Reversal of Liver Cirrhosis: A Desirable Clinical Outcome and Its Pathogenic Background

Abstract

Cirrhosis is the final stage of chronic liver damage of various etiologies. It used to be considered an irreversible lesion, but enormous advances in our understanding of hepatic cellular and molecular biology in the past 2 decades have challenged this view. There is now substantial evidence that cirrhosis can be a reversible process. This concept is supported by an increasing number of clinical reports showing the disappearance of cirrhotic lesions from liver biopsies taken from patients cured of their liver disease. The reversal of cirrhosis usually occurs in patients with short-lived liver disease, after the successful treatment of the underlying liver damage. Recently, however, we observed the spontaneous reversal of cirrhosis after the loss of hepatitis B viremia in 2 men, 21 and 28 years old, who had developed cirrhosis as young children. Several questions and controversial issues concerning the definition of advanced cirrhosis, the limitations of liver biopsy (eg, sampling, interpretation error), and the applicability of noninvasive methods to the assessment of fibrosis, are being addressed. Future prospects include the possibility of antifibrotic therapy to prevent fibrosis or favor its degradation.

INTRODUCTION

Cirrhosis is a pathological condition involving the presence, throughout the liver, of fibrous septa dividing the parenchyma into nodules ⁽¹⁾. It represents the final stage of chronic liver damage of various causes, which may be viral, alcoholic, toxic, autoimmune, metabolic, or ischemic. Its life-threatening complications include refractory ascites, variceal bleeding, encephalopathy, hyponatremia, and renal dysfunction, and it represents a major health problem worldwide ^(2,3).

Until recently, cirrhosis was considered an irreversible lesion ⁽⁴⁾, but in the past decade, both clinical reports and advances in basic science have challenged this concept ^(4–11). Reversal of cirrhosis has been described in a wide range of liver diseases ^(12–22), and the results of earlier studies, mainly concerning small series of patients, are now supported by findings emerging from large-scale trials in patients with chronic hepatitis C treated with pegylated interferon (IFN) and ribavirin ^(21–23). These biopsy-based studies have demonstrated the reversibility of cirrhosis after the clearance of viremia in successfully treated patients, emphasizing that the elimination of the underlying chronic injury is fundamental to the reversal of cirrhosis.

Key cellular and molecular mechanisms of liver fibrogenesis that lead to cirrhosis have been extensively investigated, focusing especially on the central role of the hepatic stellate cells (HSC) in the production and degradation of the extracellular matrix (ECM) ⁽²⁴⁾. There is now a substantial body of evidence to show that liver fibrosis and cirrhosis are active, dynamic processes that reflect the imbalance between the deposition and degradation of connective tissue, and that ECM deposition is far more reversible than was previously believed ^(25,26).

This review focuses on the reversal of cirrhosis as a clinical and histological event that has now been described in a variety of liver disorders and on rational scientific grounds.

PATHOGENESIS OF CIRRHOSIS

Three major mechanisms are central to the onset of cirrhosis: cell death, ECM deposition, and vascular modifications ^(27–29). The fibrotic process is characterized by both quantitative and qualitative changes in the composition of the hepatic ECM with excessive deposition (of collagen, in particular) in the portal tracts and the replacement of low-density type IV collagen with high-density types I and III collagen in the space of Disse, causing sinusoidal capillarization. The morphogenesis of cirrhosis is related to the underlying disease and reflects the topographic distribution of the liver damage and the contribution of different cells involved in the fibrogenic process ⁽³⁰⁾. In biliary diseases fibrosis deposition involves the portal tracts and then progresses to the formation of portal-portal septa. In chronic viral hepatitis different mechanisms contribute to disrupting the cellular architecture (ie, longstanding interface hepatitis, responsible for the development of portal-portal septa, and more extensive necroinflammatory lesions, such as portal-central bridging necrosis, leading to the onset of portal-central vein fibrous septa). Central-central septa are usually associated with outflow disorders. In both alcoholic and nonalcoholic steatohepatitis, fibrosis surrounding groups of hepatocytes—the so-called chicken-wire pattern—in the area around the central veins is a key step in the development of cirrhosis ⁽³¹⁾.

