Deficiency of cold‐inducible RNA‐binding protein exacerbated monocrotaline‐induced pulmonary artery hypertension through Caveolin1 and CAVIN1

Abstract Cold‐inducible RNA‐binding protein (CIRP) was a crucial regulator in multiple diseases. However, its role in pulmonary artery hypertension (PAH) is still unknown. Here, we first established monocrotaline (MCT)‐induced rat PAH model and discovered that CIRP was down‐regulated predominantly in the endothelium of pulmonary artery after MCT injection. We then generated Cirp‐knockout (Cirp‐KO) rats, which manifested severer PAH with exacerbated endothelium damage in response to MCT. Subsequently, Caveolin1 (Cav1) and Cavin1 were identified as downstream targets of CIRP in MCT‐induced PAH, and the decreased expression of these two genes aggravated the injury and apoptosis of pulmonary artery endothelium. Moreover, CIRP deficiency intensified monocrotaline pyrrole (MCTP)‐induced rat pulmonary artery endothelial cells (rPAECs) injuries both in vivo and in vitro, which was counteracted by Cav1 or Cavin1 overexpression. In addition, CIRP regulated the proliferative effect of conditioned media from MCTP‐treated rPAECs on rat pulmonary artery smooth muscle cells, which partially explained the exceedingly thickened pulmonary artery intimal media in Cirp‐KO rats after MCT treatment. These results demonstrated that CIRP acts as a critical protective factor in MCT‐induced rat PAH by directly regulating CAV1 and CAVIN1 expression, which may facilitate the development of new therapeutic targets for the intervention of PAH.

strategies in diagnosis and treatment of the disease is severally restricted. 6 Cold-inducible RNA-binding protein (CIRP) is a widely expressed RNA-binding protein, involved in various important cellular processes such as gene expression, genomic stability, apoptosis and proliferation. 7 So far, stimuli such as cold stress and short-term exposure to hypoxia have been found to increase CIRP expression, whereas heat stress and long-term exposure to hypoxia downregulate its expression. 8 Therefore, it has been reported that CIRP plays important roles in various diseases. 7 However, the effect of CIRP on PAH has not been investigated.
To determine the expression and function of CIRP in monocrotaline (MCT)-induced PAH and to investigate its downstream regulatory molecules and the related mechanisms in this process, we established MCT-induced rat PAH model, in which endothelium damage was the main cause of the disease. 9,10 CIRP was downregulated in endothelium of MCT-induced PAH, and severer disease phenotype manifested in Cirp-knockout (KO) rats. Moreover, Caveolin1 (Cav1) and Cavin1 were validated as downstream targets of CIRP, the decreased expression of these two genes aggravated apoptosis and injury of pulmonary artery endothelium. These data suggested that CIRP is critical for the homeostasis of endothelium and may serve as new therapeutic target for PAH.

| Animals
Cirp-KO rats were generated in sprague dawley (SD) background, which has been described in detail previously, 11  The systole and diastole blood pressure of rats in quiet environments without outside stimulations was measured noninvasively with coda monitor, and mean artery pressure (MAP) was calculated according to the following formula: MAP = 1/3 systole pressure + 2/3 diastole pressure.
At week four after MCT injection, the animals were anaesthetized, and right ventricular systolic pressure (RVSP) was measured by inserting a pressure-volume catheter into the right ventricle.
Subsequently, rats were sacrificed; after effectively cleaning the remaining blood with heparin saline, the lungs were harvested, snapfrozen in liquid nitrogen, and then stored at −80°C for RNA and protein extraction. To assess the right ventricular hypertrophy, right ventricle (RV) and left ventricle plus septum (LV + S) were dissected and weighed, respectively.

| Cell culture and transfection
Rat pulmonary artery endothelial cells (rPAECs) and rat pulmonary artery smooth muscle cells (rPASMCs) purchased from Creative Bioarray Co. were cultured in DMEM with 5% FBS (Excell) and used for experiments between passage three and nine.
Scramble siRNA was used as negative control (si-NC). The siRNA sequences used are listed in Table 1.
For overexpression, Cirp, Cav1 and Cavin1 cDNAs were cloned into pCMV6 tagged with Myc and DDK (Origene), and verified by Sanger sequencing. The constructs were transfected into rPAECs with Lipofectamine 3000 (Invitrogen).

| Monocrotaline pyrrole (MCTP) treatment
MCTP was extracted and purified in accordance with protocol provided by Mattocks et al 12

| LDH release assay
The medium of cultured rPAECs was harvested and centrifuged. The supernatant was collected and incubated with LDH release regents (Beyotime) at 37°C for 1 hour, then mixed with the LDH reaction mixture. The absorbance was recorded at 490 nm wavelength on plate spectrophotometer.

