International Journal of Molecular Sciences Article Evaluation and Characterization of Post-Stroke Lung Damage in a Murine Model of Cerebral Ischemia Júlia Faura 1 , Laura Ramiro 1 , Alba Simats 1 , Feifei Ma 1 , Anna Penalba 1 , Teresa Gasull 2 , Anna Rosell 1 , Joan Montaner 1,3,4 and Alejandro Bustamante 1,5, * 1 Neurovascular Research Laboratory, Valld’Hebron Research Institute (VHIR), Universitat Autònoma de Barcelona (UAB), 08035 Barcelona, Spain; juliafaura10@gmail.com (J.F.); laura.ramirop@gmail.com (L.R.); simatsalba@gmail.com (A.S.); mafeifeivhir@gmail.com (F.M.); anna.penalba@vhir.org (A.P.); anna.rosell@vhir.org (A.R.); jmontaner-ibis@us.es (J.M.) 2 Cellular and Molecular Neurobiology Research Group, Department of Neurosciences, Germans Trias i Pujol Research Institute (IGTP), 08916 Badalona, Spain; tgasull@igtp.cat 3 Stroke Research Program, Institute of Biomedicine of Seville, IBiS/Hospital Universitario Virgen del Rocío/CSIC, University of Seville, 41013 Seville, Spain 4 Department of Neurology, Hospital Universitario Virgen Macarena, 41009 Seville, Spain 5 Stroke Unit, Hospital Universitari Germans Trias i Pujol, 08916 Badalona, Spain * Correspondence: abustamanter.germanstrias@gencat.cat Abstract: After stroke and other brain injuries, there is a high incidence of respiratory complications such as pneumonia or acute lung injury. The molecular mechanisms that drive the brain-lung interaction post-stroke have not yet been elucidated. We performed transient middle cerebral artery occlusion (MCAO) and sham surgery on C57BL/6J mice and collected bronchoalveolar lavage fluid (BALF), serum, brain, and lung homogenate samples 24 h after surgery. A 92 proteins-panel developed by Olink Proteomics® was used to analyze the content in BALF and lung homogenates. Citation: Faura, J.; Ramiro, L.; Simats, A.; Ma, F.; Penalba, A.; Gasull, T.; MCAO animals had higher protein concentration levels in BALF than sham-controls, but these levels Rosell, A.; Montaner, J.; Bustamante, did not correlate with the infarct volume. No alteration in alveolar-capillary barrier permeability was A. Evaluation and Characterization observed. A total of 12 and 14 proteins were differentially expressed between the groups (FDR < 0.1) of Post-Stroke Lung Damage in a in BALF and lung tissue homogenates, respectively. Of those, HGF, TGF-α, and CCL2 were identified Murine Model of Cerebral Ischemia. as the most relevant to this study. Their protein expression patterns were verified by ELISA. This Int. J. Mol. Sci. 2022, 23, 8093. study confirmed that post-stroke lung damage was not associated with increased lung permeability https://doi.org/10.3390/ or cerebral ischemia severity. Furthermore, the dysregulation of HGF, TGF-α, and CCL2 in BALF ijms23158093 and lung tissue after ischemia could play an important role in the molecular mechanisms underlying Academic Editor: Anuska V. stroke-induced lung damage. Andjelkovic Keywords: stroke; inflammation; pulmonary complications Received: 29 June 2022 Accepted: 19 July 2022 Published: 22 July 2022 Publisher’s Note: MDPI stays neutral 1. Introduction with regard to jurisdictional claims in After acute brain injuries such as stroke, traumatic brain injury (TBI), and subarachnoid published maps and institutional affil- hemorrhage, there is a high incidence of lung-related complications, such as pneumonia, iations. acute lung injury (ALI), acute respiratory distress syndrome (ARDS), and neurogenic pulmonary edema [1,2]. The high frequency of lung damage after a brain injury suggests a very close interaction between these organs after the episode. This interaction seems Copyright: © 2022 by the authors. bidirectional, though the exact mechanisms are not yet fully elucidated. After brain injury, Licensee MDPI, Basel, Switzerland. proinflammatory cytokines are secreted