Kidney International, Vol. 50 (1996), pp. 1753—1 766

NEPHROLOGY FORUM

Goodpasture's syndromePrincipal discussant: W. KLINE BOLTON

University of Virginia Health Sciences Center, Charlottesville, Virginia, USA

A 59-year-old white man with a history of recurrent pleural effusions,coal worker's pneumoconiosis, and a positive PPD presented with hemop-tysis beginning on the morning of admission. The patient had coughed upsmall amounts of dark blood totaling approximately 4 ounces; althoughthis was his first episode of hemoptysis, he had experienced epistaxisapproximately once weekly in the previous year. The patient was married,had smoked one pack of cigarettes/day for 30 years, but had quit four yearsago. He occasionally drank alcoholic beverages, and he was a retired coalminer who had worked directly in the mines. He denied fever, chills, nightsweats, weight loss, decreased appetite, nausea, vomiting, shortness ofbreath, or cough. The patient complained of a transient nonspecific rash inthe previous weeks that had resolved.

The medical history was notable for a recurrent left-sided pleuraleffusion 5 years previously with a negative workup for collagen-vasculardisease, malignancy, and infection. He had had nephrotic-range protein-uria four years prior to admission; renal biopsy then demonstrated focalsegmental glomeruloselerosis. Treatment with an angiotensin-converting-enzyme inhibitor reduced the proteinuria to below 400 mg/24 hours. Thefamily history was negative.

Physical examination revealed a well-appearing white man in noapparent distress. His temperature was 37°C; respirations, 20/mm; pulse,80 beats/mm; and blood pressure, 120/60 mm Hg. The head, eyes, ears,nose, and throat were normal, as was the neck. The lungs were clear toauscultation. Cardiac examination was unremarkable. His abdomen wassoft, nontender, and his extremities showed no evidence of clubbing,cyanosis, or edema. Neurologic examination was normal.

Laboratory evaluation revealed a white blood cell count of 9700/mm3;hematocrit, 41%; and platelets, 192,000/mm3. Blood urea nitrogen (BUN)was 40 mgJdl and serum creatinine concentration 2.3 mg/dl; two monthsago, the BUN was 25 mg/dl and the serum ereatinine, 1.4 mg/dl. Total

The Nephrology Forum is funded in part by grants from Amgen,Incorporated; Merck & Co., Incorporated; Dialysis Clinic, Incorporated;and R & D Laboratories.

© 1996 by the International Society of Nephrology

serum protein and albumin were normal. Blood gases obtained while thepatient was breathing room air were: pH, 7.37; PaCO2, 33 mm Hg; andPa02, 80 mm Hg. A chest radiograph revealed a new right upper lobeinfiltrate (Fig. 1). Urinalysis showed 3+ protein, large blood, dysmorphicred blood cells, granular casts, and red cell casts.

The patient was admitted to the hospital for further workup. Heunderwent bronchoscopy with broneho-alveolar lavage (BAL) as well asventilation/perfusion scan. Culture and cytologic examinations of the BALfluid were negative, and the sean was interpreted as a low probability forpulmonary embolus. The assay for anti-neutrophil eytoplasmic antibody(ANCA) was positive at 1:1280 with a perinuclear pattern of fluorescence.Antibodies to the glomerular basement membrane (GBM) were negativeby indirect fluorescence assay. Renal biopsy contained 13 glomeruli, 5 ofwhich were globally sclerotic and obsolescent. The remaining 8 hadvarying degrees of hypercellularity and segmental sclerosis; two demon-strated epithelial crescents (Fig. 2). The interstitium showed a moderateamount of tubular atrophy and fibrosis with a mild mononuclear cellinfiltrate. No vasculitis was seen. Immunofluorescent examination re-vealed IgG in a moderately intense linear pattern along the GBM and asimilar pattern of staining for kappa and lambda light chains (Fig. 3). C3was found only in the mesangium without GBM staining. Fibrin waspresent in glomerular crescents (Fig. 3). Electron microscopy revealed noimmune deposits. Anti-GBM assay by ELISA was positive against therecombinant rs3(IV) NC1 domain.

The patient was treated with prednisone, 70 mg/day, and 200 mg dailyof eyclophosphamide, which was tapered over six months. Over a period ofone month, the patient's serum creatinine concentration returned to 1.4mg/dl with a BUN of 36 mg/dl. At that time, the urinary sediment showedonly an occasional red cell with no red cell casts.

Discussion

DR. W. KUNE BOLTON (Chief Division of Nephrology, andProfessor of Medicine, University of Virginia Health Sciences Center,Charlottesville, Virginia): The present case was selected to illus-trate some of the typical features seen with Goodpasture'ssyndrome (GPS), as well as some of the atypical features that areof major clinical importance to the practicing nephrologist. Thispatient also illustrates some new ideas in diagnosis, prognosis, andtherapy. The term "GPS" is an eponym derived from a report in1919 by Ernest Goodpasture, who described the clinical syndromeof pulmonary hemorrhage associated with influenza infection andthe histologic finding of acute creseentic glomerulonephritis (GN)[1, 21. We do not, as a matter of fact, know that the patientsdescribed by Goodpasture had anti-GBM disease. Over the years,the terminology has been used in different ways by differentindividuals, some including all causes of pulmonary hemorrhagewith renal dysfunction as GPS [3]. Others limited the term GPS topatients with pulmonary hemorrhage associated with anti-GBMantibodies, as opposed to GN with anti-GBM antibodies butwithout pulmonary hemorrhage [4]. Yet others espouse theconcept of anti-type-TV collagen disease rather than GPS [5].Although most of the antibody activity is directed to the

1753

EditorsJORDAN J. COHENJOHN T. HARRINGTONNIc0LAOs E. MADIAS

Managing EditorCHERYL J. ZUSMAN

Tufts University School of Medicine

Case presentation

1754 Nephrology Fonim: Goodpasture r syndrome

Fig. 1. Chest roentgenogram showing right upperlobe infiltrate which appeared coincident withhemoptysis.

a3(IV)NCI domain of type-IV collagen, antibody to other chainsand other constituents of the GBM can occur [6—8]. Indeed,studies both in humans and animal models demonstrate that manydifferent anti-GBM antibody specificities are found clinically, andthat although all these antibodies fix to the GBM, the resultantclinical phenotypes can be quite different. Since the target anti-gen(s) (GBM) appear to be the same regardless of the clinicalpresentation, and anti-GBM-associated clinical findings can rangefrom normal renal function with asymptomatic GBM depositswithout pulmonary symptoms, to pulmonary hemorrhage with orwithout GN, to GN with or without pulmonary involvement, itseems impractical to subdivide thc clinical presentation based onlyon the presence or absence of pulmonary hemorrhage. I thereforeconsider any patient with anti-GBM antibodies deposited in tissueas having GP disease, and patients with GP disease and GN orhemoptysis or both as having GP syndrome. Thus the antibodiesand target epitopes are the common link among the variousclinical presentations, and it is the environmental and secondaryeffector systems that determine the clinical presentation. In thiscontext, patients who have had renal transplants and recurrentanti-GBM antibody on the basement membrane without pulmo-nary hemorrhage and without GN are still classified as part of thespectrum of GP disease, as are symptomatic and asymptomaticallografted patients with Alport's syndrome who have de-novoanti-GBM antibody deposits. It seems likely that, as we gain abetter understanding of the pathogenesis of this disease, we willbe able to define why the clinical spectrum is so variable and will

be able to develop more rational and specific forms of therapy forGPS.

Epidemiology and clinical manifestations

Goodpasture's syndrome is uncommon, having an incidence ofapproximately 0.1 cases per million population. In our experience,it accounts for approximately 1% to 2% of renal biopsy specimens[9]. The distribution by gender is approximately equal, with a veryslight preponderance of males to females. The age at presentationcan range from the first to the ninth decade, so the commonly heldconcept that this is a disease of young males is of little clinicalpracticality. There is, however, a bimodal distribution, with agreater number of individuals presenting around age 30 and asecond peak at around age 60 [10]. This latter group has a slightergreater tendency for female distribution and GN rather thanpulmonary hemorrhage, but these tendencies are of little clinicalimportance. The disease is more common in whites than inAfrican-Americans and may be more common in certain otherracial groups, such as the Maoris in New Zealand [111. Goodpas-ture's syndrome can present year round, but its incidence appearsto increase in the spring and early summer; some localizedoutbreaks have been associated with infection. In our experience,approximately one-half to three-quarters of patients have sometype of pulmonary presentation, such as cough, or upper respira-tory tract infection, plus fever, rash, myalgia, malaise, headaches,weight loss, or other vague signs and symptoms.

In patients with renal involvement, an active urinary sediment

Nephrolo Fomm: Goodpasture 's syndrome 1755

Fig. 2. Light microscopic photograph showingJical segmental glomerular sclerosis. A, firstbiopsy; B, a glomerulus with a crescent (arrow),second biopsy.

with dysmorphic red cells and red cell casts is extremely common.If urinary findings are not present, one should consider otherdiseases. Proteinuria usually is present, hut not in the ncphroticrange, although an occasional patient will have heavy proteinuria.Significant hypertension is usually not a feature of the disease, butmild increases in blood pressure consistent with age do occur.Ultrasonography reveals normal-sized kidneys with normal echo-texture and without anatomic abnormality. Renal function canrange from normal to rapidly decreasing over a few weeks or

months [9, 12—14]. We define rapidly progressive GM (RPGN),frequently seen in GPS, as a halving of GFR or doubling of serumcrcatinine within three months or less [91.

