSSAO TG mice
The hSSAO TG mice were made by the Biocytogen Corporation (Beijing, China), as shown in Fig. 4a. The hSSAO TG mice were identi ed by PCR with DNA primers (F1: 5′-TGTGCTGAGGACTTGTGGTGTTAACC-3′; R1: 5′-CTTGGGAGGCAG CTGCAACTCCA-3′) generating a 567-bp product. Ethical approval for these animal studies was obtained from the Ethics Committee at the Capital Medical University, China.

Body mass index of mice
To generate the growth curve, the weights of the mice were measured once per week. Body weight and length (from the nose to anus) were measured at 8 weeks. Body mass index (BMI) was calculated as body weight (g)/body length 2 (cm) 37 .
Hemorrhagic areas of mice stained by H&E Frozen brain samples were sectioned at 15 µm thickness and stained with hematoxylin & eosin (H&E) 38 .
Quanti cation of brain vascular rupture by detecting brain hemoglobin Brain tissues were homogenized in Drabkin's color reagent according to the manufacturer's instructions ( # D5941, Sigma-Aldrich, USA). The extravagated blood in the tissue homogenates was quanti ed at 540 nm with a spectrophotometer ( # Multi-Mode Microplate Reader, SpectraMax i3, Molecular Devices, USA).
Hemorrhagic stroke of mice examined by MRI MRI measurements were made with a PharmaScan 7T small-animal MR scanner and a four-channel mouse brain surface coil ( # Bruker Biospin, Ettlingen, Germany) as described previously 39 .

SSAO expression in blood vessels stained by immuno uorescence
The immuno uorescence was carried out as previous report 40

Establishment of acute and chronic animal models
Acute models Three groups of adult male wild-type C57/BL-6J mice (n = 12, per group) were intraperitoneally injected with the treatments (saline; high-salt: 8% NaCL daily 5 ; FC: 5‰, 0.2 mL/10 g 14 ; and mixed solution of NaCL and FC) for 2 weeks.
Chronic models The hSSAO TG (n = 10) and wild-type mice (n = 10) were fed high -fat and/ -salt diets or folic acid (oral gavage, 1mg/kg/day weeks) for 8 consecutive weeks, as described previously 41 . Blood pressure measurement Blood pressure monitors ( # BP-2010 Series, Softron, Japan) were used to measure systolic blood pressure.

Statistical Analyses.
All data were tested for normality with the Kolmogorov-Smirnov test. When data were normally distributed, the statistical signi cance of differences was assessed with the unpaired t test and one, oneway or two-way ANOVA, analyzed by Sidak's multiple comparisons test. When the data were not normally distributed, the statistical signi cance of differences was assessed on the basis of P values with the Mann-Whitney U test. The human serum biochemical index was assessed with Student's unpaired t-test. The sex of participants was assessed with the chi-squared test. Human blood pressures were presented as the median, and p-values were assessed with the Mann-Whitney U-test. Correlations between the levels of serum SSAO/formic acid and hemorrhagic degree were assessed with the Pearson correlation coe cient, both without adjustment and with adjustment for sex and age. Statistical signi cance was set to p < 0.05. Analyses were performed in GraphPad Prism 6 software (GraphPad PRISM software, version 6.01, USA).

