These authors contributed equally to this work.
Yes-associated protein regulates the hepatic response after bile duct ligation†
Article first published online: 8 AUG 2012
Copyright © 2012 American Association for the Study of Liver Diseases
Volume 56, Issue 3, pages 1097–1107, September 2012
How to Cite
Bai, H., Zhang, N., Xu, Y., Chen, Q., Khan, M., Potter, J. J., Nayar, S. K., Cornish, T., Alpini, G., Bronk, S., Pan, D. and Anders, R. A. (2012), Yes-associated protein regulates the hepatic response after bile duct ligation. Hepatology, 56: 1097–1107. doi: 10.1002/hep.25769
Potential conflict of interest: Nothing to report.
- Issue published online: 28 AUG 2012
- Article first published online: 8 AUG 2012
- Manuscript Accepted: 31 MAR 2012
- Manuscript Received: 14 SEP 2011
- Department of Defense. Grant Number: NF093145
- National Institutes of Health. Grant Number: R01DK081417
- Howard Hughes Medical Institute
Human chronic cholestatic liver diseases are characterized by cholangiocyte proliferation, hepatocyte injury, and fibrosis. Yes-associated protein (YAP), the effector of the Hippo tumor-suppressor pathway, has been shown to play a critical role in promoting cholangiocyte and hepatocyte proliferation and survival during embryonic liver development and hepatocellular carcinogenesis. Therefore, the aim of this study was to examine whether YAP participates in the regenerative response after cholestatic injury. First, we examined human liver tissue from patients with chronic cholestasis. We found more-active nuclear YAP in the bile ductular reactions of primary sclerosing cholangitis and primary biliary cirrhosis patient liver samples. Next, we used the murine bile duct ligation (BDL) model to induce cholestatic liver injury. We found significant changes in YAP activity after BDL in wild-type mice. The function of YAP in the hepatic response after BDL was further evaluated with liver-specific Yap conditional deletion in mice. Ablating Yap in the mouse liver not only compromised bile duct proliferation, but also enhanced hepatocyte necrosis and suppressed hepatocyte proliferation after BDL. Furthermore, primary hepatocytes and cholangiocytes isolated from Yap-deficient livers showed reduced proliferation in response to epidermal growth factor in vitro. Finally, we demonstrated that YAP likely mediates its biological effects through the modulation of Survivin expression. Conclusion: Our data suggest that YAP promotes cholangiocyte and hepatocyte proliferation and prevents parenchymal damage after cholestatic injury in mice and thus may mediate the response to cholestasis-induced human liver disease. (HEPATOLOGY 2012;56:1097–1107)
Human cholestatic liver diseases are characterized by cholangiocyte proliferation, hepatocellular injury, and, ultimately, liver fibrosis. Chronic cholelithiasis, primary sclerosing cholangitis (PSC), primary biliary cirrhosis (PBC), and biliary atresia are examples of cholestatic human liver diseases, as reviewed elsewhere.1, 2 Cholestasis results in bile duct injury caused by the accumulation of toxic hydrophobic bile acids, which also damages the periportal hepatocytes.3 In response to injury, both biliary epithelial cells (BECs) and hepatocytes will proliferate to compensate for their respective loss. Repeated injury-proliferation cycles evoke inflammation and collagen deposition, which further damage the liver and lead to fibrosis.4 Therefore, understanding the mechanism that controls hepatocyte and BEC proliferation and survival may suggest novel therapeutic targets to improve recovery from chronic cholestatic diseases. Although several regulatory mechanisms have been reported,5–9 the understanding of the hepatic response after biliary injury remains incomplete.
