Introduction
Our knowledge of the hepatitis C virus (HCV) has increased greatly since its discovery through a "tour de force" of molecular biology in 1989.[1] These advances have facilitated the study of HCV biology, led to the identification of different HCV genotypes, and given us the tools to diagnose with certainty patients who are infected with this virus. Currently, efficacy of treatment is assessed by determining the presence and quantity of circulating HCV RNA. Building on the experience with chronic hepatitis B, the antiviral effect of interferon alfa in patients with chronic hepatitis C was quickly recognized.[2,3] Numerous studies have since used different types of interferons at different dosages for different periods to treat patients with this chronic disease.[4] The end points of these studies have been normalization of serum alanine transaminase (ALT) levels and the clearance of virus from blood at both the end of treatment and at 6 to 12 months after treatment (ie, sustained response). Because interferon has limited efficacy and many adverse effects, and is expensive, it is important that other major benefits associated with its not be overlookedthe most significant of which may be its antifibrotic effect. Patients with chronic hepatitis C may manifest evidence of progressive fibrosis gradually over many years; this hepatic fibrosis disrupts the hepatic architecture. Indeed, chronic and progressive hepatitis C is associated with significant morbidity and mortality and is a leading cause of cirrhosis. Progressive fibrosis leads to cirrhosis and its fatal consequences. Evidence suggests that interferon may not only halt the deposition of fibrotic tissue but may actually reverse it. Data demonstrating the antifibrogenic effects of interferon alfa in experimental models of liver injury have been confirmed by clinical trials in patients with chronic hepatitis C. Therefore, it is possible that interferon may change the natural history of hepatitis C independent of its antiviral effect.

Hepatic Fibrosis: How Does it Occur?

Many tissues accumulate scar tissue as a nonspecific response to wound healing. In the liver, this process, which can be triggered and sustained by many different chronic injuries, disrupts the normal hepatic architecture and leads to cirrhosis. The imbalance between fibrogenesis and fibrolysis affects not only the quantity of the extracellular matrix but also its quality. The usually loose subendothelial matrix becomes much thicker as fibrilforming type I collagen accumulates. Cells "sense" with special receptors (cellmatrix adhesion molecules, mainly the integrins) the extracellular milieu in which they are embedded and alter their functions in response. The "capillarization" of the sinusoids (deposition of scar tissue within the space of Disse) impairs access to hepatocellular plasma membranes. Activation of hepatic stellate cells within the Disse space plays a crucial role in fibrogenesis.[5] These cells proliferate and change their phenotype from pericytes (which store vitamin A) to contractile myofibroblasts (which secrete new components into the matrix).[6] Fibrotic tissue is dynamic and constantly turning over, suggesting the existence of enzymes that are able to digest extracellular matrix proteins. Several metalloproteinases that have this function have been described, and the activity of these enzymes is tightly regulated by inhibitors.[7] Among the cytokines that appear to orchestrate fibrogenic changes, transforming growth factor beta (TGFbeta) appears to be predominant.[8] When the activity of TGFbeta has been altered experimentally by gene therapy, hepatic fibrosis has been prevented[9] and even reversed.[10]

Hepatic Fibrosis: A Crucial Parameter in the Natural History of Chronic Hepatitis C

Chronic HCV affects about 1% of the population in western countries and even more in developing countries. The virus is usually transmitted through contact with contaminated blood virtually 100% of intravenous drug users are infected with HCV. After an episode of acute hepatitis, 60% to 80% of patients will develop chronic infection. Chronic disease is characterized by persistence of the virus in the blood and elevated serum ALT levels.[11] Fulminant hepatitis due to hepatitis C is extremely rare.[12] The major cause of morbidity and mortality associated with HCV is progression to cirrhosis, with the resultant complications of liver insufficiency, portal hypertension, and hepatocellular carcinoma. Although the median time of progression from infection to liver cirrhosis has been estimated to be 30 years,[13] the rate varies greatly among patients. Some patients will never develop cirrhosis, whereas others will become cirrhotic within a few years of infection. Host factors that are associated with an accelerated rate of progression include male sex, infection after the age of 40 years, and alcohol excess.[13] The high mortality rate of patients with chronic hepatitis C occurs almost exclusively in those who have developed cirrhosis.[14,15]