Result
Positive correlation of formic acid levels and blood pressure/hemorrhagic areas in stroke patients.
Previous studies have shown that SSAO is present in the blood and vascular smooth muscle cells 19 , which can catabolize methylamine into ammonia, formaldehyde (FA, HCHO), and hydrogen peroxide (H 2 O 2 ) 12,13 ; then the nal product, formic acid (FC), is formed after FA oxidation (Fig. 1a, b) 14 . We quanti ed the concentrations of serum SSAO, FA, H 2 O 2 , and FC, and found that serum FA and H 2 O 2 levels were lower in 42 stroke patients than 42 age-matched controls (Fig. 1c, d); while FC levels in stroke patients were higher than that of controls (Fig. 1e).
To con rm that these patients suffered from hemorrhagic stroke, low levels of serum Apolipoprotein C-III (AOP-C3) as a marker of hemorrhagic stroke 20 , were detected. As expected, AOP-C3 levels in 42 stroke patients were lower than controls ( Supplementary Fig. S1a), and the serum SSAO levels were negatively correlated with the AOP-C3 levels in 84 participants ( Supplementary Fig. S1b). Furthermore, there was a positive correlation between serum FC levels and mean blood pressure (MBP, MBP=(DBP + SBP)/2) ( Fig. 1f and Supplementary Fig. S1c, d). Notably, abnormal systolic blood pressure (SBP > 140 mmHg) and diastolic blood pressure (DBP > 90 mmHg) were observed in these stroke patients ( Supplementary  Fig. S2).
Stroke cases were further conformed by computed tomography (CT) and hemorrhagic areas analyzed by the ABC/2 method (Fig. 1g, top). There was a positive correlation between hemorrhagic areas and the levels of serum SSAO, H 2 O 2 , and FC, respectively ( Fig. 1g-i). Hence, SSAO-derived FC is closely related with hypertension and hemorrhage (Fig. 1j).
Formic acid-induced cell swelling, hypertension and hemorrhage in wild-type mice. To determine which factor is the possible trigger of cerebral hemorrhage, we examined the changes of the up-and downstream molecules of SSAO in stroke patients. High-fat diets upregulates microRNA-134 9 , and enhances SSAO expression (Fig. 2a) 11 . Hence, we examined these molecules, such as miR134 by Q-PCR, SSAO tested by human SSAO kits, and FC detected by FC kits. We found that the levels of miR134, SSAO, and FC in the patients after stroke onset 2 ~ 3 hours were higher than that after 4 ~ 6 months (Fig. 2a-d).
Excessive FC could directly elicit blood pressure elevation 15 , suggesting that FC may be a candidate trigger. The results showed that FC incubation led to cell swelling of human aortic smooth muscle cells (HASMC) at 6 and 12 hours (Fig. 2e and Supplementary Fig. 3a); while high salt (NaCL, a positive control, a trigger of hypertension and hemorrhage) with FC synergistically exacerbated cell swelling and death compared with cells receiving only NaCL or FC treatment (Fig. 2f). Similarly, high salt with FC synergistically also enhanced cell swelling and death in the cultured human vascular endothelial cells (HVEC) (Supplementary Fig. S3b, c). In addition, NaCL accelerated FC release from HASMC when the culture medium contained FC at 6 h (Fig. 2g).
To address whether FC directly induced hypertension in vivo, FC was intraperitoneally injected into wildtype mice for consecutive two weeks, blood pressure was monitored by electric blood pressure meter ( Supplementary Fig. S4), hemorrhagic areas were quanti ed by H&E, and brain vascular rupture was examined by tissue hemoglobin kits. Unexpectedly, an elevation in systolic blood pressure (SBP) was observed in these mice injected with FC, NaCL, or FC plus NaCL; while high salt accelerates FC-induced hypertension ( Fig. 2h and Supplementary Fig. S5).
Of note, FC as well as high slat (NaCL) directly elicited hemorrhagic stroke; while FC combined with NaCL resulted in greater hemorrhagic areas than did only FC or NaCL (Fig. 2i). Hemoglobin as a marker of cerebral hemorrhage is based on that hemoglobin ows out from the ruptured blood vessels 21 . In this study, FC directly caused a marked elevation in brain hemoglobin levels (Fig. 2j), indicating brain vascular rupture was occurred.
Upregulation of SSAO by high-fat diet-derived miR134 in the cultured HASMC, wild-type mice, and hSSAO TG mice. To address whether SSAO overexpression is mediated by miR134/NF-kB p65 pathway ( Fig. 3a) 22,23 , we transfected mimics of miR134 into cultured HASMC. The changes in mRNA levels of NF-kB p65 and SSAO were examined by Q-PCR; their protein contents were quanti ed by immuno uorescence (I.F.) (Fig. 3b). As expected, mimics of miR134 improved mRNA levels of both NF-kB p65 and SSAO (Fig. 3c, d), and increased the uorescence intensity of NF-kB p65 and SSAO than controls ( Fig. 3e-g and Supplementary Fig. S6). Thus, miR134 can enhance SSAO expression in vitro.
To test whether miR134 and SSAO contributes the occurrence of hemorrhagic stroke in vivo, we generated transgenic mice with overexpressing human SSAO gene (hSSAO TG ) (Fig. 4a, and Supplementary Fig. S7-9). In fact, high-fat diets or high-fat and -salt diets markedly increased body weights body mass index (BMI), and serum triglyceride (TG) in hSSAO TG mice than wild-type mice (Fig. 4b, c). Notably, high-fat diets obviously increased miR134 levels in wild-type mice; however, high-fat and -salt diets induced higher levels of miR134 in hSSAO TG mice than wild-type mice (Fig. 4d).
Through immuno uorescence with anti-SSAO, we found that high-fat diet or high-fat and -salt diets increased SSAO expression in the vascular walls and vascular abnormalities in thoracic aorta and cerebral microvessels (Fig. 4e, f, and Supplementary Fig. S10). Taken together, high-fat diet enhances miR134-mediated SSAO expression (Supplementary Fig. S11).
Formic acid-a direct trigger of hypertension/hemorrhage in h SSAO TG Mice and wild-type mice. To nd out the roles of endogenous FC in hypertension cerebral hemorrhage in wild-type mice and hSSAO TG mice fed with high-fat diet or high-fat and -salt diets for 8-weeks, we investigated the relationship between FC levels and blood pressure/hemorrhage. The results showed that the levels of serum H 2 O 2 and FA were declined (Fig. 5a, b); while serum FC levels were elevated in these two kinds of mice fed with different diets than those mice treated with the standard diet (Fig. 5c). Furthermore, we found that high-fat and -salt diets induced higher SBP and lower DBP in hSSAO TG mice than wild-type mice (Fig. 5d-f). Especially, serum FC levels were positively correlated with SBP in these six groups' mice (Fig. 6a).
Remarkably, the results of H&E staining revealed that high-fat and -salt diets caused greater hemorrhagic areas in hSSAO TG mice than wild-type mice (Fig. 6b, c). Meanwhile, more severe brain vascular rupture was occurred in the former than the latter, because the former had higher levels of brain hemoglobin (Fig. 6d). Similarly, this conclusion was con rmed by magnetic resonance imaging (MRI), in which two black cycle shadows were observed in the brains of hSSAO TG mice fed with high-fat and -salt diets compared with wild-type mice (Fig. 6e).
Further statistical analysis showed that serum FC levels were positively correlated with hemorrhagic degree detected by H&E or tissue hemoglobin kits, respectively (Fig. 6f). Notably, the higher levels of FC lead to the earlier onset of stroke in hSSAO TG mice than the wild-type mice fed with the same high-fat and -salt diets (Stroke onset: 3 weeks vs. 8 weeks) (Fig. 6g). Based on the above data, we concluded that endogenous FC is a crucial trigger of hypertension and hemorrhage.
Rescue effects of folic acid on hypertension/hemorrhage in acute or chronic formic acid-related model mice. If FC is a direct trigger of hypertension and hemorrhage, reduction of FC or enhancement of its metabolism should alleviate hemorrhagic stroke. Folic acid was selected to reduce FC-induced vascular toxicity, because it is a promoting agent of FC metabolism, which transfers FC to CO 2 and H 2 O by activating the enzyme of 10-formyl tetrahydrofolate synthetase (THF) 17 .
As expected, 2-week injection of folic acid decreased SBP in FC-or NaCL-injected wild-type mice (Fig. 7a, b and Supplementary Fig. S12). Hemorrhagic areas were also reduced in those mice after folic acid treatment ( Fig. 7c and Supplementary Fig. S13). Similarly, oral gavage of folic acid for 8 weeks could restore SBP (Fig. 7d and Supplementary Fig. S14), reduced hemorrhagic areas (Fig. 7e), and decreased brain vascular rupture in hSSAO TG mice fed with high-fat diets (Fig. 7f).