Yes-associated protein (YAP) is a transcription coactivator partnered with multiple transcription factors, including the p53 family member, p73,10 the Runt family member, Runx2,11 and the N-terminal TEA domain/transcription enhancer factor family transcription factors,12 among others. The activity of YAP is regulated through the phosphorylation of a conserved serine residue (S112 in mice and S127 in humans) located within the 14-3-3 binding motif, HxRxxS.13 The phosphoylation of YAP is controlled by the Hippo-signaling pathway, a kinase cascade that is conserved in Drosophila and mammals, as reviewed elsewhere.14 Upon phosphorylation, YAP translocates from the nucleus into the cytoplasm, where its transcriptional coactivator activity is turned off.13 The nuclear form of YAP is oncogenic because it can induce the expression of a class of genes that promote cell proliferation and inhibit cell death, such as the inhibitor-of-apoptosis protein (IAP) family member, BIRC5/Survivin,13 the secreted Cysteine-rich protein connective tissue growth factor (CTGF),15 and the epidermal growth factor (EGF) family member, amphiregulin.16 Overexpression of the YAP oncoprotein or ablation of the tumor suppressors in the Hippo pathway results in nuclear YAP accumulation, which, in turn, induces marked tissue overgrowth and frequently leads to tumorigenesis in mice.13, 17–23 Amplification of the Yap gene locus has been reported in several cancers,24–31 and overexpression of YAP has been frequently found in common solid tumors.13, 32 The correlation between YAP dysregulation and tumorgenesis has attracted intensive investigation; however, the function of YAP in non-neoplastic diseases has not been explored.
Previously, we showed that liver-specific Yap deficiency in the embryo affected bile duct development,21 which prompted us to investigate whether YAP is dysregulated in biliary disorders. In this study, we showed that YAP activity is increased in both human chronic cholestatic disorders and mice after bile duct ligation (BDL). Using the inducible Mx1-Cre (Cre recombinase) system, we deleted YAP in adult mice and performed BDLs. We found that Yap deficiency compromises BEC proliferation and blunts the regenerative response of hepatocytes. The mechanism accounting for loss of BEC proliferation was not associated with a change in Notch, Hedgehog, or Wnt signaling, but rather with loss of Survivn expression, whereas other hepatocyte-specific genes, such as c-Myc and alpha-fetoprotein (AFP), remained unchanged. Thus, these experiments point to a newly uncovered mechanism in controlling the hepatic response after biliary injury.
Materials and Methods
The use of human tissue was approved by the Johns Hopkins University (Baltimore, MD) Institutional Review Board. Formalin-fixed, paraffin-embedded liver sections from patients undergoing orthotopic liver transplantation for advanced stage (3-4)33 biliary disease consisted of 4 patients with PSC and 7 with PBC. The 3 control healthy livers were from patients undergoing resection of liver metastasis. All material was retrieved from pathology archives in Johns Hopkins University School of Medicine.
All animals were handled according to National Institutes of Health guidelines and protocols approved by the institutional animal care and use committee. Yapflox/flox mice have been described previously21 and were maintained on a C57Bl/6J background. To achieve liver-specific gene deletion in the adult phase, Yapflox/flox mice were injected with adenovirus expressing Cre or bred with transgenic (Tg) mice expressing Cre under the interferon-alpha-inducible Mx1 promoter (Tg[Mx1-cre]1Cgn/J; Jackson Laboratories).34 All experiments were performed in male mice and paternal inheritance of Mx1-Cre.
The adenoviruses, Cre-expressing adenovirus (Ade-Cre) and AdeGFP/LacZ, were prepared by David C. Johns (Johns Hopkins University). Animals were injected through the retro-orbital vein with 3 × 109 plaque-forming units. Activation of the Mx1 promoter was induced by three intraperitoneal (IP) injections of 600 μg of polyinosinic and polycytidylic acid (polyIC) (catalog no.: P1530; Sigma-Aldrich, St. Louis, MO) every other day to 5-week-old mice. One week after polyIC injection, BDL was performed as described previously.35, 36 Liver samples were harvested at indicated time points. For Fas studies, mice were injected IP with 0.165 μg/g weight of Jo-2 monoclonal antibody (catalog no.: 554255; BD Pharmingen, San Diego, CA), and the serum and liver were harvested 6 hours later.
Primary Cell Isolation and Culture.
Histology and Immunostaining.
Freshly dissected liver was fixed, processed, and paraffin-embedded in the Department of Pathology Reference Histology lab according to standard protocols. Five-micron paraffin-embedded sections were stained with hematoxylin and eosin (H&E) or processed further for immunostaining. Immunohistochemical (IHC) and immunofluorescent staining were performed according to the protocols provided by the manufacturers of the respective antibodies. Antibodies that were used are listed in Supporting Table 1. The DAB+ (catalog no.: 00-2014; Invitrogen, Carlsbad, CA) visualization system was used for IHC.
|Ki67||DAKO, M7249, 1/25|
|CK19||DSHB, Troma III, 1/50|
|CK7||DAKO, M7018, 1/50|
|YAP (for human liver)||Epitomics, 2060-1, 1/200|
|YAP (for mouse liver)||Cell signaling, 4912, 1/100|
|Envision anti-rabbit||DAKO, P0450, 1/50|
|Rabbit anti-Rat||DAKO, P0450, 1/50|
|Alexa488 conjugated goat anti-rat||Invitrogen, A11006, 1/200|
|Alexa568 conjugated goat anti-rabbit||Invitrogen, A11001, 1/200|
Quantification of Parenchymal Necrosis Area and Number of BECs After BDL.
