0
Original Research: Pulmonary Vascular Disease |

Discordant Regulation of microRNA Between Multiple Experimental Models and Human Pulmonary HypertensionmicroRNA Discordance in Pulmonary Hypertension OPEN ACCESS

Kenny Schlosser, PhD; Mohamad Taha, BSc; Yupu Deng, MD; Baohua Jiang, MD, PhD; Duncan J. Stewart, MD
Author and Funding Information

From the Regenerative Medicine Program (Drs Schlosser, Deng, Jiang, and Stewart and Mr Taha), Ottawa Hospital Research Institute; and the Department of Cellular and Molecular Medicine (Mr Taha and Dr Stewart), University of Ottawa, Ottawa, ON, Canada.

CORRESPONDENCE TO: Duncan J. Stewart, MD, The Ottawa Hospital, Critical Care Wing, Room 6117 - Box 511, 501 Smyth Rd, Ottawa, ON, K1H 8L6, Canada; e-mail: djstewart@ohri.ca


FUNDING/SUPPORT: This work was supported by the Canadian Institutes of Health Research [Grant MOP57726 to D. J. S.], CREST Ontario Research Excellence Fund [Grant GL2-01-042 to D. J. S.), Heart and Stroke Foundation of Canada [K. S.], and the Entelligence Young Investigator’s grant from Actelion Pharmaceuticals US, Inc [K. S.].

This is an open access article distributed under the terms of the Creative Commons Attribution-Noncommercial License (http://creativecommons.org/licenses/by-nc/3.0/), which permits unrestricted use, distribution, and reproduction to noncommercial entities, provided the original work is properly cited. Information for reuse by commercial entities is available online.


Chest. 2015;148(2):481-490. doi:10.1378/chest.14-2169
Text Size: A A A
Published online

BACKGROUND:  The dysregulation of microRNA (miRNA) is known to contribute to the pathobiology of pulmonary arterial hypertension (PAH). However, the relationships between changes in tissue and circulating miRNA levels associated with different animal models and human pulmonary hypertension (PH) have not been defined.

METHODS:  A set of miRNAs that have been causally implicated in PH, including miR-17, -21, -130b, -145, -204, -424, and -503, were measured by reverse transcription-quantitative polymerase chain reaction in the plasma, lung, and right ventricle of three of the most common rodent models of PH: the rat monocrotaline and SU5416 plus chronic hypoxia (SuHx) models and the mouse chronic hypoxia model. Plasma miRNA levels were also evaluated in a cohort of patients with PAH and healthy subjects.

RESULTS:  Several miRNA showed PH model-dependent perturbations in plasma and tissue levels; however, none of these were conserved across all three experimental models. Principle component analysis of miR expression changes in plasma revealed distinct clustering between rodent models, and SuHx-triggered PH showed the greatest similarity to human PAH. Changes in the plasma levels of several miRNA also correlated with changes in tissue expression. In particular, miR-424 was concordantly increased (1.3- to 1.5-fold, P < .05) in the plasma, lung, and right ventricle of hypoxic mice and in the plasma of patients with PAH.

CONCLUSIONS:  miRNAs with established etiologic roles in PH showed context-dependent changes in tissue and circulating levels, which were not consistent across rodent models and human PAH. This suggests different miRNA-dependent mechanisms may contribute to experimental and clinical PH, complicating potential diagnostic and therapeutic applications amenable to these miRNAs.

Figures in this Article

microRNAs (miRNAs) are short (approximately 22 nt) non-protein-coding RNA molecules that regulate gene expression at the posttranscriptional level by directing the translational inhibition or degradation of cognate messenger RNAs. To date, > 2,500 mature human miRNAs have been reported,1 which are predicted to control the majority of protein-coding genes.2 The dysregulated cellular expression of miRNAs has been implicated in the etiology of many diseases,3 including, more recently, pulmonary arterial hypertension (PAH).4 This progressive disease is characterized by changes in the structure and function of the pulmonary vasculature, which precipitates right-sided heart failure.5

A number of miRNAs, including miR-17, miR-21, miR-130, miR-145, miR-204, miR-424, and miR-503, have been demonstrated to be causally involved in the development of pulmonary hypertension (PH).612 The aberrant expression of these miRNAs in lung vascular cells is known to contribute to the dysregulation in proliferation, apoptosis, and inflammation that underlie PH.4 Moreover, these miRNAs have been proposed as potential targets for therapeutic intervention.612 Indeed, strategies designed to reverse pathologic changes in lung miRNA expression have previously shown promise in attenuating hemodynamic and/or vascular remodeling abnormalities in preclinical animal models of PH.612 Although experimental PH induced by the combination of SU5416 and chronic hypoxia (SuHx) in rats has been suggested to most closely model human PAH,13 miRNAs have mainly been assessed in the monocrotaline (MCT) or chronic hypoxia (Hx) models. Although these models can reproduce important features of human PAH, they generally lack the severe pulmonary arteriopathy observed in patients with PAH. Whether the regulation of miRNA is conserved between these different experimental models has not yet been established, but this may provide important clues to help differentiate model-specific alterations from changes that may relate to more central mechanisms in the pathobiology of PH.

We analyzed the expression of these seven miRNAs across three of the most commonly used rodent models of PH, including MCT rats, SuHx rats, and Hx mice. The expression patterns were assessed in the lung as well as two other biologic sources that have largely been overlooked in previous studies, the right ventricle and plasma. Of note, plasma has become increasingly recognized as a key source of extracellular miRNAs, which have the potential to serve as practical noninvasive biomarkers of disease.14 This application would be particularly useful in PAH given the difficulty associated with the early diagnosis of this condition.15 Moreover, the convenience of plasma collection circumvents some of the inherent challenges of obtaining clinically relevant biologic specimens from patients, which can facilitate broader studies on the pathobiology of PAH and provide direct insight into the human disease. We, therefore, also examined the expression pattern of these seven miRNAs in a cohort of patients with PAH and healthy subjects.

