Chronic liver disease is characterized by decreased regeneration of hepatocytes and increased formation of fibrous tissue. These characteristics may be the sequel of various chronic processes such as cholestasis, viral infections, toxin exposure, and metabolic disorders. Dogs have complex liver diseases such as hepatitis and cirrhosis which are highly comparable with the human counterparts. Moreover, coding sequences of dogs proved highly homologous to the human sequences 1, especially compared to the rodent genome. Thus, dogs may fulfill a role as a spontaneous animal model in between toxin-induced or surgical models in rodents, and spontaneous diseases in man. The complex interplay of many factors active in chronic liver disease makes it difficult to unravel the roles of different individual pathogenetic pathways. Dogs display liver diseases, which are potentially valuable models to compare complex with simple pathologic entities.

We have chosen these two congenital dog diseases for comparative analysis of liver growth/regeneration, fibrosis, and hepatic homeostasis: congenital portosystemic shunt (CPSS) and primary portal vein hypoplasia (PPVH). CPSS is characterized by an abnormal single large communication between the portal vein and a major systemic vein (cava or azygos). This results in the virtual absence of portal vein perfusion to the liver from birth onwards. Liver growth remains nearly absent but there is essentially no liver pathology 23. PPVH is a developmental abnormality in which the terminal vein branches are not or only partially present and, in most cases, in combination with congenital portal fibrosis, but without inflammation 4. PPVH is associated with portal hypertension and reduced liver growth. Thus, these two congenital diseases represent relatively simple models for reduced liver growth associated with fibrosis (PPVH) or without fibrosis (CPSS). Both diseases have a decrease in liver growth due to differences in portal perfusion which results in a massive reduction of liver size.

Because hepatocyte growth factor (HGF) is one of the most important genes involved in liver growth/regeneration 567, abnormal expression of HGF could play a major role in the decreased liver size in CPSS or PPVH. Therefore, treatment of dogs with HGF could be a possible therapeutic approach. A pre-requisite for treatment is that HGF signaling components are unaffected in those dogs. Consequently, we focused on measuring gene products involved in signaling of HGF and counteracting transforming growth factor β1 (TGF-β1). All biological responses induced by HGF are elicited by binding to its receptor, a transmembrane tyrosine kinase encoded by the MET proto-oncogene (c-MET). The signaling cascade triggered by HGF begins with phosphorylation of the receptor and is mediated by concomitant activation of different cytoplasmic effectors that bind to the same multifunctional binding site. The c-MET mediated response includes two key pathways involved in cell survival and mitogenesis 8. The first; protein kinase B (PKB/Akt) is activated by phosphoinositide 3-kinase (PI3K) and elicits cell survival 910. The second; ERK1/2 (also known as p42/44 MAPK), a member of the mitogen-activated protein (MAP) kinase family, is activated by the RAS-RAF-MEK pathway and is responsible for mitogenesis 11. A third response of HGF is the branching morphogenesis which next to the PKB and ERK pathways requires involvement of the signal transducer and activator of transcription (STAT) 3 pathway 12.

It is well established that an increase of TGF-β1 in liver promotes the formation of extracellular matrix (ECM) components and suppresses hepatocyte proliferation 1314. Prolonged overexpression of TGF-β1 in non-parenchymal cells causes hepatic fibrosis in humans and experimental animals. In several fibrosis models, fibrotic lesions are associated with an increase in collagens and TGF-β1 mRNAs 15. The intracellular pathway that is activated by TGF-β1 receptors is mediated by Smads. Smad2 is activated via carboxy-terminal phosphorylation by TGF-β1 type I receptor kinases. When bound with co-Smads, they act as TGF-β1-induced transcriptional activators of target genes 16.

Cell homeostasis is the result of balance between cell death, cell proliferation, and growth-arrest. Therefore we investigated expression levels of pro-apoptotic Fas ligand and caspase-3, anti-apoptotic Bcl-2 17, cell-cycle stimulating TGFα, and cell-cycle inhibitor p27kip. All of these gene-products are regulated directly or indirectly by PKB 9.