Clinical Assessment of Hepatic Fibrosis

The continuing deposition of fibrotic tissue in the hepatic parenchyma is central to the development of cirrhosis. The previous belief that cirrhosis resulted from the collapse of existing stroma due to parenchymal cell necrosis has been proven incorrect. Rather, cirrhosis actually results from the deposition of newly formed collagen.[16] The ability to stop this process or even reverse it would change the natural history of hepatitis C because of the regenerative capacity of the liver. The reversal of fibrosis has been reported in patients with autoimmune hepatitis who were successfully treated with steroids.[17,18] Although the availability of a simple, accurate method to determine the extent of hepatic fibrosis would be a great asset in the management of patients with chronic liver diseases, it remains a challenge. Several serum markers have been proposed as surrogates for collagen deposition and/or resorption. However, none is specific for the liver, and these serum markers could reflect matrix turnover in other organs such as bone. These markers correlate more with the formation or degradation of extracellular matrix than with its actual amount and do not provide information about the degree of architectural disruption.

The Nterminal propeptide of type III collagen (NPIIIP) could reflect fibrogenesis and fibrolysis, because it is released by the conversion of type III procollagen into collagen.[19] However, the correlation between serum levels of NPIIIP and hepatic levels of TGFbeta and procollagen messenger RNAs (mRNAs) suggests that this proposed marker primarily reflects fibrogenesis.[20] Clinically, NPIIIP correlates with histologic activity,[2123] a finding that supports this interpretation. By contrast, elevated serum levels of hyaluronic acid, a highmolecularweight polysaccharide of the extracellular matrix, correlate well with extensive hepatic fibrosis.[24] Other proposed serum markers of hepatic fibrosis are prolyl hydroxylase,[25] an enzyme involved in collagen crosslinking; type IV collagen, an important component of basement membrane[26,27]; and laminin.[2830] The formation and degradation of extracellular matrix is regulated by cytokines, which can be measured in tissue or serum. TGFbeta is one of the most important fibrogenic mediators and functions to increase the production of extracellular matrix proteins by hepatic stellate cells.[31] Serum levels of matrix metalloproteinase 2 and tissue inhibitor of matrix metalloproteinase 1 may serve to measure the degree of hepatic fibrosis and predict response to interferon in patients with chronic hepatitis C.[32]Along the same lines, serum NPIIIP, hyaluronic acid, and type IV collagen levels have been shown to correlate with hepatic fibrosis and to predict response to interferon therapy.[33,34]

Unfortunately, these noninvasive measures have significant limitations. For example, elevated serum levels of NPIIIP could be the result of decreased hepatic uptake and removal by sinusoidal endothelial cells rather than to collagen deposition.[35] Therefore, these biochemical tests have not yet replaced the percutaneous liver biopsy, which remains the "gold standard" for quantifying hepatic fibrosis. Liver biopsy also provides information about the severity of tissue inflammation, a histologic feature that does not correlate very well with the elevation of the serum ALT levels. There are 2 potential problems when liver biopsy is used to determine the extent of hepatic fibrosis. First is sampling artifact. A percutaneous liver biopsy specimen samples approximately 1 of 60,000 of the liver parenchyma and may not be representative of the whole organ. Hepatic cirrhosis may be missed in 24% of cases with a blind biopsy.[36] In one study of 3 consecutive samples obtained through a singleentry site, histologic findings were concordant for cirrhosis in all 3 specimens in only 50% of the 75 patients studied.[37] In yet another study, the amount of fibrosis on biopsy specimens was often not representative of the amount found at autopsy.[38] Second, it is difficult to grade hepatic fibrosis accurately. In the widely used Knodell system, the scoring of fibrosis is not linear: fibrous portal expansion is scored as 1, there is no score of 2, bridging fibrosis is scored as 3, and cirrhosis as 4.[39] A scoring system that evaluates hepatic fibrosis more precisely has recently been developed and validated the Metavir system. This system defines a score of 2 (rare septa formation) and is linear.[40] Although there is little interobserver variation in the interpretation of fibrosis with the Metavir system, morphometry with quantitative image analysis correlates best with serum markers for liver fibrosis when stains such as Masson trichrome or picrosirius red are used.[40,41] However, such methods are tedious and rarely used. These limitations should be kept in mind when the literature on the antifibrotic properties of interferon alfa is reviewed.