We now report a surprising lack of concordance in the pattern of miRNA expression between different rodent models of PH and between experimental and human PH. These findings have important implications for the potential usefulness of miRNAs identified in experimental PH for application as biomarkers and therapeutic targets in human PAH. In addition, our results provide new molecular evidence supporting SuHx-induced PH as a more relevant model of human PAH.

Peripheral blood samples from patients with PAH and healthy subjects were obtained with informed consent, under a protocol approved by the Ottawa Hospital Research Ethics Board (#2011470-01H), as previously described.16 All animal procedures were approved by the University of Ottawa’s Animal Care Ethics Committee and complied with the principles and guidelines of the Canadian Council on Animal Care. A detailed description of the inclusion/exclusion criteria for human subjects, animal models, plasma isolation, total RNA extraction, RNA quality control assessment, reverse transcription-quantitative polymerase chain reaction, reference control characterization (e-Fig 1), and statistics is provided in e-Appendix 1.

Basal miRNA Expression Patterns Largely Conserved Between Species

Plasma levels of individual miRNA varied by nearly three orders of magnitude and generally followed the rank order: miR-204 < miR-503 < miR-130b < miR-424 < miR-17 < miR-145 < miR-21 (Fig 1A). Interestingly, this pattern was largely conserved across humans, rats, and mice under physiologic conditions, with the exception of miR-145, which ranked below miR-424 in human plasma (Fig 1A). miRNA expression patterns in the lung and right ventricle were also comparable between rats and mice (Fig 1A), and only modest differences in the rank order of miRNA expression were evident between plasma and tissues. Overall, a strong positive correlation was observed between plasma miRNA levels and their corresponding expression levels within the lung (Pearson r = 0.89-0.90, P < .01 for rats; and Pearson r = 0.95, P < .01 for mice) and right ventricle (Pearson r = 0.87-0.88, P < .05 for rats; and Pearson r = 0.80, P < .05 for mice) (Fig 1B). This correlation was consistently observed across rats (both Fischer and Sprague-Dawley strains) and mice.

Figure Jump LinkFigure 1 –  A, B, Relative levels of miRNAs in plasma, lung, and RV across humans, rats, and mice. A, Plasma miRNA levels were measured by reverse transcription-quantitative polymerase chain reaction (RT-qPCR) and normalized to internal reference controls (miR-142-3p and miR-320a for humans and miR-24-3p for rats and mice). Data were scaled to the lowest miRNA in each group (mean level from approximately two- to 21-fold above assay detection limit), which was set to one. miRNA levels in the lung and RV were normalized to RNU6-2 and scaled to the lowest miRNA in each group (mean level at least 50-fold above the assay detection limit). Data are presented as mean ± range for n = 13 healthy human subjects, n = 10 control CDF rats, n = 13 control S-D rats, and n = 11 control C57Bl6J mice. B, Correlation between relative miRNA levels in plasma and either lung or RV. Pearson correlation coefficients (r) are shown for log transformed data (*P < .05, **P < .01). Each symbol represents the mean level of one of the seven miRNAs under investigation. CDF = Fischer; miRNA = microRNA; RV = right ventricle; S-D = Sprague-Dawley.Grahic Jump Location
Predominant miRNA Downregulation in MCT Rats

Rats received a single intraperitoneal injection of MCT (or vehicle control), and right ventricular systolic pressure (RVSP) and right ventricular hypertrophy (right ventricle to left ventricle plus septum [RV/(LV + S)] mass ratio) were significantly increased 3 weeks post MCT injection (Figs 2A, 2B). The plasma levels of both miR-322 (the rat homolog of human miR-424) and miR-503 were significantly decreased in MCT rats relative to control rats (1.3-fold down, P < .01 and 1.8-fold down, P < .001, respectively) (Fig 2C), which mimicked a significant decline in lung tissue expression (3.7-fold down, P < .0001 and 5.6-fold down, P < .0001, respectively). miR-204 was decreased in lung tissue in MCT rats (1.9-fold down, P < .01) and showed a trend (P = .097) toward lower levels in the plasma. miR-145 was also decreased in the lung (2.0-fold, P < .0001) and right ventricle (1.3-fold, P < .05) of MCT rats, whereas no change in plasma level was detected. miR-130b was decreased in the lung of MCT rats (1.7-fold down, P < .01) but showed a trend (P = .07) toward higher levels in plasma. Only miR-21 was increased, showing trends toward higher levels in plasma (P = .07) and a significant increase in lung tissue expression (3.2-fold up, P < .0001). The plasma, lung, and right ventricle levels of miR-17 were not altered in MCT rats.

Figure Jump LinkFigure 2 –  A-C, MCT rat model of pulmonary hypertension. A, Experimental outline. Male CDF rats were given a single intraperitoneal injection of MCT or vehicle control (n = 10-11/group). Plasma, lung, and RV tissue were collected 3 wk later. B, RVSP and RV/(LV + S) mass ratio. Data are presented as boxplots with whiskers showing minimum-maximum range. C, Relative miRNA levels were measured by RT-qPCR. Plasma miRNA levels were normalized to an internal reference control (miR-24-3p), and expressed relative to the control group. miRNA levels in the lung and RV were normalized to RNU6-2 and expressed relative to the control group. Data are presented as mean ± SEM #P < .1, *P < .05, **P < .01, ***P < .001, ****P < .0001 vs control group. miR-322 is the rat homolog of human miR-424. Con = control; IP = intraperitoneal; LV = left ventricle; MCT = monocrotaline; RV/(LV + S) = right ventricle to left ventricle plus septum; RVSP = right ventricular systolic pressure. See Figure 1 legend for expansion of other abbreviations.Grahic Jump Location
Modest miRNA Changes in SU5416/Hx Rats