Regardless of the cause, sinusoidal capillarization is an early event with remarkable effects on the metabolic exchanges between hepatocytes and blood circulation. Further impairment in liver function stems from the formation of novel intrahepatic vessels, via porto-portal and porto-central collaterals, that shunt the blood away from the hepatocytes.

Fully developed cirrhosis has diverse morphological features, with more or less numerous, slender or broad septa and parenchymal nodules of varying sizes and shapes. It has been suggested that the severity of cirrhosis be classified according to the characteristics of the septa ⁽³²⁾. More recently, a combination of nodule size and septal thickness has been proposed for staging the severity of portal hypertension. Basing the severity of cirrhosis on histological criteria ⁽³³⁾ may help us to predict its potential for reversal.

The anatomic classification distinguishes between micronodular cirrhosis, typically associated with alcohol abuse and characterized by small uniform nodules a few millimeters in size; macronodular, or posthepatitis, cirrhosis, with large bulging nodules of varying sizes; and mixed cirrhosis, typical of primary biliary cirrhosis and primary sclerosing cholangitis. Most cases of advanced cirrhosis acquire a mixed pattern, however, because of the ongoing regenerative and fibrogenic process. Incomplete septal cirrhosis, characterized by slender fibrous septa that do not connect portal tracts to portal tracts and/or central veins, is regarded as a histological feature of regression of cirrhosis ⁽³⁴⁾. The practical value of the above classification is to remind pathologists that it may be difficult to recognize macronodular and incomplete septal cirrhosis in liver biopsy specimens ⁽³⁵⁾.

Cellular Effectors of Fibrogenesis

The HSC have been considered the key cellular source of collagen ^(36–38). In the normal liver, HSC reside in a quiescent status in the space of Disse, between hepatocytes and sinusoidal endothelial cells, and they store vitamin A and other retinoids. In response to liver injury and the related production of cytokines, HSC undergo a phenotypic transformation into proliferative, fibrogenic, and contractile myofibroblasts (cells not found in the normal liver). The proliferation of myofibroblasts, stimulated especially by platelet-derived growth factor-β, amplifies the number of fibrogenic cells. The direct fibrogenic activity of the myofibroblasts, stimulated particularly by transforming growth factor-β-1, gives rise to a dramatic increase in the synthesis and deposition of ECM. Myofibroblast contractility, primarily in response to endothelin-1, leads to increased portal resistance by constricting individual sinusoids and contracting the liver as a whole. In addition, HSC produce tissue inhibitor of matrix metalloproteinases (TIMPs), which inhibits collagen-degrading matrix metalloproteinases and tips the balance between ECM synthesis and degradation in favor of ECM synthesis and fibrogenesis.

Hepatic cell types other than HSC may also have fibrogenic potential ^(39–41). In an experimental model of bile duct ligation, fibroblasts normally residing around vessels and bile ducts may acquire a myofibroblastic phenotype and start depositing collagen in the portal zones. Together with vascular smooth muscle cells, fibroblasts around centrilobular veins or in the liver capsule may also change into myofibroblasts, thus contributing to fibrogenesis. More recent studies in humans have suggested that mesenchymal cells in the portal tract have a significant role in the development of liver fibrosis and cirrhosis.

Hepatic fibrosis pathways are largely similar, regardless of the cause of the liver injury, but some disease-specific fibrogenic mechanisms are under investigation, including the direct stimulation of HSC by viral infection in hepatitis C, high leptin levels, and increased leptin signaling by HSC in nonalcoholic steatohepatitis.

Factors Interfering With Fibrogenesis

The rate of progression of fibrosis can vary considerably. Progression tends to be rapid in some cases, as in infants with congenital hepatic fibrosis, patients coinfected with hepatitis C virus (HCV) and HIV, and patients with recurrent hepatitis C after liver transplantation. Risk factors such as alcohol consumption, a high body mass index, or immunosuppressive therapy may also accelerate progression ⁽⁴²⁾. Genetic factors are likely to influence the deposition and elimination of fibrosis, but the complex mechanisms of fibrogenesis, involving a large number of cytokines, make the study of gene polymorphisms rather difficult.

