fax: (46) 40-337041
Liver Failure and Liver Disease
Differential detection of PAS-positive inclusions formed by the Z, Siiyama, and Mmalton variants of α1-antitrypsin
Article first published online: 14 OCT 2004
Copyright © 2004 American Association for the Study of Liver Diseases
Volume 40, Issue 5, pages 1203–1210, November 2004
How to Cite
Janciauskiene, S., Eriksson, S., Callea, F., Mallya, M., Zhou, A., Seyama, K., Hata, S. and Lomas, D. A. (2004), Differential detection of PAS-positive inclusions formed by the Z, Siiyama, and Mmalton variants of α1-antitrypsin. Hepatology, 40: 1203–1210. doi: 10.1002/hep.20451
- Issue published online: 14 OCT 2004
- Article first published online: 14 OCT 2004
- Manuscript Accepted: 6 JUL 2004
- Manuscript Received: 4 MAY 2004
- Swedish Research Council
- Lund University
- Medical Research Council (UK)
- Wellcome Trust
- Alpha-one foundation
- Papworth NHS Trust
Several point mutations of α1-antitrypsin cause a perturbation in protein structure with consequent polymerization and intracellular accumulation. The retention of polymers of α1-antitrypsin within hepatocytes results in protein overload that in turn is associated with juvenile hepatitis, cirrhosis, and hepatocellular carcinoma. The detection of α1-antitrypsin polymers and understanding the molecular basis of polymer formation is of considerable clinical importance. We have used a monoclonal antibody (ATZ11) that specifically recognizes a conformation-dependent neoepitope on polymerized α1-antitrypsin to detect polymers within hepatocytes of individuals with α1-antitrypsin deficiency. Paraffin-embedded liver tissue specimens were obtained from individuals who were homozygous for the Z (Glu342Lys), Mmalton (52Phe del), and Siiyama (Ser53Phe) alleles of α1-antitrypsin that result in hepatic inclusions and profound plasma deficiency. Immunohistological staining with a polyclonal anti-human α1-antitrypsin antibody showed hepatic inclusions in all 3 cases, while ATZ11 reacted with hepatic inclusions formed by only Z α1-antitrypsin. Polymers of plasma M and Z α1-antitrypsin prepared under different conditions in vitro and polymers of recombinant mutants of α1-antitrypsin demonstrated that the monoclonal antibody detected a neoepitope on the polymerized protein. It did not detect polymers formed by a recombinant shutter domain mutant (that mirrors the effects of the Siiyama and Mmalton variants), polymers formed by cleaving α1-antitrypsin at the reactive loop, or C-sheet polymers formed by heating α1-antitrypsin in citrate. In conclusion, the ATZ11 monoclonal antibody detects Z α1-antitrypsin in hepatic inclusions by detecting a neoepitope that is specific to the polymeric conformer and that is localized close to residue 342. (HEPATOLOGY 2004;40:1203–1210.)
α1-Antitrypsin is a proteinase inhibitor that is synthesized in the liver and secreted into the circulation.1 The average concentration of α1-antitrypsin in plasma in individuals with 2 normal M alleles (Pi M) is approximately 1.3 mg/mL, with a half-life of 3 to 5 days. Biochemical and crystallographic studies have revealed that the structure of α1-antitrypsin is based on three β-sheets (A-C) and an exposed mobile reactive loop that presents the P1-P1′ methionine-serine residues as a pseudosubstrate for the target proteinase2 (Fig. 1). The proteinase cleaves at the P1-P1′ bond and is swung from the upper to the lower pole of the molecule in association with insertion of the reactive loop as an extra strand in β-sheet A.3 The reactive loop/β-sheet A interaction is crucial for α1-antitrypsin to function as an antiproteinase, but this conformational flexibility can also be detrimental because point mutations facilitate the sequential insertion of the reactive site loop into a β-sheet of another molecule, thereby leading to formation of polymers.1, 4, 5 It is this polymeric α1-antitrypsin that is retained within hepatocytes in association with the severe Z (Glu342Lys),4, 6 Siiyama (Ser53Phe),7 and Mmalton (52Phe del)8 deficiency variants. The accumulation of α1-antitrypsin as polymers within the endoplasmic reticulum of hepatocytes causes protein overload that is associated with juvenile hepatitis,9 cirrhosis, and hepatocellular carcinoma.10 It is generally believed that the lack of circulating α1-antitrypsin results in uncontrolled proteolytic attack and early-onset panacinar emphysema.11
Most experimental data indicates that polymers of mutant α1-antitrypsin form by the sequential insertion of the reactive center loop of one molecule into β-sheet A of another (Fig. 1).4, 7, 12–16 However, there is evidence from crystallographic studies,17, 18 biophysical analysis,19–21 and the Mmalton variant,8 that polymers may also form by linkage between the reactive center loop of one molecule and β-sheet C of another (Fig. 1). Another alternative is that polymers form by linkage of the reactive center loop as strand 7A.22, 23 Polymers linked by β-sheet C or as strand 7A have been reported in crystal structures, but these were not stable when dissolved in aqueous solution.17, 18, 22, 23 We have used the monoclonal antibody ATZ11, which specifically recognizes an epitope on polymerized α1-antitrypsin,6 to explore the conformation of polymers of different mutants of α1-antitrypsin within hepatocytes. The specificity of the antibody has been characterized using polymers of plasma α1-antitrypsin formed under different conditions and polymers of recombinant α1-antitrypsin induced by different mutations.