In patients who also have pulmonary involvement, the presen-tation is nonspecific; GPS is but one of many causes of pulmonaryhemorrhage (Table 1). These causes include vasculitis (includingmost importantly, microscopic polyarteritis nodosa, Wegener'sgranulomatosis, and Churg-Strauss vasculitis), certain collagen-vascular diseases, and a variety of different types of GN. However,

1756 Nephrology Forum: Goodpasture's syndrome

Table 1. Renal disease with pulmonary hemorrhage

GlomerulonephritisGoodpasture's syndromeImmune-complex glomeruloncphritisAcute silicoprotcinosisGlomerulonephritis with CHFD-penicillamine toxicity, trimellitic anhydride toxicity

VasculitisPolyarteritis nodosaMicroscopic polyarteritis nodosaWegener's granulomatosisHenoch-Schönlein purpuraMixed IgG-IgM cryoglobulinemiaChurg-Strauss allergic vasculitisLymphoid granulomatosisNecrotizing sarcoid granulomatosisHypersensitivityBehcet syndromeHughes-Stovcn syndromeGiant cell arteritisHypocomplementemic urticariaTakayasu's arteritis

Collagen vascular diseaseSystemic lupus erythematosusProgressive systemic sclerosisRheumatoid arthritis

Cardiovascular diseaseAdvanced uremia with superimposed CHFMitral stenosisPulmonary embolism with infarction and underlying renal vein

thrombosisInfection

Legionnaire's diseaseNecrotizing pneumonitisBronchitisTuberculosisFungusLung abscessBronchiectasis

Malignancy with associated glomerulopathyBlood dyscrasiasThrombotic thrombocytopenic purpuraHemolytie-uremic syndromeIdiopathic hemosiderosis

Fig. 3. Immunofluorescence microscopy illustrating (A) linear deposits ofIgG along GBM and (B) fibrin within a crescent.

the differential diagnosis must include infection, malignancy, andany cause of pulmonary hemorrhage with decreased renal func-tion, as can occur with various antibiotics and unrelated acuterenal failure. Pulmonary hemorrhage can precede the develop-ment of GN by several years, or it can develop after renal diseaseis manifest. The intensity of pulmonary hemorrhage provides fewclues to the presence of GPS and can range from blood-streakedsputum to massive fatal pulmonary hemorrhage. Quite clearly,temporal relationships exist between hemoptysis and environmen-tal exposure, including various hydrocarbons, especially cigarettesmoking, and episodes of infection [15, 16]. Patients often com-plain of shortness of breath and cough, but usually not chest painor pleurisy, as occurs with pulmonary embolism.

Examination of the sputum can demonstrate hemosiderin-containing macrophages, a nonspecific finding indicating pulmo-nary hemorrhage; chest radiographs reveal variable results. Alve-olar infiltrates spreading out from the hilum are common, but

nodular densities and isolated segmental infiltrates do occur, asseen in today's patient. Pulmonary function can be decreased, andepisodes of pulmonary hemorrhage are associated with increaseduptake of carbon monoxide, which can be used as a test ofpulmonary hemorrhage and to document intercurrent episodes ofpulmonary hemorrhage [17]. Increased uptake occurs because ofbinding of carbon monoxide to extravasated blood within the lungand formation of carboxyhemoglobin in alveoli.

Laboratory findingsThe sine qua non for the diagnosis of GPS is demonstration of

bound anti-GBM antibodies in the glomeruli of the kidneys. Thisfinding usually is reflected by the presence of circulating anti-GBM antibodies, most commonly of the IgG class, althoughoccasionally IgA or 1gM [18]. The detection method for thiscommon laboratory test is an extremely important consideration.Although circulating anti-GBM antibodies are present in morethan 90% of patients with anti-GBM disease, indirect fluorescentassays using kidneys from various species as substrate, includingmonkey kidney, can be falsely negative in 10% of cases or more.In the present case, this method was used because it is less

Nephrology Forum: Goodpasture's syndrome 1757

Anti-GBM diseaseGoodpasture's syndromeIdiopathic RPGNMembranous nephropathy with secondary anti-GBM diseaseANCA + vaseulitis

Non-anti-GBM diseaseDiabetic nephropathyRenal allograft cx-vivo perfusionFocal sclerosis (rare)PolyarteritisPost-infectious glomerulonephritis (rare)Minimal-change disease (rare)Ankylosing_spondylitis_(rare) -

expensive than are radioimmunoassays (RIA) and ELISAs, whichare more specific and have less than 5% false-negative rates. Onthe other hand, there are few false-positive results with theseassays. Indirect fluorescence tests, as well as RIA and ELISA, arefalsely positive in probably 1% or less of positive tests. False-positive tests can arise from antibodies directed against otherconstituents of type-IV collagen, from inexperience in interpreta-tion of indirect fluorescence, and from lack of proper controls.When pulmonaiy hemorrhage, GN, and a negative RIA or ELISAassay are found, the chances of GPS are remote.

In my opinion, the diagnosis must be made on the basis of renalhistologic examination unless biopsy is contraindicated. We per-form a renal biopsy on an emergency basis in any patient whomight have GPS, regardless of the presence or absence ofanti-GBM antibodies in the circulation. The rationale for thisaggressive diagnostic approach involves the need for assessing thedegree of crescent involvement and the amount of sclerosis andfibrosis at presentation, and to ascertain whether another diseaseprocess is present [19]. Time is critical, because prognosis andresponse to therapy are closely tied to duration of disease andlevel of renal function prior to diagnosis and treatment. Thepresence of linear GBM deposits does not necessarily indicateGPS (Table 2), and determination of whether antibodies in thecirculation or on the GBM in vivo are specific for GBM is anecessary confirmation for the diagnosis of GPS. If circulatinganti-GBM antibody is present, this is sufficient; if not, micro-elution of renal tissue sections is necessary. Specific antibodieseluted from glomeruli by indirect fluorescence fix to normalhuman GBM, whereas antibodies in non-anti-GBM disease asso-ciated with linear deposits do not fix to normal human GBM.

In summary, the differential diagnosis of GPS encompasses thespectrum of any type of GN associated with any degree or causeof pulmonary hemorrhage. The broader definition of GPS utilizedin this discussion makes a renal biopsy imperative in virtually allpatients with any renal manifestations, including an active urinarysediment or abnormal urinary protein, to clarify the underlyingpathology. The pathologic diagnosis has major importance in thechoice of appropriate therapy.

IVatural course and prognosis

The clinical course of untreated, and even treated, GPS isbleak; this disease generally is associated with an extremely poorprognosis [10, 20, 21]. When we reviewed the natural course andprognosis of GPS in the mid-1980s, 89% of patients reported inthe literature who were treated with various types of immunosup-

VaseulitisWegener's granulomatosisChurg-Strauss allergic vasculitis

Microscopic PANAnti-GBM diseaseCrohn's diseaseUlcerative colitisFelty's syndromeSjogren's syndromePolymyositisSystemic lupus erythematosusHenoch-Schdnlein purpuraRheumatoid arthritisHemodialysisChronic liver diseaseHIV infectionOther acute/chronic infection

pressive therapy, but not pulse methylprednisolone or plasmaexchange, progressed to death or dialysis; only 10% experiencedsome type of improvement [9]. The observation that anti-GBMantibodies appear to cause the disease led to the use of plasmaexchange as the appropriate therapeutic modality [22, 231. Withthis type of treatment, approximately 50% of patients improved.However, patients with oligoanuria and those on dialysis uni-formly fail to respond to plasma exchange, pulse methyipred-nisolone, or other therapy [9, 19, 23, 24]. Of note are rare reportsof patients with GPS who apparently recovered spontaneouslyeven after acute renal failure and the requirement for dialysis, andthose in whom withdrawal of the offending agent, such as hairspray, was associated with recovery [25—27]. In general, however,the consensus has been that the prognosis is extremely poor, butperhaps improving because the defined spectrum of the disease isincreasing (thus, embracing less ill patients), and because bothspecific and general medical therapy have improved. We shouldremember, however, that the grim figures on prognosis refer toGPS patients with the clinical presentation of RPGN.

Other factors also are important in the overall prognosis ofGPS. A clear genetic influence exists. Patients who carry anHLA-DR W2, especially with HLA-B7, appear to have a moremalignant course of ON [23, 281. A high degree of crescentformation (that is, more than 50%), the presence of tubulointer-stitial disease with fibrosis, scarring, and glomerulosclerosis allbode ill for the patient [19, 29]. In patients who have a serumereatinine of 6 mg/dl, oliguria, or anuria, and in those requiringdialysis, response to therapy is extremely poor and recovery rare.