H&E-stained liver sections were used to measure the areas of necrosis using ImageJ software (National Center for Biotechnology Information [NCBI], Bethesda, MD). Five 2× objective fields were randomly chosen, imaged, and the percentage of necrosis area/total area was then calculated.
Liver sections were stained with cytokeratin (CK)19 to highlight BECs. To exclude the difference between dilated and undilated bile ducts, we measured the perimeter of each bile duct to evaluate the BEC numbers. The perimeter of each bile duct was measured with ImageJ software (NCBI). Five 4× objective fields were randomly chosen, imaged, and the bile duct perimeters were calculated by adding the respective numbers of each field.
RNA Isolation and Real-Time Polymerase Chain Reaction.
Cellular RNA was extracted using the RNeasy kit (catalog no.: 74104; Qiagen, Venlo, The Netherlands), reverse-transcripted, and subjected to real-time quantitative polymerase chain reaction (PCR), as described in the Supporting Materials.
Protein Lysate and Western Blotting Analysis.
Tissues, isolated hepatocytes, or bile ducts were lysed in radioimmunoprecipitation assay buffer (150 mM of NaCl, 50 mM of Tris-HCl [pH7.4], 1% NP-40, 0.5% sodium deoxycholate, and 0.1% sodium dodecyl sulfate) with protease inhibitor (catalog no.: 10925700; Roche, Mannheim, Germany) and phosphatase inhibitor (catalog no.: 78420; Thermo Scientific, Rockford, IL). Proteins were separated on 4%-12% Bis-Tris gels (catalog no.: NP0322BOX; Invitrogen) and transferred onto polyvinylidene fluoride membranes. Blots were probed with antibodies against YAP (catalog no.: 4912; Cell Signaling Technology, Inc., Danvers, MA) and phospho-YAP (Ser127) (catalog no.: 4911; Cell Signaling Technology) and normalized by glyceraldehyde 3-phosphate dehydrogenase (catalog no.: G9545; Sigma-Aldrich). Signals were detected and quantified by the Molecular Imager Gel Doc XR System (Bio-Rad, Hercules, CA).
Serum Alanine Aminotransferase Measurements.
Serum levels of alanine aminotransferase (ALT) were measured using commercially available kits (product nos. 68-D) from Biotron Diagnostic Inc. (Hemet, CA), according to the manufacturer's protocol.
Bile Ductular Reactions in Human PSC and PBC Diseased Livers Show Increased Nuclear YAP Expression and Activity.
Liver sections from control patients without chronic cholestatic liver disease and patients with advanced stage (3-4) PSC or PBC were examined for YAP protein expression. In the healthy human liver, YAP staining was localized to the portal tract bile duct epithelium and the endothelium of the hepatic artery (Fig. 1, top panel). There was no specific staining of YAP in the hepatocytes. We looked in more detail at the interlobular bile ducts and found strong YAP staining on the plasma membrane of BECs, but little YAP staining in the nucleus, as revealed by colocalization with BEC membranous marker CK7. However, in diseased cholestatic PSC and PBC livers, we found the uniform strong expression of YAP in the periportal bile ductular reactions associated with each of the patient samples (N = 11). In contrast to the predominant membrane localization of YAP in BECs of the healthy human liver, the BECs in the ductular reactions of the cholestatic livers showed uniformly YAP staining throughout the cell, including the nucleus (Fig. 1, middle and bottom panels). Furthermore, YAP is expressed in small individual cells in the periportal region. It is possible that these represent liver-progenitor–like cells. Given the role of nuclear YAP in promoting active transcription,13 these findings suggest that YAP activity in BECs may participate in the bile ductular reactions observed in human cholestatic liver diseases.
Induction of YAP in the Murine Liver After Experimentally Induced Cholestasis.
