Abstract
Testing the blood for evidence of hepatic damage and dysfunction frequently involves measuring several blood constituents simultaneously to screen for disease. While useful, this approach occasionally leads to apparent disparities between the blood test results, and the results of other diagnostic tests such as histology. In part, these perceived discrepancies may stem from a lack of appreciation for tissue, cellular, and molecular factors that affect the appearance of hepatic disease biomarkers in the blood. Further confusing the matter is that in some instances the mechanisms responsible for the appearance of diagnostic compounds in blood are only partially understood. Many of the known factors that affect hepatic biomarkers are similar to those affecting other tissue markers, while others are unique to the liver, such as those involved with cholestasis. Disease conditions can also cause misleading results by affecting tissue concentrations of test compounds, hepatic mass, and the clearance rate of compounds from the blood. Knowledge of the factors affecting the blood concentrations of biomarkers, as well as investigations into the mechanisms behind changes to hepatic biomarker concentrations, may allow for a better interpretation of blood test results and fewer inconsistencies between diagnostic results.
Introduction
Clinical pathology can be used to test for the presence of hepatic injury by sampling the peripheral blood. Blood sampling is a noninvasive technique that can be done without sacrificing or even harming the test subject and for that reason can be valuable in screening for hepatic injury, even repeatedly, over the course of a study. The liver is a complex organ, containing several cell types and carrying out a diverse group of functions. It can therefore become diseased or fail in numerous ways, and no one biomarker can serve as a universal hepatic screening test. Upon occasion, clinical pathology tests can give confusing, conflicting, or perceived erroneous results that appear to not agree with other methods such as necropsy and histopathology findings. It is very likely that when such discrepancies occur, some unpredicted change has happened to the biomarker that was induced by the pathophysiologic state. In toxicological studies it is common to have unintended effects that might alter the interpretative power of clinical pathology tests. If discrepancies occur between clinical pathology testing and other methods it is frequently important to determine what may be the cause. The location, concentration, form, and degree of access of biomarkers to the blood can all change and could be responsible for unpredictable changes and resulting difficulties in data interpretation. Changes to the blood clearance of hepatic biomarkers could also have an effect. Furthermore, species differences affect the value of any one biomarker.
There is no single reason for incongruent clinical pathology test results and multiple factors could be responsible for discrepancies in any single case. However, there are logical places to look for causes when discrepancies occur. By so doing, illogical results may become coherent, and the significance of any changes better interpreted. This manuscript will examine several of the factors that affect the ability of serum biomarkers to correctly detect hepatic injury. Although the exact mechanisms responsible for the appearance of hepatic disease biomarkers in the blood is frequently not known, an understanding of the known factors can often help in the interpretation and understanding of clinical pathology tests, which can be vital in cases of disparate results.
Pathways Taken by Biomarkers from the Liver to the Blood
Hepatocytes are polar epithelial cells and as such have apical and basolateral surfaces. The basal surfaces of hepatocytes face the space of Disse, which contains interstitial fluid and few cells. Hepatic sinusoids contain the perfusing blood and are separated from the space of Disse by a lining of endothelial and Kupffer cells. Substances leaving the hepatocytes therefore cross the space of Disse and endothelial cells to gain access to the blood. Diffusion across the interstitial space is unlikely to impede the appearance of enzymes in the blood. Attenuations of the endothelial cell cytoplasm form fenestrations. Most of these fenestrations are 100 to 150 nm in diameter, and tend to cluster into sieve plates (Wisse, 1970; Tanikawa, 1979). They are dynamic structures whose diameters are affected by changes to luminal blood pressure, vasoactive substances, drugs, and toxins. These fenestrations likely regulate the passage of large structures such as cells, but allow free exchange of even large proteins such as solubilized enzymes. Hence, like the space of Disse, the endothelial cells do not offer an appreciable obstruction to the passage of solubilized hepatic biomarkers into the blood. The escape of biomarkers from the hepatocytes and bile is therefore more important to their appearance in the blood than is altered diffusion across the space of Disse or devitalization of endothelial cell function.