Rats received a single subcutaneous injection of the vascular endothelial growth factor receptor-2 inhibitor, SU5416 (or vehicle control) and were exposed to hypoxic conditions (10% oxygen) for a period of 3 weeks, followed by 5 weeks of normoxic conditions (Fig 3A). Plasma, lung, and right ventricle miRNA levels were measured after 8 weeks, at which time there was a significant increase in RVSP and RV/(LV + S) mass ratio (Fig 3B). Surprisingly, despite the severity of the PH phenotype, levels of these specific miRNAs were not significantly altered in plasma or lung between SuHx and control groups (Fig 3C), and only miR-145 showed a trend toward increased levels in the lung. The only significant changes in tissue miR expression were seen in the right ventricle of SuHx rats, with a twofold increase in miR-21 (P < .01) and twofold decrease in miR-204 (P < .05) (Fig 3C).

Figure Jump LinkFigure 3 –  A-C, SuHx rat model of pulmonary hypertension. A, Experimental outline. At d 0, male S-D rats were given a single SC injection of SU5416 or vehicle control (n = 9-13/group). Rats were housed under hypoxic conditions for the first 3 wk and then held in normoxic conditions for another 5 wk until study end point. Plasma, lung, and RV tissue were collected at end point (8 wk after SU5416 injection). B, RVSP and RV/(LV + S) mass ratio. Data are presented as boxplots with whiskers showing minimum-maximum range. C, Relative miRNA levels were measured by RT-qPCR. Plasma miRNA levels were normalized to an internal reference control (miR-24-3p) and expressed relative to the control group. miRNA levels in the lung and RV were normalized to RNU6-2 and expressed relative to the control group. Data are presented as mean ± SEM. #P < .1, *P < .05, **P < .01 vs control group. miR-322 is the rat homolog of human miR-424. O2 = oxygen; SC = subcutaneous; SuHx = SU5416 plus chronic hypoxia. See Figure 1 and 2 legends for expansion of other abbreviations.Grahic Jump Location
miRNA Upregulation in Hx Mice

Mice were exposed to Hx for 3 weeks (Fig 4A), after which time there were modest increases in RVSP and RV/(LV + S) mass ratio (Fig 4B). In contrast to the MCT model, a general pattern of upregulation was observed among the miRNAs that were altered in hypoxic mice. One miRNA in particular, miR-322 (the mouse/rat homolog of human miR-424), was shown to be significantly and concordantly higher in the plasma (1.3-fold, P < .05), lung (1.4-fold, P < .001), and right ventricle (1.4-fold, P < .05) of hypoxic mice vs control mice (Fig 4C). The expression of miR-503 was also significantly higher in the lung and right ventricle of Hx mice, but plasma levels were not significantly altered. The plasma levels of miR-17 were significantly elevated in Hx mice (1.4-fold, P < .01), whereas no changes in expression were observed in the lung or right ventricle. A trend (P < .1) toward higher plasma levels of miR-204 was evident in Hx mice, but no significant difference in lung or right ventricle tissue was detected. Both miR-21 and miR-145 were increased in the lungs of Hx mice (1.6- and 1.4-fold, respectively; P < .001), with no change in plasma or right ventricle levels. No significant changes in the plasma, lung, or right ventricle levels of miR-130b were observed.

Figure Jump LinkFigure 4 –  A-C, Hx mouse model of pulmonary hypertension. A, Experimental outline. Male C57Bl6J mice were housed under hypoxic or control normoxic conditions (n = 10-11/group) for 3 wk after which time plasma, lung, and RV tissue were collected. B, RVSP and RV/(LV + S) mass ratio. Data are presented as boxplots with whiskers showing minimum-maximum range. C, Relative miRNA levels were measured by RT-qPCR. Plasma miRNA levels were normalized to an internal reference control (miR-24-3p) and expressed relative to the control group. miRNA levels in the lung and RV were normalized to RNU6-2 and expressed relative to the control group. Data are presented as mean ± SEM #P < .1, *P < .05, **P < .01, ***P < .001 vs control group. miR-322 is the mouse homolog of human miR-424. Hx = chronic hypoxia. See Figure 13 legends for expansion of other abbreviations.Grahic Jump Location
Principle Component Analysis Underscores Differences Between Models

To gain further molecular insight into the relative similarity or dissimilarity between experimental models, principle component analysis (PCA) was conducted on miRNA expression changes in plasma, lung, and right ventricle. The 21-dimensional miRNA expression space (defined by changes in the seven target miRNAs across three different biologic sources) was linearly transformed and mapped into just two dimensions (ie, principle components) to facilitate visualization and comparison of the variability among individual rodents within and between each PH model (Fig 5A). The distance between any pair of individuals in the PCA plot is related to the degree of similarity between the two individuals in high-dimensional space. The animals in each experimental model formed distinct clusters, with some overlap observed among rodents exposed to hypoxia alone or in combination with SU5416. Animals in the MCT and hypoxia models showed the greatest separation in principle component space, which contrasted with their relative closer proximity in a biplot comparison of their phenotypic abnormalities in RVSP and RV/(LV + S) mass ratio (Fig 5B).

