Introduction

Alpha-1-antitrypsin (a1AT) deficiency liver disease has a complex pathophysiology, and is highly variable between patients. A1AT is a serum glycoprotein synthesized in the liver and secreted into the serum, where its function is to protect host tissues by inhibiting neutrophil proteases released non-specifically during inflammation [1••, 2–4]. Over 100 variant alleles of the a1AT gene (SERPINA1) have been described, but the majority of patients with liver disease are ZZ homozygotes, referred to as ZZ or “PIZZ” in World Health Organization nomenclature [5]. The mutant Z gene directs the synthesis of a mutant protein which accumulates within hepatocytes rather than being efficiently secreted. Within the hepatocyte, the accumulated Z protein accumulates in the endoplasmic reticulum (ER), and may attain an unusual, polymerized conformation. ZZ adults have an increased risk of developing emphysema due to insufficient circulating a1AT in the lung to inhibit connective tissue breakdown. ZZ homozygous children and adults may also develop liver disease, cirrhosis, and hepatocellular carcinoma, resulting from the intracellular accumulation of a1AT mutant Z protein, which triggers cell death and chronic liver injury [6, 7].

Clinical Liver Disease

ZZ homozygous patients may exhibit highly variable disease [1••, 3, 8]. Infants may develop the “neonatal hepatitis syndrome,” which includes cholestatic jaundice, pruritus, poor feeding, poor weight gain, hepatomegaly, and splenomegaly [9, 10]. This can include elevated total or conjugated bilirubin, elevated serum transaminases, hypoalbuminemia, or coagulopathy due to vitamin k deficiency or to liver synthetic dysfunction. Liver biopsy findings may be highly variable in infants including giant cell transformation, lobular hepatitis, significant steatosis, fibrosis, hepatocellular necrosis, bile duct paucity, or bile duct proliferation [1••, 8, 11]. Differentiation from other cholestatic liver diseases of infancy by liver biopsy alone is not reliable. Globular, eosinophilic inclusions in some but not all hepatocytes are usually seen under conventional H&E stain, which represent dilated ER membranes engorged with polymerized a1AT mutant Z protein [12]. Staining with periodic acid-Schiff (PAS), followed by digestion with diastase to stain glycoproteins red, is used to highlight the “globules” (“PAS-positive”). Examination of liver biopsies for PAS-positive globules should be done with caution, as similar structures have sometimes been described in other liver diseases. Furthermore, the globules are not present in all hepatocytes or can be small and “dust-like,” or even absent, in small infants.

ZZ a1AT deficiency in older children may present as asymptomatic chronic hepatitis, failure to thrive, poor feeding, or as isolated hepatomegaly or splenomegaly. These various signs and symptoms of liver disease appear in less than 50 % of ZZ children, meaning that most patients escape diagnosis during childhood [9]. The risk of life threatening liver disease in childhood may be as low as 5 % [3, 13, 14]. Many children are completely healthy, without evidence of liver injury, except for mild and usually clinically insignificant elevations of serum AST or ALT. The liver biopsy findings in later childhood often become more classic with easy to identify, large globules in many but not all hepatocytes, steatosis, possible lobular inflammation, and possible fibrosis. Rarely, children with previously unrecognized chronic liver disease and cirrhosis present with ascites, gastrointestinal bleeding, or hepatic failure. Some children with severe liver disease in the first few months or years of life may enter a “honeymoon period” with few signs or symptoms and normal growth, before entering a period of renewed progressive injury and decompensation as teenagers. However, many patients with established cirrhosis and portal hypertension remain stable and grow normally for years or decades with minimal intervention [8, 15–17].

Progressive liver disease is uncommon in previously well young and middle-aged ZZ adults, although the risk of liver disease increases with advancing age [3, 4, 13, 18]. Adults may develop chronic hepatitis, with or without cirrhosis [4]. The biochemical and histopathologic findings in ZZ adult may be similar to those of alcoholic liver disease, which may lead to diagnostic confusion if specific serum testing for a1AT deficiency is not performed in patients undergoing evaluation for unexplained liver disease. Liver biopsy findings in adults may include lobular inflammation, variable hepatocellular necrosis, fibrosis, cirrhosis, steatosis, and PAS-positive, diastase-resistant globules in some, but not all, hepatocytes [8]. These findings can be similar to those of alcoholic liver disease. There also appears to be an increased risk of hepatocellular carcinoma in ZZ adults, although the magnitude of the risk is unclear [4, 11, 19]. Autopsy studies show that histologically significant, but possibly clinically undetected, liver injury and cirrhosis may be present in 40–50 % of elderly ZZ adults [4]. As middle-aged emphysema is more effectively treated, or prevented altogether, as a result of decreased smoking, it is possible that more older adults with ZZ liver disease will come to medical attention.