in the brain, the brain-blood barrier (BBB) is This article is an open access article impaired, facilitating leukocyte infiltration into the brain. These events involve several distributed under the terms and systemic alterations that can damage the peripheral organs, including the lungs. This local conditions of the Creative Commons peripheral damage, in turn, exacerbates the primary damage in the brain. This bidirectional Attribution (CC BY) license (https:// relationship between the brain and the lungs has been termed the “double hit model” [3]. creativecommons.org/licenses/by/ The local effect of stroke on the lungs has become the focus of recent research efforts. In 4.0/). an experimental rat model of cerebral ischemia, both pulmonary edema and ultrastructural Int. J. Mol. Sci. 2022, 23, 8093. https://doi.org/10.3390/ijms23158093 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2022, 23, 8093 2 of 12 changes in the lung parenchyma were observed after ischemia. In addition, levels of proinflammatory cytokines, including interleukin-6 (IL-6) and tumor necrosis factor-α (TNF-α), were elevated in the plasma, brain, and bronchoalveolar lavage fluid (BALF) [4]. In contrast, Austin et al. [5] did not observe lung damage (i.e., edema or impaired lung function) after cerebral ischemia in mice. However, they did observe inflammation in this organ through the expression of pro-inflammatory cytokines in both lung tissue and BALF. Also, the immune niche of the lung is dysregulated after cerebral ischemia, as characterized by a decrease in lymphocytes and an increase in alveolar macrophages and neutrophils [6]. Stroke-associated pneumonia (SAP) is one of the most frequent complications after a stroke, with an incidence of around 12% [7], and increases the risk of mortality and disability in stroke patients and their clinical length-of-stay [8]. Nowadays, there is a lack of preventive therapies for SAP in clinical practice. Clinical trials testing antibiotic prophylaxis have not succeeded either in reducing the SAP incidence or improving the prognosis of stroke patients [9]. As stroke-induced immunosuppression is one of the major factors for the occurrence of SAP, immunomodulating therapies have also been proposed for SAP prevention, with great success in experimental studies but still moderate translation into human-based studies [10]. Hence, identifying novel therapeutic targets and therapies for SAP prevention is urgently needed in this field of study. Therefore, this study aims to elucidate the mechanisms that can drive stroke-induced lung damage by characterizing BALF and lung protein content to identify potential thera- peutic targets for the prevention of SAP. 2. Results 2.1. BALF Characterization after Cerebral Ischemia To characterize the impact of stroke on the lungs, we evaluated the BALF protein content in 42 ischemic and 11 sham-control mice. The total protein concentration increased in ischemic compared with sham-control animals (275.7 (interquartile range (IQR): 218.4–400.6) µg/mL vs. 215.8 (174.6–266.1) µg/mL), p = 0.009, N = 42 vs. 11) (Figure 1A). No correlation was observed between BALF protein content and the severity of the ischemia, as evaluated by the infarct volume (in mm3 ) (Figure 1B). Next, we aimed to characterize the protein content in BALF to identify molecular alter- ations caused by cerebral ischemia. For this, we analyzed protein content by simultaneously measuring 92 molecules in the BALF samples from 42 ischemic and 11 sham-control animals. Proteins with more than 40% of their values below the detection limit were excluded, and of the 49 remaining, 12 were differentially expressed (FDR < 0.1) between the two groups. All of these differentially expressed proteins were upregulated in the ischemic animals. The most significant proteins were hepatocyte