Figure Jump LinkFigure 5 –  A, B, PCA mapping of miRNA expression changes in three rodent models of pulmonary hypertension. A, Hx (), SuHx (), and MCT () denote individual animals from each experimental model. For each animal, the change in seven target miRNAs measured across three different biologic sources (plasma, lung, and RV) were mapped to a position in two-dimensional space (ie, PCs) for easy visualization. The distance between any pair of animals in the PCA plot is related to the degree of similarity between the two observations in higher dimensional space. A total of 47.1% of the entire information content of the 21-dimensional miRNA expression space is captured in the two-dimensional PCA plot, with 32.4% and 14.7% of the variability in data explained by PC1 and PC2, respectively. Ellipses denote 1.5 SDs, and line vectors connect to the centroid. B, Biplot showing distribution of individual animals from various pulmonary hypertension models according to their RV/(LV + S) and RVSP. Data are presented as fold change to mean baseline level in control animals. PC = principle component; PCA = principle component analysis. See Figure 14 legends for expansion of other abbreviations.Grahic Jump Location

To gain insight into whether the discordance in miRNA regulation between models was also reflected in the regulation of downstream messenger RNA targets, we examined two specific messenger RNAs (mRNAs) (Fgf2 and Rhob) that are known to be direct targets of miR-424/503 and miR-21, respectively7,8 (e-Fig 2). Fgf2 transcript levels were significantly elevated in lung tissue from both rat models of PH (1.4-fold in SuHx and 1.5-fold in MCT, P < .05), but no significant change was observed in the lungs of hypoxic mice vs control mice. The mRNA level of Rhob was downregulated in SuHx rat lungs (1.2-fold, P < .05), whereas no significant change was observed in the lungs of MCT rats or Hx mice. Of note, the increased Fgf2 level in the lungs of MCT rats was consistent with the observed decrease in miR-322/503 in this model. However, this expected inverse relationship between mRNA target and miRNA was not directly observed in other models or between Rhob and miR-21 (e-Fig 2).

PCA Exposes Relative Similarity Between Human and SuHx-Triggered PH

To gain insight into which experimental trigger of PH most closely models the molecular changes associated with human PAH, target miRNAs were also measured in plasma from 14 patients with PAH (58 ± 10 years, 57% women) and 13 healthy control subjects (45 ± 13 years, 69% women) (Fig 6A). The PAH patient cohort was composed of both idiopathic (43%) and associated forms (57%) of the disease, of which the majority (79%) were on PH-specific therapy at the time of blood sampling. Detailed clinical characteristics of this patient cohort have been described previously16 (e-Table 1). The plasma level of miR-17 was decreased in patients with PAH (1.4-fold, P < .05), and miR-145 also showed a trend toward lower levels (1.3-fold, P < .1). An increase in miR-424 was observed in patients with PAH (1.5-fold, P < .05), whereas levels of miR-21, -130b, -204, and -503 were not significantly different between patients with PAH and healthy control subjects. However, it should be noted that these miRs were originally linked to PH by interrogation of other biologic samples, including cell and lung tissue specimens,8,9 or by comprehensive in silico network analyses.7 PCA of the plasma changes in these seven miRNAs exposed a relative similarity in the underlying miR regulation between patients with PAH and rodents with SuHx-triggered PH (Fig 6B). In principle component space, SuHx rats showed the closest overlap with patients with PAH and also exhibited a similar degree of heterogeneity in the miRNA response among individuals. In comparison, rodents from the MCT and Hx models were distributed in tighter clusters within the PCA plot.

Figure Jump LinkFigure 6 –  A-C, Assessment of target miRNAs in plasma of patients with PAH. A, miRNAs were measured by RT-qPCR in total RNA extracted from plasma of 13 healthy control subjects and 14 patients with PAH. miRNA levels were normalized to the geomean of two internal reference controls (miR-142-3p and -320a) and then expressed relative to the control group. #P < .1, *P < .05 vs control group. Data are represented as mean ± SEM. B, PCA of plasma miRNA changes in patients with PAH (), Hx mice (), SuHx rats (), and MCT rats (). A total of 50.9% of the information content associated with the seven-dimensional plasma miRNA space is captured in the two-dimensional PCA plot. Ellipses denote 1 SD, and line vectors connect to the centroid. C, Receiver operating characteristic curves of miR-424 and miR-17 plasma levels showing the trade-off between sensitivity and specificity in distinguishing patients with PAH from healthy control subjects. PAH = pulmonary arterial hypertension. See Figure 15 legends for expansion of other abbreviations.Grahic Jump Location
Potential Diagnostic Biomarkers of PAH

For miRNAs that were altered in the plasma of patients with PAH, receiver operating characteristic (ROC) curve analysis was conducted to gauge their potential diagnostic usefulness. miR-424 and miR-17 both exhibited significant areas under their corresponding ROC curves (area under the ROC curve of 0.76 and 0.73, respectively; P < .05) (Fig 6C) and exhibited the capacity for high sensitivity (0.93 for both miR-424 and miR-17; 95% CI, 0.66-1.00) and moderate specificity (0.69; 95% CI, 0.39-0.91 and 0.62; 95% CI, 0.32-0.86, respectively) in discriminating between patients with PAH and healthy participants. However, no significant correlation was observed between plasma miRNA level and various physiologic indexes of disease severity in this small cohort, including mean pulmonary artery pressure, pulmonary vascular resistance index, and the 6-min walk distance test (data not shown).

This study provides one of the most extensive assessments of PH-induced changes in miRNA expression reported to date. The expression patterns of seven miRNAs were analyzed in the plasma, lung, and right ventricle of three of the most commonly used rodent models of PH and in plasma from patients with PAH. We focused specifically on a set of miRNAs that have previously been implicated causally in PAH, based on supporting evidence that included loss- and/or gain-of-function experiments in cells and/or specific animal models.610,12 In contrast to previous studies that have investigated the specific molecular mechanisms of individual miRNAs, the current study now helps to answer broader questions about the nature of their regulation across different biologic sources, species, and experimental vs clinical forms of PH. This may provide new perspectives on the molecular pathobiology of PH and important insight on the translational potential of animal studies and the related merits of therapeutic strategies aimed at reversing pathologic changes in miRNA expression.

