Intrahepatic heteropolymerization of M and Z alpha-1-antitrypsin

Abstract

The α-1-antitrypsin (or alpha-1-antitrypsin, A1AT) Z variant is the primary cause of severe A1AT deficiency and forms polymeric chains that aggregate in the endoplasmic reticulum of hepatocytes. Around 2%-5% of Europeans are heterozygous for the Z and WT M allele, and there is evidence of increased risk of liver disease when compared with MM A1AT individuals. We have shown that Z and M A1AT can copolymerize in cell models, but there has been no direct observation of heteropolymer formation in vivo. To this end, we developed a monoclonal antibody (mAb2H2) that specifically binds to M in preference to Z A1AT, localized its epitope using crystallography to a region perturbed by the Z (Glu342Lys) substitution, and used Fab fragments to label polymers isolated from an MZ heterozygote liver explant. Glu342 is critical to the affinity of mAb2H2, since it also recognized the mild S-deficiency variant (Glu264Val) present in circulating polymers from SZ heterozygotes. Negative-stain electron microscopy of the Fab2H2-labeled liver polymers revealed that M comprises around 6% of the polymer subunits in the MZ liver sample. These data demonstrate that Z A1AT can form heteropolymers with polymerization-inert variants in vivo with implications for liver disease in heterozygous individuals.

Keywords: Diagnostics; Genetic diseases; Genetics; Hepatology; Structural biology.

Figure 1. Identification of a mAb specific for the WT M A1AT.

(A) Anti-A1AT mAb_3C11 (nonconformationally selective, left panel) or mAb_2H2 (right panel) antibodies were used to probe purified M (blue) and Z (red) A1AT in either the monomeric (dashed lines) or heat-induced polymeric (solid lines) forms by antigen ELISA. Recognition of the samples by mAb_3C11 was approximately equal, but mAb_2H2 showed a preference for the M variant. (B) Interaction between immobilized mAb_2H2 and plasma-purified monomeric M (blue) or Z (red) A1AT variants in either the native or reactive loop-cleaved form. The relative maximal response above baseline was calculated from progress curves recorded at each concentration and is proportional to the mass of the material captured by the chip-bound antibody. Data are shown as ± SD (n = 3). The curves correspond with a hyperbolic function used to derive the K_D values for M A1AT (solid lines); this was not possible for the Z A1AT samples due to the limited binding observed over the concentration range (dashed lines). (C) Evaluation of mAb_2H2 specificity by immunofluorescence in cells. CHOK1 cells expressing either M or Z A1AT were seeded on coverslips, induced with doxycycline for 48 hours, permeabilized, and stained with anti–total A1AT mAb_3C11, anti–polymer mAb_2C1, or mAb_2H2. Cells expressing Z A1AT showed punctate staining with mAb_2C1 but no signal with mAb_2H2; conversely, cells expressing M A1AT were negative to mAb_2C1 and showed strong recognition by mAb_2H2. Both variants were well recognized by the control mAb_3C11. Scale bars: 15 μm.

Figure 2. Characterization of the 2H2 epitope.

(A) Central panel: the A1AT-Fab_2H2 complex (PDB accession 6I3Z) is shown, with the Fab heavy chain colored blue; the light chain colored green; β-sheets A, B, and C colored red, salmon, and yellow, respectively; and the site of the Z mutation indicated by a red ellipse. Arrows denote regions disordered in the crystal structure; none of these occur near the binding site. Left panel: the cleaved A1AT component of the complex is shown as surface-on-cartoon, with the Fab_2H2 binding site colored blue. Right panel: detail of interactions at the site of the Z mutation, with Lys290 at the center of a cluster of polar residues. (B) Detail of residues at the interface between A1AT and the Fab_2H2 heavy chain (V_H, left panel) or light chain (V_L, right panel). (C) Electrophoretic mobility shift assay using M, S, or Z A1AT incubated with an equimolar ratio of mAb_3C11 or mAb_2H2. The samples were resolved by nondenaturing PAGE and revealed by Coomassie blue staining. The A1AT monomer, mAb-bound A1AT, and noncomplexed mAbs are denoted by gray, black, and white arrowheads, respectively. Structural figures were prepared with PyMOL (Schrodinger).