ANCA and anti-GBM antibodies

Today's patient had circulating ANCA in a perinuelear pattern(pANCA), which is associated with antibodies to myeloperoxi-dase. Although numerous diseases are associated with ANCA(Table 3), we usually associate its presence with vasculitic syn-dromes [301. We used to think that coexistent GPS and vasculitiswere rare, and isolated reports of this association were consideredinteresting oddities [31, 32]. In 1989, however, O'Donoghue et alreported three patients with ANCA-positive serum and anti-GBMdisease [33]. Since that time a number of additional patients havebeen described, yielding a surprisingly high prevalence of ANCAwith GPS [34—36]. Approximately 30% of patients with circulatinganti-GBM antibodies will have circulating ANCA at some time

Table 2. Linear GBM deposits Table 3. Positive ANCA-associated conditions

1758 Nephro1oy Forum: Goodpasture s syndrome

during the course of the disease. The ANCA can be positive priorto or after the development of anti-GBM disease. In patients withanti-GBM disease, the pattern for thc ANCA is perinuclear in atleast three-fourths of the patients and cytoplasmic in the otherone-quarter. The clinical course of disease in patients withanti-GBM disease and a positive ANCA appears to be somewhatdifferent than that in those with pure anti-GBM disease. First,only about one-half the "double-positive" patients require dialy-sis. In contradistinction to the scenario with GPS, a substantialnumber of ANCA-positive patients have recovered sufficient renalfunction to stop dialysis, although the exact percentage is not clear[34, 351; it might be as high as 30% to 50% judging from the smallseries that are available. Many ANCA-positive patients also haveother vague systemic findings, including rash, fever, and myalgia,which are more suggestive of a vasculitie process than pureanti-GBM disease, although such findings can occur with GPS.Finally, there appears to be an inverse correlation betweenanti-GBM titers, ANCA titers, and disease course; that is, patientswho have higher ANCA titers and lower anti-GBM titers tend tohave a course that is more suggestive of vasculitis and respond totherapy and even recover from dialysis; those with the highestanti-GBM titers pursue a more devastating course without returnof renal function. Another clinical feature of "double-positive"patients is the tendency for frequent relapses, seldom seen inpatients with pure anti-GBM disease. However, informationregarding this clinical entity is scant at present. Further studies areneeded to clarify the interrelationship between anti-GBM andANCA, and to determine what therapy is most appropriate andwhat factors determine the response to therapy. Because ANCAis a relatively new laboratory test, it is not known how manypatients reported in the literature over the last several decadeswith anti-GBM disease had concomitant ANCA positivity andhow that affected the results of previous studies.

PathogenesisUnderstanding the pathogenesis of a disease is of paramount

importance for developing specific therapy. The classic studies byLerner, Glassock, and Dixon nearly 30 years ago were among thefirst demonstrations of a GN in which the pathogenesis was known[37]. In those studies, nephritic kidneys obtained from patientswith GPS were acid-eluted and the eluted antibody purified. Theantibody then was injected intravenously into unilaterally ne-phrectomized monkeys. Microscopic examination of the GBMrevealed intense staining for human IgG associated with fulmi-nant GN, pulmonary hemorrhage, and progression to renal failurewith proteinuria. The recent observations that GBM, lung, cho-roid plexus, and basement membranes within the eye containadditional collagen chains other than the a1 and a2 typically foundin type IV collagen, and documentation that most of the circulat-ing antibodies in patients with GPS are directed to the a3non-cotlagenous (NCI) domain of type IV collagen, has addedgreatly to our understanding of the pathogenesis of GPS [7, 8, 38,391. Although controversial, much evidence suggests that thecirculating antibodies in patients with GPS are targeted mainly tothe a3NC1 domain, with much of that reactivity to the last 36amino acids of the chain; these have been termed "Goodpasture'sepitope" [40]. Further evidence of the specificity of these antibod-ies for this domain has been derived from the studies by Neilsonand coworkers, who expressed recombinant proteins for the firstfive NC1 domains of type IV collagen and showed that the only

one with significant binding with Goodpasture's sera was a3 [41].Other laboratories have reported that Goodpasture's sera, whileshowing broad reactivity to various components of GBM, includ-ing other NC1 domains, nonetheless have most reactivity to thea3NCI domain [6—81.

As with other autoimmune diseases, patients have a strongsusceptibility based on genetic background and, as I noted, certainHLA antigens are clearly associated with susceptibility to GPSand possibly to a worse phenotype. This association is similar tothe demonstration in various animal species of strain and speciesdifferences in susceptibility to disease [42]. Much evidence sup-ports the concept that insult to the basement membrane orcross-reactivity with exogenous epitopes can initiate the process ofautoimmunity resulting in the phenotype of GPS (Table 4).Different initiating events, such as infection, exposure to toxins,renal injury (including underlying glomerulopathy, as in today'spatient), ischemia, membranous nephropathy, and various typesof neoplasia could expose cryptic antigens by damaging the GBM.In addition, urine contains antigens that conceivably could resultin an autoimmune response if they gained access to the circula-tion. Finally, the model of most interest is that of Alport's familialnephritis. These patients' GBMs are nonreactive with Goodpas-ture's sera as opposed to normal GBM, and they lack the epitopeto which Goodpasture's sera react [43, 44]. Many different abnor-malities of a3(IV) collagen, which have been associated withAlport's syndrome, might result in incorrect assembly of thea3NCI domain in the GBM, thus accounting for the lack of theGoodpasture's epitope. Transplantation in patients who haveend-stage renal disease from Alport's syndrome has been associ-ated with the development of circulating anti-GBM antibodiesand their deposition in the graft, resulting in loss of the graft fromanti-GBM disease in a small number of cases [45—48]. Patientswho have this latter syndrome appear to be those most likely to bereported in the literature, as the actual occurrence with conse-quent graft loss may be small. In a combined series of 75 patientswith Alport's disease who received renal transplants, 8 developedanti-GBM antibodies, and 2 patients lost their grafts [45—47].Retransplantation in these 2 patients produced good graft sur-vival, however. Thus, it seems obvious that some patients doindeed develop de-novo anti-GBM disease in the normal allograftin the setting of Alport's syndrome, but this is relatively uncom-mon. Whether these patients were of a nonsusceptible MHC typeor whether other factors were involved is not clear. Patients withGPS have an exuberant cellular infiltrate with macrophages andT-cells and an intense secondary effector response, which appearsresponsible for the final phenotypic expression [discussed in detailin Refs. 23, 49—53].

Table 4. Events associated with the induction of anti-GBM diseasein humans

Infections

Toxin exposureRenal injuryNeoplasiaNeoantigensEndogenous antigens

InfluenzaBacterial endocarditisHydrocarbons, tobacco, drugsIschemia, membranous ncphropathyLymphomas, adenomasAlport's syndromeGBMLungUrineChoroid plexus

Nephrology Fomm: Goodpasture's syndrome 1759

Experimental studies show that GBM contains antigens thatcan induce experimental autoimmune GN in several species,including sheep, goats, rats, rabbits, and mice [54—58]. Thenephritogenic component that causes the proliferative GN ap-pears to be present in dimers containing a3, a4, and a5, and not inthose dimers containing a1 and a2 [57]. Recent studies fromKalluri and colleagues show that dimers of NC1 type IV collagencontaining mostly a3 but also some components of a1, a2 and a4can induce GN in rabbits and mice [57, 58]. The work from ourown laboratory and that of Sado shows that GBM collageninduces a model that appears to parallel GPS in every respect,complete with proteinuria, proliferative necrotizing GN withcrescents, circulating and bound anti-GBM antibodies, and pul-monary hemorrhage [55, 56, 59, 601.

It remains to be seen whether the nephritogenic antigeninvolved in GPS is the same as that in experimental models ofGPS. Although a3NCI-enriched dimers produce disease, and typeIV collagen lacking a3NC1 does not, immunization of rats withGoodpasture's epitope is likewise unproductive of disease [57—59,61]. Whether this latter outcome results from lack of conforma-tional changes that are pathogenic, whether other modificationsof the molecule are required for nephritogenicity, or whetherdifferent effector mechanisms are involved with synthetic peptidesrather than with native protein are among possibilities that remainto be evaluated. It is clear that patients with GPS have manydifferent circulating antibodies to a variety of components ofbasement membrane, and that these antibodies can react withdifferent peptides along the a3NCI chain. If the true nephrito-genic peptides can be identified, then direct intervention in GPSis feasible. However, patients and animals alike can demonstratecirculating and bound anti-GBM antibody without disease, andpatients who had GPS originally can receive transplants withoutdeveloping GPS or even GP disease [62, 63]. Why these differ-ences occur and what accounts for them are intriguing questions,the answers to which will further refine our ability to modelspecific therapies for this devastating clinical syndrome.

Treatment

Pulmonary hemorrhage associated with GPS usually respondsrapidly to pulse methylprednisolone or to plasma exchange [9, 23,64]. In some patients, neither modality is of benefit; we have hada few patients with intractable pulmonary hemorrhage who havedied because plasma exchange, pulse methyiprednisolone, andcytotoxic therapy all were ineffective.