We found that each of the target miRNAs was detectable in the plasma of mice, rats, and humans. Moreover, the relative rank order of plasma miR levels was highly conserved across species, suggesting that similar mechanisms may control the basal release and clearance of these miRNAs from circulation across these species. Of note, the plasma miR levels in three separate animal models consistently showed strong positive correlations with miRNA levels in the lung. Although not definitive evidence, this correlation does suggest that circulating miR levels may be driven at least in part by the magnitude of lung expression, consistent with the fact that the lung represents the largest vascular bed with the greatest opportunity to release or clear miRNAs across the pulmonary circulation.

Interestingly, none of the PH-associated changes in plasma miRNA levels were common between two or more of the experimental models (Table 1). However, these discordant patterns were not limited to circulating miRNAs, as both lung and RV miRNA expression showed distinct model-specific patterns of change. The known mRNA targets of two miRNAs were also examined in the lungs of each model. Although confirming the expected inverse regulation between miR-322/503 and Fgf2 in the MCT model, there was also evidence of discordance in other models and for other miRNA targets. This suggests that in these cases miRNA may not be the sole or even the most important mechanism of regulation for these targets. In previous studies, miRNA expression changes associated with experimental PH have commonly been evaluated in lung tissue, and our assessment of miRNA expression changes were generally quite comparable to previous reports611,17 (e-Table 2); however, there were a few exceptions that we speculate could be attributed to differences in the experimental time course and severity of PH or possibly other study-specific factors (eg, animal strain) (e-Table 2).

Table Graphic Jump Location
TABLE 1 ]  Summary of Changes in miRNA Levels Associated With Human and Experimental PH

Fold increase or decrease (vs control group). n denotes sample size/group. P < .05 unless otherwise noted. miRNA = microRNA; ns = not significant; PH = pulmonary hypertension; RV = right ventricle.

a 

P < .1 (trend).

Of note, within each model system we observed similar changes in the PH-associated miRNA levels in plasma and tissues. A general pattern of downregulation was observed among miRNAs that were altered in the MCT model (ie, five of six altered miRs), which contrasted with a general pattern of upregulation among the miRNAs that were altered under hypoxic conditions. The divergence in miR regulation between these models was further supported by PCA, in which rodents with MCT-induced PH clearly formed a distinct cluster separate from rodents with hypoxia-induced PH. It is intriguing that the same miRNAs are differentially regulated in these two models of PH. Certainly, Hx and MCT produce distinct PH phenotypes that likely involve different molecular mechanisms; however, it remains possible that some changes may reflect off-target effects of the specific stimuli used to induce PH. In contrast, we found only modest changes in the mean miRNA levels in the SuHx model of PH, despite far greater abnormalities in hemodynamics and RV remodeling. Nevertheless, the distribution of individual animals/patients revealed in the PCA plot of changes in plasma miRNA suggests this model may more closely resemble human PAH than the MCT and hypoxia models. This is consistent with the similarities in cellular and histologic abnormalities (ie, plexiform lesions) and progressive irreversible phenotype that have previously supported the SuHx model as most relevant to the human disease.18

miR-424 was the only miRNA to show concordant changes in the plasma of humans and at least one experimental model of PH, Hx. In hypoxic mice, elevated plasma levels of miR-322 (the mouse analog of human miR-424) also mirrored increased expression within the lungs and right ventricle, suggesting that altered plasma levels may be a surrogate marker of disease activity in damaged/remodelled tissues. To our knowledge, the only other miRNA previously reported to be concordantly altered in the plasma of human PAH and a rodent model of PH is miR-26a, which was decreased in the plasma, lung, and right ventricle of MCT rats,16 and in the plasma16 and cellular buffy coat19 of patients with PAH. However, unlike miR-424, it remains to be established whether miR-26a is causally involved in the development of PH in the MCT model, which is currently an area of investigation in our laboratory.

PAH is a rare disease with prevalence estimated at only 15 to 26 cases/million,20,21 making large cohort studies inherently challenging. Although the sample size of this human patient cohort is limited, it nonetheless can provide some relevant insight into the relative performance of these miRs as potential biomarkers of PH. Of the seven miRNAs investigated in this study, only miR-424 and miR-17 exhibited significant differences in plasma levels between patients with PAH and healthy control subjects. The plasma levels of both miRNAs showed significant areas under ROC curves, indicative of a good balance between sensitivity and specificity in discriminating patients with PAH from healthy subjects.

This study has some limitations, including the small sample size and heterogeneity of subjects in the PAH cohort. The diagnostic usefulness of possible miRNA-based biomarkers will clearly need to be validated in larger prospective cohort studies. It is also uncertain what effect patients with different forms of PAH and therapy may have on comparisons with the more well-defined experimental forms of PH induced in the animal models. Finally, it is important to note that cell-specific changes in miRNA (or mRNA) expression, which may underlie disease activity, could be obscured by the complex cellular background in tissue measurements.

Despite their previously established roles in the pathobiology of experimental PH, the majority of miRNAs investigated in this study showed limited potential as circulating biomarkers of this disease. In addition, three separate experimental models of PH showed marked disparities in their pattern of miRNA expression changes in tissue and plasma. These differences have important implications for how preclinical models are leveraged to gain insight into the molecular mechanisms of human PH.

Author contributions: K. S. and D. J. S. are guarantors of the manuscript. K. S. contributed to the study conception/design, data acquisition/analysis/interpretation, writing/revision of the manuscript, and final approval of the manuscript; M. T., Y. D., and B. J. contributed to study design, data acquisition/analysis, manuscript revision, and final approval of the manuscript; and D. J. S. contributed to study design, data interpretation, manuscript writing/revision, and final approval of the manuscript.