Mechanisms of Fibrolysis

Experimental models in rats and transgenic mice and cell culture studies have enhanced our knowledge of the critical events and processes in fibrosis and cirrhosis reversal. In the normal liver the ECM accumulation produced by HSC and other fibrogenic cells is balanced by a proportional increase in its degradation. Our knowledge of the cell types and of enzymes that degrade collagen in the liver is still limited. Matrix metalloproteinases are zinc-dependent enzymes expressed mainly by HSC and Kupffer cells; they are capable of degrading different matrix components ⁽⁴³⁾. In the damaged liver increased mitochondrial processing peptidase activity is the major mechanism behind the regression of fibrosis. This activity is regulated by TIMPs. Recent data support the hypothesis that TIMPs play a predominant part in regulating fibrosis by protecting fibrotic ECM against degradation by collagenase ^(43,44). In rodents the reversal of fibrosis is accompanied by an increase in the apoptosis of activated HSC, which simultaneously reduces ECM and tissue inhibitors of metalloproteinases. Apoptotic mechanisms, however, would be different between rodent and human cells, and fully activated human HSC proved resistant to proapoptotic stimuli ^(45,46). Clearly, at some point (probably coinciding with the onset of the clinical symptoms of cirrhosis), fibrosis becomes irreversible. Animal models suggest that this event coincides with a significant collagen cross-linking by tissue transglutaminase, leading to the production of an insoluble matrix ⁽⁴⁷⁾.

CIRRHOSIS REVERSAL IN CLINICAL PRACTICE

Regression of fibrosis and cirrhosis in humans is not a novel concept. Anecdotal reports published more than 30 years ago recorded an improvement in patients with cirrhosis treated for hemochromatosis and Wilson disease ^(48,49). More recently, regression of fibrosis and cirrhosis (Table 1) has been documented in the entire spectrum of chronic liver diseases ^(11–21) (ie, in autoimmune hepatitis and primary biliary cirrhosis after effective immunosuppressive therapy, in biliary obstruction after surgical decompression, in thalassemia after iron depletion, in hepatitis B after lamivudine therapy, and in hepatitis D during long-term follow-up after IFN treatment). Larger studies have been conducted in patients with HCV-related cirrhosis. Poynard et al ⁽²¹⁾ examined liver biopsy specimens taken before and after therapy from 153 patients with HCV-related cirrhosis treated with different pegylated IFN and ribavirin regimens. Using the METAVIR scoring system, they found that the extent of liver fibrosis had improved in 75 (49%) patients: from stage 4 to stage 3 in 23 patients, to stage 2 in 26 patients, to stage 1 in 23 patients, and to a virtually normal histological appearance in 3 patients. No such improvements were recorded in the control group of patients treated with IFN monotherapy. Reversal of cirrhosis was more common among younger patients.

TABLE 1: Cirrhosis reversal in relation to cause and treatment: data published in the past decade

In the above-mentioned studies the favorable outcome of liver disease always followed some kind of treatment, whereas we have recently reported on the spontaneous reversal of cirrhosis in 2 young patients in a cohort of 99 Italian children with chronic hepatitis B ^(50,51). In these 2 patients biopsy-proven micronodular cirrhosis had developed by the time they were 2 and 8 years old, respectively. The source of infection was vertical in 1 case and unknown in the other. They had a short-lived hepatitis Be antigen (HBeAg) antigen-positive phase with active cytolysis but without any clinical or biochemical features of liver failure. After e antibody 10 HBeAg seroconversion, they became hepatitis B virus DNA negative by conventional hybridization techniques, and 1 eventually cleared HBsAg. Further liver biopsies performed during their childhood confirmed the histological diagnosis, but liver biopsy specimens obtained at the ages of 21 and 28 years, respectively, showed only mild periportal fibrosis. Liver histology was reviewed by the same expert pathologist on adequate bioptic samples. The results of biochemical tests and ultrasound were consistent with a normal liver. In these patients initial active liver damage may have caused the early evolution to cirrhosis, whereas a subsequent spontaneous cessation of virus replication, before the onset of clinical complications, probably enabled the reversal of cirrhosis.