As I said earlier, conventional treatment—defined as prednisoneand other types of immunosuppression in standard doses—hasproduced little beneficial effect. The presence of circulatinganti-GBM antibodies has served as an obvious rationale forremoval of these antibodies by plasmapheresis or immunoglobu-lin-adsorbent columns. Although the literature contains limitedinformation about patients who have anti-GBM disease withnormal kidneys or minimal renal disease, it seem obvious that thisgroup of patients will have a good prognosis with either notherapy, withdrawal of the offending agent, or standard immuno-suppressive therapy [12, 13, 261. In the setting of RPGN or activecrescentic involvement on renal biopsy, however, aggressive ther-apy is required in those patients with the potential for benefit. Incontrast, patients who have the poor prognostic findings I men-tioned are unlikely to respond and should be spared the risks ofaggressive therapy.

I would like to address the issue of plasmapheresis. The largestseries to date is that from Lockwood et al, who have treated morethan 100 patients with anti-GBM disease, most of whom havebeen treated with plasma exchange and immunosuppression [10,23, 24, 651. Six of seven patients with normal renal function whowere treated all improved, and one patient with normal renalfunction improved without treatment. Three-fourths of 20 pa-tients with serum creatinine levels of � 7 mgldl improved withplasma exchange, but only one of 12 with creatinine levels > 7mg/dl showed improvement. No improvement was observed inany of the 58 patients who were on dialysis.

Plasma exchange effectively removes circulating antibodiesdirected to GBM [22, 64]. Standard therapy consists of 14exchanges of 4 liters each over two weeks, combined with pred-nisone and immunosuppressive drugs such as cyclophosphamideor azathioprine. This approach diminishes circulating antibodylevels without rebound after cessation of plasma exchange as longas immunosuppressive therapy is continued. However, there is nogood correlation between circulating antibodies at the time ofpresentation and the histologic severity of GN, nor does a goodcorrelation exist between a decrease in antibody titers and re-sponse to therapy in the patients who do not require dialysis [22,66]. Only one prospective randomized controlled study has beenaccomplished. In this study, Johnson and coworkers compared theeffects of therapy with pulse methyiprednisolone immunosuppres-sion alone to immunosuppression combined with plasma ex-change (Fig. 4) [64]. Although antibody levels were reduced to amuch greater extent, as one would expect with plasma exchange,there was no significant difference in the clinical outcome betweenthe two groups. Analysis of clinicopathologic findings at studyentry suggested that the percentage of crescents on initial renalbiopsy and the entry serum creatinine concentration correlatedbetter with outcome than did the type of therapy utilized. Theauthors noted that the patients with a lower percentage ofcrescent formation and well-preserved renal function did betterwith either type of treatment than did patients with severe

16

1412a)

C 10

a)0E

a)Ca

CW20

O =1o =I&PE• = DIALYSISD = NON-DIALYSIS* = NO BIOPSY TISSUE

0

0 0Do

S

.

.

Do

U

S0

0*

10 20 30 40 50 60 70 80 90 100% Crescents

Fig. 4. Influence of entty creatinine and percentage of crescents in patientstreated with either immunosuppression (Ij) or immunosuppression plusplasma exchange (I& PE). Solid symbols represent patients proceeding todeath or dialysis. (*) indicates no biopsy or no glomeruli on biopsy. Usedwith permission [64].

1760

>

50%Crescents

Nephrology Forum: Goodpasture's syndrome

Serum creatinine 5 mg!dl >

Series

Briggs [66]Johnson [64]

Simpson [68]

Bolton [19]

Conventional!N pulse therapy

1917

19 017

I m mu nosu ppressionand plasma exchange

*Fajled initial pulse—came off dialysis @Solid — treatment failureOpen — Response to therapyPx — Plasma exchange

4 mc., serum creatinine 4.0 mg/dl

Fig. 5. Influence of ently serum creatinine and percentage of crescents in 4 series with individually analysable patients treated with plasmapheresis or othertherapy. Used with permission [67].

crescentic involvement and impaired GFR. Although this is theonly prospective randomized trial that has been done, it had only17 patients. Our own experience and that of others has suggestedthat patients with mild anti-GBM disease may well respond totherapies other than plasma exchange [52, 67]. Figure 5 graphi-cally displays the results of four series in which adequate infor-mation allowed assessment in individual patients of the impact ofserum creatinine and percentage of crescent involvement [19, 64,66, 68]. Patients were treated with plasma exchange, pulse meth-ylprednisolone, or conventional therapy (defined as no therapy orany type of immunosuppression other than plasma exchange or

pulse methylprednisolone) [91. Patients treated with plasma ex-change received variable numbers of exchanges but also receivedprednisone and cyclophosphamide. The individual patients fromthe four series are differentiated by symbols. As Figure 5 illus-trates, patients with anti-GBM disease generally segregate atentry into the medical system into those who have a serumcreatinine of � 5 mg/dl and < 50% crescents or those who havea greater involvement of glomeruli, with > 50% crescents and aserum creatinine > 5 mgldl. In the former group, 78% of patientstreated with conventional therapy or pulse methylprednisolonewere stable or improved (11/14) compared with 11 patients

Non-Px Px

0/36 0/5

1*• U®®A A A** • I®A A A* • •®A A*• ®A A* • A A

A A

Non-Px Px

11/14 11/11

A 0•0 *

A 0 *

r pulse

ll pulse

Nephrology Fomm: Goodpasture's syndrome 1761

treated with plasma exchange, all of whom were stable or im-proved. Conversely, in the patients with serum creatinine levels of> 5 mg/dl or > 50% crescents, none of the 36 patients treatedwith any type of immunosuppression alone showed any improve-ment, nor did any of the 5 patients treated with plasma exchange.These findings are consistent with the large series from Lockwoodin which these latter types of patients showed no response [231.

While the recommendation always has been that patients withanti-GBM disease should be treated first and foremost withplasma exchange, several issues need to be considered. Patientswith lesser degrees of crescentic involvement and with lessadvanced renal failure appear to do very well without plasmaexchange. In patients who have GPS and who receive transplants[62, 63, 69], anti-GBM deposits recurred in approximately one-half of them but only one-third of the patients with anti-GBMantibodies lost their transplants because of recurrent disease. Allthose patients were, of course, receiving immunosuppressivetherapy for their allograft and were temporally distant from theoriginal initiating process. In addition, in patients with Alport'ssyndrome who are known to have an increased risk of developinganti-GBM disease, the de-novo development of anti-GBM diseaseassociated with linear deposits along the GBM on neo-antigens inthe allograft is uncommonly associated with loss of the transplant[45—47]. Finally, plasma exchange removes circulating antibodies,which do not correlate well with the clinical course, but it alsoinduces a state of immunosuppression, causes "decomplementa-tion," and decreases fibrinogenic products. Substances that mighthave immunoregulatory activities, such as anti-idiotypic antibod-ies, are removed as well. We have in fact seen several patients withminimal renal dysfunction who were treated with plasma ex-change and who appeared to have an acceleration of their diseasethat resulted in an eventual need for dialysis and transplantation.These observations suggest that the question of appropriatetherapy for anti-GBM disease is still not entirely settled.

Nonetheless, it is necessary to have a therapeutic stratagem forapproaching the patient with GPS in the clinic. Table 5describesour therapeutic protocols, and the algorithm in Figure 6 presentsour approach to the diagnosis and therapy of GPS. The corner-stones in decision-making are the serum ANCA studies, the renalbiopsy, the anti-GBM assay, and the patient's renal function. InANCA-negative patients with serum creatinine levels � 6 mg/dland � 50% crescents, we generally use plasma exchange andimmunosuppression. For patients with > 50% crescents and > 6mgldl serum creatinine, oligoanuria, or those on dialysis, we treatpulmonary hemorrhage but utilize only a short course of plasmaexchange and immunosuppression or none at all. If the ANCA ispositive, we treat for vasculitis with or without plasma exchange.For non-dialysis-dependent patients with high-titer anti-GBMantibodies, we use plasma exchange and immunosuppression.Although plasma exchange appears of no benefit for dialysis-dependent patients with GPS, there is evidence that it is useful ifvasculitis is the cause of dialysis dependence [701. Therefore weuse plasma exchange and immunosuppression for dialysis-depen-dent, "double-positive" patients. For low-titer anti-GBM/ANCA-positive patients, we lean more toward using therapy directedagainst vasculitis, that is, pulse methylprednisolone or oral pred-nisone and cyclophosphamide, although others use intravenouscyclophosphamide [71, 72]. We might be swayed to add plasma-pheresis if an RPGN course or a highly active histologic lesionwere present. When plasma exchange is used, one could make an

Table 5. Therapeutic protocols

A. Pulse methylprednisoloneVolume repletion ascertainedNo diuretics 3 hours before pulse or 24 hours after

30 mg/kg methylprednisolone given IV over 20 minutes everyother day for 3 doses

Single dose not to exceed 3 gMonitor blood pressure during and 4 hours after infusion

B. Oral prednisone, given every other dayDose" (mg/kg) Duration/months of therapy'

2.0 0.51.75 1.01.50 3.01.25 6.01.00 6.00.75 6.00.50 6.00.25 6.00.125 12.00.0625 12.0

C. Cyclophosphamide2.0 mg/kg daily for 3 months1.5 mg/kg daily for 3 months1.0 mg/kg daily for 3 months0.5 mg/kg daily for 3 monthsDecrease dose 50% for Ccr 10 mg/mm"Hold dose if WBC 3500/mm3, or platelets 100,000/mm3

D. Plasma exchange4 liter exchanges 5/7 days or every other day X 14 exchanges

minimumReplacement with 5% normal human serum albuminFresh-frozen plasma if bleeding problemsAlways be certain cyclophosphamide is used when patients receive

plasma exchange (immunoglobulin rebound suppression) -a Patients � 60 years of age should receive 75% of oral prednisone dose

on alternate days; for very obese patients, use ideal body weight.h If full remission of proteinuria, normal function for one month,

decrease to next dose step.Ccr = creatinine clearance.

argument for plasma replacement rather than albumin in patientswith ANCA, because plasma can have beneficial effects even invasculitic patients refractory to standard vasculitis therapy [73]. Inall cases, treatment is individualized by clinical judgment in anattempt to provide the most effective treatment while not puttingthe patient at undue risk from therapy. These considerationstempered the treatment decision for the patient presented here.