The present study was designed to describe the differential gene-expression of the above indicated crucial pathways involved in growth/regeneration, fibrosis, and cellular homeostasis in liver tissues of dogs with CPSS (reduced growth/regeneration without fibrosis) and PPVH (reduced growth/regeneration and fibrosis) in comparison with healthy animals. These simple congenital dog models may be used to unravel the roles of different gene products in those pathways. These well-defined large animal models are intended to serve as the first spontaneous liver diseases to investigate novel regenerative/anti-fibrotic therapies, such as HGF treatment. This study may also serve as a basis for future comparison with more complex diseases like chronic hepatitis and cirrhosis.

In order to analyze the possibility of growth factor therapy, two congenital canine liver diseases were molecularly dissected. The expression of a total of 17 gene products involved in liver growth/regeneration, fibrosis, ECM, and cellular homeostasis was measured and normalized to the average amount of two reference genes (Q-PCR). Western blot analysis confirmed the quantitative mRNA results and, furthermore, showed activated pathways. These two independent techniques provided insight into the effects of portal venous hypoperfusion in two canine hepatic diseases; congenital portosystemic shunt (CPSS) without fibrosis and primary portal vein hypoplasia (PPVH) with fibrosis. Taken together, the obtained data provided insights in the feasibility for HGF-treatment.

The normalization performed in this study was obtained by averaging the amount of two different reference genes, glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and hypoxanthine phosphoribosyl transferase (HPRT). No samples were more than 5 percent apart from the individual measured reference genes levels (data not shown). This normalization strategy, using the average amount of two reference genes, is taken as a prerequisite for accurate Q-PCR expression profiling which enables us to measure small expression differences and allows the study of their biological relevance 19.

It is well known that HGF plays an essential role in development 20 and regeneration of the liver, and increases hepatocyte viability. The found decrease in gene-expression of both HGF and its receptor agrees with the reduced liver size in these canine disorders. However, and in contrast to the c-MET levels which correlate nicely with the found protein levels, the amount of HGF mRNA does not seem to reflect protein levels. This can be contributed to HGF which can be a paracrine but also an endocrine factor. Extra-hepatic HGF could have been present in the pancreas or intestinal tract 21.

Although HGF and c-MET mRNA levels were decreased, downstream targets of this tyrosine cascade signaling pathway were still active. Downstream targets, such as Fas ligand and p27kip, were chosen as direct or indirect targets of the HGF-cMET-PI3K-PKB axis. Fas ligand transcription is regulated by FOXO (forkhead box, sub-group "O" transcription factors). Therefore, the decrease in Fas ligand can be explained by an active PKB which directly phosphorylates FOXO 22. A similar result can be seen in the reduced levels of p27kip mRNA, as this is down-regulated at the gene-transcription level by active PKB 23. Combined, this indicates that PKB is active in both diseases, which was confirmed by Western blot analysis. It remains to be seen whether other receptor tyrosine kinases (e.g., EGF receptor or insulin receptor) activate this pathway in these dogs 24. Next to the activated PKB pathway, we have analyzed other c-MET mediated responses in CPSS and PPVH. ERK1/2 showed to be activated in both diseases to a similar level as the healthy controls. The significance of the slightly reduced phosphorylated STAT3 in PPVH, which is phosphorylated by HGF on serine 727 25, needs to be further investigated. Taken together, the pathways which elicit all major biological functions of c-MET showed to be active in CPSS and PPVH.

Prolonged or overexpression of TGF-β1 acts to suppress cell proliferation, and induces a deposition of ECM proteins, resulting in fibrosis in major organs such as liver 2627. We showed that in PPVH the TGF-β1 pathway through Smad2 is activated, consistent with the fibrosis seen in PPVH. Measurements on fibrosis related gene products revealed no elevated activity of the TGF-β1 pathway in CPSS. Gene expression levels related to the TGF-β1 pathway, including its receptors, and the proteolytic activator of TGF-β1 (uPA) were elevated in PPVH, thus indicating an active Smad pathway that could subsequently lead to fibrosis. Western blot analysis confirmed found TGF-β1 levels. Measurements on collagen gene-expression, especially collagen I, confirm the current paradigm of TGF-β1 signaling in fibrous tissues like PPVH 28. Contrary, non-fibrotic CPSS did not show any alterations in collagen expression. The observation of phosphorylated Smad2 in healthy liver tissue showed that the phosphorylation of Smad2 is a dynamic process and has already been described in other publications 2930.