Financial/nonfinancial disclosures: The authors have reported to CHEST that no potential conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Role of sponsors: The sponsors had no role in the design of the study, the collection and analysis of the data, or the preparation of the manuscript.

Additional information: The e-Appendix, e-Figures, and e-Tables can be found in the Supplemental Materials section of the online article.

Hx

chronic hypoxia

MCT

monocrotaline

miRNA

microRNA

mRNA

messenger RNA

PAH

pulmonary arterial hypertension

PCA

principle component analysis

PH

pulmonary hypertension

ROC

receiver operating characteristic

RV/(LV + S)

right ventricle to left ventricle plus septum

RVSP

right ventricular systolic pressure

SuHx

SU5416 plus chronic hypoxia

Kozomara A, Griffiths-Jones S. miRBase: annotating high confidence microRNAs using deep sequencing data. Nucleic Acids Res. 2014;42(database issue):D68-D73. [CrossRef] [PubMed]
 
Friedman RC, Farh KK, Burge CB, Bartel DP. Most mammalian mRNAs are conserved targets of microRNAs. Genome Res. 2009;19(1):92-105. [CrossRef] [PubMed]
 
Jiang Q, Wang Y, Hao Y, et al. miR2Disease: a manually curated database for microRNA deregulation in human disease. Nucleic Acids Res. 2009;37(database issue):D98-D104. [CrossRef] [PubMed]
 
White K, Loscalzo J, Chan SY. Holding our breath: the emerging and anticipated roles of microRNA in pulmonary hypertension. Pulm Circ. 2012;2(3):278-290. [CrossRef] [PubMed]
 
Schermuly RT, Ghofrani HA, Wilkins MR, Grimminger F. Mechanisms of disease: pulmonary arterial hypertension. Nat Rev Cardiol. 2011;8(8):443-455. [CrossRef] [PubMed]
 
Pullamsetti SS, Doebele C, Fischer A, et al. Inhibition of microRNA-17 improves lung and heart function in experimental pulmonary hypertension. Am J Respir Crit Care Med. 2012;185(4):409-419. [CrossRef] [PubMed]
 
Parikh VN, Jin RC, Rabello S, et al. MicroRNA-21 integrates pathogenic signaling to control pulmonary hypertension: results of a network bioinformatics approach. Circulation. 2012;125(12):1520-1532. [CrossRef] [PubMed]
 
Kim J, Kang Y, Kojima Y, et al. An endothelial apelin-FGF link mediated by miR-424 and miR-503 is disrupted in pulmonary arterial hypertension. Nat Med. 2013;19(1):74-82. [CrossRef] [PubMed]
 
Courboulin A, Paulin R, Giguère NJ, et al. Role for miR-204 in human pulmonary arterial hypertension. J Exp Med. 2011;208(3):535-548. [CrossRef] [PubMed]
 
Caruso P, Dempsie Y, Stevens HC, et al. A role for miR-145 in pulmonary arterial hypertension: evidence from mouse models and patient samples. Circ Res. 2012;111(3):290-300. [CrossRef] [PubMed]
 
Yang S, Banerjee S, Freitas Ad, et al. miR-21 regulates chronic hypoxia-induced pulmonary vascular remodeling. Am J Physiol Lung Cell Mol Physiol. 2012;302(6):L521-L529. [CrossRef] [PubMed]
 
Bertero T, Lu Y, Annis S, et al. Systems-level regulation of microRNA networks by miR-130/301 promotes pulmonary hypertension. J Clin Invest. 2014;124(8):3514-3528. [CrossRef] [PubMed]
 
Abe K, Toba M, Alzoubi A, et al. Formation of plexiform lesions in experimental severe pulmonary arterial hypertension. Circulation. 2010;121(25):2747-2754. [CrossRef] [PubMed]
 
Gupta SK, Bang C, Thum T. Circulating microRNAs as biomarkers and potential paracrine mediators of cardiovascular disease. Circ Cardiovasc Genet. 2010;3(5):484-488. [CrossRef] [PubMed]
 
Traill EM. Pulmonary arterial hypertension: an early diagnosis makes a difference. JAAPA. 2011;24(5):30-34. [PubMed]
 
Schlosser K, White RJ, Stewart DJ. miR-26a linked to pulmonary hypertension by global assessment of circulating extracellular microRNAs. Am J Respir Crit Care Med. 2013;188(12):1472-1475. [CrossRef] [PubMed]
 
Caruso P, MacLean MR, Khanin R, et al. Dynamic changes in lung microRNA profiles during the development of pulmonary hypertension due to chronic hypoxia and monocrotaline. Arterioscler Thromb Vasc Biol. 2010;30(4):716-723. [CrossRef] [PubMed]
 
Maarman G, Lecour S, Butrous G, Thienemann F, Sliwa K. A comprehensive review: the evolution of animal models in pulmonary hypertension research; are we there yet? Pulm Circ. 2013;3(4):739-756. [CrossRef] [PubMed]
 
Wei C, Henderson H, Spradley C, et al. Circulating miRNAs as potential marker for pulmonary hypertension. PLoS ONE. 2013;8(5):e64396. [CrossRef] [PubMed]
 
Peacock AJ, Murphy NF, McMurray JJ, Caballero L, Stewart S. An epidemiological study of pulmonary arterial hypertension. Eur Respir J. 2007;30(1):104-109. [CrossRef] [PubMed]
 
Humbert M, Sitbon O, Chaouat A, et al. Pulmonary arterial hypertension in France: results from a national registry. Am J Respir Crit Care Med. 2006;173(9):1023-1030. [CrossRef] [PubMed]
 