Materials and Methods
Liver sections from 5 individuals with α1-antitrypsin deficiency were included in the study. Two patients had the Z allele (1 homozygous Pi Z and 1 heterozygous Pi MZ), 2 had the Mmalton mutation (1 homozygous Mmalton and 1 heterozygous M-Mmalton), and 1 individual was homozygous for the Siiyama mutation. In addition, 2 liver specimens from individuals with the normal Pi M α1-antitrypsin phenotype were used as controls. DNA from all of these individuals was isolated from peripheral lymphocytes and the α1-antitrypsin gene was sequenced as described previously.24
The liver tissue specimens were fixed in formalin and embedded in paraffin. Serial sections were stained with either hematoxylin or periodic acid-Schiff (PAS) with and without treatment with diastase. Immunohistochemistry was performed according to the standard peroxidase-antiperoxidase method25 with a commercial polyclonal anti–α1-antitrypsin antibody (1:5000; Dako, Denmark) or the monoclonal ATZ11 (1:100) antibody.6
Human plasma M and Z α1-antitrypsin were purified as detailed previously26 or purchased from Calbiochem (Boston, MA),. Polymers of M α1-antitrypsin were prepared by incubating the protein at 1 mg/mL in 0.015 mol/L Tris-HCl containing 0.15 mol/L NaCl, pH 7.4, or 20 mmol/L acetate buffer, pH 4, or phosphate buffered saline (PBS) at 60°C for 3 hours; polymers of Z α1-antitrypsin were prepared by incubating the protein in PBS at 57°C for 4 hours. M α1-antitrypsin cleaved at the reactive center loop (Fig. 1B) was prepared by incubating α1-antitrypsin in a 50:1 ratio with StaphylococcusAureus V8 protease at 37°C for 2 hours. Complete cleavage of the protein was confirmed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis. C-sheet polymers (Fig. 1E) were prepared by heating M α1-antitrypsin in 0.7 M citrate adjusted to pH 8 at 60°C or 70°C for 3 hours as detailed previously.27 Reactive loop cleaved α1-antitrypsin polymers (Fig. 1G) were prepared by incubating plasma M α1-antitrypsin at 1 mg/mL with 0.01 mg/mL of papaya proteinase IV (PP4; CN Biosciences, Nottingham, UK), a glycine-specific cysteine protease that cleaves α1-antitrypsin at the P9-P10 bond of the reactive loop, at 37°C for 2 hours.28 Recombinant Glu342Ala and His334Ala α1-antitrypsin were expressed and purified as described previously.29 Polymers of these mutants were prepared by heating at 0.5 mg/mL and 50°C (His334Ala α1-antitrypsin) or 55°C (Glu342Ala α1-antitrypsin) in 0.1 mol/L Tris containing 5 mmol/L ethylenediaminetetraacetic acid, pH 7.4, for 30 minutes.
Western Blot Analysis.
Nondenaturing polyacrylamide gel electrophoresis (PAGE) was performed in a Mini-PROTEAN II electrophoresis cell (Bio-Rad Laboratories, Hercules, CA) on 7.5% wt/vol acrylamide gels as described previously.27 Electrophoretically separated samples were transferred to a polyvinylidene difluoride membrane (Imobilon-P; Millipore Corporation, Bedford,MA) in transfer buffer (25 mmol/L Tris, 190 mmol/L glycine) using a semidry blot electrophoretic transfer system (Trans-Blot SD; Bio-Rad Laboratories). After the transfer, the membranes were blocked with 5% wt/vol nonfat dried milk in PBS containing 0.1% vol/vol Tween 20 at 4°C overnight. Blots were incubated for 2 hours with a polyclonal rabbit antibody against human α1-antitrypsin (Sigma-Aldrich, St Louis, MO; 1:10,000) or the mouse monoclonal antibody ATZ11 (1:50 or 1:5,000). The protein bands were visualized by incubation with horseradish peroxidase-conjugated secondary antibody against rabbit or mouse immunoglobulins (1:25,000), followed by enhanced chemiluminescence.