Many types of therapy not discussed here have been tried inGPS without success. Data from experimental animals suggestthat a variety of new agents may be beneficial in GPS. Theseinclude IL-i receptor antagonists, TNFa receptor protein, anti-bodies to various cell-surface adhesion molecules, deoxyspergua-un, and immunoregulating proteins such as CTLA4-Ig [74—77].Removal of anti-GBM antibodies by immunosorbent columns hasyielded no benefit [78]. Perhaps one of the most exciting potentialtherapies is the use of immunosorbent chromatography using theGoodpasture epitope to remove circulating antibodies. Prelimi-nary evidence from Boutaud et al demonstrates that anti-GBMantibodies can be absorbed from the serum of patients with GPSusing recombinant a3(IV) NC1 domain [79].

Finally, as in the case of the anti-GBM assay, additionalimportant considerations these days include the cost of therapy.At our institution, a 14-day course of plasma exchange is approx-imately $34,000 plus the cost of hospitalization. The cost of pulse

History & physicalexamination

Hematologic,chemical and

serologic studies

Renal biopsy Other

Crescentic Otherglomerulonephritis

Fig. 6. Algorithmic approach to diagnosis and therapy of patients with pulmonaty renal syndrome.

methyiprednisolone is $156. However, since the cost of dialysis isabout $20,000/year, it may be cost effective to use plasmapheresisin all patients with a serum creatinine level < 5—6 mg/dl.Appropriately conducted prospective trials obviously are required[80]. I believe that the diagnostic and therapeutic approachprovided here represents a reasonable and cost-conscious regi-men until parameters can be delineated that predict whichpatients will develop the most aggressive GPS requiring the mostaggressive treatment, and what that treatment should be.

Questions and answersDR. JOHN T. HARRINGTON (Dean ad interim, Tufts University

School of Medicine, Boston, Massachusetts): Why does anti-GBMantibody cause disease in patients with "real" Goodpasture'sdisease, whereas the same anti-GBM antibodies in the patient

with Alport's syndrome who has received a transplant do not eatup the glomerular basement membrane?

DR. BOLTON: That is a terrific question. I do not know theanswer. The antibody specificity by immunoblot appears to be thesame. Whether the difference is related to transplant immunosup-pression or different epitope specificities in the de-novo glomer-ulonephritis is not known.

DR. AJAY SINGH (Division of Nephroloçy, New England MedicalCenter, Boston): My first question is about the nature of theantibody response in patients with Goodpasture's syndrome. Asyou know, McPhaul et al, in their classic study [141, and others I'msure, have documented that IgG is the predominant isotype inpatients with Goodpasture's. However, I recall that some studiesalso have suggested a unique subclass distribution among patientswith Goodpasture's syndrome as compared to sera from normal

1762 Nephrology Forum: Goodpasture's syndrome

/Anti-GBMantibodies

Serumcreatinine

�6 mg/dl

�50% crescents

>6 mg/dl

> 50% crescents

Plasma exchangeprednisonecyclophosphamide

-Supportive care-Pulse steroids

orplasma exchange forpulmonary hemorrhage-Transplantation

Prednisonecyclophosphamide

Nephrology Forum: Goodpasture's syndrome 1763

subjects. This would suggest a regulatory defect in the immuneresponse. Could you comment on this? Further, do you know ofstudies that have explored whether certain idiotypes are morefrequently expressed in the anti-GBM antibody response?

DR. BOLTON: Segelmark examined antigen restriction and IgGsubclass of anti-GBM antibodies using purified monomers of a1,a2, a3, and a4 chains. There was a preference for IgG 1 isotypesin Goodpasture's syndrome but a subset, mostly females, also hadhigh-titer IgG 4 antibodies [18]. Lewis et al also have noted arelative absence of IgG 3 isotype and anti-GBM disease and largeamounts of IgG 4, a pattern of distribution different from normalserum components [81]. Rees et al have reviewed the clinical andexperimental data in great detail regarding idiotype responses andimmune regulation [23].

DR. SINGH: As you know, David Isenberg and Yehuda Shoen-feld and others have reported the presence of certain idio-types—in particular 16/6—shared among different patients withlupus [82, 83]. Indeed, with respect to the 16/6 idiotype, Shoenfeldhas reported that when antibodies expressing this idiotype areinjected into normal laboratory mice, they develop disease [84].Has this sort of work been performed in Goodpasture's syn-drome?

DR. BOLTON: Several animal models have been used to studythese questions and are reviewed in the excellent chapter by Reesthat I mentioned [23].

DR. SINGH: My other question is directed towards the geneticsof Goodpasture's syndrome. A close association between DR2and Goodpasture's syndrome has been reported in the literature[15, 85]. However, as you know, the Goodpasture's antigen(COL4A5) has been cloned, and maps to the q36 to q37 site onchromosome 2 [86], whereas the HLA antigen, DR2, maps tochromosome 6. These observations would suggest a nonstructuralexplanation for the relationship of DR2 with Goodpasture's.What do you think might be the association?

DR. BOLTON: T agree. Presumably immune responsiveness to theGoodpasture's antigen is linked to the DR2, that is, peptideprocessing and presentation to T-cell receptors.

DR. NIcoLAos E. MADIAS (Chief Divison of Nephrology, New

England Medical Center): My question concerns the Goodpas-ture's epitope. Can you expand on what aspects of the nature andsuprastructure of Goodpasture's epitope might account for itscryptic properties?

DR. BOLTON: The epitope being bound by Goodpasture's serumis sequestered inside the hexamer, which is made up of theglobular regions of the various noncollagenous domains. One canincrease the binding between the noncollagenous domain andGoodpasture's antibody up to 50-fold by revealing the epitope;this can be done by treating the hexamer with acid which unfoldsit. Binding is also diminished by manipulations that affect tertiarystructure, such as reduction and alkylation, digestion with trypsinand cyanogen bromide, and disruption of lysines by acetylation.Thus, the antigen is cryptic or hidden and discontinuous.

DR. MIAs: Is the stoichiometry and assembly of the subunitsof the epitope identical at the various sites where Goodpasture'sepitope exists?

DR. BOLTON: I don't know. I suspect that it is not. Basementmembrane in the lung and kidney have different dimer/monomerratios for a3, with kidney being mostly dimers and some mono-mer, but lung being mostly monomer and some dimer [87]. This

suggests that there are differences even though the amino acidsare the same.

DR. HARRINGTON: I have a followup question regarding thesame amino acid epitope but at a clinical level. Smoking, hairspray, and other pulmonary toxins frequently are invoked asstimuli to the production of Goodpasture's syndrome. If that istrue, why does it happen so extraordinarily infrequently?

DR. BOLTON: We are beginning to see a wide spectrum ofanti-GBM disease ranging from patients who are entirely asymp-tomatic to those who have RPGN. It is not known why thisspectrum exists. It is probably related to the intensity of exposureto the antigen, or to the epitopes involved, or to geneticallyinfluenced immune responsiveness.

DR. ANDREW J. KING (Division of Nephrology, New EnglandMedical Center): Does Goodpasture's syndrome originate in thekidney? If so, what is the sequence of events? Specifically, is thererelease of collagen or some other process that initiates thedisease?

DR. BOLTON: We think that it must originate from kidneys, butI do not know that anybody has really shown that. Ray Steblayshowed years ago that you can cause Goodpasture's syndrome inanimals by immunizing with lung basement membrane [88]. Thelist of things that cause release of basement membrane constitu-ents includes a wide range, from cortical necrosis to membranousnephropathy to other underlying disease. There are all sorts ofinsults in the kidney that have been associated with it. Not so inthe lungs. Maybe that points to the kidney as being the primarysite. But I don't know that for sure.

DR. KING: Frequently these patients present Friday evening inthe emergency room. We rely on ANCA and indirect immuno-fluorescence as our guidelines as to whether the patient has thistype of disease. Why do you think the renal biopsy is essential inthe management of patients who are not dialysis dependent? Doyou ever rely on these serologic markers?