Experimental Evidence for an Antifibrogenic Effect of Interferon

Interferons have antifibrotic properties independent of their antiviral effect. For example, they have been shown to inhibit collagen synthesis by human fibroblasts.[42] More than a decade ago, interferon gamma was found to have antifibrogenic properties; it was shown to effectively inhibit collagen deposition in a murine model of schistosomiasis.[43] In culture, interferon gamma inhibited hepatic stellate cell proliferation and collagen synthesis[44]; in vivo, it reduced the hepatic expression of type I and type II collagens, decreased the number of activated hepatic stellate cells, but had no effect on the hepatic expression of TGFbeta.[45] Interferon gamma, however, is not used to treat chronic viral hepatitis. Interferon alfa, which is used in this setting, inhibits the proliferation of cultured human hepatic stellate cells and inhibits collagen secretion.[46] This antifibrogenic effect of interferon was first clinically observed in dermatology patients (interferon is commonly used to treat keloids)[47] and later in the hepatology setting. Interferon affects collagen degradation by activating collagenases. When rats were rendered fibrotic by 15 days of bile duct ligation, treatment with interferon alfa for the subsequent 15 days partially reversed the hepatic fibrosis.[48] Interferon alfa treatment increased the capacity of the liver to degrade type I and type III collagen and also increased the activity of plasminogen activator (plasminogens are thought to play a key role in balancing the proteolytic activities that mediate extracellular matrix integrity).[49] A doseresponse study using this design found that the lowest dose of interferon alfa (25,000 IU/d) had the highest antifibrotic effect.[50] There was no significant effect observed when bile ductligated rats were treated daily with 100,000 IU of interferon alfa for 4 weeks.[51] However, the daily administration of 100,000 IU of interferon alfa for 9 weeks prevented the development of liver fibrosis in another animal modelrats exposed for a long time to carbon tetrachloride and phenobarbitone.[51] The antifibrotic effect of interferon alfa may be mediated by cytokines. In patients with chronic hepatitis C, treatment with interferon alfa has been shown to reduce the hepatic expression of TGFbeta, and this reduction is associated with histologic improvement.[52]

Effect of Interferon Therapy on Progression of Hepatic Fibrosis in Chronic Hepatitis C

The primary goal of interferon therapy is eradication of HCV. However, it is possible that interferon may improve the natural history of hepatitis C independent of its antiviral activity by, for example, decreasing hepatic inflammation and fibrosis and reducing the risk of hepatocellular carcinoma. Numerous studies report that interferon decreases the severity of parenchymal and portal inflammation. This effect is most apparent in biochemical and virological responders but also occurs in the absence of a virologic or biochemical response.[53,54] In the first 2 randomized studies on the use of interferon alfa in the treatment of chronic hepatitis C, no decrease in extent of fibrosis was observed. However, in these studies, fibrosis was measured by the Knodell system, and this scale is not as sensitive as other scoring systems in patients with only moderate degrees of fibrosis.[2,3] In a landmark study by Castilla and colleagues,[20] patients with chronic hepatitis C were found to have elevated levels of TGFbeta mRNA in their livers. This finding correlated with levels of type I procollagen hepatic mRNA and with serum levels of type III procollagen. These markers of hepatic fibrosis decreased during therapy with interferon alfa,[20] but the study authors could not detect an effect on the fibrosis component of the Knodell score. However, the lack of linearity and sensitivity of the Knodell system at lower levels of fibrosis may have made it difficult to detect small changes in fibrosis. Other studies do show a decrease in fibrosis in similar settings. In a metaanalysis of the histologic effect of interferon alfa that included 460 patients in 10 randomized controlled trials, there was a significant decrease in fibrosis detected.[53]