The diagnosis of a1AT deficiency does not require liver biopsy, although biopsy remains the gold standard to document the degree of liver injury. The gold standard for the diagnosis of a1AT deficiency is the genotype analysis of genomic DNA, or the analysis of the “phenotype” of a1AT protein in a patient’s serum. Some clinicians advise the use of a serum a1AT level as a screening test and then perform the gold standard test if the result is outside the normal range (see Table 1). This approach is common in some liver clinics, although the results should be interpreted with caution as a1AT is an acute phase reactant. Although a ZZ patient would not be expected to ever produce a level in the normal range, this author has seen SZ patients with liver disease occasionally have a1AT levels reported in the normal range during systemic inflammation. Care should also be taken not to obtain serum for a level or phenotype if the patient has recently had a plasma transfusion.

Table 1 a1AT phenotypes and corresponding typical a1AT serum levels
Full size table

Heterozygotes and Other Genotypes

Individuals carrying one normal M allele and one mutant Z allele (“PIMZ” or “MZ”), who represent approximately 2 % of US and European populations are usually asymptomatic and healthy with regard to liver disease [3]. However, data from retrospective, referral center studies show a 3- to 5-fold over-representation of MZ heterozygous patients in groups with chronic liver disease, sometimes in association with concurrent viral hepatitis [3, 15]. The most widely accepted explanation is that the MZ heterozygous state likely represents a genetic modifier of other liver diseases. There are anecdotal case reports of rare MZ adults developing liver disease, including the development of PAS-positive globules in hepatocytes, without other apparent risk factors for liver disease, although the possible genetic or environmental causes remain controversial [3, 20–22]. An MZ phenotype result is not readily accepted as the cause of otherwise unexplained liver disease without extensive further evaluation.

The S allele is another mutant commonly found in North America and Europe. SS individuals are well and free of lung and liver disease, although SZ individuals may develop liver disease identical to ZZ patients, including PAS-positive, diastase-resistant globules. While not well documented, the risk of liver disease in SZ is likely of lower magnitude than in ZZ [3, 23, 24]. Over 100 other rare mutations in the a1AT gene have been described, some of which yield a gene produce with a normal M migration on the phenotype gel, but, when present in the heterozygous state with a Z allele, can accumulate within the liver cause liver disease [3, 25, 26]. Such patients are usually recognized by a profoundly low a1AT level in peripheral blood that is not in keeping with low normal level expected by an “MZ” phenotype result. Serum deficiency states caused by null genes, or other unusual alleles, which do not direct the synthesis of a protein product which accumulates within the liver, are not associated with liver disease [3].

Management

Management of a1AT-associated liver diseases is based on standard liver disease supportive care, in order to prevent complications such as bleeding, ascites, pruritus, malnutrition, fat-soluble vitamin deficiency, infection, hepatocellular carcinoma, and growth disturbances, or to attenuate the effects if they do occur [1••, 3, 8]. There is no specific treatment of a1AT liver disease, at present. Some clinicians prescribe ursodeoxycholic, although the results of small clinical reports are inconclusive [27]. It is common practice for ZZ patients to be followed annually by a physician knowledgeable in liver disease, although many patients have normal health and can be monitored conservatively. Careful attention should be paid to many aspects of the patients’ health, not only liver symptoms. Patients should be urgently cautioned to avoid personal smoking and second-hand smoke, and to have regular lung function monitoring [3, 28, 29]. Laboratory tests should not exclusively focus on serum transaminase levels, as some patients with cirrhosis can have normal or nearly normal ALT and AST. The recommendations of the American Association for the Study of Liver Diseases (AASLD) guidelines for surveillance of liver cancer recommend liver ultrasound every 6 months for all liver disease patients with >2 % risk per year of liver cancer. Although data are lacking to firmly apply this recommendation to ZZ patients, a conservative approach would be to obtain liver ultrasound every 6 months in any patient with cirrhosis, significant fibrosis, or persistently large AST or ALT elevations. Patients with advanced liver disease should abstain from alcohol completely, while the recommendations are less clear for patients without advanced disease. Adult liver data in patients with hepatitis C but without significant injury suggest that up to the equivalent of the 3 alcoholic drinks per week may be safe (see AASLD guidelines). Studies in animal models of a1AT liver disease show that non-steroidal anti-inflammatory drugs (NSAIDS) may be uniquely toxic to the ZZ liver. NSAIDS in model systems increase a1AT mutant Z protein synthesis, increase the hepatic burden of accumulated mutant protein, and potentiate liver injury [30]. Acetaminophen is not known to have the same effect. It may be prudent to limit the intake of NSAIDS in ZZ patients. Exogenous a1AT replacement, which is commonly used to treat a1AT-associated emphysema, has no effect on the development of liver disease. Children with ZZ a1AT deficiency do not develop emphysema, although they may be at increased risk for childhood asthma [31, 32]. ZZ children are commonly referred to an adult pulmonologist at age 18 years for a baseline evaluation, unless asthma or other respiratory symptoms are present in childhood.