growth factor (HGF), NAD kinase (NADK), and protein phosphatase inhibitor 2 (PPP1R2) (Table 1, Figure 2A). Table 1. Top table of BALF (left) and lung homogenate (right) protein expression. Significant proteins are shown. The p-value column corresponds to the unadjusted p-value. logFC: logarithmic fold-change, FDR: false discovery ratio. BALF Lung Homogenate Uniprot Uniprot Protein logFC p-Value FDR Protein logFC p-Value FDR Code Code Protransforming growth Hepatocyte growth factor 0.740 >0.001 0.002 Q08048 factor alpha 1.296 >0.001 >0.001 P48030 (HGF) (TGFA) NAD kinase C-C motif chemokine 2 1.237 >0.001 0.002 P58058 1.115 >0.001 >0.001 P10148 (NADK) (CCL2) N(G),N(G)- Protein phosphatase dimethylarginine inhibitor 2 0.873 >0.001 0.005 Q9DCL8 dimethylaminohydro- 0.392 0.001 0.010 Q9CWS0 (PPP1R2) lase 1 (DDAH1) Int. J. Mol. Sci. 2022, 23, 8093 3 of 12 Table 1. Cont. BALF Lung Homogenate Uniprot Uniprot Protein logFC p-Value FDR Protein logFC p-Value FDR Code Code Poly [ADP-ribose] Caspase-3 1.209 >0.001 0.005 P70677 polymerase 1 0.891 0.001 0.010 P11103 (CASP3) (PARP1) CNN family member 1 Calsyntenin-2 0.767 0.001 0.006 P18406 0.691 0.001 0.011 Q9ER65 (CYR61) (CLSTN2) Glial cell line-derived Integrin beta-1-binding neurotrophic factor 0.609 0.001 0.006 P48540 protein 2 −0.873 0.002 0.026 Q9R000 (GDNF) (ITGB1BP2) Tumor necrosis factor receptor superfamily Interleukin-1 alpha 0.524 0.001 0.006 P25446 −0.928 0.003 0.029 P01582 member 6 (IL1A) (FAS) Integrin beta-6 Peroxiredoxin-5 0.573 0.002 0.015 Q9Z0T9 0.510 0.004 0.033 P99029 (ITGB6) (PRDX5) Disintegrin and metalloproteinase Peroxiredoxin-5 0.897 0.007 0.030 P99029 domain-containing 0.182 0.008 0.063 Q9R1V7 (PRDX5) protein 23 (ADAM23) Synaptosomal-associated Growth-regulated alpha protein 29 1.007 0.007 0.030 Q9ERB0 protein 1.600 0.008 0.063 P12850 (SNAP29) (CXCL1) Dihydropteridine Follistatin reductase 0.634 0.007 0.030 Q8BVI4 0.333 0.015 0.096 P47931 (FST) (QDPR) Legumain Matrilin-2 0.554 0.007 0.030 O89017 −0.441 0.017 0.096 O08746 (LGMN) (MATN2) NAD kinase −0.256 0.017 0.096 P58058 (NADK) Transforming growth factor beta-1 proprotein 0.618 0.018 0.096 P04202 (TGFB1) An increase in the HGF levels in the BALF of the ischemic animals compared with sham animals was confirmed using another technique (ELISA) (2541.3 (2051.35–3491.26) pg/mL vs. 1887.31 (1283.22–2496.97) pg/mL, p = 0.021) (Figure 3A). HGF levels measured by both techniques (Proximity Extension Assay (PEA) and Enzyme-Linked ImmunoSorbent Assay (ELISA) positively correlated (R = 0.56, p < 0.0001) (Figure 3B). Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW 3 of 13 Int. J. Mol. Sci. 2022, 23, 8093 4 of 12 Evaluation of Figure1.1.Evaluation Figure of the the protein concentration concentration in in BALF. BALF.(A) (A)Boxplot Boxplotcomparing comparingthe theBALF BALF protein pro- concentration tein concentrationbetween between ischemic ischemicandand sham sham mice mice(N(N==53) 53)(B) (B) Scatter representing the Scatter plot representing theBALF BALF protein proteinconcentration concentration(X(Xaxis) axis)and andthethe infarct volume infarct volume(Y axis) of the (Y axis) ischemic of the ischemicanimals (N =(N animals 42).=(C) 42). Two representative (C) Two images representative of cerebral images ischemia of cerebral measured ischemia by TTC measured staining. by TTC The boxplot staining. in (A)in The boxplot shows increased (A) shows protein increased concentration protein concentrationin MCAO in MCAO animals compared animals compared to to sham, sham,while whilethe thescatterplot scatterplot in B showed Int. J. Mol. Sci. 2022, 23, x FOR PEER REVIEW the lack of correlation between infarct volume and protein concentration. MCAO: in B showed the lack of correlation between infarct volume and protein concentration. MCAO: 4middle of 13 middle cerebral artery