| 5-ethynyl-2'-deoxyuridine (EdU) assay
To test the effects of rPAECs-conditioned medium on rPASMCs proliferation, rPAECs were transfected with specific siRNA or plasmids and treated with MCTP for 48 hours. rPASMCs were subjected to 24 hours of growth arrest in serum-free medium, then treated with conditioned medium from rPAECs. After 48 hours, rPASMCs were fixed for EdU staining in accordance with manufacture's instruction (Thermo).

| Permeability assay
Rats were injected 1% evans blue (EB) dye (Sigma) (6 mL/kg) in PBS/4% bovine serum albumin (BSA) via tail vein. One hour later, animals were sacrificed and the lungs were dissected after completely flushing out residual intravascular dye with PBS. Dissected lungs were photographed and homogenized in formamide (Sigma), then incubated at 37°C for 24 hours. After centrifuge, the supernatant was collected and measured at 620 nm on plate spectrophotometer.
To assess endothelial monolayer permeability, transwells

| TUNEL assay
After blocking with PBS/6% goat serum, the lung sections (4 µm) and rPAECs were incubated with TUNEL reaction mixture (Roche) at 37°C in humid box for 1 hour in the dark. Nuclei were stained with DAPI (Sigma). The images were captured with fluorescence microscope (Leica).

| RNA immunoprecipitation (RIP)
RNA-IP was performed according to the Magna RIP RNA-binding Protein Immunoprecipitation Kit protocol (Millipore). Briefly, rPAECs at 80%-90% confluency in culture dishes were lysed in RIP lysis buffer, and immunoprecipitated with anti-CIRP antibody (Proteintech) or normal rabbit IgG (Millipore) conjugated on magnetic beads, then incubated with proteinase K (Millipore) to isolate the immunoprecipitated RNA.
Finally, CIRP-binding RNAs were extracted by acid-phenol-chloroform and analysed by qPCR. The primers used are listed in Table 2.

| Western blotting
Total protein was extracted with RIPA buffer containing proteinase inhibitor (Roche) and quantified with BCA protein assay reagent (Beyotime), then denatured at 95°C. Proteins were separated on 10% SDS-polyacrylamide gels and transferred to Immobilon-P polyvinylidene fluoride (PVDF) membranes. The membranes were blocked with TBS/5% non-fat milk, then incubated with the following primary antibodies against CAV1 (Abcam), CAVIN1 (Abcam), CIRP (Proteintech) and β-ACTIN (Santa Cruz) at 4°C overnight and probed with secondary antibodies at room temperature for 1 hour.
The signals were detected on Odyssey imager. The primers used are listed in Table 2.

| Statistics
All data shown are the mean ± SD from three independent experiments. Statistical analyses were conducted by Student's t test, Oneway ANOVA and Tukey-Kramer multiple comparisons test.

| Cirp was down-regulated in the lung of MCTinduced PAH rat
To evaluate the expression of Cirp in PAH, we established rat PAH model by intraperitoneal injection of MCT. Remarkably, at week four, both mRNA and protein level of CIRP were significantly decreased in the lung of PAH rats revealed by qPCR and Western blotting, respectively ( Figure 1A,B). By immunohistochemistry, we found CIRP was widely expressed in the lung, and notably, the reduced expression of CIRP upon MCT administration predominantly localized in the endothelium of pulmonary artery ( Figure 1C), indicating the potential role of CIRP in endothelial cell function. To further explore the possible mechanism of CIRP in PAH, we analysed CIRP expression in cultured rPAECs and rPASMCs respectively by immunofluorescence and Western blotting, which detected significantly higher expression of CIRP in rPAECs than in rPASMCs ( Figure 1D,E).
As MCTP is the active metabolite of MCT in vivo, we treated both rPAECs and rPASMCs with MCTP. Upon MCTP stimulation, the pro-

| Cirp deletion exacerbated MCT-induced PAH
Next, to investigate the role of CIRP in PAH, we generated Cirp-KO rats with seven bases deletion in exon three by transcription activator-like effector nuclease (TALEN). The Cirp gene deletion in Cirp-KO rats was confirmed by Western blotting, in which CIRP protein was barely detected in the lung ( Figure S1A). Both Cirp-KO and WT rats were normal in appearance and behaviour.
Although there was no difference in RVSP and RV/LV + S be- In our previous research, Cirp-KO rats presented arrhythmia at baseline. 11 We wonder whether the effect of CIRP on MCT-