Discussion
The rst important result obtained in this study is that SSAO-derived FC endogenously induces hypertension and hemorrhage, and exogenous high-salt intake accelerates FC-induced vascular rupture. However, reduction of FC by folic acid could alleviate hypertension cerebral hemorrhage (Fig. 8).
Blood vessel breakage in the brain is the direct reason for hemorrhagic stroke 24 . High-fat diets have been found to change the pro le of microRNAs including miR134 9 , and to enhance SSAO expression in blood vessels. MiR134 activates NF-kB 22 , and the phosphorylation of NF-kB p65 upregulates SSAO. In the present study, we provided evidence that miR134 upregulated SSAO in the cultured HASMC, wild-type mice and hSSAO TG mice fed with high-fat diet, especially, in hemorrhagic stroke patients.
Another interesting nding was that serum FC (> 0.47 µg/mL) caused hypertension and hemorrhage in both mice and humans. Endogenous FC is mainly generated from the metabolism of SSAO-derived FA and/or methanol. Unsurprisingly, clinical cases show that exogenous methanol can induce hypertension 25,26 , and hemorrhage 27,28 . Actually, acute or chronic FC increased SBP in wild-type mice and hSSAO TG mice. Especially, higher levels of FC were associated with earlier stroke onset; and FC still remained relatively high levels after stroke. Hence, scavenging FC may contribute to stroke recovery.

Figure 8
The enhancement of formic acid metabolism by folic acid alleviates hypertension cerebral hemorrhage.
Schematic model describing the molecular mechanisms of SSAO-derived formic acid-induced hypertension cerebral hemorrhage. Folic acid (a promoting agent of FC metabolism, which transfers