Figure 3. The structure and composition of MZ liver polymers.

(A) A1AT polymers were extracted from the explant liver tissue of a ZZ homozygote and an MZ heterozygote. The purified material was resolved by nondenaturing PAGE with M A1AT monomer and heat-induced polymer for reference and visualized by immunoblot with the anti-A1AT polymer mAb_2C1 (right panel) and anti–total A1AT polyclonal antibody after stripping and reprobing the membrane (left panel). (B) The purified heat-induced M polymers, as well as MZ and ZZ liver polymers, were imaged by uranyl acetate negative-stain EM in the absence (top panels) and presence (bottom panels) of complexed Fab_2H2. Representative micrographs are shown. Scale bars: 60 nm. (C) Polymers with (blue) or without (red) at least 1 Fab_2H2 protuberance were classified according to shape and the number of constituent subunits recorded. The mean polymer length is indicated by the central bar ± SD; linear polymers with and without a detectable M component were 7.4 ± 4.0 and 6.5 ± 4.2 subunits in length, respectively, and circular polymers had 8.1 ± 2.6 (with) and 6.6 ± 2.6 (without) subunits (±SD). Polymer length differences in the presence (n = 53) or absence (n = 159) of detectable M subunits were not statistically significant by a Mann-Whitney U test. (D) Single-particle analysis of micrograph images of Fab_2H2-labeled heat-induced polymers, showing class sums representing the average of 111–624 dimer particle images each (columns 1 and 3) and the corresponding optimally oriented 3-dimensional structures (columns 2 and 4). The A1AT subunits are shown in blue, the Fab heavy chain in red, and the light chain in green. (E) The relationship between the dihedral angle defined by the centers of mass of the 2 Fab_2H2 molecules and A1AT molecules in the dimer is shown, along with the distance between the A1AT centers of mass, as obtained from the structures in D.

Figure 4. In vivo heteropolymerization with the Z variant.

(A) Plasma samples from individuals with MM, SS, ZZ, MS, MZ, or SZ genotypes were analyzed by sandwich ELISA. Total polymeric content was determined using anti-polymer mAb_2C1 as the capture antibody and the nonconformation-specific mAb_3C11-HRP as the detection antibody. Analysis by 1-way ANOVA with a Bonferroni’s multiple comparisons test (degrees of freedom = 84) showed ZZ plasma had a significantly higher polymer content than for the other genotypes (****P < 0.0001) and SZ plasma higher than MM and MS (^####P < 0.0001) and SS (^####P < 0.001). Each point is the average of 3 independent experiments on 1 sample, and data are shown as mean ± SD for individuals of the same genotype (n = 16, 17, 3, 20, 20, and 14 for MM, MS, SS, MZ, SZ, and ZZ, respectively). (B) The same plasma samples were analyzed in parallel using the same capture antibody and mAb_2H2-HRP as the detection antibody. Each point is the average of 3 independent experiments, and the mean ± SD of all samples of the same genotype are shown, with the same number of individuals as in A. Analysis by 1-way ANOVA with a Bonferroni’s multiple comparisons test showed the SZ plasma had a significantly higher recognition by mAb_2H2 than for the other genotypes (****P < 0.0001). (C) A model of M/Z A1AT heteropolymerization. It has been established (26, 40, 41) that, following expression, A1AT (rectangle) folds via a polymerization-prone monomeric intermediate (denoted by an asterisk) before adopting the native conformation (ellipse). The M variant (blue) is normally efficiently folded and secreted (lower panel); in the presence of Z A1AT (red), our data are consistent with the sequestration of a fraction of this material into polymers in a manner that permits further polymer extension (upper panel).