| Identification of Cav1 and Cavin1 as CIRP's targets in PAH regulation
We then set to investigate the underline targets of CIRP in MCT- and Cavin1 attracted our attention, both of which predominantly expressed in pulmonary artery endothelium and were involved in the progress of PAH. 17 It is worth noting that endothelial cell-specific knockout of Cav1 or Cavin1 respectively could directly induce PAH in C57 mice, 18,19 and CAV1 is a well-known mutant gene in idiopathic pulmonary artery hypertension (IPAH) patients. 20 We performed RIP on rPAECs, which revealed dramatic enrichment of CIRP on both Cav1 and Cavin1 mRNAs, verifying these two genes as direct targets of CIRP ( Figure 3D,E). Furthermore, the binding of CIRP to either Cav1 or Cavin1 was significantly reduced after MCTP stimulation ( Figure 3D,E).
We then analysed the expression of CAV1 and CAVIN1 in the lungs of Cirp-KO rats four weeks after MCT treatment. Compared with WT rats, significantly lower amount of both CAV1 and CAVIN1 mRNA ( Figure S2A,B) and protein ( Figure 3C) was detected in Cirp-KO rats. By immunohistochemistry, both CAV1 and CAVIN1 mainly expressed in pulmonary artery endothelium, and decreased drastically after MCT treatment especially in CIRP deficiency ones ( Figure 3F,G). These data suggest that CAV1 and CAVIN1 may mediate CIRP in maintaining homeostasis of endothelium.

| CIRP regulated MCTP-induced rPAECs apoptosis and permeability through CAV1 and CAVIN1
Since MCTP directly targets endothelium in vivo and induces excessive PAECs apoptosis, which is detrimental to vascular function, endothelium is considered to be the primary target in this process, eventually leading to PAH. [21][22][23] By TUNEL assay, we found Cirp depletion indeed aggravated rPAECs apoptosis in response to MCT treatment in vivo ( Figure 4A). Meanwhile, in cultured rPAECs, MCTPinduced pronounced apoptosis indicated by increased cleaved CASPASE 3, which was intensified by Cirp knockdown ( Figure S3A) and counteracted by Cirp overexpression ( Figure S3B).
One aspect of endothelium dysfunction contributing to PAH is the continuous damage of PAECs, resulting in increased vessel permeability. 24 We therefore injected EB dye via tail vein to assess the endothelium integrity in vivo. Though no vessel leakage manifested in both unstimulated WT and Cirp-KO rats ( Figure 4B), elevated extravasation of EB dye into lungs was detected following MCT injection with Cirp-KO rats presenting much higher permeability two weeks after MCT treatment ( Figure 4B). These results indicated that CIRP deficiency deteriorated MCT-induced endothelium damage.
Based on the previous data that Cav1 and Cavin1 may be down-  Figure 4G) or down-( Figure S4E) regulation respectively.

| CIRP regulated the proliferative effect of conditioned media from MCTP-treated rPAECs on rPASMCs
Although the regulatory function of CIRP in MCT-induced PAH is mainly by protecting endothelium from injury, we also observed exceedingly thickened intimal media of pulmonary artery in Cirp-KO rats ( Figure 5D,E). In PAH, the damaged endothelium could secret growth factors and vasoactive mediators, which induced intimal media proliferation. To test the effects of damaged rPAECs on rPASMCs, we cultured MCTP-treated rPAECs, collected the conditioned media and applied it onto rPASMCs. As shown in Figure 5A

| D ISCUSS I ON
In this study, Cirp-KO rats showed aggravated endothelium damage in MCT-induced PAH. Protein mass spectrometry analysis provided evidence that CAV1 and CAVIN1 are potential targets of CIRP in this process. In subsequent functional analysis, we observed increased apoptosis and hyperpermeability in MCTP-treated rPAECs with Cirp depletion, which was caused by reduced expression of CAV1 or CAVIN1, indicating that CIRP mediates MCT-induced PAH through regulating the level of CAV1 and CAVIN1 in pulmonary artery endothelium.
Drugs or toxins exposure often causes devastating damage to pulmonary artery endothelium, eventually leading to PAH. 9,25,26 MCT is an alkaloid that is often used to induce PAH in rodents, which could mimic a great deal of pathophysiological features in human PAH. 27 It has been reported that MCTP induces LDH release and apoptosis of endothelial cells. 22 Although rPAECs manifested increased apoptosis and hyperpermeability when treated with MCTP, 22,28 rPASMCs presented minimal changes. 21 Thus, in MCT-induced PAH, endothelium is widely recognized as the primary target of MCTP. Our data demonstrated the profound role of rPAECs in CIRP's regulation of MCT-induced PAH.