Mechanisms of Appearance of Cholephilic Biomarkers in the Blood
The biliary system begins at the bile canaliculi, denoting the apical surfaces of hepatocytes (Jones, 1996; Crawford, 1999; Tolman and Rej, 2001). The canaliculi form a band around each hepatocyte. Tight junctions separate the bile canaliculi from basolateral surfaces. These normally obstruct movement of substances between the bile and blood. Pulsatile contractions of the bile canaliculi help to propel the bile through the biliary tree and ultimately into the duodenum through the sphincter of Oddi (Watanabe et al., 1991). Decreased bile flow, i.e., cholestasis, results in increased concentrations of biliary constituents in the blood. Cholestasis can occur from a variety of hepatic insults including chemically induced changes to bile flow as well as physical obstruction of flow through the biliary tree (Eakins, 1978). The precise mechanisms by which biliary constituents gain access to the blood are still not known, but there has been substantial progress made in understanding the basic mechanisms, and they appear to not be as simple as once thought.
There is some evidence to suggest that the regurgitation of small molecular weight compounds, such as bile salts and bilirubin, occurs through leaky tight junctions caused by hepatic injury including bile duct obstruction and exposure to cholestatic toxins (Boyer, 1983; Alpers et al., 1990). Experimental evidence of leaky tight junctions includes observations that the rare earth metal, lanthanum, has access to the blood when infused in a retrograde fashion into the biliary tree of animals with cholestatic disease (Toyota et al., 1984). Lanthanum perfusion fixation of livers in situ also leads to the appearance of the lanthanum in the bile canaliculi, but only if the tight junctions have been functionally compromised by a hepatic insult (Takakuwa et al., 2002). Hence, the regurgitation of small molecular weight compounds through leaky tight junctions is generally accepted to occur, particularly in severe or prolonged forms of cholestasis (Takaki et al., 2001).
Cellular Redistribution of Cholephilic Enzyme Activity
Several pathways for the escape into the blood of large molecular weight biliary constituents such as enzymes have been proposed including, paracellular regurgitation of bile through leaky tight junctions, and by transcellular transport pathways (Alpers et al., 1990). Paracellular bile regurgitation of enzymes has been proposed to occur by mechanisms similar to those of small molecular weight compounds and have been supported by studies showing the movement of horseradish peroxidase through tight junctions (Lowe et al., 1988; Konno et al., 1992). Retrograde infusion of ferritin and polymeric and secretory forms of IgA have all shown the possibility of reversed transcytosis from the apical to the basolateral surfaces of membrane vesicles with cholestasis (Carpino et al., 1981; Jones et al., 1984). Such movement would be an alternate transport pathway of large molecules moving into the blood and could occur without the necessity of distortion to tight junctions. Nevertheless, several features of much of the cholephilic enzyme activity found in the blood are not entirely answered by either of these pathways.
These features include the fact that much of the cholephilic enzyme activity found in the blood is relatively low molecular weight and hydrophilic, while almost all of the activity found in bile is hydrophobic (Inoue et al., 1980; Toda et al., 1980; Tsuji et al., 1980; Selvaraj et al., 1984), and the hydrophobic forms found in blood frequently present within large particles that have biochemical characteristics consistent with their originating from basolateral, not apical, membranes (De Broe et al., 1985; Kihn et al., 1991). These particles are not found in bile and are believed to be membrane fragments shed from the basolateral surfaces of liver cells (Deng et al., 1996b; Van Hoof et al., 1997). Hence, cholephilic enzymes likely have at least 2 subcellular origins in cholestasis; from the apical surfaces, perhaps by regurgitation or reverse transcellular transport, and from the basolateral surfaces by membrane fragment formation. In certain forms of hepatic injury it is very likely that the basolateral surfaces are actually the preponderate source of cholephilic enzymes in serum. The recognition of the importance of the basolateral surfaces of liver cells as a source of cholephilic enzymes includes that it could explain instances where there is an increase in serum cholephilic enzyme activity in noncholestatic disease. Indeed, increased enzyme induction may be all that is necessary for increased activity in serum. Any drug or compound that induces de novo synthesis of the cholestatic biomarkers could therefore result in their increased serum activity without cholestasis. The diagnostic impact of this is that serum cholephilic enzyme activity can increase in diseases other than cholestasis.