DR. BOLTON: In the situation that you describe, in which thepatient arrives at off hours, we go ahead and treat if we cannot getthe pathology. We use clinical judgment and give pulse methyl-prednisolone and wait until we get the definitive answer. We usethe biopsy because of the inability of the clinical course to revealthe degree of crescent formation and cellularity, and thus revers-ibility of crescents, and the degree of fibrosis and scarring. If thereis severe sclerosis in the glomeruli or interstitium, I would tend tobe much less aggressive in terms of treatment and have a muchlower threshold for stopping.

DR. HARRINGTON: How do you treat the patient who is anti-GBM positive? The data you showed revealed a strong correlationbetween a serum creatinine of less than 5 mg/dl and a percentageof crescents of less than 50%, and vice versa. In view of those dataand the low cost of measuring the serum creatinine, versus a renalbiopsy, why don't we believe the serum creatinine and act on it?

DR. B0LT0N: In today's cost-conscious environment, we mightbe headed in this direction, I feel more secure in knowing howmuch sclerosis is present to tell the patient what I think theprognosis is going to be. I cannot tell the patient as much aboutprognosis by looking at the serum creatinine.

DR. HARRINGTON: Do you mean that if the serum creatinine isI or 6 mg/dl, you can't predict the outcome?

DR. BOLTON: As Figure 5 shows, histology is very accurate inpredicting prognosis. The numbers are too small for serumcreatinine <5 mg/dI and < 50% crescents to know whether serum

1764 Nephrology Forum: Goodpasturev syndrome

creatinine alone is predictive. If the serum creatinine is > 6 mg/dl,however, there is probably much less chance of the patientresponding.

DR. MADIAS: I would like to return to Dr. Harrington's questionabout Alport's syndrome. As you mentioned, most cases ofX-linked Alport's syndrome have defects of the a5(IV) chain,which can affect the cx3(IV) chain [89]. I would have expected thatmost anti-GBM antibodies post grafting would be anti-5. Do thefew patients who develop aggressive anti-GBM disease havepredominantly anti-cs3 antibodies as opposed to anti-cs3, and is itknown whether their genetic defect is different? Is there arelationship between aggressive anti-GBM disease and pre-exist-ing rejection episodes? Under what circumstances should onemonitor levels of anti-GBM antibodies?

DR. BOLTON: In the one case examined, the de-novo antibodieswere to a3 and not a5 or other chains [44]. 1 am not aware ofanalysis of genetic defects in those with and without de-novodisease, but that is an interesting question. I am also not aware ofan association with previous rejection, but the numbers are verysmall. I do not believe prospective monitoring of anti-GBMantibodies is warranted.

DR. KING: Many of the patients who present to us are dialysisdependent. Do we treat for possible pulmonary hemorrhage? Doyou believe in screening patients for pulmonary hemorrhage byusing carbon monoxide diffusion capacity and other modalities?

DR. BOLTON: We do not routinely do that. if the patient doesnot have any kind of clinical pulmonary symptoms, we do notscreen or treat.

DR. SINGH: Work from Kalluri and colleagues presented inabstract form at the recent American Society of Nephrologymeeting suggests that there is variable phenotypic expression ofanti-GBM disease in murine models [58]. In this regard, humansalso demonstrate phenotypic variability, from anti-GBM nephritiswithout pulmonary abnormalities to the full-blown Goodpasture'ssyndrome [3]. Could you comment on the potential role of geneticfactors in modulating the clinical expression of anti-GBM dis-ease?

DR. BouroN: There are differences in the clinical phenotypethat occur based on the underlying immune response. Thisresponse is very complex in terms of the type of cytokine response,type and degree of antibody and cellular immune response, andother effector systems. Presumably genetic influence in all theseeffector systems dictates the final clinical expression.

DR. HARRINGTON: I don't understand the patients who aredouble positive, that is, ANCA and anti-GBM positive. Is theirdisease closer to Goodpasture's or Wegener's? Do they havedouble disease or is this one disease? Does it make a difference inclinical outcome whether they are P-ANCA versus C-ANCApositive?

DR. BOLTON: I really do not know. A few years ago ANCApositivity with anti-GBM disease was considered rare. Flowevcr,now there are a number of papers indicating 30% association.Most arc P-ANCA (75%) and titers of ANCA and anti-GBMantibodies tend to be inversely related. I am not convinced thatANCA causes disease, but it could. It is appealing to think thatthese patients have vasculitis that damages the kidney, whichreleases basement membrane and then secondary anti-GBMantibodies are formed. I do not know whether these secondaryantibodies are pathogenic. Thus, some might be closer to Wege-ner's, while others are more like Goodpasture's syndrome.

DR. MADIAS: Do any of these patients have evidence of systemicvasculitis?

DR. B0L1oN: They have many vague manifestations suggestiveof vasculitis. Histologic lesions of vasculitis are uncommon, how-ever.

DR. MAD!AS: About a year ago in a Nephrology Forum, Dr.Robert Atkins discussed the role of cytokines, specifically IL-i, inthe pathogenesis of the entity and presented data on the beneficialeffects of the IL-l receptor antagonist (IL-Ira) in a rat model ofaccelerated anti-GBM disease [90]. He concluded his Forum bysaying that he and his colleagues were setting up an internationalmulticenter trial to study the use of IL-ira as an adjunct toconventional immunosuppression in patients with RPGN includ-ing anti-GBM disease. Do you know whether this trial is inprogress?

DR. BOLTON: I talked to Bob at the meetings this past month.He did not say anything about it going forward. I do not knowwhat the status is.

DR. MADtAS: You listed a number of futuristic therapies. Doyou think gene therapy will be included among them in an effortto alter cytokine production?

DR. BOLTON: This is certainly possible, certainly futuristic, andraises many questions. How do you target the gene to its site ofaction? How do you control it? What are the long-term effects?How does it perturb other regulatory processes? What are theunexpected consequences? The leaps between theories, genes,animal models, and the bedside are significant. Gene therapy is anappealing approach to the control of inflammation, but from apractical point we need much more information.

Reprint requests to: Dr. W. K. Bolton, Chief Division of Nephrology,University of Virginia Health Sciences Center, Box 133, Charlottesville,Virginia, 22908, USA

Acknowledgment

The author thanks C.V. Cook for secretarial assistance. This work wassupported in part by U.S. Public Health Service grant DK44107 from theNational Institute of Diabetes and Digestive and Kidney Diseases.

References

1. GOODPASTURE EW: The significance of certain pulmonary lesions inrelation to the etiology of influenza. Am J Med Sci 158:863—870, 1919

2. STANTON MC, TANGC JD: Goodpasture's syndrome (pulmonaryhaemorrhage associated with glomerulonephritis). Aust Ann Med7:132—144, 1958

3. 1-IOLDSWORTI-I SR, Boven N, ThoMsoN NM, ATKINS RC: The clinicalspectrum of acute glomcrulonephritis and lung haemorrhage (Good-pasture's syndrome). Q J Med 55:75—86, 1985

4. COHEN AH, GLASSOCK RJ: Anti-GBM glomerulonephritis includingGoodpasturc's syndrome, in Renal Pathology, edited by TISHER CC,BRENNLR BW, Philadelphia, Lippincott, 1989, p 494

5. HUDSON BG: Anti-type TV collagen diseases. The immunologic basisof renal disease: Paradigms for the 21st century: TheAmerican Societyof Nephrologyc Advances in Basic Science Conference, San Diego, CA,1995

6. J0IINAN5sON C, HELLMARK T, WIESLANDIR J: One major epitope inGoodpasture's syndrome. .1 Am Soc Nephrol 4:608, 1993

7. KEFALIDES NA, OHNO N, WILSON CB: Heterogeneity of antibodies inGoodpasture syndrome reacting with type IV collagen. Kidney mt43:85—93, 1993

8. MAISUKIJRA II, BUTKOWSKI RJ, FIsh AJ: The Goodpasture antigen:Common epitopes in the globular domains of collagen IV. Nephron64:532—539, 1993

Nephrolo Forum: Goodpasture's syndrome 1765

9. BOLTON WK: The role of high dose steroids in nephritic syndromes:The case for aggressive use, in Controversies in Nephrology andHypertension, edited by NARIN5 RG, New York, Churchill-Living-stone, 1984, p 421

10. SAVAGE COS, PUSEY CD, BOWMAN C, REE5 AJ, LOCKWOOD CM:Antiglomerular basement membrane antibody mediated in the BritishIsles 1980—4. Br Med J 292:301—304, 1986

11. TEAGUE CA, DOAK PB, SIMPSON Ii, RAINER SF, HERD5ON PB:Goodpasture's syndrome: An analysis of 29 cases. Kidney mt 13:492—504, 1978

12. BAILEY RR, SIMPSON IJ, LYNN KL, NEALE TJ, DOAK PB, MCGIVENAR: Goodpasture's syndrome with normal renal function. Clin Neph-rot 15:211—215, 1981

