Antonio Núñez-Roldán, Servicio de Inmunología, Hospital Universitario Virgen del Rocío, Avda. Manuel Siurot s/n, 41013 Sevilla, Spain. E-mail: email@example.com
Four patients of 283 liver-transplant recipients (1·4%) developed de novo immune-mediated hepatitis approximately 2 years after transplantation. Antibodies showing an unusual liver/kidney cytoplasmic staining pattern were detected in the sera of all four patients and one of them was used to screen a human liver cDNA expression library with the aim of identifying the antigenic target of these newly developed antibodies. After cloning and sequencing the gene, it was identified as the gene encoding the glutathion-S-transferase T1 (GSTT1), a 29-kD molecular weight protein, expressed abundantly in liver and kidney. Sera from the other three patients also contained anti-GSTT1 antibodies, two of them demonstrated by immunoblot analysis against the recombinant antigen and the other, which was negative by immunoblot, gave a positive reaction when used directly to screen the same library, suggesting it to be directed to a conformational epitope. The GSTT1 enzyme is the product of a single polymorphic gene that is absent from 20% of the Caucasian population. When we analysed the GSTT1 genotype of the four patients described above, we found that this gene is absent from all of them. Three donor paraffin embedded DNA samples were available and were shown to be positive for GSSTT1 genotype. In accordance with these results, we suggest that this form of post-transplant de novo immune hepatitis, that has been reported as autoimmune hepatitis by others, could be the result of an antigraft reaction in individuals lacking the GSTT1 phenotype, in which the immune system recognizes the GSTT1 protein as a non-self antigen, being the graft dysfunction not the result of an autoimmune reaction, but the consequence of an alo-reactive immune response.
Autoimmune hepatitis (AIH) is a chronic liver disorder of unknown aetiology associated with circulating autoantibodies. The distinction between autoimmune hepatitis and other autoimmune liver diseases − primary biliary cirrhosis and primary sclerosing cholangitis − is generally based on an array of clinical, serological and immunological features, following a proposed scoring system . A number of autoantibodies, such as antinuclear (ANA), antismooth muscle (SMA), antiliver/kidney microsomal antigen (LKM), antiformiminotransferase cyclodeaminase (LC-1) and antisoluble liver antigen/liver-pancreas antigen (SLA/LP), have been associated with AIH and are in the basis of the classification into different AIH categories .
Recently, a different type of AIH has been described in patients, mainly children, who underwent liver transplantation, characterized by histological changes of chronic hepatitis, a high concentration of IgG, presence of non-organ-specific antibodies and response to immunosuppressive therapy [3,4].
In this study, we report four new cases of this form of de novo post-transplant immune hepatitis, clinically indistinguishable from those reported previously, and show that cytoplasmic antibodies present in our patients are directed against glutathione S-transferase T1 (GSTT1), an enzyme which belongs to a family of enzymes involved in the cellular detoxification of toxic reactive electrophiles, that is absent in 20% of the population . We raise the hypothesis that the production of these antibodies and the subsequent liver dysfunction could be the result of an antibody-mediated immune response in which GSTT1 acts on the transplanted liver as a non-self protein in liver recipients with GSTT1-negative phenotype.
Materials and methods
We studied four adult patients (three females and one male, aged from 26 to 62 years) who had suffered liver failure of diverse origin and had undergone orthotopic liver transplantation. The original disease causing hepatic failure was HCV-related cirrhosis in two cases, liver failure secondary to ebrotidine treatment in one case and liver failure of unknown origin in one case. They were all given immunosuppression therapy with cyclosporin. Antibodies associated with liver diseases studied by IIF as ANA, AMA, SMA and LKM were absent before transplantation. During the third year of post-transplant outcome, they were diagnosed as having de novo autoimmune hepatitis, following the scoring system of the International Autoimmune Hepatitis Group . They all had liver biopsies with the typical histological features of AIH, increased serum IgG, presence of new cytoplasmic antibodies and good response to treatment with steroids. Post-transplant sera from these patients showed an unusual liver/kidney immunofluorescence pattern.