Although all enzyme biomarkers of cholestasis are membrane bound, the majority of serum cholephilic enzyme activity is believed to be normally hydrophilic (Price and Sammons, 1974; De Broe et al., 1975; Tsuji et al., 1980).
While this could imply the production of specialized secretory forms of these enzymes, it has actually been shown that the enzymes are synthesized in their normal form but then become solubilized through the loss of their hydrophobic membrane anchoring motifs (Solter and Hoffmann, 1995, 1999). The anchors are cleaved by either proteases or, in the case of enzymes with GPI anchoring domains, by glycosyl phosphatidyl inositol (GPI) phospholipase D (Deng et al., 1996a). The solubilization of cholestatic enzymes can occur while bound either to basolateral membranes, or to enzymes regurgitated from bile. For solubilization to take place, enzymes must be adequately solubilized from adjoining membranes to allow access of the cleavage enzymes to the cleavage site (Low and Huang, 1991).
This is an important concept because it has been shown that cholephilic enzymes attached to intact membrane are not susceptible to cleavage enzyme activity. This is most likely due to steric hindrance. For release from the basolateral membranes, increased bile acids concentrations are considered essential for partially solubilizing the membranes through detergent action, which relieves steric hindrance and allows access of the enzymes to the cleavage sites (Van Hoof et al., 1997). Without detergents, there is little release of enzyme activity from cytoplasmic membranes (Low and Huang, 1991; Deng et al., 1996a; Solter and Hoffmann, 1999). Because hepatic bile acids concentrations increase with cholestasis (Greim et al., 1972; Hatoff and Hardison, 1981; Solter et al., 1994), there exists the appropriate conditions to allow enzyme release.
The release of increased amounts of cholephilic enzymes from the basolateral surfaces of hepatocytes is also facilitated by an increased expression and redistribution of the cholephilic enzymes to the basolateral surfaces of the hepatocytes and on biliary epithelial cells (Bulle et al., 1990; Solter and Hoffmann, 1995; Solter et al., 1997). This is likely due to a combination of events including induction of increased de novo synthesis of cholephilic enzymes in the case of alkaline phosphatase (ALP, EC3.1.3.1) (Hatoff and Hardison, 1979; Solter et al., 1994; Wiedmeyer et al., 2002a, 2002b). The basolateral appearance of cholephilic enzymes is also facilitated by the fact that after synthesis, all apical enzymes are believed to first be transported to the basolateral surfaces before arriving at their final site on the bile canaliculi (Bartles et al., 1987; Schell et al., 1992; Maurice et al., 1994). Increased basolateral surface enzyme activity allows access of the cholephilic enzymes to the blood without the necessity of bile regurgitation.
Evidence to support this pathway includes experimental choledococaval shunts in rats which cause the continuous recirculation of bile through the liver (Putzki et al., 1989; Ogawa et al., 1990). This model has shown that serum cholephilic enzyme activity will increase in the absence of increased biliary pressures, alterations to tight junctions, or changes in biliary permeability. Hence, increased biliary pressure and tight junction permeability are not necessary for the release of cholephilic enzymes. This has been most extensively shown with ALP, whose serum activity correlates with the amount and type of hepatic bile acid load (Ogawa et al., 1990). Bile acids likely increase enzyme activity through increased enzyme synthesis and release (Hatoff and Hardison, 1981, 1982). Increased serum ALP activity also occurs in canine glucocorticoid hepatopathy, however serum ALP activity increases without concurrent increases in hepatic bile acids concentrations or other evidence of cholestasis (Solter et al., 1994). Bile acid perfusion of the liver during the enterohepatic circulation of bile acids is thought to provide the detergency necessary for enzyme release, while glucocorticoids cause induction of enzyme synthesis (Solter et al., 1997).