13. ZIMMERMAN SW, VARANASI UR, HOFF B: Goodpasture's syndromewith normal renal function. Am J Med 66:163—171, 1979

14. MCPitAuI. JJ, DIXON FJ: Characterization of human anti-glomerularbasement membrane antibodies eluted from glomerulonephritie kid-neys. J C/in Invest 49:308—317, 1990

15. DONAGHY M, REES AJ: Cigarette smoking and lung haemorrhage inglomerulonephritis caused by autoantibodies to glomerular basementmembrane. Lancet 2:1390—1392, 1983

16. WHITWORTH JA, MOREL-MAROGER L, MIGNON F, RICHET G: Thesignificance of extracapillary proliferation. Nephron 16:1—19, 1976

17. EWAN PW, JONES HA, RHODES CG, HUGHES JMB: Detection ofintrapulmonary hemorrhage with carbon monoxide uptake: Applica-tion in Goodpasture's syndrome. N Engl J Med 295:1391—1396, 1976

18. SEGELMARK M, BUTKOWSKI R, WIESI.ANDER J: Antigen restriction andIgG subclasses among anti-GBM autoantibodies. Nephrol Dial Trans-plant 5:991—996, 1990

19. BO!:'oN WK, STURGILL BC: Methylprednisolone therapy for acuteereseentic rapidly progressive glomerulonephritis. Am J Nephrol9:368—375, 1989

20. WALKER JF, WATSON AJ, GARRETt P, CRONIN CJ, BEOGII B, DoNo-HOE JF, CARMODY M, O'DWYER WF: Goodpasture's syndrome—7years experience of two Dublin renal units. IrMedJ75:328—333, 1982

21. STRUTZ F, NEILSON EG: The role of lymphocytes in the progression ofinterstitial disease. Kidney mt 45:S105—S1 10, 1994

22. PETERS DK, REES AJ, LocKwooD CM, PUSEY CD: Treatment andprognosis in antibasement membrane antibody-mediated nephritis.Transplant Proc 14:513—521, 1982

23. REES AJ, LocKwoot CM: Antiglomerular basement membrane an-tibody-mediated nephritis, in Diseases of the Kidney, edited bySCHRIER RW, GOTrSCHALK CW, Boston, Little, Brown, 1988, p 2091

24. LOCKWOOD CM, PEARSON TA, REES AJ, EVANS DJ, PETERS DK:Immunosuppression and plasma exchange in the treatment of Good-pasture's syndrome. Lancet 1:711—715, 1976

25. COHEN LH, WILSON CB, FREEMAN RM: Goodpasture syndrome:Recovery after severe renal insufficiency. Arch Intern Med 136:835—837, 1976

26. BERNIS F, HAMELS J, QUOIDBACH A, MAHIEU P, BOUVY P: Remissionof Goodpasture's syndrome after withdrawal of an unusual toxin. C/inNephrol 23:312—317, 1985

27. STRAUCII BS, CIARNEY A, DOCTOROUFF 5, KASHGARIAN M: Good-pasture's syndrome with recovery after renal failure. JAIVL4 229:444,1974

28. REES AJ, PETERS OK, COMI'SION DAS, BATCHELOR JR Strongassociation between 1-ILA DRW2 and antibody mediated Goodpas-ture's syndrome. Lance! 1:966—968, 1978

29. GLASSOCK RJ: Clinieo-pathologie spectrum and therapeutic strategiesin "rapidly progressive" glomerulonephritis. Proc Dialysis TransplantForum 5:109—116, 1975

30. WILKOWSKI MJ, BOLTON WK: Autoantihodies and other autoimmuneserologic abnormalities in dialysis patients. Scmin Dial 8:226—231,1995

31. Wt MJ, RAJARAM R, SHELP WD, BEIRNE GJ, BURKHOLDER PM:Vasculitis in Goodpasture's syndrome. Arch Paihol Lab Med 104:300—302, 1980

32. WAHI.S TL, BoNsIB SM, SCHUSTER VL: Coexistent Wegener's granu-lomatosis and anti-glomerular basement membrane disease. HumanPat/lot 18:202—205, 1987

33. O'DONOGHUE DJ, SI IORT CD, BluNrI II,IY C, LAWI.ER W, BALLARDIEFW: Sequential development of systemic vasculitis and anti-neutro-

phil eytoplasmic antibodies complicating anti-glomerular basementmembrane disease. C/in Nephrol 32:25 1—255, 1989

34. JAYNE DRW, MARSHALL PD, JONES SJ, LocKwooD CM: Autoanti-bodies to GBM and neutrophil cytoplasm in rapidly progressiveglomerulonephritis. Kidney In! 37:965—970, 1990

35. BOSCH X, MIRAPEIX E, FONT J, BORRELLAS X, RODRIGUEZ R,LoPEz-SoTo A, INGELMO M, REVERT L: Prognostic implication ofanti-neutrophil cytoplasmic autoantibodies with myeloperoxidasespecificity in anti-glomerular basement membrane disease. C/in Neph-rol 36:107—113, 1991

36. SHORT AK, ESNAULT VL, LOCKWOOD CM: ANCA and anti-GBMantibodies in RPGN. Adv Exp Med Biol 336:441—444, 1993

37. LERNER RA, GLASSOCK RJ, DIxoN FJ: The role of anti-glomerularbasement membrane antibody in the pathogenesis of human glomer-ulonephritis. J Exp Med 126:989—1004, 1967

38. HUDSON BG, WEISLANDER J, WISDOM BJ, NOELKEN ME: Goodpas-ture syndrome: Molecular architecture and function of basementmembrane antigen. Lab Invest 61:256—269, 1989

39. BUTKOWSKI RI, LANGEVEID JPM, WIESLANDER J, HAMILTON J, HUD-SON BG: Localization of the Goodpasture epitope to a novel chain ofbasement membrane collagen. J Biol Chem 262:7874—7877, 1987

40. KALLURI R, GUNWAR 5, REEDERS ST, MORRISON KC, MARIYAMA M,EBNER KE, NOELKEN ME, HUDSON BG: Goodpasture syndrome.Localization of the epitope for the autoantibodies to the earboxyl-terminal region of the a3(IV) chain of basement membrane collagen.J Biol Chem 266:24018—24024, 1991

41. NEILSON EG, K.ALLURI R, SUN MJ, GUNWAR S, DANOFF T, MARIYAMAM, MYERS JC, REEDERS ST, HUDSON BG: Specificity of Goodpastureautoantibodies for the recombinant noneollagenous domains of hu-man type IV collagen. J Biol Chem 268:8402—8405, 1993

42. SADO Y, NAITO 1, AKITA M, OKIGAKI T: Strain specific responses ofinbred rats on the severity of experimental autoimmune glomerulo-nephritis. J C/in Lab Immunol 19:193—199, 1986

43. MCCOY RC, JOHNSON HK, STONE WJ, WILSON CB: Absence ofnephritogenic GBM antigen(s) in some patients with hereditarynephritis. Kidney Int 21:642—652, 1982

44. HUDSON BG, KALLURI R, GUNWAR 5, WEBER M, BALLESTER F.HUDSON JK, NOELKEN ME, SARRAS M, RICHARDSON WR, SAUS J.ABRAHAMSON DR, GLICK AD, HARALSON MA, HELDERMAN JH,STONE WJ, JACOBSON HR: The pathogenesis of Alport syndromeinvolves type IV collagen molecules containing a3(IV) chain: Evi-dence from anti-GBM nephritis after renal transplantation. Kidney mt42:179—187, 1992

45. MILlINER DS, PIERIDES AM, HOLLEY KE: Renal transplantation inAlport's syndrome. Mayo C/in Proc 57:35—43, 1982

46. GOBEL J, OLBRICHT J, OFFNER G, HELMCHEN U, REPP H, KOCH KM.

FREI U: Kidney transplantation in Alport's syndrome: Long termoutcome and allograft anti-GBM nephritis. Kidney In! 38:299—304.1992

47. QUERIN 5, NOEL LH, GRUNFELD JP, DROZ D, MATHIEU F, BERGER J,KREIS H: Linear glomerular IgG fixation in renal allografts: incidenceand significance in Alport's sydrome. C/in Nephrol 25:134—140, 1986

48. SHAH B, FIRST MR, MENDOZA NC, CLYNE DH, ALEXANDER 1W,WIESS MA: Alport's Syndrome: Risk of glomerulonephritis inducedby anti-glomerular-basement-membrane antibody after renal trans-plantation. Nephron 50:34—38, 1988

49. Borroa WK, INNEs DJ, STURGILL BC, KAISER DL: T-cells andmaerophages in rapidly progressive glomerulonephritis: Clinicopath-ologic correlations. Kidney Int 32:869—876, 1987

50. SALANT DJ: Nephrology Forum: Immunopathogenesis of creseentieglomerulonephritis and lung purpura. Kidney In! 32:408—425, 1987

51. B0I:IoN WK: Mechanisms of glomerular injury: Injury mediated bysensitized lymphocytes. Semin Nephrol 11:285—293, 1991

52. COUSER WG: Rapidly progressive glomerulonephritis: Classification,pathogenetic mechanisms, and therapy. Am J Kidney Dis 11:449—464,1988