Pathophysiology of a1AT ZZ Liver Disease

The initiating event in liver injury in a1AT deficiency is the accumulation of the abnormal, a1AT mutant Z protein within liver cells (see Fig. 1) [33]. During biosynthesis, the a1AT mutant Z gene is appropriately transcribed and translated, followed by translocation of the nascent polypeptide chain into the ER lumen. In the ER, the mutant Z protein molecule folds slowly and inefficiently into its final, secretion-competent conformation [34, 35]. A system of quality control proteins in the ER recognizes these mutant Z molecules as abnormal and directs them to a series of proteolytic pathways rather than allowing secretion [36–41]. For reasons that are still not clear, some of the mutant Z molecules escape proteolysis and may attain a variety of abnormal conformations including a unique state in which multiple molecules aggregate to form large protein “polymers” [35, 42, 43]. This polymer conformation is highly thermodynamically stable and links large groups of mutant Z molecules together with non-covalent bonds. Accumulations within hepatocytes of the polymerized mutant Z protein may be large enough to be seen under light microscopy and represent the “globules” observed in hepatocytes in the ZZ liver. Some evidence suggests that accumulation of mutant Z protein in this specific polymer conformation is especially important in trigging liver injury [20, 21, 28, 44–46].

Fig. 1
figure 1

Hypothetical liver injury pathway in ZZ a1AT deficiency. The a1AT mutant Z protein is synthesized, but then retained in the ER of hepatocytes rather than efficiently secreted. Quality control processes within the cells direct most of the mutant Z protein molecules into intracellular proteolysis pathways (ERAD). However, some of the mutant Z protein molecules escape proteolysis and attain a unique, polymerized conformation forming inclusions in the ER. Autophagic degradation is upregulated as a response to mutant Z polymer accumulation. A small population of hepatocytes develops especially large accumulations of polymerized mutant Z protein and undergoes apoptosis. The hepatocytes with a smaller burden of mutant Z protein proliferate, possibly with the input of a liver stem cell population, to maintain the functional liver mass. The chronic process of injury, cell death, and compensatory proliferation leads to end organ fibrosis, cirrhosis, and HCC. Given the variable nature of clinical liver injury between individuals with the same genotype, and the usually slow disease progression, there are likely to be important environmental and genetic disease modifiers affecting the rate and magnitude of these processes

Full size image

Once the mutant Z protein is retained in the ER, the cell employs a variety of proteolytic processes in an attempt to reduce the intracellular mutant Z protein burden and reduce injury. These include ubiquitin-dependent and ubiquitin-independent proteasomal pathways, as well as other mechanisms sometimes referred to as “ER-associated degradation” (ERAD) [37, 39, 44, 45]. It is thought that the proteasomal pathways as a part of ERAD are the primary route for degradation of a1AT mutant Z in the non-polymerized conformation. Although many of the mechanistic steps in the degradation process, and their specific sequence, are still under investigation, previous work has shown that two molecules present in the ER, calnexin and ER manosidase I (ERmanI), are likely to be critical points of control. Calnexin is a transmembrane ER chaperone which binds a1AT mutant Z, becomes targeted for degradation by linkage to ubiquitin, and then is degraded as this trimolecular complex (a1AT mutant Z-calnexin-ubiquitin) by the proteasome [40]. Studies in human fibroblast cell lines established from ZZ homozygous patients show that patients susceptible to liver disease have less efficient ER-associated degradation of a1AT mutant Z protein than ZZ patients without liver disease [24, 41]. The reduced efficiency of degradation in the liver disease patients presumably leads to a greater steady state burden of mutant Z protein within liver cells and increased liver injury. Studies of the enzyme ERmanI also suggest that it may also have a critical role in directing a1AT mutant Z molecules to the proteasome for degradation. These data raise the possibility that allelic variations in these genes, or in other genes involved in the quality control or proteolytic systems, might alter susceptibility to liver injury by changing the efficiency of degradation [36, 46]. There has been a report that susceptibility to liver disease might also be related to allelic variations in the a1AT gene itself, which would not otherwise be considered disease-associated mutations [47, 48]. Another important proteolytic pathway appears to be autophagy. This is a highly conserved degradation system in which specialized vacuoles degrade abnormal proteins and larger structures such as senescent organelles. Studies show that the accumulation of the polymerized a1AT mutant Z protein within cells induces an autophagic response, and that autophagy is an important route for the degradation of a1AT mutant Z polymers. In experimental systems, liver injury can be reduced by increased autophagic degradation of mutant Z polymer protein [49••, 50••].

