Implication of overexpression of dishevelled-associated activator of morphogenesis 1 (Daam-1) for the pathogenesis of human Idiopathic Pulmonary Arterial Hypertension (IPAH)

All autopsy cases with a diagnosis of PAH (IPAH or APAH) recorded at Toho University Omori Medical Center during the period from 1958 to 2011 were included in this study. Age-matched controls without pulmonary arterial abnormalities were selected separately. This study was approved by the Ethics Committee of Toho University School of Medicine (approval no. 2709425035).

Built-in controls for Dvl-2 and Daam-1 were set in bronchial epithelium. The control for Wnt-11 was set in the gastric fundic gland, as indicated in The Human Protein Atlas [11].

Formalin-fixed and paraffin-embedded (FFPE) tissues of lungs from autopsy subjects were cut into sections (thickness, 3 μm) and mounted on slide glasses. After deparaffinization, samples were stained with hematoxylin and eosin and Elastica van Gieson. FFPE stomach tissues from 2 selected IPAH patients were prepared in the same manner.

FFPE lung and stomach tissues were cut into sections (thickness, 3 μm). After deparaffinization, samples were immersed in 0.1% trypsin solution and heated to 95 °C by water bath, for antigen retrieval. Staining was done with the universal immunoenzyme polymer method, a double staining method developed by Nichirei Biosciences, Tokyo, Japan [12]. After blocking with 3% peroxidase methanol for 10 min, 2 drops of the primary antibody were added for 30 min. The samples were then washed and Simple Stain MAX-PO MULTI (Nichirei Biosciences, Tokyo, Japan) was added as a common secondary antibody for another 30 min. Anti–Wnt-11 antibody (dilution 1:50, Atlas Antibodies, Stockholm, Sweden), anti–Dvl-2 antibody (dilution 1:50, Santa Cruz Biotechnology, Dallas, TX, USA), and anti–Daam-1 antibody (dilution 1:200, Proteintech, Chicago, IL, USA) were used as primary antibodies. After color development with diaminobenzidine, samples were stained with hematoxylin.

During optical microscopic observation, the vessel diameter of PA was classified into 2 categories, according to Brenner’s classification of pulmonary arteries and arterioles defined at a diameter of 100 μm [13]. Cellular components of the pulmonary arterial walls were classified into 3 categories: endothelial cells, myofibroblasts (for PAH cases), and smooth muscle cells (SMC). All slides were evaluated by 3 independent, board-certified pathologists.

The rates of positive expression of Dvl-2 and Daam-1 were evaluated as a single trend, to identify expression patterns. Thus, the comparison of 2-means method was used to compare the odds ratios for positive expression rates of Dvl-2 and Daam-1 between groups. In this analysis, control and IPAH specimens, and control and APAH specimens, were compared separately.

The salient characteristics of the autopsy cases are shown in Table 1.

Wnt-11 was expressed in 100% (2/2) of control samples. The positive expression rate of Dvl-2 was 56% (5/9) for IPAH cases, 71% (5/7) for APAH cases, and 69% (11/16) for age-matched controls. The positive expression rate for Daam-1 was 100% (9/9 and 7/7) for IPAH and APAH cases and 94% (15/16) for age-matched controls.

Because of the low positive expression rate for Dvl-2 for the built-in controls, pulmonary arteries were evaluated only for built-in control samples that were positive for both Dvl-2 and Daam-1.

Although more than 90% of IPAH cases are sporadic, familial accumulation of IPAH has long been recognized [14]. Such cases were classified as familial pulmonary (arterial) hypertension (FPAH) as early as 1973 [15]. However, after serial reports of mutations in BMPR2 and other TGFβ-related genes in FPAH patients in the early 2000s [5, 6, 16], the term HPAH replaced FPAH in the Dana Point classification of 2009. HPAH is used to refer to patients with newly diagnosed IPAH and genetic mutations and those previously classified as having FPAH [17]. According to the Nice classification of 2013, which succeeded the Dana Point classification, up to 80% of HPAH patients present with BMPR2 mutations, and an additional 5% present with mutations in other TGFβ superfamily genes, such as activin receptor-like kinase 1 (ALK1), endoglin (ENG), SMAD4, SMAD8, and caveolin1 (CAV1) [5]. Among TGFβ superfamily receptors, the TGF receptors are believed to promote proliferation and maturation of pulmonary arterial SMC (PASMC), whereas BMP receptors appear to suppress proliferation of PASMC and apoptosis of arterial endothelial cells [18, 19]. Thus, a mutation in any TGFβ superfamily receptor gene could trigger an imbalance between TGF and BMP receptors, thereby leading to aberrant contraction and proliferation of PASMC and, ultimately, IPAH [20, 21].