The Appearance of Cytosolic Enzymes in Blood
It is generally accepted that increased cytosolic enzyme activity in the blood occurs secondary to hepatic damage or necrosis. Unlike cholephilic enzymes, these enzymes are not membrane bound and therefore do not require release from membrane anchors. However, healthy plasma membranes should be impermeable to macromolecules such as enzymes. Therefore, plasma membrane integrity must be compromised in some way for cytosolic enzymes to appear in the blood. There are two basic hypotheses regarding what degree of cell integrity must be lost before cytosolic enzymes will escape from cells in general (Kristensen, 1994; Mair, 1999). First, the release of cytosolic enzymes occurs only in instances of irreversible cell damage and therefore their appearance in blood is always indicative of cell death. The second hypothesis states that cells release cytosolic enzymes during both the reversible and irreversible phases of cell injury and therefore appearance in blood does not necessarily indicate cell death. Of the two, the latter hypothesis seems currently most accepted and is supported by experimental studies as well as the evolution of a better understanding of the pathophysiological mechanisms of reversible cell injury (Piper et al., 1984; Cotran et al., 1999). From an interpretive standpoint, whether or not only necrotic cells can release enzymes is not inconsequential, as it has a direct effect on the clinical explanation of what increased serum cytosolic enzyme activity implies about cell viability and recovery from hepatic injury.
General observations favoring enzyme release during reversible cell damage include the apparent lack of histologic evidence of necrosis in spite of increased serum enzyme activity (Frederiks et al., 1983). Conceivably however, this may also reflect a lack of sensitivity of histologic techniques to detect cell necrosis when it is patchy, early on, or involves only a small number of cells. Nevertheless, additional findings made over the previous 2 decades have led to a growing acknowledgement that the escape of enzymes from cells likely includes mechanisms other than cell death (Gores et al., 1990; Kamiike et al., 1989; Kristensen, 1994; Mair, 1999). The term “cell leakage” is frequently used to describe this phenomenon and, although rarely defined, implies an escape of cytosolic enzymes through either perforations or tears. However, the actual process by which cytosolic enzymes enter the blood from reversibly damaged cells is unlikely to occur in this way. While increased membrane permeability is a well-known outcome of irreversible cell damage, it is unlikely that cells sustaining the formation of perforations or tears of adequate size to allow the leakage of macromolecules such as enzymes could maintain adequate viability to recover. Rather, such conditions would be anticipated to result in cell rupture. Hence, a mechanism other than pore formation or tears most likely occurs if cytosolic enzymes can escape from cells during periods of reversible cell damage. The mechanism proposed to explain the appearance of cytosolic enzymes in blood with reversible damage is by the formation of membrane blebs that detach and allow the cell membrane to reseal without cell death (Lemasters et al., 1983; Gores et al., 1990; Kristensen, 1994; Mair, 1999). Rupture of these membrane blebs most likely occurs in the space of Disse or blood to release the enzyme contents.
Cell membrane bleb formation has been well described following hypoxic insults to several tissues including hepatocytes (Gores et al., 1990; Kristensen, 1994; Cotran et al., 1999; Mair, 1999) and likely reflects the development of 2 sequential events. Namely, decreased cell energy stores followed by increased cytosolic calcium concentrations (Cotran et al., 1999). The impaired energy metabolism from hypoxia results in depletion of ATP. This allows an influx of sodium and chloride and an efflux of potassium from cells by osmotic diffusion gradients across the cytoplasmic membrane. Lactate and other anaerobic metabolites, including adenosine, inosine, phosphates, and intermediates of glycolysis increase the osmotic activity of the cell. The low energy stores and increased osmotic load result in cell swelling and damage to the mitochondria. At some stage during cell injury, the cell cannot exclude calcium ions and there is an influx of calcium that increases cytosolic calcium concentrations. This activates intracellular phospholipases, endonucleases, and proteases resulting in the formation of eicosinoids and adversely affecting the membrane lipid content and cytoskeletal architecture through proteolysis and disruption of the phosphorylation state of cytoskeletal proteins. The combination of cell swelling, alteration of membrane lipid composition and disruption of cytoskeletal proteins results in the formation of membrane blebs. If damage to the cell membrane is more severe, rupture of the cell membrane occurs and irreversible cell damage ensues.