DNA sequencing of the entire α1-antitrypsin gene confirmed the Z (Glu342Lys in exon V), Mmalton (deleted 52Phe in exon II), and Siiyama (Ser53Phe in exon II) alleles in affected individuals, but no mutations were found in the α1-antitrypsin gene from the 2 Pi M controls. The histopathological evaluation of the liver sections with light microscopy revealed features of established cirrhosis in the Pi Z homozygote, the Pi MZ heterozygote, and the Mmalton homozygote. The M-Mmalton heterozygote showed features of chronic active hepatitis while sections from the liver from the Siiyama homozygote showed a mild nonspecific reactive hepatitis and signs of chronic venous congestion in a context of substantially preserved lobular architecture. Examination of tissue sections from all the patients with α1-antitrypsin deficiency showed round eosinophilic inclusions in parenchymal liver cells in hematoxylin-stained preparations. The inclusions were positive with periodic acid-Schiff staining after diastase digestion (PASD; Fig. 2) and strongly reacted with the polyclonal anti–α1-antitrypsin antibody (Fig. 3). In keeping with previous observations,30 the amount of immunostained intrahepatic α1-antitrypsin was much higher when compared to that detected with either hematoxylin or PASD. Serial sections from the same livers with the monoclonal ATZ11 antibody showed strong positive staining of the PASD α1-antitrypsin inclusions in hepatocytes from the Pi Z and Pi MZ α1-antitrypsin individuals (Fig. 3; Table 1). In contrast, inclusions from both the Mmalton homozygote and the Siiyama homozygotes did not stain with this antibody. The 2 Pi M α1-antitrypsin control livers showed no inclusions of α1-antitrypsin. There was no PASD staining or staining with either the polyclonal or monoclonal anti–α1-antitrypsin antibody.
|Mutant or Conformer of α1-Antitrypsin||Detected by ATZ11 Monoclonal Antibody*|
|Z α1-antitrypsin (Glu342Lys) in liver biopsy||+++|
|Siiyama α1-antitrypsin (Ser53Phe) in liver biopsy||−|
|Mmalton (52Phe del) α1-antitrypsin in liver biopsy||−|
|M α1-antitrypsin monomer in vitro||−|
|M α1-antitrypsin A sheet polymers in vitro||++|
|M α1-antitrypsin C sheet polymers in vitro||−|
|Z α1-antitrypsin (Glu342Lys) monomer in vitro||+|
|Z α1-antitrypsin (Glu342Lys) polymers in vitro||+++|
|Recombinant Glu342Ala α1-antitrypsin polymers||+++|
|Recombinant His334Ala α1-antitrypsin polymers||−|
|Reactive loop cleaved α1-antitrypsin polymers||−|
|Reactive loop cleaved α1-antitrypsin||−|
The cause of the difference in immunostaining between inclusions formed by the Z and Siiyama or Mmalton variants was investigated by characterizing the binding of the monoclonal antibody to polymers prepared under defined conditions in vitro. Reactive loop–β-sheet A polymers of α1-antitrypsin can be prepared by heating M α1-antitrypsin at raised temperatures or by incubating Z α1-antitrypsin under physiological conditions.4, 26 C β-sheet polymers form by incubating α1-antitrypsin at pH 4,21 or heating in citrate,27 and are structurally distinct from A β-sheet polymers (Figs. 1C and E). In the present study, loop–β-sheet A polymers were prepared by heating purified wild type α1-antitrypsin at 1 mg/mL and 60°C in 0.015 mol/L Tris-HCl containing 0.15 mol/L NaCl, pH 7.4, and loop–β-sheet C polymers were formed by heating M α1-antitrypsin at 1 mg/mL and 60°C in 0.7 mol/L citrate buffer, pH 8. The resulting polymers were analyzed by 4% to 20% wt/vol nondenaturing PAGE followed by immunoblotting with the polyclonal anti–α1-antitrypsin or the ATZ11 monoclonal antibody. Our data clearly show that both the ATZ11 monoclonal antibody and the polyclonal antibody recognize polymers of α1-antitrypsin that are formed by β-sheet A linkage (Figs. 4A and B, lanes 1 and 3; Table 1). However, in contrast to the polyclonal antibody, the ATZ11 monoclonal antibody did not detect polymers formed in citrate (Figs. 4A and B, lane 2) or at low pH (data not shown). Moreover, the monoclonal antibody did not detect polymers of α1-antitrypsin prepared by heating in 0.7 mol/L citrate, pH 8, at 70°C for 2 hours (data not shown). The finding that ATZ11 antibody failed to react with citrate polymers of α1-antitrypsin provides evidence that the neoepitope recognized by ATZ11 antibody on polymerized α1-antitrypsin at physiological pH does not exist on polymers formed under nonphysiological conditions. Further analysis showed that the ATZ11 monoclonal antibody detected polymers of both M and Z α1-antitrypsin, but the Z α1-antitrypsin polymers were detected at lower concentrations of antibody (Figs. 5 and 6A; Table 1).