Nine studies now report a significant improvement in hepatic fibrosis after interferon alfa therapy in patients with chronic hepatitis C (Table 1).[5563] All studies using a morphometric approach reported that interferon had an antifibrotic effect that was not detected when the Knodell scoring system was used.[5557] Of the 8 studies that have addressed the effect of interferon therapy on serum markers for hepatic fibrosis, all found a significant effect in responders (Table 2).[20,30,5961,6466]

Of at least equal importance is the fact that the antifibrotic effect of interferon may not be limited to patients who have sustained biochemical and virologic responses.[55,56,58,63] Interferon therapy can reduce hepatic inflammation in the absence of viral eradication. A 50% reduction in the sum of the inflammatory components of the Knodell score at the end of the treatment regimen was observed in 75% of biochemical responders and in 38% of biochemical nonresponders.[67] This finding was elegantly demonstrated in a metaanalysis that reported that as many as 51% of patients have histologic improvement despite failure to normalize ALT levels.[68] Because progression of fibrosis is driven by inflammation, a decrease in inflammation may itself lead to a longterm improvement in fibrosis. The beneficial effect of interferon on fibrosis in hepatitis C was underscored in a recent study by Sobesky and colleagues.[63] These investigators demonstrated that fibrosis stage (as determined by the Metavir scoring system) improved in 29% of treated responders, 18% of treated nonresponders (as determined by elevated 3month ALT activity), and in only 8% of controls; fibrosis worsened in 22%, 22%, and 56% of these patients, respectively. Moreover, the rate of progression of fibrosis was halted by interferon in both responders and nonresponders.[63]

This important antifibrogenic effect of interferon has not been recognized until recently. The fact that some trials combining interferon alfa2b with ribavirin for the treatment of hepatitis C have not detected an effect on fibrosis does not preclude it.[69,70] There are several reasons that may account for this seemingly contentious finding. First, scoring systems such as the Knodell scale may not be sensitive enough to detect small changes in fibrosis. The Knodell system divides most patients into only 2 categories: fibrous portal expansion (score = 1) and bridging necrosis (score = 3). Studies that use more complete and sensitive scoring systems[58,59,63] or those that use quantitative morphometry[5557] have found significant effects. Second, fibrosis is a dynamic process. However, its regression takes time and necessitates long duration of treatment and followup.[17,71] Third, until recently, attention has focused primarily on biochemical and virological outcomes. When histology was studied, only changes in inflammation and liver cell necrosis were carefully measured. In addition, histologic improvement can occur in a significant proportion of patients who are not sustained biochemical or virologic responders.

By reducing the severity of inflammation, repetitive periods of treatment may halt the progression of fibrosis or even allow its regression.[72] This effect may last for several months after completion of therapy,[64] suggesting that periods of treatment alternating with intervals without may suffice to prevent development of cirrhosis. The antifibrotic effect of interferon should be considered in the planning and analysis of studies that combine interferon or pegylated interferon with other antiviral drugs.

Conclusion

Interferon has an antifibrogenic effect that may be independent of its antiviral effect. Experimental data indicate that interferon can reduce and prevent deposition of fibrotic tissue in the liver. Clinical studies have confirmed this effect using both histologic evaluation and serum markers of hepatic fibrosis. Although interferon is cost effective in the treatment of hepatitis C because of its antiviral effect,[73,74] its antifibrogenic effect may increase its therapeutic indications and require re-evaluation of its applications in this setting. Because the longterm morbidity and mortality from this common chronic disease relates directly to degree of hepatic fibrosis, it is important to recognize that its antifibrotic effect does not always correlate with the conventional end points normalization of serum ALT levels and clearance of HCV RNA from blood. Given the slow turnover of fibrotic tissue, longterm treatment may be necessary. It will be important to determine if lower doses or shorter periods of treatment may be sufficient to maintain interferon's antifibrogenic effect and whether new pharmacologic forms of interferon with apparently higher antiviral efficacy and longer biologic halflives will also have greater antifibrogenic effects.