The ultrastructural changes of reversible cell damage confirm plasma membrane blebbing, which has been well described under several conditions such as ischemia/ hypoxemia, shock, or toxins that inhibit energy metabolism (Gores et al., 1990). Direct membrane damage has also been hypothesized to result in reversible membrane leakage through a variety of mechanisms including the depletion of energy stores (Kristensen, 1994). For example, the presence of detergents or activation of phospholipases could also modify or degrade the phospholipid bilayer resulting in loss of function of membrane ion pumps and channels that begins the cascade of events leading to cell swelling and the formation of cytoplasmic blebs. Energy depletion could exacerbate the effects of membrane damage and a net loss of ATP could occur through attempted reparation of membrane damage.
The degree to which blebbing causes cytoplasmic enzyme activity to increase is controversial but is likely less than following cell necrosis. For example, it has been estimated that the amount of enzyme that can be released by membrane blebs is likely only a few percent of the total cellular content (Kristensen, 1994). Hence, greater increases of serum enzyme activity likely reflect principally irreversible cell damage and necrosis while mild increases may indicate mostly membrane blebbing and therefore, reversible cell damage. Such a model is consistent with most clinical and pathologic interpretations of the relative degree of increased serum enzyme activity.
Molecular Weight and the Appearance of Enzymes in Blood
There is no definitive evidence suggesting that molecular weight affects the appearance of hepatic enzymes in the blood during the reversible stage of cell injury, although it has been rarely tested in an empirical manner. In theory, molecular weight could affect the ability of compounds to escape from the liver through pores occurring in damaged membranes, by affecting the diffusion of enzymes through interstitial spaces, or by affecting passage through intercellular tight junctions. Enzyme leakage through tight junctions has already been discussed, but there is little evidence to suggest that enzyme size affects sieving through the compromised tight junctions. Some selectivity of leakage based on size has been described for hepatocytes, but was only a minor factor among many including, intercellular and intracellular location, enzyme induction, and the state of liver health and type of injury (Schmidt and Schmidt, 1987). In other tissues, notably perfused rat heart, molecular weight does not affect the rate of release of soluble enzymes from cells. Indeed, the effect of enzyme size on the appearance of myocardial enzymes in blood has been deemed, “overemphasized and overestimated” (Mair, 1999). Once plasma membrane integrity is lost, cytosolic enzymes begin to appear in the interstitium at the same rate, independent of size (Altona et al., 1984; Diederichs et al., 1986). In contrast to membrane pore formation, there is substantial empirical evidence for plasma membrane blebbing with hepatocellular injury, but this would be anticipated to offer at most a minor size exclusion effect for soluble cytoplasmic enzymes. In contrast to release from cells, there is a slight effect of molecular weight on diffusion through the interstitial spaces as observed in some tissues, such as the heart (Mair, 1999). However, given the direct access of hepatic enzymes to the blood, these time differences are likely inconsequential relevant to hepatic injury.
Effects of Cellular Location and Intracellular Compartmentalization
In order for serum enzyme activity to increase, the enzyme must have access to the blood. Without such access, enzymes will expel into nonblood areas. For example, gamma glutamyltransferase (GGT, EC2.3.2.2) is a membrane-bound enzyme found in substantial quantities on the luminal side of the brush border of renal epithelial cells in the proximal convoluted tubule (Albert et al., 1964). Renal GGT is therefore mostly excreted into urine, not blood (Price, 1982). Hepatic GGT activity however, has a broader cellular location including biliary epithelial cells, and over the basolateral surfaces of hepatocytes where it has direct access to blood (Lanca and Israel, 1991). Hence, most of the serum GGT activity in blood is from the liver.