To study the potential role of YAP in a cholestatic liver, we adopted the well-established BDL animal model of human cholestatic liver disease.39 We performed BDL on mice and harvested the liver tissue immediately after ligation (day 0), limiting any physiological change within the liver. We also harvested livers at 5 days post-BDL, which corresponds to the peak of BEC and hepatocyte proliferation, as well as 15 days post-BDL, when the ductular reactions have ceased.40 We examined the hepatic YAP expression in the liver after BDL. Western blotting analysis detected an increase of YAP protein levels in the whole liver, hepatocytes, and BECs 5 days after BDL (Fig. 2A,B). This increase of YAP protein was not the result of increased transcription, because Yap messenger RNA (mRNA) was unchanged (Fig. 2D). In contrast, mRNA levels of BEC-enriched marker osteopontin (OPN) and epithelial cell adhesion molecule (EpCAM)21 increased steadily in the whole liver from days 0 to 15 post-BDL (Fig. 2E,F). Because BECs showed significantly higher OPN and EpCAM mRNA levels than hepatocytes at all time points (Fig. 2E,F), the steady increase in OPN and EpCAM mRNA levels in the whole liver likely reflects an increase in the number of BECs. Therefore, OPN and EpCAM mRNA levels from whole liver can serve as an indirect measure of BEC mass. Even though there was an overall increase in YAP activity, reflected in the significant increase in total YAP protein 5 days post-BDL, the Hippo pathway's kinase cascade was unperturbed, because the ratio of phosphorylated YAP to total YAP protein remained unchanged (Fig. 2C). Consistent with this finding, immunostaining showed that YAP protein levels were increased in both the nucleus and cytoplasm of hepatocytes and BECs 5 days post-BDL, compared to those at days 0 and 15 post-BDL (Supporting Fig. 1).
Yap Deficiency Compromises Bile Ductular Reaction After BDL.
To test whether this increased YAP activity is critical for the hepatic response after BDL, we performed BDL in mice with liver-specific deletion of Yap in adult mice. The previously reported Alb-Cre;Yapflox/flox is not suitable for these studies, because these mice develop an abnormal embryonic biliary system.21 We achieved deletion of Yap in the adult liver by injecting Ade-Cre into Yapflox/flox mice or injecting polyIC into Mx1-Cre;Yapflox/flox mice. Both methods resulted in the efficient suppression of YAP expression in the adult liver, with no observable phenotypic changes in the liver (Supporting Figs. 2 and 3). These data suggest that the deletion of YAP in the adult mouse does not induce baseline changes in BECs or hepatocytes. Because Mx1-Cre;Yapflox/flox mice show slightly better suppression of Yap mRNA and avoid the potential complication of adenovirus-induced hepatitis,41 all subsequent experiments detailed polyIC-treated Mx1-Cre;Yapflox/flox with Yapflox/flox littermate controls. After BDL, it has been shown that there is proliferation of BECs at the periportal interface between the biliary tree and hepatocytes,40, 42 which can be highlighted using CK19 IHC staining. We found that the number of BECs increased in both Yapflox/flox and Mx1-Cre;Yapflox/flox livers 5 days post-BDL and were comparable in both genotypes (Fig. 3A,B and Supporting Fig. 4 5d). However, at 15 days post-BDL, there were significantly fewer BECs in Mx1-Cre;Yapflox/flox livers, compared to the control Yapflox/flox livers (Fig. 3A,B and Supporting Fig. 4 15d). In agreement with the histological and IHC analysis, there was a progressive increase in mRNA levels of BEC markers EpCAM and OPN from days 5 to 15 post-BDL in control Yapflox/flox, but not in Mx1-Cre;Yapflox/flox livers (Fig. 3C). Thus, loss of YAP activity compromises the bile ductular reaction after BDL.
YAP Is Required for BEC Proliferation After BDL.
One of the possible sources for the bile ductular reactions is the proliferation of preexisting BECs.43 It has been demonstrated that BEC proliferation peaks during the first 5 days post-BDL and remains at a low level thereafter.40 We therefore sought to measure whether there was reduced BEC proliferation in the Mx1-Cre;Yapflox/flox liver after BDL. We found that the number of Ki67-positive BECs was significantly reduced in Mx1-Cre;Yapflox/flox livers at day 5 post-BDL, compared to the control Yapflox/flox livers (Fig. 4A,B). In contrast, no apoptotic BECs were seen in both Yapflox/flox and Mx1-Cre;Yapflox/flox livers at either day 5 or 15 post-BDL (Supporting Fig. 5). To confirm that the compromised BEC proliferation post-BDL is a direct result of Yap deficiency, we isolated and cultured primary BECs from control and Mx1-Cre;Yapflox/flox livers and compared their proliferation response to EGF stimulation. Compared with wild-type (WT) BECs, Yap-deficient BECs showed a significantly decreased proliferation rate (Fig. 4C). These findings suggest that Yap deficiency compromises cholestasis-induced bile ductular reaction as a result of impaired BEC proliferation, rather than enhanced apoptosis.
