The characterization of p170 antigen recognized by PNP sera has proved to be a challenging task. Our study unravels the identity of this protein targeted by PNP autoantibodies as the protease inhibitor A2ML1. This conclusion is based on several lines of evidence: 1) the tryptic mass profile of the immunoprecipitated p170 has a significant match with that expected for A2ML1; 2) p170 is recognized by anti-A2ML1 antibodies and can be immunoprecipitated from culture media of human keratinocytes, in the same manner as A2ML1; 3) PNP sera immunoprecipitate recombinant A2ML1 from cell extracts, whereas binding to A2ML1 was never observed with sera obtained from normal volunteers (n
52) as well as patients with autoimmune bullous diseases of the skin (n
56); 4) p170-reactive PNP sera selectively labelled transfected cells expressing recombinant A2ML1; 5) pre-incubation of p170-reactive PNP sera with recombinant A2ML1 selectively abrogated reactivity of the PNP sera with p170 by immunoblot or immunoprecipitation, and further reduced the labelling of the epidermal granular cell layers, where A2ML1 is predominantly expressed. Together, our data unravel a novel class of proteins targeted by autoantibodies in patients suffering from this devastating multiorgan disease.
A2ML1 is a broad range protease inhibitor belonging to the A2M protease inhibitor family. It binds different classes of proteases and inhibits their activity or reduces the substrate spectrum by a “trap mechanism” in which the inhibitor covalently binds the protease and creates a sterical hindrance within the active site 
. During this process, the conformational change releases the COOH-terminal extension of A2ML1, which then can be recognized by lipoprotein receptor-related protein 1 (LRP1) receptor to allow the internalization and clearance of the complex protease inhibitor-protease 
. A2ML1 is expressed in many tissues such as the epidermis, thymus, and testis, while A2ML1 ESTs reported in UniGene mainly arise from normal and tumoral stratified epithelia 
. Furthermore, the EST expression profile (Hs.620532, NCBI Unigene, EST profile viewer) suggests the presence of high transcript levels of A2ML1 in oesophagus, mouth, pharynx, intestine, and muscle. Interestingly, comparison of the sequence of A8K2U0 and CAD48670, which displays the highest score with p170 tryptic mass profile suggests the existence of splice variant(s) of A2ML1.
The exact function of A2ML1 is not yet defined. By analogy to A2M which has been proposed to be an element of the innate immunity 
A2ML1 may participate in defense mechanisms by binding to inflammatory cytokines, growth factors and by targeting a broad range of proteases 
. Furthermore, A2ML1 is likely to be directly implicated in the maintenance of epidermal homeostasis based on its ability to form covalent complexes with the kallikrein KLK7 in vitro 
, a protease involved in proteolysis of intercellular structures and in desquamation process 
Our findings provide an explanation for the failure to identify p170 so far. First, the full length cDNA sequence of A2ML1 in humans was reported in 2004 
, while its functional characterization was carried out in 2006 
. Therefore, indexation and theoretical profiling of A2ML1 in databases have been only available in the past few years. Second, the biochemical properties of A2ML1 constitute an additional challenge for its identification. A2ML1 is a secreted glycoprotein with a secondary structure constrained by multiple disulfide bridges 
. Furthermore, over-heating of A2ML1 in reducing-denaturing conditions results in its cleavage in two polypeptides of 120 kDa and 60 kDa (
and personal observations), which reduces the signal intensity at the expected electrophoretic migration. Finally, our analyses demonstrate that PNP anti-A2ML1 autoantibodies recognize conformational epitopes, explaining the two-decade failure to detect p170 by Western blot under denaturing conditions 
. It is likely that production of A2ML1 in an eukaryotic expression system ensuring proper folding and posttranslational modifications facilitates its detection.
There is no ortholog of human A2ML1 in mouse, rendering the direct in vivo
demonstration of the pathogenicity of anti-A2ML1 antibodies by passive transfer studies impossible. Accordingly, gene targeting experiments are not feasible. Furthermore, no hereditary human disorder has been mapped to the A2ML1 gene locus so far. Nevertheless, there are a number of indirect observations providing support to the idea that auto-antibodies against p170 are involved in the initiation or the perpetuation of tissue damage, since: 1) in a significant number of reported PNP patients, including patient 9 of the present study, autoantibody reactivity against p170 is found alone or with few additional reactivities, at least by immunoprecipitation 
; 2) PNP sera binding to p170 can be detected already at an early stage of the disease 
; 3) A2ML1 is not only expressed in skin but also in other organs affected in PNP (Hs.620532, NCBI Unigene, EST profile viewer); 4) acquired or genetic defects of protease inhibitors may cause a variety of muco-cutaneous diseases with systemic symptoms 
. For example, mutations in SPINK5 gene encoding the serine protease inhibitor lympho-epithelial Kazal-type inhibitor (LEKTI) cause Netherton syndrome 
associated with chronic skin inflammation and skin barrier defects. LEKT1 is thought to regulate kallikrein activity 
The impact of auto-antibodies on biochemical properties and function of A2ML1 could not be determined, due to difficulties to obtain sufficient amounts of purified A2ML1. However, two studies document the negative impact of the binding of autoantibodies to protease inhibitors. First, in acquired autoimmune angioedema, autoantibody binding to C1-inhibitor (C1-inh) facilitates its cleavage by its target proteases and result in a non-functional truncated circulating form of C1-inh 
. Second, in rheumatoid arthritis, auto-antibodies to serpin E2 diminish the inhibitory activity of serpin on urokinase plasminogen activator serine protease 
. In analogy, anti-A2ML1 antibodies may either destabilize A2ML1 or prevent its interaction with its target proteases, inflammatory cytokines or its receptor LRP1 and thereby affect the activity of extracellular proteases or amplify tissue damage. In this context, it should be noted that our domain mapping results show that the NH2
-terminal portion of A2ML1 is almost systematically targeted by PNP autoantibody. Since this domain is implicated in the recognition of target proteases, PNP autoantibodies to A2ML1 may prevent the formation of protease-protease inhibitor complex. The identification of physiologically important targets of A2ML1 will be extremely useful to further understand the pathological involvement of anti-A2ML1 autoantibodies.
Dissecting the mechanisms underlying the association of PNP with distinct neoplasia is probably key for our understanding of the onset of autoimmunity in PNP. In the case of Castleman disease, tumor resection results in remission of PNP symptoms 
. Castleman, thymoma or follicular dendritic cell sarcoma cells have been shown to produce autoantibodies reactive with various PNP autoantigens 
. Noteworthy, reactivity with p170 is found in up to 76% of patients with PNP associated with Castleman disease 
. Future systematic studies with prospective cohorts of patients and detailed analysis of the immunological profile are needed to assess whether presence of anti-A2ML1 autoantibodies is associated with a particular PNP phenotype and organ involvement as well as a specific type of neoplasia. In this context, the retrospective nature of our study precluded a reliable analysis. Nevertheless, based on the tissue distribution profile of A2ML1 and its lack of expression in pulmonary epithelium 
, it is unlikely that autoantibodies to A2ML1 contribute to bronchiolitis obliterans and respiratory failure, a frequent cause of death in PNP.
Our study thus puts an end to a relentless search for p170 and identifies PNP as a first example of an autoimmune multiorgan syndrom in which autoantibodies to a protease inhibitor might contribute to tissue damage by aggravating and precipitating inflammation.