The difference in staining of hepatocytes between the individual with the Z α1-antitrypsin mutation and those expressing the Mmalton or Siiyama mutation (Fig. 3), and the preference of the antibody for Z rather than M α1-antitrypsin polymers, suggests that the ATZ11 monoclonal antibody detects a neoepitope that is exposed close to amino acid 342, which is mutated from a glutamate to a lysine in Z α1-antitrypsin. To explore this in more detail, recombinant α1-antitrypsin was prepared in which the 342Glu residue was mutated to Ala (it is not possible to purify recombinant Glu342Lys α1-antitrypsin from Escherichia coli because it readily forms aggregates). Mutants of α1-antitrypsin were also prepared in which His334 was mutated to Ala. This mutation disrupts an important hydrogen bond network within the shutter domain and mimics the effects of the Mmalton and Siiyama mutations.29 Both of these substitutions generate a mutant of α1-antitrypsin that favors polymer formation more readily than the wild type protein. The mutants were polymerized by heating at 0.5 mg/mL and 50°C (His334Ala α1-antitrypsin) or 55°C (Glu342Ala α1-antitrypsin) for 30 minutes, and the resulting polymers were assessed by nondenaturing PAGE and Western blot analysis. The monoclonal antibody detected polymers formed by the Glu342Ala mutation (Fig. 6A, lane 6; Table 1) but not those formed by shutter domain mutants (Fig. 6A, lane 5). The antibody did not detect monomeric Glu342Ala α1-antitrypsin when assessed alone (data not shown) or as a mixture with polymeric Glu342Ala α1-antitrypsin (Fig. 6A, lane 6). The neoepitope that is revealed upon polymer formation is not merely expansion of β-sheet A because the monoclonal antibody does not recognize reactive loop cleaved α1-antitrypsin (Fig. 1B and Fig. 6A, lane 8). Moreover, the antibody did not detect polymers of α1-antitrypsin formed by the cleaved reactive loop inserting into β-sheet A of another α1-antitrypsin molecule (Fig. 1G and Fig. 6A, lane 7). Taken together, the different patterns of detection seen in the livers of individuals with the Z, Siiyama, and Mmalton variants are due to the monoclonal antibody detecting a neoepitope that is revealed on polymer formation and located close to residue 342.
Over 75 allelic variants of α1-antitrypsin have been reported and classified according to their migration in isoelectric focusing analysis.31 However, only 3 of these variants, Z,32 Siiyama,33 and Mmalton,34 are associated with the formation of intracellular hepatic inclusions in association with profound plasma deficiency. Although the mutations underlying these severe deficiency variants are in different regions of the molecule, they are linked by their propensity to form intracellular polymers.1, 5 Indeed it is now possible to correlate the rate at which mutants of α1-antitrypsin form polymers with the severity of the intracellular inclusions and concomitant plasma deficiency.35
We have assessed the inclusions formed in association with the Z, Siiyama, and Mmalton mutations by immunohistochemistry with a polyclonal anti– α1-antitrypsin antibody and the ATZ11 monoclonal antibody. The monoclonal ATZ11 antibody was prepared by immunizing mice with globular α1-antitrypsin inclusions purified from the liver of an individual with Z α1-antitrypsin deficiency.36 It is able to detect the α1-antitrypsin in the plasma of individuals who are homozygous or heterozygous for the Z allele and to identify Z α1-antitrypsin in liver tissue sections. Indeed a combination of a polyclonal antibody to α1-antitrypsin and the monoclonal antibody ATZ11 have been used to determine the prevalence of Z α1-antitrypsin in liver biopsies.37 We have confirmed these findings by showing that sections from the livers of individuals with Z α1-antitrypsin stain intensely with both the polyclonal and the ATZ11 monoclonal antibody.