Enzymes of diagnostic importance may also be present within intracellular organelles or otherwise bound to cell constituents that restrict their ability to freely move into the interstitial spaces following loss of cell membrane integrity (Altona et al., 1984; Schmidt and Schmidt, 1987; Mair et al., 1994). The mitochondrion is of particular clinical importance in this regard, and enzymes present in the mitochondrial matrix are not as readily released into the blood as cytosolic enzymes. For example, plasma membrane blebbing does not appear to be adequate for the loss of enzymes from the mitochondrial matrix (Kamiike et al., 1989). This can have important implications on the use of such enzymes as markers of hepatic injury. A substantial amount of hepatocellular aspartate aminotransferase (AST; EC 2.6.1.1) is present in mitochondria, ranging from approximately 81% to 85% in rat hepatocytes and from approximately 30% to 40% in canine hepatocytes (Baumber and Doonan, 1976; Pappas, 1980, 1986; Keller, 1981). Assessment of AST release from rat liver following ischemia has shown that hepatocytes do not lose mitochondrial AST (mAST) until there is appreciable loss of cell integrity, probably necrosis (Kamiike et al., 1989). Cells with minimal loss of integrity, resulting in the formation of only membrane blebs do not lose mAST. In fact, ischemic liver does not lose mAST until almost all cytosolic AST is lost. It has also been observed that following treatment in rats with the hepatotoxin, carbon tetrachloride, more cytoplasmic AST enters the blood than mAST in spite of a greater synthesis of mAST, implying that more severe forms of hepatic injury are necessary for the release of mitochondrial matrix enzymes than cytosolic enzymes (Pappas et al., 1984; Pappas, 1986).
Glutamate dehydrogenase (GLDH, EC1.4.1.3) is a second enzyme found in the mitochondrial matrix (Schmidt and Schmidt, 1988). It is considered hepatospecific, and in early studies, increased serum activities was considered more dependent on irreversible cell damage than are cytosolic enzymes (Clampitt and Hart, 1978; Schmidt and Schmidt, 1988). However, the findings of more recent studies have been less clear-cut, suggesting that GLDH may appear in blood without appreciable evidence of hepatocellular necrosis (O’Brien et al., 2002). The reason is not known, but it was hypothesized that this may reflect residual cytoplasmic GLDH from enzyme synthesis, or mitochondrial material within released cytoplasmic membrane blebs. These results may also imply a loss of mitochondrial membrane integrity with some forms of hepatic injury which allows the leakage of GLDH into the cytoplasm and into the blood.
Enzyme Concentration Gradients and Tissue Mass can Change with Hepatic Injury
The concentration of an enzyme within a tissue and the overall mass of each organ can have substantial effect on the degree to which serum enzyme activity may change following its release from cells (Henderson and Moss, 2001). The relative size and high enzyme content of the liver gives substantial advantage to development of enzyme biomarkers for hepatic injury over that of several other tissues. One graphic example that has been given is comparing the release of enzymes from the prostate and liver in man (Henderson and Moss, 2001). The concentration gradient of prostatic acid phosphatase between the prostate and blood is 10,000:1, while the concentration gradient of ALT (EC2.6.1.2) between liver and blood is 100,000:1. Therefore, substantially fewer cells need be damaged in the liver than in the prostate to detect an increase in serum enzyme activity. In addition, the prostate weighs approximately 20 grams, while the liver is over 1,000 grams. Hence, extensive destruction of the liver will result in tremendously abnormal serum enzyme levels relative to extensive prostatic disease.
The accurate assessment of blood concentrations of biomarkers of hepatic injury is frequently dependent on a stable tissue concentration between the controls or pretreatment individuals and the diseased individuals. In reality, this is often not the case. The concentration of enzymes may be affected by both induction and inhibition of enzyme synthesis and this can be affected by disease. For example in dogs treated with the glucocorticoid, prednisone, there is an almost 4-fold increase of hepatic GGT activity per gram of tissue (Solter et al., 1994). The total hepatic GGT activity increases even further due to an approximate doubling in hepatic mass with prednisone treatment. Diminished tissue enzyme concentrations secondary to disease may also affect the ability to correctly interpret serum enzyme levels. This has been shown to be true when using serum ALT activity as an indicator of hepatic damage following exposure to microcystin-LR, a hepatotoxin found in harmful algal blooms (Solter et al., 2000). Hepatic ALT synthesis decreases in rats following treatment with microcystin-LR. This results in decreased hepatic ALT concentrations and causes serum ALT activity to be an insensitive indicator of hepatocellular damage relative to other hepatic enzymes, such as sorbitol dehydrogenase (SDH; EC1.1.1.14).