Commercially available liver, kidney and stomach rat tissues (Biosystem, Barcelona, Spain) were used as substrate. Primary antibody was incubated on slides for 0·5 h in a humid atmosphere at room temperature and washed extensively with phosphate-buffered saline to remove any unbound antibody. Bound antibodies were detected with fluorescein isothiocyanate-conjugated rabbit antihuman IgG, M, A (Dako, Glostrup, Denmark). After washing, the slides were read on an epifluorescence microscope.
SDS-PAGE and immunoblotting
Immunoblot analysis techniques were carried out by using extracts derived from liver tissue. Briefly, a small portion of liver tissue was lysed in 250 µl of NET-2F (50 mm Tris-HCl pH 7·4/150 mm NaCl/5 mm EDTA/0·1% SDS/Nonidet P-40/0·5% sodium deoxycholate/0·02% NaN3/1 mm phenyl-methylsulphonyl fluoride/0·1% iodoacetamide) for 20 min at 4°C and then centrifuged to eliminate cell debris. Electrophoresis of cell lysates was performed in a 15% polyacrylamide gel, as described by Laemmli . Proteins were transferred to nitrocellulose filters, as described by Towbin et al. . After transfer, nitrocellulose strips were blocked with 3% (wt/vol) non-fat dry milk in TBST (10 mm Tris-HCl pH 7·5/150 mm NaCl/0·05% Tween 20) for 1 h at RT. Primary antibody was then incubated with the strips for 1 h at RT and washed extensively with TBST. Bound antibodies were detected by incubating the nitrocellulose strips with alkaline phosphatase-conjugated antihuman IgG immunoglobulins (Dako) diluted in 3% non-fat dry milk in TBST. After washing, the membrane was developed with NBT/BCIP (Boehringer Mannheim, Germany) in AP buffer (100 mm Tris-HCl pH 9·6/100 mm NaCl/5 mm MgCl2).
CDNA library screening
A human liver cDNA library, cloned in Uni-ZAP XR vector (Stratagene, La Jolla, CA, USA), was used. Clones were selected by immunological screening as described by Young and Davis . Those bacteriophages reacting with antiserum were subsequently purified to homogeneity. Before screening, serum was extensively adsorbed against bacterial proteins and wild-type λZAP phage to eliminate background. Bound antibodies were detected by immunoblot analysis as described above.
DNA sequencing was carried out on an ABI PRISM 310 Genetic Analyser (Applied-Biosystems, Foster City, CA, USA). Computer analysis of nucleic acid was performed with the University of Wisconsin Genetics Group Sequence Analysis Software Package .
Affinity purification of antibodies
Affinity purification of antibodies from Uni-ZAP XR clones was performed by using IPTG impregnated nitrocellulose filters to induce the production of the recombinant protein in standard LB plates. After incubating the plates overnight, the filters were blocked, probed with primary antibody and washed as described for immunoblot analysis . Bound antibody was eluted from the filters and concentrated in Centricon-30 microconcentrators (Amicon, Beverly, MA, USA).
A single bacterial colony expressing the recombinant protein was transferred into 5 ml of LB medium containing ampicillin (100 µg/ml), incubated overnight at 37°C with vigorous shaking, and induced to produce the protein with 10 mm IPTG for 1 h. Bacteria were harvested by centrifugation, resuspended in 1 ml of 50 mm glucose/10 mm EDTA/25 mm Tris-HCl, pH 8·0/lysozyme (4 mg/ml) and incubated at RT for 5 min. After centrifugation, the spheroplasts were resuspended in 500 µl of NET-2F lysis buffer prechilled to 4°C, incubated for 30 min on ice with occasional mixing, and centrifuged at 12 000 g for 10 min at 4°C to eliminate debris. Electrophoresis of bacterial lysates, transfer of proteins and immunological detection were performed as described above.
Characterization of the GSTT1 locus polymorphism
Genomic DNA was isolated from blood samples following standard procedures or from paraffin embedded tissues. PCR reactions using the primers TTCCTTACTGGTCCTCACATCTC and TCACCGGATCATGGCCAGCA were carried out as described . As a positive control, primers of exon 6 from the human coagulation factor IX gene, rendering a fragment of similar size (400 bp), was used. In the case of paraffin embedded samples, a second round of amplification with the primers 5′-GCTAGTTGCTGAAGTCCTGCTT-3′ and 5′-TTGGGTCGGCCTTCGAAGACTT-3′ was needed, rendering a PCR fragment of 200 bp when the GSTT1 gene was present.