Several studies using animal models have reported interaction of BMPR2 and TGFβ superfamily receptors, which supports the hypothesis of TGF–BMP imbalance [22‐24]. One report showed depletion of BMPR2 with sustained expression of TGFβ2 receptors in human tissues [23]. However, BMPR2 mutation is present in only 10% to 40% of IPAH patients, and only 20% of individuals with a BMPR2 mutation develop IPAH during their lifetime [5, 6, 16]. These findings regarding BMPR2 mutations in IPAH patients and development of IPAH among individuals with BMPR2 mutations suggest that unrevealed signal transduction pathways are responsible for IPAH pathogenesis, either in cooperation with, or independent of, dysfunction in TGFβ systems [9].

A 2008 study confirmed the high reproducibility of pulmonary arterial remodeling and elevated systolic pressure in the right ventricle, which closely mimics IPAH, after repeated intratracheal injection of S. chartarum in otherwise normal male ddY mice [7]. This fungus species is ubiquitous in our environment, and very few cases of S. chartarum infection have been reported [7, 8, 25‐28].

A 2012 study reported the results of molecular biological analyses of a mouse model of IPAH induced by repeated intratracheal injection of S. chartarum. In that study, lung RNA expression profiles of the PAH model mice were assayed with a DNA microarray technique for gene ontology and pathway analyses [9]. Candidate signal transduction pathways were then compared with the results of a DNA microarray assay analysis of human IPAH, which were published in 2009 [10]. The signal expression patterns of the mouse PAH model and human IPAH were very similar. Commonalities in fluctuations in signal transduction pathways in human IPAH and the mouse PAH model were confirmed in the following pathways: upregulation of Janus kinase/signal transducers and activators of the transcription (JAK/STAT) pathway, the hemostasis pathway, the estrogen receptor pathway, and the serotonin receptor pathway, and down-regulation of the vascular endothelial growth factor (VEGF) pathway, the platelet-derived growth factor (PDGF) pathway, apoptosis, and the BMP signaling pathway. Because of the marked similarities in signal expression patterns between human IPAH and our PAH model animal prepared by S. chartarum, there were only 3 signal transduction pathways extracted that were uniquely affected in human IPAH PA in higher hierarchical pathways, namely, upregulation of the Wnt/PCP pathway and the succeeding Ras homolog gene family, member A/Rho-associated, coiled-coil–containing protein kinase (RhoA/ROCK) pathway, and down-regulation of the TGFβ pathway. The RhoA/ROCK pathway induces contraction and proliferation of medial SMC [29, 30] and is positively regulated by the Wnt/PCP and TGFβ pathways. This suggests that upregulation of the Wnt/PCP pathway leads to subsequent upregulation of the RhoA/ROCK pathway and that down-regulation of TGFβ pathway is affected by down-regulation of its higher hierarchical pathway or by negative feedback from the RhoA/ROCK pathway [31, 32]. Recent studies of fasudil, a RhoA/ROCK inhibitor, also indicate that the RhoA/ROCK pathway is an important factor in IPAH development, as IPAH improved after fasudil was given to experimental animals and human IPAH patients [33‐35].

The present study aimed to directly confirm expression and localization of major cascading proteins of Wnt/PCP pathway, namely Wnt-11, Dvl-2, and Daam-1, on pulmonary arterial walls. Because biopsy specimens of pulmonary arteries are not easy to obtain, all samples were obtained in the form of FFPE from autopsy, and IHC was chosen as a simple means of examination. Because of the unstable nature of IHC, we used a painstaking 2-step procedure to confirm our results for Dvl-2 and Daam-1. In the first step, a built-in control was observed for reactivity with the primary antibody, and slides were selected for further evaluation only when positive reactivity was confirmed for the built-in control.

Because Wnt-11 was not observed in any PA specimen but was confirmed in parietal cells of gastric fundic glands, our evaluation focused on the expression patterns for Dvl-2 and Daam-1. Observation of endothelial cells and myofibroblasts showed no characteristic expression patterns for Dvl-2 and Daam-1 under any experimental conditions. However, in SMC, there was a clear difference in expression patterns between experimental conditions. Whereas the positive expression rate for Daam-1 was lower than that for Dvl-2 in control specimens without pulmonary arterial abnormalities and in APAH specimens, the positive expression rate for Daam-1 was higher than that for Dvl-2 in IPAH specimens (Fig. 4). In other words, the expression pattern for IPAH contrasted with those for APAH and control specimens. When the ratio of positive expression rates was compared in SMC, the Daam-1/Dvl-2 ratio was 1.5 for small and medium-sized arteries in IPAH but less than 1.0 for control and APAH, and the value was even lower for small arteries (Fig. 5).