Figures

Figure Jump LinkFigure 1 –  A, B, Relative levels of miRNAs in plasma, lung, and RV across humans, rats, and mice. A, Plasma miRNA levels were measured by reverse transcription-quantitative polymerase chain reaction (RT-qPCR) and normalized to internal reference controls (miR-142-3p and miR-320a for humans and miR-24-3p for rats and mice). Data were scaled to the lowest miRNA in each group (mean level from approximately two- to 21-fold above assay detection limit), which was set to one. miRNA levels in the lung and RV were normalized to RNU6-2 and scaled to the lowest miRNA in each group (mean level at least 50-fold above the assay detection limit). Data are presented as mean ± range for n = 13 healthy human subjects, n = 10 control CDF rats, n = 13 control S-D rats, and n = 11 control C57Bl6J mice. B, Correlation between relative miRNA levels in plasma and either lung or RV. Pearson correlation coefficients (r) are shown for log transformed data (*P < .05, **P < .01). Each symbol represents the mean level of one of the seven miRNAs under investigation. CDF = Fischer; miRNA = microRNA; RV = right ventricle; S-D = Sprague-Dawley.Grahic Jump Location
Figure Jump LinkFigure 2 –  A-C, MCT rat model of pulmonary hypertension. A, Experimental outline. Male CDF rats were given a single intraperitoneal injection of MCT or vehicle control (n = 10-11/group). Plasma, lung, and RV tissue were collected 3 wk later. B, RVSP and RV/(LV + S) mass ratio. Data are presented as boxplots with whiskers showing minimum-maximum range. C, Relative miRNA levels were measured by RT-qPCR. Plasma miRNA levels were normalized to an internal reference control (miR-24-3p), and expressed relative to the control group. miRNA levels in the lung and RV were normalized to RNU6-2 and expressed relative to the control group. Data are presented as mean ± SEM #P < .1, *P < .05, **P < .01, ***P < .001, ****P < .0001 vs control group. miR-322 is the rat homolog of human miR-424. Con = control; IP = intraperitoneal; LV = left ventricle; MCT = monocrotaline; RV/(LV + S) = right ventricle to left ventricle plus septum; RVSP = right ventricular systolic pressure. See Figure 1 legend for expansion of other abbreviations.Grahic Jump Location
Figure Jump LinkFigure 3 –  A-C, SuHx rat model of pulmonary hypertension. A, Experimental outline. At d 0, male S-D rats were given a single SC injection of SU5416 or vehicle control (n = 9-13/group). Rats were housed under hypoxic conditions for the first 3 wk and then held in normoxic conditions for another 5 wk until study end point. Plasma, lung, and RV tissue were collected at end point (8 wk after SU5416 injection). B, RVSP and RV/(LV + S) mass ratio. Data are presented as boxplots with whiskers showing minimum-maximum range. C, Relative miRNA levels were measured by RT-qPCR. Plasma miRNA levels were normalized to an internal reference control (miR-24-3p) and expressed relative to the control group. miRNA levels in the lung and RV were normalized to RNU6-2 and expressed relative to the control group. Data are presented as mean ± SEM. #P < .1, *P < .05, **P < .01 vs control group. miR-322 is the rat homolog of human miR-424. O2 = oxygen; SC = subcutaneous; SuHx = SU5416 plus chronic hypoxia. See Figure 1 and 2 legends for expansion of other abbreviations.Grahic Jump Location
Figure Jump LinkFigure 4 –  A-C, Hx mouse model of pulmonary hypertension. A, Experimental outline. Male C57Bl6J mice were housed under hypoxic or control normoxic conditions (n = 10-11/group) for 3 wk after which time plasma, lung, and RV tissue were collected. B, RVSP and RV/(LV + S) mass ratio. Data are presented as boxplots with whiskers showing minimum-maximum range. C, Relative miRNA levels were measured by RT-qPCR. Plasma miRNA levels were normalized to an internal reference control (miR-24-3p) and expressed relative to the control group. miRNA levels in the lung and RV were normalized to RNU6-2 and expressed relative to the control group. Data are presented as mean ± SEM #P < .1, *P < .05, **P < .01, ***P < .001 vs control group. miR-322 is the mouse homolog of human miR-424. Hx = chronic hypoxia. See Figure 13 legends for expansion of other abbreviations.Grahic Jump Location
Figure Jump LinkFigure 5 –  A, B, PCA mapping of miRNA expression changes in three rodent models of pulmonary hypertension. A, Hx (), SuHx (), and MCT () denote individual animals from each experimental model. For each animal, the change in seven target miRNAs measured across three different biologic sources (plasma, lung, and RV) were mapped to a position in two-dimensional space (ie, PCs) for easy visualization. The distance between any pair of animals in the PCA plot is related to the degree of similarity between the two observations in higher dimensional space. A total of 47.1% of the entire information content of the 21-dimensional miRNA expression space is captured in the two-dimensional PCA plot, with 32.4% and 14.7% of the variability in data explained by PC1 and PC2, respectively. Ellipses denote 1.5 SDs, and line vectors connect to the centroid. B, Biplot showing distribution of individual animals from various pulmonary hypertension models according to their RV/(LV + S) and RVSP. Data are presented as fold change to mean baseline level in control animals. PC = principle component; PCA = principle component analysis. See Figure 14 legends for expansion of other abbreviations.Grahic Jump Location
Figure Jump LinkFigure 6 –  A-C, Assessment of target miRNAs in plasma of patients with PAH. A, miRNAs were measured by RT-qPCR in total RNA extracted from plasma of 13 healthy control subjects and 14 patients with PAH. miRNA levels were normalized to the geomean of two internal reference controls (miR-142-3p and -320a) and then expressed relative to the control group. #P < .1, *P < .05 vs control group. Data are represented as mean ± SEM. B, PCA of plasma miRNA changes in patients with PAH (), Hx mice (), SuHx rats (), and MCT rats (). A total of 50.9% of the information content associated with the seven-dimensional plasma miRNA space is captured in the two-dimensional PCA plot. Ellipses denote 1 SD, and line vectors connect to the centroid. C, Receiver operating characteristic curves of miR-424 and miR-17 plasma levels showing the trade-off between sensitivity and specificity in distinguishing patients with PAH from healthy control subjects. PAH = pulmonary arterial hypertension. See Figure 15 legends for expansion of other abbreviations.Grahic Jump Location

Tables

Table Graphic Jump Location
TABLE 1 ]  Summary of Changes in miRNA Levels Associated With Human and Experimental PH

Fold increase or decrease (vs control group). n denotes sample size/group. P < .05 unless otherwise noted. miRNA = microRNA; ns = not significant; PH = pulmonary hypertension; RV = right ventricle.

a 

P < .1 (trend).