Four liver transplant recipients from a total of 283 over a period of 10 years of observation developed a de novo post-transplant immune hepatitis during the third post-transplant year. Antibodies that produced a rare pattern of immunofluorescence staining in liver and kidney tissue, different to the known pattern of AMA or LKM-1 antibodies, were detected at that time in the sera of all four patients. In rat liver tissue the antibodies stained the cytoplasm of perivenous hepatocytes giving a zonation image, stronger adjacent to the vessels and decreasing toward the periphery (Fig. 1a). In rat kidney tissue, proximal tubules and different segments of distal tubules were stained (Fig. 1b). In HEp2 cells antimitochondrial pattern was absent. These patients were treated successfully with steroids as soon as they presented with graft dysfunction in association with autoimmune features showing a clinical and biochemical improvement. In one case the antibody titre dropped and another became negative.
In order to characterize the target antigen(s) recognized by these new antibodies we selected serum from one of the patients with the unusual IIF staining pattern and screened a human liver cDNA λ-ZAP expression library. Two clones were isolated, both containing an identical insert of 1·1 Kb. Comparative sequence analysis showed that they encode the GSTT1, a 29-kD molecular weight protein. To investigate the presence of anti-GSTT1 antibodies in sera from the other three liver-transplanted patients that showed the new staining pattern by IIF, bacterial cell extract containing IPTG-induced GSTT1 recombinant protein was used in immunoblot assay. The sera corresponding to two of them recognized the 29 kD band (Fig. 2). Serum from the patient that did not give a positive signal with the recombinant protein was used to screen the same expression library. We obtained one clone that contained an identical insert of 1·1 Kb coding for the enzyme GSTT1, a result that was consistent with the IIF staining data. Therefore, the antibodies developed by this patient are probably recognizing a conformational epitope of the same antigen. Affinity-purified antibodies reproduced the IIF staining pattern when used with rat liver, kidney and stomach substrates, and reacted on immunoblot with the recombinant protein giving the 29 kD band described for the prototype serum (Fig. 2). As control, 20 liver-transplanted patients without clinical manifestations of autoimmune hepatitis were also tested and showed to be negative for this IIF pattern and for the 29 kD GSTT1 band (data not shown).
Since this immune response could be the result of an antibody mediated reaction against the GSTT1 gene product present in the donor liver, and taking into account that 20% of the general population have the GSTT1-null genotype, we attempted to analyse the GSTT1 locus in the transplanted patients and in the corresponding donors in order to establish the donor-recipient GSTT1 matching. Seven samples, four corresponding to the recipients and three from the donors, were available and were studied by PCR. None of four recipient DNAs produced a fragment of the expected size of 480 bp (Fig. 3a), and were therefore asigned as GSTT1-negative individuals bearing the GSTT1-null genotype. In the case of the three donors, from whom we could obtain DNA from paraffin-embedded liver biopsy samples, a band of 200 bp corresponding to the GSTT1 gene (Fig. 3b) was obtained. Therefore, as expected, the four samples from the recipient individuals were negative and the three from the donors were GSTT1 positive.
In this study we have described the presence of anti-GSTT1 antibodies and their association with de novo immune hepatitis in four liver-transplanted patients. The four patients represent 1·4% of 283 liver-transplanted patients over a period of 10 years. None of them were diagnosed as having autoimmune hepatitis before transplantation and were referred to our laboratory because of signs of liver dysfunction secondary to de novo autoimmune hepatitis according to the scoring system of the International Autoimmune Hepatitis Group, such as typical histological features in liver biopsies and increase of serum IgG. All four patients responded to steroid treatment. The presence of new antibodies was detected first in routine immunofluorescence analysis of post-transplant sera from these patients. The fluorescence pattern obtained with all four sera on rat liver was uncommon, staining the cytoplasm of perivenous hepatocytes. Then, in order to identify the structural target of these antibodies, a human liver cDNA expression library was used for screening with one serum, and two identical clones were obtained. Direct DNA sequencing demonstrated that the isolated clones corresponded to the GSTT T1 gene that encodes the GSTT1 detoxification enzyme. Subsequent studies demonstrated that all four sera recognized the GSTT1 recombinant protein, one of them in only the non-denaturated form, whereas the other three were positive by immunoblot analysis. This indicates that the epitopes recognized by different sera on the GSTT1 molecule might be distinct.