The relative ratios of different enzymes in serum may also change due to changes in tissue concentrations, even within the same tissue. With chronic liver disease in man for example, hepatic ALT concentrations decline more than AST, resulting in an increase in the serum AST/ALT ratio as disease progresses (Schmidt and Schmidt, 1993). The tissue concentrations of hepatic biomarkers may also differ between species, and affect the efficacy of an assay to determine hepatic injury. Hepatic ALT activity for example, is relatively low in sheep, and is well known as an insensitive test for hepatic injury in that species (Clampitt and Hart, 1978).
Alterations to Blood Clearance Rates with Hepatic Injury
Serum half-life has a significant effect on the concentrations of hepatobiliary compounds in blood. Enzymes with very short serum half-lives may be virtually useless as diagnostic enzymes because their appearance in blood is too transient. For example, on a per gram basis the small intestinal mucosa contains higher ALP activity than liver (Clampitt and Hart, 1978). However in dogs virtually none of the serum ALP activity is intestinal in origin (Solter et al., 1994). In large part this is due to the extremely short serum half-life of the intestinal isoenzyme, which is measured in minutes in dogs (Hoffmann et al., 1999). In contrast to the intestinal isoenzyme, the canine hepatic isoenzymes of ALP have blood half-lives measured in days, and they make up the majority of total serum ALP activity in dogs. In large part, the reason for the difference in half-life between ALP isoenzymes is the nature of their glycosylation (Kuhlenschmidt et al., 1991). ALP is removed from the serum by asialoglycoprotein receptors in the liver. These receptors recognize terminal galactose residues, which are covered by sialic acid. An ALP isoenzyme of liver origin is only cleared from the blood after it has had the sialic acid residues removed by endogenous neuraminidase activity while in the circulation. In contrast, canine ALP isoenzyme of intestinal origin contains relatively little sialic acid, which is removed much more rapidly than liver ALP. Hence, the duration that an intestinal isoenzyme of ALP spends in the blood is much shorter than ALP originating from the liver, and this affects the total serum activity of each.
Changes to the rate of clearance with hepatic injury of drugs, cholephilic organic anions and some endogenous compounds, such as bile acids, have also been exploited as biomarkers of hepatic injury. The concept of drug clearance was first applied to the physiology of the disposition of therapeutic drugs, but was later found to also apply to how drug disposition affects them as indicators of hepatic function (Branch, 1982; Papich and Davis, 1985; Boothe, 1990). The concepts likely apply to the clearance of organic anions and endogenous cholephilic compounds, such as bile acids, as well. Hepatic clearance has been well described elsewhere (Wilkinson and Shand, 1975; Williams, 1983; Papich and Davis, 1985) and is a measure of the efficiency of an organ to permanently remove a substance from the blood perfusing that organ and is expressed in terms of the volume of blood from which the substance is removed per unit of time. Although systemic clearance is the sum of the clearances of each individual organ, in essence, only the liver clears substances used as liver function tests. Therefore for such compounds, hepatic clearance (CLH) is synonymous with systemic clearance. At steady state, CLH is dependent upon both liver blood flow (Q) and the efficiency of hepatic extraction, which can be stated as the hepatic extraction ratio (E).The terms are related through the equation:
where Ci is the concentration of the compound in the blood entering the liver (i.e., portal vein and hepatic artery), and Co is the concentration of the compound in the blood leaving the liver (via the hepatic vein). Hence, the hepatic extraction ratio, or the fractional difference between Ci and Co, measures the efficiency of the liver to eliminate a substance from the blood as it perfuses the liver. The hepatic extraction ratio of a compound can vary between 0 and 1. For substances with an E value of nearly one, CLH is principally limited by blood flow through the liver, i.e., CLH = Q and clearance is said to be “flow-limited” (Papich and Davis, 1985; Boothe et al., 1992). Flow-limited compounds, which have high intrinsic clearances include lidocaine and propranolol, which have been implicated as potential hepatic function test compounds (Branch, 1982). Cholephilic dyes and endogenous organic anions, such as bromosulphthalein, indocyanine green, and bile acids can also be considered flow-limited compounds in most species (Gilmore and Thompson, 1980, 1981; Branch, 1982; Zhao et al., 1993) (although perhaps not indocyanine green in the dog (Ketterer et al., 1960; Boothe et al., 1992)). In theory, intrinsic metabolic processes should not be the major factor influencing the clearance of these compounds, whereas changes in hepatic blood flow should have substantial effects. In fact, excluding the rare instances of congenital defects affecting hepatocellular uptake of specific organic anions, even with severe forms of hepatocellular failure, the hepatic clearance of such compounds can be considered largely dependent on hepatic blood flow.