The expressions of the pro-apoptotic genes Fas ligand and caspase-3 were clearly decreased. Bcl-2 gene-expression was elevated two times in PPVH; but not in CPSS (Figure 4). Western blot analysis showed that the unprocessed form of caspase-3 was present in lesser amount in CPSS and PPVH; however, the amount of processed or active bands compared to healthy control was higher in the diseases compared to healthy controls. This indicates that although the total amount of caspase-3 is lower, there is more cleavage of the caspase-3 to its active forms in the diseases, possibly leading to an increase in apoptosis.

Both HGF and TGF-β1 need extracellular processing to become biologically active. The serine protease HGF activator is responsible for activation of proHGF 31. Our studies revealed that HGF activator gene-expression was doubled in dogs with CPSS and halved in case of PPVH. This indicated an increased HGF activation in CPSS. Although levels of HGF activator were reduced in PPVH, this does not necessarily indicate a lack of extracellular processing of HGF. Interestingly uPA, the activator of TGF-β1, was expressed at an increased level in dogs with PPVH. This may, via active TGF-β1-receptor interaction, indicate an activation of Smads and thus the formation of collagens.

Differential gene expression measurements on hepatic diseases have been performed in the past; nevertheless, little is known about levels of genes that play an important role in fibrosis. There have been measurements on cirrhosis in man and rat that indicate an up or down-regulated expression of several proteins 32. Although these results might be significant in severe forms of fibrosis, these data depict an end-point of the disease whereas earlier stages may be more informative.

Regeneration with recombinant HGF has been achieved in rodent models of liver failure 3334. Moreover, besides its regenerative capacity, HGF is known to have an antifibrogenic effect 3536 and thus reduces or prevent fibrosis in PPVH. TGF-β1 intervention to halt the progression of liver fibrosis and positively effect regeneration, has been applied successfully 37 even in cirrhosis 38. The measured gene products involved in fibrosis in PPVH make it a good spontaneous animal model to investigate new therapeutic strategies to influence the HGF and/or TGF-β1 pathways in vivo. Furthermore, most fibrogenic models are induced by toxins such as dimethylnitrosamine (DMN), CCl₄, or thioacetamide 39. The canine PPVH model is not drug-induced; therefore, may be better to compare with human diseases and thus fill the gap between induced rodent models and human diseases.

This study is the first to measure expression profiles of crucial pathways of liver growth/regeneration, fibrosis, and hepatic homeostasis in spontaneous canine liver diseases. The present findings in two diseases with relatively simple pathogenesis may also serve as basis for evaluation of more complex diseases like hepatitis and cirrhosis. Evaluation of such complex diseases in dogs is highly suitable for comparative studies on the roles of different pathways in the pathogenesis of liver diseases in man. Two further conclusions can be deduced from the data presented here. First, the pathophysiological differences between CPSS and PPVH can nicely be explained by the Q-PCR measurements and Western blots. Second, although c-MET levels were reduced, downstream signaling seemed to be functional and provides a rational background to design controlled studies for HGF-supplementation in CPSS and PPVH.

The author(s) declare that they have no competing interests.

BS performed most Q-PCR measurements and wrote the manuscript. LP participated in the setup of Q-PCR measurements and helped to draft the manuscript. TI histochemically examined samples described in this manuscript. BA helped perform the0 Western blot experiments. JIJ histochemically examined samples described in this manuscript. FS helped collect all samples. JR participated in the study design and helped to draft the initial manuscript. All authors read and approved the final manuscript.