OPEN QUESTIONS AND CONTROVERSIES

Despite the mounting clinical evidence, the potential for the reversal of cirrhosis remains a debated issue, the methods used in the clinical studies are still criticized, and even the terminology is a matter of controversy. In their recent review ⁽¹¹⁾, Friedman and Bansal proposed standardizing the use of the terms “reversal” and “regression” in this context: “reversal” would preferably indicate the complete restoration of normal liver architecture after recovery from liver damage, whereas “regression” would simply indicate a reduction in the amount of fibrosis, irrespective of its initial extent.

Sampling Error

Sampling error, defined as the deviation of a sample from the tissue from which it was taken, is an issue in the diagnosis of cirrhosis, in particular of the macronodular and incomplete septal types. A needle biopsy covers a small portion of the whole liver, representing approximately 1/50,000 of the organ. The available data consistently indicate the risk of underestimating fibrosis on the basis of semiquantitative assessments on excessively small samples ^(52–55). The sampling error issue has been reviewed in detail elsewhere ⁽⁵²⁾. The strategy for reducing the risk of sampling error is to obtain “adequate” samples, which should contain no less than 11 complete portal tracts for the staging of chronic hepatitis ⁽⁵⁶⁾. Caution is consequently needed when a reversal (or regression) of fibrosis/cirrhosis is diagnosed on substandard biopsy samples. It is also important to emphasize that a histological diagnosis of cirrhosis relies both on the extent of fibrosis and on the detection of architectural derangement. In the hands of an expert pathologist, various subtle histological changes reflecting architectural disruption may point to a confident diagnosis of cirrhosis even in small samples and/or in the macronodular type.

Scoring Systems and Intra- and Inter-observer Variation

At least 3 scoring methods are commonly used in the staging of liver fibrosis. They are based on the collagen staining of biopsy samples and take into account the typical progression of fibrosis from periportal to septal, and ultimately to nodule formation. In the Knodell and METAVIR scores ^(57,58) fibrosis is staged as 0 to 4, where 4 indicates cirrhosis. The Ishak ⁽⁵⁹⁾ system scores fibrosis on a scale from 0 to 6, where 5 means incomplete or early cirrhosis and 6 means established cirrhosis. Recent studies have found good intra- and interobserver reproducibility for the staging of fibrosis regardless of the scoring system used ⁽⁵⁴⁾. All of these systems are semiquantitative and nonlinear, however, and have been designed to stage the deposition of fibrosis rather than its regression, which is likely to be a dynamic process taking months or years to complete, so a biopsy taken shortly after treatment of the liver disease may fail to show any significant improvement. By contrast, the invasiveness of liver biopsy makes a close histological monitoring of fibrosis unfeasible, especially in pediatric patients, hence, the investigations into noninvasive methods for assessing liver fibrosis. Several putative serological markers of fibrosis have been studied in adults, and encouraging results have been obtained by combining different markers. Several recent reviews are available, and many studies in progress are attempting to validate such noninvasive methods for clinical practice ⁽⁶⁰⁾.

CONCLUSIONS AND FUTURE DIRECTIONS

Some messages seem to emerge unequivocally from the literature:

1. Reversal of cirrhosis is more likely to occur in patients with a relatively short-lived history of disease and well-preserved liver function.
2. An improvement in, or the disappearance of, the underlying chronic liver injury is a necessary condition for the reversal of cirrhosis.
3. Reversion of cirrhosis usually follows specific treatment but may also be a spontaneous event.
4. Reversal is probably a slow process, which starts soon after the liver injury has been overcome and may take several years; a short-lived liver disease and stronger regenerative activity could accelerate the reversal of cirrhosis in some pediatric subgroups.

Several aspects of the reversal of cirrhosis deserve further discussion and clarification, particularly concerning the definition of irreversible cirrhosis based on clinical signs and symptoms, and the concept of histological severity of cirrhosis. The identification of reliable noninvasive markers for assessing liver fibrosis is highly desirable to overcome the drawbacks of liver biopsy, especially in children.

Removing the cause of liver damage remains the best treatment for cirrhosis ^(61,62), but it is anticipated that pinpointing efficient targets for antifibrotic therapy, identified amidst the cascade of events leading to the progression of fibrosis, will definitely change the prognosis of chronic liver disease in the future.