F I G U R E 4
Cirp knockdown exacerbated MCTP-induced rPAECs apoptosis and permeability that was counteracted by Cav1 or Cavin1 overexpression. A, TUNEL assay of lungs isolated from WT and Cirp-KO rats at week four after saline or MCT injection. The apoptotic index was measured by counting puncta of TUNEL signal in randomly selected CD31-positive cells (n ≥ 3 per group). Scale bar, 46 µm. B, Images of EB dye staining in lungs isolated from WT and Cirp-KO rats at different time points after MCT injection respectively, and quantitative assessment of EB dye content in the lungs. C, Representative immunoblots and densitometric analysis of cleaved CASPASE-3, CAV1 and CAVIN1 in Cirp knockdown and Cav1 or Cavin1 overexpressing rPAECs following MCTP treatment. β-ACTIN served as loading control. (n = 3 per group). D, TUNEL assay of Cirp knockdown and Cav1 or Cavin1 overexpressing rPAECs following MCTP treatment. DAPI stains nuclei. Scale bars, 50.4 µm. E, The apoptotic index was measured by counting puncta of TUNEL positive signal. F and G, Permeability and cytotoxicity of Cirp knockdown and Cav1 or Cavin1 overexpressing rPAECs following MCTP treatment were assessed by quantifying absorbance of FITC-dextran (F), LDH release (G). WT, wild type; KO, Cirp-knockout; MCT, monocrotaline; MCTP, monocrotaline pyrrole; TUNEL, Terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labelling; LDH, lactate dehydrogenase; CD31, Platelet endothelial cell adhesion molecule-1, PECAM-1/CD31; DAPI, 4′,6-diamidino-2-phenylindole; The data are shown as mean ± SD from three independent experiments. **** P <.0001, ***P <.001, ** P <.01, * P <.05. ns, not significant At present, reduced CIRP expression was reported when cells were exposed to heat stress and hypoxia etc, indicating that CIRP served as a protective factor against stressful conditions. 29 For the first time, we discovered that CIRP expression decreased, particularly, in endothelium of MCTP-induced PAH, and genetically depletion of Cirp aggravated endothelium apoptosis and permeability, consistent with its protective role in other stress and diseases.
Knockdown of Cirp could promote apoptosis in various cells in different disease models. For instance, knocking down Cirp could aggravate apoptosis in cardiomyocytes. 30  The predominant function of CIRP is to stabilize mRNAs of target genes. 33 Liu et al showed CIRP binds to 3' untranslated region (UTR) of Cav1 and coding sequence (CDS) of Cavin1, respectively.

Down-regulation of Cirp decreased mRNA expression of Cav1 and
Cavin1. 16 Similarly, our study confirmed the binding of CIRP to Cav1 and Cavin1 mRNAs in rPAECs, which was significantly reduced after MCTP stimulation. Interestingly, we found that in the absence of MCT, the Cav1 and Cavin1 mRNAs in Cirp-KO rats were less than those of WT rats ( Figure S2A,B); however, at protein level, little difference was observed. The protein abundance regulation credibly reflects biological roles of specific genes. However, approximately one-to two-thirds of the variance in steady-state protein level can be explained by partial correlation of mRNA transcript and protein abundances. 34 The drift of mRNA transcript and protein abundance may be compensated by translational and protein-degradational regulation. 34 Thus, we speculated that CIRP acted as one of stabilizers In response to stimulation, damaged endothelial cells in PAH could release various vasomediators and growth factors inducing vasoconstriction and smooth muscle cell proliferation and eventually leading to pulmonary vascular luminal obliteration and resistance. 41 From the results obtained in this research, we assumed that MCTP could induce rPAECs to secret factors to promote rPASMCs proliferation, which is fundamental to pulmonary vascular remodelling. As the proliferative effect of conditioned media was also correlated to CIRP, CAV1 and CAVIN1 expression in MCTP-treated rPAECs, it is intriguing to identify these proliferation factors and investigate the underlying mechanisms, that is, how CIRP regulates their expression, which may facilitate the development of new targets for the treatment of PAH.
The loss of CIRP aggravated endothelial injury by downregulating CAV1 and CAVIN1, which are essential in maintaining endothelium homeostasis. The protective role of CIRP in MCTinduced endothelium damage indicated that overexpression of CIRP in endothelium may serve as a protective factor in toxininduced PAH. In the future, gene therapy targeting CIRP may become a novel intervention against PAH in clinical practice.
However, because CIRP is widely expressed in tissues and organs, the molecular mechanisms that CIRP regulates CAV1 and CAVIN1 in endothelium need further investigation in order to circumvent any possible side effects.
To summarize, we characterized CIRP's critical role in MCTinduced PAH. In Cirp-KO rats, MCT induced aggravated PAH with severe endothelium damage. The main function of CIRP in MCTinduced PAH is to protect endothelial cells from apoptosis and to maintain endothelium integrity, which is mediated by its downstream targets CAV1 and CAVIN1.

ACK N OWLED G EM ENTS
We thank the great support from Prof. Yi-Han Chen; Dr Dandan Liang, Yi Liu, Honghui Ma, Dan Shi and Liang Xu for critical discussion on the manuscript. Jian Yang is funded by the National Natural Science Foundation of China (31871491).

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.