occlusion. cerebral artery occlusion. Next, we aimed to characterize the protein content in BALF to identify molecular alterations caused by cerebral ischemia. For this, we analyzed protein content by simul- taneously measuring 92 molecules in the BALF samples from 42 ischemic and 11 sham-control animals. Proteins with more than 40% of their values below the detection limit were excluded, and of the 49 remaining, 12 were differentially expressed (FDR < 0.1) between the two groups. All of these differentially expressed proteins were upregulated in the ischemic animals. The most significant proteins were hepatocyte growth factor (HGF), NAD kinase (NADK), and protein phosphatase inhibitor 2 (PPP1R2) (Table 1, Figure 2A). Figure 2. Volcano plot of significant proteins in the BALF (A) and lung (B) analyses of ischemic Figure 2. Volcano plot of significant proteins in the BALF (A) and lung (B) analyses of ischemic vs. shamsham vs. animals. animals. The The X-axis X-axis represents represents the logFC the logFC and and the Y-represents the Y-represents the logarithm the logarithm in base in base 10 of10the of the False False Discovery Discovery Ratio Ratio (FDR). (FDR). TheThe horizontal horizontal line line symbolizesananFDR symbolizes FDR<0.1 < 0.1and andthe thevertical vertical lines lines symbolize an symbolize an FDR FDR >> |0.25|. |0.25|.Proteins Proteinsover overor orunder underthethementioned mentionedcut-off cut-off points points for for FDR FDR and and FCFC represent thedifferentially represent the differentiallyexpressed expressed proteins proteins andand are are marked marked in brown. in brown. A totalA of total of 12 proteins 12 proteins (BALF) (BALF) and 14showed and 14 (lung) (lung) differential showed differential expressionexpression betweenand between ischemic ischemic and sham sham animals. animals. LogFC: LogFC: logarithmic logarithmic fold-change,fold-change, FDR: falseratio. FDR: false discovery discovery ratio. Table 1. Top table of BALF (left) and lung homogenate (right) protein expression. Significant pro- teins are shown. The p-value column corresponds to the unadjusted p-value. logFC: logarithmic fold-change, FDR: false discovery ratio. BALF LUNG HOMOGENATE (TGFB1) An increase in the HGF levels in the BALF of the ischemic animals compared with sham animals was confirmed using another technique (ELISA) (2541.3 pg/mL (2051.35–3491.26) vs. 1887.31 (1283.22–2496.97) pg/mL, p = 0.021) (Figure 3A). HGF levels Int. J. Mol. Sci. 2022, 23, 8093 measured by both techniques (Proximity Extension Assay (PEA) and Enzyme-Linked 5 of 12 ImmunoSorbent Assay (ELISA) positively correlated (R = 0.56, p < 0.0001) (Figure 3B). HGF,CCL2 Figure3.3.HGF, Figure CCL2and andTGF-α TGF-αmeasurements measurementsby byELISA. ELISA.(A) (A)Boxplot Boxplotcomparing comparingHGFHGFlevels levelsinin ® panel BALF of BALF of MCAO MCAO vs. vs. sham shammice mice(N(N= =49)49) (B)(B) Dotplot Dotplotrepresenting representingHGFHGF measurement by Olink measurement by Olink ® (X-axis) panel and measurement (X-axis) and measurementby ELISA (Y axis) by ELISA (N = (N (Y axis) 53).= (C) 53).Boxplot comparing (C) Boxplot CCL2 comparing levels CCL2 in lung levels in lung of MCAO of MCAO vs. sham vs. sham micemice (N = (N 30)=(D) 30)Dot (D)plot Dot representing plot representingCCL2CCL2 measurement measurement by ® by Olink Olink ® panel panel (X and (X axis) axis)ELISA and ELISA measurement measurement (Y axis) (Y axis) (N = (N30).= (E) 30).Boxplot (E) Boxplot comparing comparing TGF-α TGF-α levels levels in in lung lung of MCAO vs. sham mice (N = 30) (F) Dot plot representing TGF-α measurement of MCAO vs. sham mice (N = 30) (F) Dot plot representing TGF-α measurement by the Olink® by the Olink ® panel panel(X (Xaxis) axis)and andELISA ELISAmeasurement measurement(Y (Yaxis) axis)(N (N==30). 