The finding of globular inclusions of abnormal α1-antitrypsin in the rough endoplasmic reticulum of hepatocytes is a characteristic feature of the accumulation of Z α1-antitrypsin, but it also seen with other rare variants, such as Mmalton (Cagliari) and Siiyama. In this and in previous studies,24 we have shown that the globular inclusions in hepatocytes from Z α1-antitrypsin individuals stained with both the polyclonal and the ATZ11 monoclonal antibody, while Mmalton liver sections reacted with the polyclonal, but not with the monoclonal antibody. The differences are not a nonspecific effect of differing severity of tissue damage because the sections from both the Z and Mmalton homozygotes showed features of established cirrhosis. The immunostaining of the hepatocytes from a Mmalton homozygote was mirrored by the staining of hepatocyes from an individual who is homozygous for another shutter domain mutant, Siiyama (Ser53Phe). Once again, α1-antitrypsin was stained with the polyclonal, but not the monoclonal, antibody. Recent investigations have shown that the ATZ11 monoclonal antibody detects the polymeric and enzyme-complexed conformers of α1-antitrypsin.6 However, little is known about its ability to bind to polymers of α1-antitrypsin formed by different linkages (i.e., via interaction with β-sheet A or C) or in association with different mutations. We therefore investigated the cause of the difference in immunostaining between inclusions formed by the Z, Siiyama, or Mmalton variants by characterizing the binding of the monoclonal antibody to polymers prepared under defined conditions in vitro.
The monoclonal antibody recognized polymers of Z α1-antitrypsin prepared in vitro. It also detected polymers of M α1-antitrypsin and monomeric Z α1-antitrypsin but at much higher concentrations of antibody. The antibody did not detect polymers formed at low pH or in citrate that are likely to be linked by β-sheet C rather than β-sheet A.21 Taken together with the data from immunostaining, it is apparent that the monoclonal antibody preferentially detects an epitope on polymers of Z α1-antitrypsin. This was explored in more detail by assessing the ability of the antibody to bind to reactive loop cleaved monomeric α1-antitrypsin (in which β-sheet A is filled with its own reactive loop cleaved at P4-P5), polymers of reactive loop cleaved α1-antitrypsin (in which β-sheet A is filled by the reactive loop of another molecule cleaved at P9-P10), and polymers formed by a mutation in the shutter domain (His334Ala) that mimics the effects of the Siiyama and Mmalton mutations. All of these polymers were recognized by a polyclonal antibody, but none were detected with the ATZ11 monoclonal antibody. The specificity was further assessed by generating recombinant α1-antitrypsin in which the glutamic acid at position 342 (the same position as the Z deficiency allele) is mutated to alanine. This mutant more readily formed polymers than did the wild type protein, and these polymers were recognized by both the polyclonal and monoclonal antibody. Monomeric Glu342Ala α1-antitrypsin was recognized by the polyclonal but not by the monoclonal antibody. Thus the neoepitope that is detected by the ATZ11 monoclonal antibody must lie close to position 342 and be revealed upon the formation of reactive-loop β-sheet A polymers. It is not the Z mutation (Glu342Lys) itself because the monoclonal antibody detected polymers formed by Glu342Ala α1-antitrypsin. Moreover, previous studies have shown that the antibody does not detect a linear peptide containing the Glu342Lys substitution.6 Our previous work has shown that the Z mutant favors the formation of polymers by perturbing the relationship between strand 5 of β-sheet A and the reactive center loop. The mutation causes a partial insertion of the reactive center loop into β-sheet A (Fig. 1H) and the opening of the lower part of β-sheet A,15 which then acts as an acceptor for the reactive center loop of another α1-antitrypsin molecule.1, 4 In contrast, the shutter domain mutants underlie the middle of β-sheet A, causing opening of the lower region but having less effect on the reactive loop (Fig. 1A). It must be this difference that accounts for the difference in specificity of the ATZ11 monoclonal antibody.
In summary, our data show that the structural perturbations that underlie the polymerization of Z α1-antitrypsin are different from those that underlie the formation of polymers of Siiyama and Mmalton α1-antitrypsin. The specificity of the ATZ11 monoclonal antibody can be used to detect polymers in the PAS-positive inclusions formed by Z α1-antitrypsin in liver biopsies from affected individuals. However, a lack of staining of PAS-positive inclusions with the monoclonal antibody that are recognized by a polyclonal anti–α1-antitrypsin antibody indicates that the underlying molecular defect results from the Mmalton or Siiyama alleles.
The authors thank Robin Carrell, University of Cambridge, for helpful advice.