Clinically, liver injury in ZZ humans is usually a slow process which takes place over years to decades, and analysis of human livers have shown that accumulation of a1AT mutant Z protein is very heterogeneous among each individual hepatocyte. In the past, it has been difficult to reconcile these clinical data with in vitro, cell biological mechanistic studies. New insights from recent studies show that a cellular injury cascade is triggered within the small population of hepatocytes with the largest a1AT mutant Z polymerized protein accumulation [6, 33, 51–54]. Hepatocytes with the largest a1AT mutant Z protein accumulation, perhaps only a few percent of the total hepatocytes, have increased caspase activation and increased susceptibility to apoptosis [38, 51, 52, 54]. There is also a newly recognized component of oxidative injury [55••]. These processes cause a low, but higher than normal, baseline rate of hepatocyte death in ZZ liver tissue compared to normal liver [2]. The cells with low polymer accumulation then proliferate to maintain the functional liver mass. Over time, the continued stress, death, and repair leads to liver fibrosis, cirrhosis, HCC, and chronic organ injury. Environmental and genetic modifiers of protein secretion, degradation, apoptosis, or regeneration would then by hypothesized to influence the progression of liver disease in an individual patient.

New Therapeutic Approaches

As understanding of the cellular mechanisms of injury in this disease has improved, so have hypotheses for new approaches to treatment. A variety of gene therapy approaches have been examined for many of the single gene defects, including a1AT, with limited success in human trials, and also with significant safety concerns. Mouse models of ZZ a1AT deficiency have been successfully treated with adeno-associated viral vectors, or other vectors, containing siRNAs, ribozyme, and other tools for inhibiting transcription or translation of the mutant gene. In one recent publication involving the mouse model, it was even possible to inhibit the mutant Z protein synthesis while generating wild-type, M, a1AT protein synthesis with an exogenous mRNA [56••]. There has also been intense interest in small molecules which might alter the folding of the a1AT mutant Z molecule to prevent retention, prevent polymerization, allow secretion, or promote degradation. If secretion could be increased and polymerization in the periphery prevented, then lung disease might also be treated by such an approach. The drug 4-phenylbutyrate, a so-called “chemical chaperone,” was shown in cell culture and in animal models to increase secretion of a1AT mutant Z protein. However, it failed to show benefit in a human trial [57].

Recently, new studies of two drugs, rapamycin and carbamazepine, aimed at decreasing the intracellular accumulation of mutant Z protein in the liver by stimulating autophagic degradation, have garnered significant attention [49••, 50••]. Autophagy is involved in the degradation of a1AT mutant Z protein polymers, as discussed above, and studies in model mice of these drugs showed reduced intrahepatic mutant Z protein, reduced hepatic fibrosis, and a reduction in other markers of liver injury. However, human studies of rapamycin have not commenced due to concerns of the potential toxicity. One study of carbamazepine in ZZ patients with end-stage cirrhosis has been undertaken, but to insure safety it will only examine a dose 1/10th as large as the effective dose observed in the animal model (Clinicaltrials.gov #NCT01379469). It is unclear if this approach will lead to benefits and, for now, use of these rapamycin and carbamazepine in a1AT are not recommended outside organized clinical trials.

Other strategies are also being explored, including drugs aimed at inhibiting points in these intrahepatic injury cascades, anti-fibrotic treatments, and hepatocyte transplant approaches. As in many other liver diseases, research into anti-fibrotic treatments and cell transplants is intense, but as yet has not led to workable clinical treatments. Cyclosporin has shown promise in animal studies in a1AT to inhibit the mitochondrial signaling during apoptosis to reduce liver damage, but it has not been tested for this use in humans [58].

Conclusions

Homozygous ZZ alpha-1-antitrypsin deficiency is an under-diagnosed, metabolic liver disease which can affect adults and children. The clinical manifestations are highly variable, with many patients remaining healthy or exhibiting only mild biochemical abnormalities until late in life. Genetic and environmental disease modifiers are thought to be improtant, but are still poorly understood. Accumulation of the a1AT mutant Z protein within hepatocytes activates an intracellular injury cascade of apoptotic liver cell death and compensatory hepatocellular proliferation leading to end organ injury. There is no specific treatment for a1AT-associated liver disease, other than supportive measures and liver transplant. Research into new treatment approaches has led to new clinical trials, and holds out hope for improved clinical outcomes in the future.