Many liver diseases have been speculated to result in changes to hepatic blood flow (Williams, 1983; Papich and Davis, 1985; Boothe, 1990; Crawford, 1999). For example, hepatic inflammation can result in large quantities of blood bypassing hepatocellular contact secondary to proliferation of blood vessels and obstruction of normal channels by collagen deposition. Chronic hepatic disease, such as cirrhosis, resulting in destruction of normal liver architecture further diminishes hepatocellular blood flow. Collateral vein formation may divert large quantities of blood away from the liver. On the other hand, conditions that result in diminished hepatic blood flow such as congestive heart failure or shock may also result in decreases in CLH , which corresponds directly with reduction in blood flow (Papich and Davis,1985). Such conditions may therefore confound the diagnosis of primary liver failure.
In contrast to flow-limited compounds, clearance of substances having a relatively low E value will be little affected by changes in blood flow (Branch, 1982; Williams, 1983; Papich and Davis, 1985; Boothe, 1990). Low-extraction compounds have been termed “capacity limited” (Papich and Davis, 1985; Boothe et al., 1992). Unlike highly extracted compounds, CLH may be affected in capacity limited compounds by plasma protein binding. The clearance of capacity limited/protein binding-sensitive compounds is limited principally by protein binding since only unbound compound can be cleared by the liver. Thus, CLH will parallel protein binding for such compounds. Capacity limited/binding insensitive compounds are not bound or are poorly bound to plasma proteins. For such compounds, liver clearance parallels intrinsic clearance, which is the inherent ability of the liver to remove a particular substance from the blood and includes, hepatocellular uptake, intracellular transport, metabolism, and biliary secretion (Branch, 1982; Boothe et al., 1992). Capacity limited/binding insensitive compounds include antipyrine, aminopyrine, and caffeine (Papich and Davis, 1985; Boothe et al., 1992). Substances in this category have been used experimentally to assess metabolic capacity of the liver and usually reflect hepatic cytochrome P450 concentration or total hepatocellular mass (Statland et al., 1973; Vesell et al., 1973; Boothe et al., 1992; Branch, 1982; Monroe et al., 1982; Renner et al., 1984; Schaad et al., 1995; Tanaka and Breimer, 1997).
The use of the different categories of compounds eliminated by the liver could theoretically allow one to quantitate different aspects of liver function, but although speculated for some time, it has not found widespread use, perhaps due to a lack of ability to distinguish clearly between different forms of hepatic injury (Wilkinson and Shand, 1975; Branch, 1982; Hofmann, 1982; Boothe, 1990; Boothe et al., 1992). What may prove more useful in the future will be the testing of the clearance of multiple compounds simultaneously as a means of gauging the activity of several cytochrome P450 enzymes simultaneously in toxicological and clinical studies (Tanaka et al., 2003).
Summary
Numerous factors affect the diagnostic interpretation of biomarkers of hepatic injury. Of the factors that could have potential impact, the ability to discern the cell of origin, the concentration gradient, and organ size, cell sub-compartments, and serum half-life are likely the most important in affecting the appropriate interpretation of altered enzyme results. Contrary to conventional wisdom, other than some effect upon diffusion through interstitial spaces and capillaries, the molecular weight of the enzymes has minimal effect on the appearance and interpretation of diagnostic enzyme results. With regard to so-called cholestatic enzymes, the redistribution and release from basolateral surfaces of hepatocytes and increased induction with various forms of drugs and disease conditions makes cholestasis, increased biliary pressures, and changes to tight junction permeability unnecessary for their appearance in the blood. Hepatic clearance is another element of hepatic injury that lends itself to clinical pathology testing and can be used to assess alterations to hepatic blood flow, hepatic mass, and specific cytochrome P450 levels. Both xenobiotics and endogenously produced substances have been used. The multiple functions and diseases of the liver make it best to use multiple tests for assessing hepatic injury and changes to function.