To confirm the specificity of signal expression patterns, we used the 2-means method to compare the odds ratios of the positive expression rates for Dvl-2 and Daam-1. We found that in small arteries the expression patterns for Dvl-2 and Daam-1 in IPAH PASMC differed from those for control PASMC (P = 0.05), whereas the difference between APAH PASMC and control PASMC was not significant (P = 0.70). For medium-sized arteries, the P-value was 0.28 for the comparison between IPAH PASMC and control PASMC and 0.98 for the comparison between APAH PASMC and control PASMC. Although both differences were nonsignificant, the lower P-value for the former might reflect the particular characteristics of the expression patterns of Dvl-2 and Daam-1 in IPAH PASMC.

The results showed a greater difference in Dvl-2/Daam-1 expression rate for small arteries than for medium-sized arteries. This tendency reflects the accepted characteristics of IPAH progression, namely, small arteries are affected first and medium-sized arteries are affected after substantial progression of lesions in small arteries [2, 4, 13, 36].

Wnt/PCP pathway is 1 of the 3 known Wnt family signal transduction pathways related to cell migration and polarity that are conserved across most vertebrates [37]. First discovered in 1980s, this pathway was initially studied by researchers in embryogenesis, then by researchers in oncology [38, 39]. Studies of embryogenesis suggest that the Wnt/PCP pathway is responsible for cardiogenesis and vasculogenesis [40, 41].

The Wnt/PCP signal transduction pathway is initiated by conjugation of the Wnt ligand to the 7-transmembrane Fz receptor, where it activates Dvl-2 in cytosol by an unknown mechanism. Activated Dvl-2 conjugates with Daam-1, which is autoinhibited in its natural state by its characteristic conformation and requires transformation into a binary formation with Dvl-2 for active function. Activated Daam-1 subsequently activates the RhoA/ROCK pathway, which leads to continuous medial SMC contraction and proliferation [37, 42].

The Kyoto Encyclopedia of Genes and Genomes [43] and the Reactome Pathway Database [44] indicate that the Wnt/PCP pathway reserves no known alternate or collateral pathways that complement pathway function in cases of impaired expression or function of cascading proteins. Therefore, depletion in any cascading protein of the Wnt/PCP pathway would likely result in nearly identical morphological defects. Recent studies have reported that double-outlet right ventricle, a major cardiovascular malformation, was found in mice with Dvl-2 or Daam-1 knockout, and 1 study reported mutation of the Daam-1 gene located on 14q23.1 in an aborted fetus with double-outlet right ventricle [45‐49]. These findings suggest that Wnt-11, Dvl-2, and Daam-1 are a set of cascading proteins in the Wnt/PCP signal transduction pathway, which cannot be replaced by another signal transduction pathway.

A recent study using Xenopus embryos with Wnt-11 or Dvl-2 knockout reported that, after direct injection of the active form of Daam-1, the embryos exhibited dorsal formation, a function expected if the Wnt/PCP pathway is normal [50]. The Daam-1 used in this study was mutated to function independently without binary formation with activated Dvl-2 and was thus expected to compensate for the function of the Wnt/PCP pathway. These findings indicate that activated Daam-1 is capable of functioning and signal transduction even in the absence of activated upstream cascading proteins.

The present study revealed a characteristic Wnt/PCP signal transduction pattern in PASMC from IPAH patients. The positive expression rate was higher for downstream Daam-1 than for upstream Dvl-2; however, in PASMC from patients without pulmonary arterial abnormalities and APAH patients, the positive expression rate was lower for Daam-1 than for Dvl-2. These findings suggest that IPAH pathogenesis differs from that of APAH, although both diseases yield very similar pathological findings, ie, extreme hypertrophy of PASMC [13, 36, 47]. In addition, our finding of elevated Daam-1 expression in the presence of depleted Dvl-2 expression in medial SMC from IPAH cases suggests that Daam-1 is activated independently of Dvl-2. As mentioned above, the Wnt/PCP signal transduction pathway has no alternate collateral pathway; thus, a mutation of Daam-1, which is located on 14q23.1 [48‐50], may subsequently activate the RhoA/ROCK pathway, resulting in aberrant medial SMC contraction and hypertrophy ultimately manifesting as IPAH [35]. Therefore, Dvl-2 underexpression in the context of Daam-1 overexpression in medial SMC from IPAH cases could be a result of a negative feedback mechanism in the signal transduction pathway. Indeed, at least 2 studies have reported negative feedback in the Wnt canonical and TGFβ superfamily pathways [31, 32].