Yap-Deficient Hepatocytes Are More Susceptible to Injury.
After BDL, the Mx1-Cre;YAPflox/flox mice developed ascites and 35% died within 15 days, whereas all control littermates showed no ascities or mortality. Liver histology revealed significantly more hepatocellular necrosis in Mx1-Cre;Yapflox/flox than in Yapflox/flox livers at day 15 post-BDL, with essentially no difference between the two groups 5 days post-BDL (Fig. 5A,B). To further test whether Yap deficiency compromised hepatocyte survival, we challenged the Mx1-Cre;Yapflox/flox and Yapflox/flox mice with Jo-2 antibody, a Fas agonist and a potent hepatocellular apoptotic stimulus.44, 45 Six hours after Jo-2 antibody injection, we observed significantly more apoptotic and terminal deoxynucleotidyl transferase dUTP nick end labeling–positive hepatocytes in Mx1-Cre;Yapflox/flox livers, compared to control livers (Supporting Fig. 6A,B). Consistently, we observed a significantly higher serum ALT level in Mx-1Cre;Yapflox/flox mice, compared to control Yapflox/flox mice (Supporting Fig. 6C). Taken together, these data support that Yap deficiency compromises hepatocyte survival, which is in agreement with our previous observation that overexpression of YAP in hepatocytes confers resistance to Jo-2-induced hepatocyte apoptosis.13, 17
Loss of YAP Delays Hepatocyte Proliferation After BDL.
At day 5 post-BDL, which normally corresponds to peak of hepatocyte proliferation,40 Mx1-Cre;Yapflox/flox livers showed a significant reduction in proliferating hepatocytes, compared to control Yapflox/flox livers (Fig. 6A,B 5d). This difference was unlikely the result of the degree of injury, because at day 5 post-BDL, hepatocyte necrosis was similar in Yapflox/flox and Mx1-Cre;Yapflox/flox livers (Fig. 5A,B). On the other hand, at 15 days post-BDL, hepatocyte proliferation was apparent in Mx1-Cre;Yapflox/flox livers, but not in control Yapflox/flox livers (Fig. 6A,B 15d). This likely represents a kinetic delay in regenerative proliferation of hepatocytes, because there is no difference in the final liver-to-body-weight ratio in Mx1Cre;Yapflox/flox (8.81% ± 0.51%) and Yapflox/flox (8.46% ± 0.86%). To study whether YAP is directly involved in modulating hepatocyte proliferation, we isolated primary hepatocyte from Yapflox/flox and Mx1-Cre;Yapflox/flox livers and cultured them in vitro in the presence of EGF. The proliferation response of primary hepatocytes to EGF was significantly attenuated as a result of Yap deficiency (Fig. 6C), which suggested that YAP plays a direct role in delaying hepatocyte proliferation after BDL.
YAP Mediates Survivin mRNA Expression After BDL.
To begin to understand how YAP might control bile duct proliferation, we examined pathways that are known to play critical roles in BEC development. We found no significant changes in the mRNA levels of the Notch2 receptor or its target gene, Hes1, at 5 or 15 days after BDL. However, both genes showed a slightly lower basal line expression in Mx1-Cre;Yapflox/flox, compared to control Yapflox/flox, mice (Fig. 7).6, 46 Similarly, there was no difference between Yapflox/flox control and Mx1-Cre;Yapflox/flox at any of the time points after BDL for the Hedgehog target genes, Gli2 or FoxL1, both of which have been reported to be critical in cholestasis-mediated bile duct proliferation (Fig. 7).7, 9
Next, we evaluated a set of genes' mRNA that have previously been reported to be up-regulated in YAP Tg livers, including Survivin, Ctgf, Afp, Gpc3, c-Myc, Sox4, Opn, and EpCam.13 We determined the mRNA expression of these genes in Yapflox/flox and Mx1-Cre;Yapflox/flox livers at days 0, 5, and 15 post-BDL and correlated the findings with YAP protein levels (Figs. 3C and 8). We surmised that the ideal YAP target gene(s) should have a peak expression that matches the induction of YAP protein levels in WT livers at day 5 post-BDL which should also be significantly suppressed in Yap-deficient livers. Among those candidate genes, only Survivin met the above-mentioned criteria. All other candidates showed peak gene expression at day 15 post-BDL, which more likely reflects progress of the bile ductular reaction (Fig. 3C for EpCam and Opn, and Fig. 8 for all others). Given Survivin's well-documented roles in enhancing proliferation and inhibiting apoptosis,47, 48 our findings suggest that Survivin may be a critical mediator in YAP-mediated regenerative response after cholestatic injury.