References

Kozomara A, Griffiths-Jones S. miRBase: annotating high confidence microRNAs using deep sequencing data. Nucleic Acids Res. 2014;42(database issue):D68-D73. [CrossRef] [PubMed]
 
Friedman RC, Farh KK, Burge CB, Bartel DP. Most mammalian mRNAs are conserved targets of microRNAs. Genome Res. 2009;19(1):92-105. [CrossRef] [PubMed]
 
Jiang Q, Wang Y, Hao Y, et al. miR2Disease: a manually curated database for microRNA deregulation in human disease. Nucleic Acids Res. 2009;37(database issue):D98-D104. [CrossRef] [PubMed]
 
White K, Loscalzo J, Chan SY. Holding our breath: the emerging and anticipated roles of microRNA in pulmonary hypertension. Pulm Circ. 2012;2(3):278-290. [CrossRef] [PubMed]
 
Schermuly RT, Ghofrani HA, Wilkins MR, Grimminger F. Mechanisms of disease: pulmonary arterial hypertension. Nat Rev Cardiol. 2011;8(8):443-455. [CrossRef] [PubMed]
 
Pullamsetti SS, Doebele C, Fischer A, et al. Inhibition of microRNA-17 improves lung and heart function in experimental pulmonary hypertension. Am J Respir Crit Care Med. 2012;185(4):409-419. [CrossRef] [PubMed]
 
Parikh VN, Jin RC, Rabello S, et al. MicroRNA-21 integrates pathogenic signaling to control pulmonary hypertension: results of a network bioinformatics approach. Circulation. 2012;125(12):1520-1532. [CrossRef] [PubMed]
 
Kim J, Kang Y, Kojima Y, et al. An endothelial apelin-FGF link mediated by miR-424 and miR-503 is disrupted in pulmonary arterial hypertension. Nat Med. 2013;19(1):74-82. [CrossRef] [PubMed]
 
Courboulin A, Paulin R, Giguère NJ, et al. Role for miR-204 in human pulmonary arterial hypertension. J Exp Med. 2011;208(3):535-548. [CrossRef] [PubMed]
 
Caruso P, Dempsie Y, Stevens HC, et al. A role for miR-145 in pulmonary arterial hypertension: evidence from mouse models and patient samples. Circ Res. 2012;111(3):290-300. [CrossRef] [PubMed]
 
Yang S, Banerjee S, Freitas Ad, et al. miR-21 regulates chronic hypoxia-induced pulmonary vascular remodeling. Am J Physiol Lung Cell Mol Physiol. 2012;302(6):L521-L529. [CrossRef] [PubMed]
 
Bertero T, Lu Y, Annis S, et al. Systems-level regulation of microRNA networks by miR-130/301 promotes pulmonary hypertension. J Clin Invest. 2014;124(8):3514-3528. [CrossRef] [PubMed]
 
Abe K, Toba M, Alzoubi A, et al. Formation of plexiform lesions in experimental severe pulmonary arterial hypertension. Circulation. 2010;121(25):2747-2754. [CrossRef] [PubMed]
 
Gupta SK, Bang C, Thum T. Circulating microRNAs as biomarkers and potential paracrine mediators of cardiovascular disease. Circ Cardiovasc Genet. 2010;3(5):484-488. [CrossRef] [PubMed]
 
Traill EM. Pulmonary arterial hypertension: an early diagnosis makes a difference. JAAPA. 2011;24(5):30-34. [PubMed]
 
Schlosser K, White RJ, Stewart DJ. miR-26a linked to pulmonary hypertension by global assessment of circulating extracellular microRNAs. Am J Respir Crit Care Med. 2013;188(12):1472-1475. [CrossRef] [PubMed]
 
Caruso P, MacLean MR, Khanin R, et al. Dynamic changes in lung microRNA profiles during the development of pulmonary hypertension due to chronic hypoxia and monocrotaline. Arterioscler Thromb Vasc Biol. 2010;30(4):716-723. [CrossRef] [PubMed]
 
Maarman G, Lecour S, Butrous G, Thienemann F, Sliwa K. A comprehensive review: the evolution of animal models in pulmonary hypertension research; are we there yet? Pulm Circ. 2013;3(4):739-756. [CrossRef] [PubMed]
 
Wei C, Henderson H, Spradley C, et al. Circulating miRNAs as potential marker for pulmonary hypertension. PLoS ONE. 2013;8(5):e64396. [CrossRef] [PubMed]
 
Peacock AJ, Murphy NF, McMurray JJ, Caballero L, Stewart S. An epidemiological study of pulmonary arterial hypertension. Eur Respir J. 2007;30(1):104-109. [CrossRef] [PubMed]
 
Humbert M, Sitbon O, Chaouat A, et al. Pulmonary arterial hypertension in France: results from a national registry. Am J Respir Crit Care Med. 2006;173(9):1023-1030. [CrossRef] [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).
Supporting Data

Online Supplement

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543