At this point in the investigation, it could be stated that the GSTT1 enzyme is recognized as the major antigen by antibodies appearing in a group of liver-transplanted patients that developed a liver dysfunction with all the characteristic features of an autoimmune hepatitis. The GSTT1 is part of a group of isoenzymes involved in the cellular detoxification of both xenobiotic and endobiotic compounds . The GSTs can be divided into several main classes: alpha, mu, kappa, sigma, pi, zeta and theta. It has become an area of increasing research because a growing number of GST genes are being recognized as polymorphic and certain alleles, particularly those that confer impaired catalytic activity, may be associated with increased sensitivity to toxic compounds .
It has been demonstrated recently that gene deletions of GSTT1 occur, resulting in a complete absence of the enzyme in about 20% of the general population, the range of GSTT1-negative individuals among different ethnic groups being 11–58% . It is possible, therefore, that a negative GSTT1 individual could mount an immune response against this enzyme when his immune system encounters this non-self protein. This could be the case for liver transplant recipients, negative for GSTT1, who received a GSTT1-positive graft. In such cases, it is reasonable to speculate that the immune response against GSTT1 could correspond to a graft rejection mechanism leading to graft dysfunction, and not to a true autoimmune hepatitis.
The fact that the four patients had the GSTT1-null genotype and the three donors whose DNA were available carry the GSTT1 gene, make conceivable our hypothesis in favour of an alloreactive response on the basis of the post-transplant liver dysfunction observed in these de novo immune hepatitis cases. Prospective follow-up studies are under way in our liver-transplanted patients in order to establish the influence of donor-recipient GSTT1 matching in the graft outcome.
Similar clinical conditions to those observed in our patients have been described in seven children  and two adults  in whom de novo autoimmune hepatitis following a liver transplantation was reported. However, in all these cases, the exact nature of the immunological target was not investigated further and therefore the question of whether these hepatitis cases were the result of an autoimmune response or the consequence of an alloreactive mechanism remains unanswered.
On the other hand, autoantibodies directed against drug-metabolizing enzymes have been detected in AIH type 2 and in drug-induced hepatitis, where these enzymes participate in the metabolism of the inducing agent , and antibodies against GST have been detected previously in the sera of individuals with AIH showing antibodies to cytosolic soluble liver antigen (SLA), a finding that allowed the authors to conclude that GST subunit proteins represent the major autoantigen in anti-SLA-positive AIH . Nevertheless, these data have been replicated by others , who have elegantly excluded a reactivity of AIH-relevant anti-SLA antibodies with GST. It is possible that these enzymes are highly immunogenic, acting as autoantigens in some situations, but also as foreign antigens in some cases of donor-recipient mismatching, although the frequency and the intensity of graft dysfunction episodes among GSTT1-negative patients receiving a GSTT1-positive graft is not currently known. We are now investigating the presence of antibodies to GSTT1 recombinant protein in the sera of individuals with different subtypes of AIH in order to gain some insight into this question.
In summary, our findings have demonstrated an association between the occurrence of de novo immune-mediated hepatitis with the presence of antibodies to GSTT1 in recipients of liver transplants bearing a GSTT1-null genotype, suggesting that these newly developed antibodies could be the consequence of an antibody-mediated immune response against a foreign protein present in the graft and not the result of an autoimmune reaction. Studies will be conducted in order to ascertain this hypothesis as well as to investigate the role of GSTT1 protein as an immunological target leading to a liver dysfunction in post-transplant as well as in other immune-mediated hepatitis.
We are grateful to Dr Jorge Fernandez Alonso for his useful histology advice. This work has been supported by grants from Fondo de Investigaciones Sanitarias, Ministerio de Sanidad y Consumo, Spain (FIS 99/0255 and 00/0566) and from Plan Andaluz de Investigación (PAI, grupo CTS-0197), Junta de Andalucía.