In this study, we provide a promising mechanism for controlling the hepatic response to biliary obstruction. We showed that YAP, the transcription coactivator of the Hippo-signaling pathway, is activated in the livers of both human patients with biliary diseases and mice with biliary obstruction. As a transcription coactivator, YAP can activate other transcription factors to induce proliferation and antiapoptosis-associated genes. Therefore, the increased YAP activity could be critical for promoting BEC proliferation, which is an important component of the repair process that occurs after damage to bile ducts during the courses of cholestatic liver diseases. The above-described hypothesis was tested using BDL in Yap-deficient livers. By investigating the repair response of Yap-deficient livers after BDL, we showed that YAP activity is important for BEC and hepatocyte proliferation and survival of hepatocyte after biliary obstruction. Our studies identify a novel therapeutic target to enhance BEC proliferation in chronic cholestatic diseases, which could ultimately be used to prevent chronic biliary diseases from progressing to bile ductopenia.
The increase of YAP protein levels could be a universal response to tissue injury. The recently published study by Cai et al.49 also showed that YAP protein levels were dramatically increased in regenerating colonic crypts of murine intestine after dextran sodium sulfate (DSS)-induced injury, and that YAP activity is required for colonic regeneration after DSS treatment. Interestingly, both our study and the study of Cai et al. show that the increase in YAP protein levels in two different tissue injury models is the result of a post-transcriptional mechanism, because neither study detected an increase of Yap mRNA levels in regenerating tissues. Future studies should reveal the molecular underpinnings of this post-transcriptional regulation of YAP protein levels and whether it represents a more widespread mechanism in other regenerative processes.
Given its potent oncogenic function, YAP activity must be carefully regulated during and after the regenerative response to injury.13, 17, 25 Consistent with this hypothesis, Cai et al.49 and Benhamouche et al.50 have demonstrated a synergistic effect of tissue regeneration and defective Hippo pathway signaling in accelerating malignant transformation in murine colon and liver, respectively. Notably, the risk for developing meningioma, a tumor that frequently harbors Nf2 mutations, is significantly increased in patients with a history of head trauma,51 suggesting that such synergy may also be relevant to human tumorigenesis.
Among the YAP-inducible genes in the liver, we found that only Survivin mRNA level correlates with YAP protein increase in WT livers after BDL and is reduced in Yap-deficient livers. Survivin, a member of the IAPs family, has been shown to suppress apoptosis through interacting with caspases52 and to promote cell division through interference with cell-cycle–related kinases and microtubule networks.53 This gene is overexpressed in many human malignancies, but is absent or present at very low levels in most non-neoplastic adult tissues.54 Therefore, Survivin appears to be a very reasonable target whose expression is induced by YAP during the regenerative response after biliary obstruction. It will be informative to investigate whether Survivin is required for the regeneration response after BDL with Survivin conditional knockout mice.55
Additional Supporting Information may be found in the online version of this article.
|HEP_25769_sm_SuppFig1.tif||1465K||Supporting Information Figure 1.|
|HEP_25769_sm_SuppFig2.tif||1336K||Supporting Information Figure 2.|
|HEP_25769_sm_SuppFig3.tif||1933K||Supporting Information Figure 3.|
|HEP_25769_sm_SuppFig4.tif||629K||Supporting Information Figure 4.|
|HEP_25769_sm_SuppFig5.tif||1434K||Supporting Information Figure 5.|
|HEP_25769_sm_SuppFig6.tif||1713K||Supporting Information Figure 6.|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.