30).The Theboxplots boxplotsshowed showedincreased increasedHGF, HGF,aa trend towards increased CCL2 and decreased TGF-a values in MCAO vs. sham animals, while the trend towards increased CCL2 and decreased TGF-a values in MCAO vs. sham animals, while the scatterplots showed positive correlations for HGF and CCL2 and a negative correlation for TGF-a. scatterplots showed positive correlations for HGF and CCL2 and a negative correlation for TGF-a. HGF: Hepatocyte Growth Factor, CCL2: C-C Motif Chemokine Ligand 2, TGF-α: Transforming HGF: Hepatocyte Growth Growth Factor α, ELISA: Factor, CCL2: ImmunoSorbent Enzyme-Linked C-C Motif Chemokine Assay,Ligand BALF: 2, TGF-α: Transforming bronchoalveolar lavage Growth fluid, NPX: Factor α, ELISA: normalized Enzyme-Linked protein ImmunoSorbent expression, MCAO: Assay, artery middle cerebral BALF:occlusion. bronchoalveolar lavage fluid, NPX: normalized protein expression, MCAO: middle cerebral artery occlusion. 2.2. Protein Expression Characterization in Lung Next, we aimed to characterize the protein content in whole lung homogenates to iden- tify new molecules deregulated in the lung after stroke. To that end, the same 92 molecules were simultaneously measured in the lung homogenates from the same 24 ischemic and 6 sham-control animals. Out of the 75 proteins included in the final analysis after filtering by expression, 14 differentially expressed between experimental conditions were found. Four of these were downregulated in MCAO mice, while the remaining ten were upregulated after stroke. Among the significantly dysregulated proteins, only two showed significant differences between the two groups in BALF and lung homogenates. These were NADK and mito- chondrial peroxiredoxin-5 (PRDX5). NADK was upregulated in the BALF of ischemic mice and downregulated in the lung, while PRDX5 was upregulated in both samples from ischemic mice. In this case, Transforming Growth Factor-α (TGF-α) in its proform (proTGF-α), C-C Motif Chemokine Ligand 2 (CCL2), and dimethylarginine dimethylaminohydrolase 1 between the two groups in BALF and lung homogenates. These were NADK and mito- chondrial peroxiredoxin-5 (PRDX5). NADK was upregulated in the BALF of ischemic mice and downregulated in the lung, while PRDX5 was upregulated in both samples from ischemic mice. Int. J. Mol. Sci. 2022, 23, 8093 In this case, Transforming Growth Factor-α (TGF-α) in its proform (proTGF-α), C-C 6 of 12 Motif Chemokine Ligand 2 (CCL2), and dimethylarginine dimethylaminohydrolase 1 (DDH1) were identified as the most significant proteins of this analysis. Table 1 shows the logFCs and p-values of the differentially-expressed proteins plotted in Figure 2B. (DDH1) were identified as the most significant proteins of this analysis. Table 1 shows the Next, the TGF-α and CCL2 findings were further verified by an alternative tech- logFCs and p-values of the differentially-expressed proteins plotted in Figure 2B. nique (ELISA) in the same lung homogenate samples. A slight increase in CCL2 was Next, the TGF-α and CCL2 findings were further verified by an alternative technique confirmed by ELISA in ischemic compared with sham animals. However, this difference (ELISA) in the same lung homogenate samples. A slight increase in CCL2 was confirmed did not reach statistical significance (2487 (±508.4) pg/mL vs. 2145 (±459.5) pg/mL, p = by ELISA in ischemic compared with sham animals. However, this difference did not reach 0.14) (Figure statistical 3C). A positive significance (2487 correlation was also (±508.4) pg/mL observed vs. 2145 between (±459.5) the pELISA pg/mL, = 0.14)and Olink (Figure ® 3C). values, with a correlation coefficient of 0.66 (p < 0.0001) (Figure 3D). A positive correlation was also observed between the ELISA and Olink® values, with a Regarding correlation TGF-α, of coefficient the0.66 ELISA (p