53. MAIN 1W, NIKOI.IC-PAIERSON Dl, ATKINS RC: T cells and macro-phages and their role in renal injury. Semin Nephrol 12:395—407, 1992

54. STEBLAY RW: Preliminary evidence for the production of nephritis byan autoimmune mechanism, in Inflammation and Diseases of Connec-tive Tissue, edited by MILLS LC, MOYER JH, Saunders, 1961, p 272

1766 Nephrology Forum: Goodpasture's syndrome

55. BOLTON WK, MAY WJ, STURGILL BC: Proliferative glomerulonephri-tis in rats: A model for autoimmune glomerulonephritis in humans.Kidney mt 44:294—306, 1993

56. SAIo Y, OKIGAKI T, TAKAMIYA H, SENO S: Experimental autoim-mune glomerulonephritis with pulmonary hemorrhage in rats. Thedose-effect relationship of the nephritogenic antigen from bovineglomerular basement membrane. J Clin Lab Immunol 15:199—204,1984

57. K..iiuu R, GAITONE VH II, NOELKEN ME, HUDSON BG: The a3chain of type IV collagen induces autoimmune Goodpasture syn-drome. Proc Nail Acad Sci 91:6201—6205, 1994

58. KALLURI R, DANOFF TM, NEILSON EG: Murine anti-a3 (IV) collagendisease: A model of human Goodpasture syndrome and anti-GBMncphritis. JAm Soc Nephrol 6:833, 1995

59. BOLTON WK, MAY WJ, STtJIWII.I, BC, LUO A, Fox PL: Study of EHStype IV collagen lacking Goodpasture's epitope in glomerulonephri-tis. Kidney In! 47:404—410, 1995

60. SADO Y, KAGAWA M, NAITO I, OKIGAKI T: Properties of bovinenephritogenic antigen that induces anti-GBM nephritis in rats and itssimilarity to the Goodpasture antigen. Virchows .4rch [B] 60:345—351,1991

61. BOLTON WK, LU0 AM, Fox P, MAY WJ, Fox i: Goodpasture'sepitope in the development of experimental autoimmune glomerulo-nephritis in rats. Kidney liii, in press

62. COUSER WG, WALLACE A, MONACO AP, LEWIS EJ: Successful renaltransplantation in patients with circulating antibody to glomerularbasement membrane: Report of two cases. Clin Nephrol 1:381—388,1973

63. WILSON CB, DixoN FJ: Antiglomerular basement membrane antibodyinduced glomerulonephritis. Kidney In! 3:74—89, 1973

64. JOhNSON JP, MOORE J JR, AUSTIN HA Ill, BALOW JE, ANTONOVYCHTT, WILSON CB: Therapy of anti-glomerular basement membraneantibody disease: Analysis of prognostic significance of clinical, patho-logic and treatment factors. Medicine 64:219—227, 1985

65. HIND CRK, LOCKWOOD CM, PETERS DK, PARASKEVAKOU H, EVANSDi, RuES AJ: Prognosis after immunosuppression of patients withcrescentic ncphritis requiring dialysis. Lancet 1:263—264, 1983

66. BRIGGS WA, JOHNSON JP, TEICHMAN 5, YEAGER HC, WILSON CB:Antiglomerular basement membrane antibody-mediated glomerulo-nephritis and Goodpasture's syndrome. Medicine 58:348—361, 1979

67. BOLTON WK: Treatment of erescentic glomerulonephritis. Nephroloffv1:257—268, 1995

68. SIMPSON IJ, DOAK PB, WILLIAMS LC, BLACKLOCK HA, HILL RS,TEAGUE CA, HERDSON PB, WILSON CB: Plasma exchange in Good-pasture's syndrome. Am J Nephrol 2:301—311, 1982

69. BEROREM H, JERVELL J, BRODWALL E, FLATMARK A, MELLBYE 0:Goodpasture's syndrome: A report of seven patients including long-term follow-up of three who received a kidney transplant. Am J Med68:54—58, 1980

70. PUSEY CD, REES AJ, EVANS Di, PETERS DK, LOCKWOOD CM: Plasmaexchange in focal necrotizing glomerulonephritis without anti-GBMantibodies. Kidney mt 40:757—763, 1991

71. FALK RJ, HOGAN S, CAREY TS, iENNETrE JC: Clinical course ofanti-neutrophil cytoplasmic autoantibody-associated glomerulone-phritis and systemic vasculitis. Ann Intern Med 114:430—431, 1991

72. WILKOWSKI MJ, BOLTON WK: Systemic vasculitis: The biological basis,in The Kidneys, edited by LEROY EC, New York, Marcel Dekker,1992, p 381

73. JAYNE DR, DAVIES MJ, Fox Ci, BLACK CM, LOCKWOOD CM:Treatment of systemic vasculitis with pooled intravenous immuno-globulin. Lancet 337:1137—1139, 1991

74. NIKOLIC-PATERSON Di, LAN HY, HILL PA, VANNICE JL, ATKINS RC:

Suppression of experimental glomerulonephritis by the interleukin-1receptor antagonist: Inhibition of intercellular adhesion molecule-iexpression. JAm Soc Nephrol 4:1695—1700, 1994

75. NISHIKAWA K, GUO YJ, MIYASAKA M, TAMATANI T, COLLINS AB, SYMS, MCCLUSKEY RT, ANDRES G: Antibodies to intercellular adhesionmolecule 1/lymphocyte function-associated antigen 1 prevent crescentformation in rat autoimmune glomerulonephritis. J Exp Med 177:667—677, 1993

76. LAN HY, NIKOLIC-PATERSON Di, ZHENG 5, VANNICE JL, ATKINS RC:

Suppression of experimental crescentic glomerulonephritis by theinterleukin-1 receptor antagonist. Kidney mt 43:479—485, 1993

77. LAN HY, ZARAMA M: Suppression of experimental crescentic glomer-ulonephritis by deoxyspergualin. JAm Soc Nephrol 3:1765—1774, 1993

78. ESNAULT VLM, TESTA A, JAYNE DRW, S0UI,ILL0U JP, GUENEL J:Influence of immunoadsorption on the removal of immunoglobulin Gautoantibodies in crescentic glomerulonephritis. Nephron 65:180—184,1993

79. BOUTAUD AA, KALLURI R, KAHSAI TZ, NOELKEN ME, HUDSON BG:Goodpasturc Syndrome: Selective removal of anti-a3(TV) collagenantibodies. A potential therapeutic alternative to plasmapheresis. ExpNephrol, in press

80. BOLTON WK: The nature of clinical trials in nephrology. Kidney in!42:1061—1069, 1992

81. LEWIS Ei, BUSCH GJ, SCHUR PH: Gamma G globulin subgroupcomposition of the glomerular deposits in human renal diseases. J ClinInvest 49:1103—1113, 1970

82. BAKIMER R, KRAUSE I, ABU-SHAKARA M, TEPLIZKI-AMITAL H, ISEN-BERG DA, ALARCON-SEGOVIA D, VILLAREAI. GM, SHOENFELD Y: Thefrequency of a common anti-DNA idiotype (16/6) in different popu-lations of patients with systemic lupus erythematosus. J Rheumatol18:1035—1037, 1991

83. BUSKILA D, SIIOENFELD Y: Anti-DNA idiotypes: Their pathogenicrole in autoimmunity. Concepts Immunopathol 8:85—113, 1991

84. SHOENFELD Y, MozuS E: Pathogenic anti-idiotype (16/6 Id) in sys-temic lupus erythematosus. Rheumatol mt 11:91—93, 1991

85. REES AJ, PETERS DK, AMOS N, WELSH KI, BATCHELOR R: Theinfluence in HLA linked genes on the severity of anti-GBM antibodymediated nephritis. Kidney mt 26:444—447, 1983

86. TURNER N, MASON PJ, BROWN R, Fox M, POVEY S, REES A, PUSEYCD: Molecular cloning of the human Goodpasture antigen demon-strates it to be the A3 chain of Type IV collagen. J Clin Invest89:592—601, 1992

87. SHEN G, BUTKOWSKI R, CHENG T, WIESLANDER i, KATZ A, CASS J,FISH Ai: Comparison of non-collagenous type IV collagen subunits inhuman glomerular basement membrane, alveolar basement mem-brane, and placenta. Connect Tissue Res 24:289—301, 1990

88. STEBLAY RW, RUDOFSKY UH: Experimental autoimmune antiglo-merular basement membrane antibody-induced glomerulonephritis. I.The effects of injecting sheep with human homologous or autologouslung basement membranes and complete Freund's adjuvant. ClinImmunol Immunopathol 27:65—80, 1983

89. HUDSON BG, KALLURI R, GUNWAR 5, WEBER M, BALLESTER F,HUDSON iK, NOELKEN ME, SARRAS M, RICHARDSON WR, SAUS J,ABRAHAMSON DR, GLICK AD, HARALSON MA, HELDERMAN JH,STONE WJ, iACOBSON HR: The pathogenesis of Alport syndromeinvolves type IV collagen molecules containing a3 (IV) chain: Evi-dence from anti-GBM nephritis after renal transplantation. Kidney In!42:179—187, 1992

90. ATKINS R: Nephrology Forum: Interleukin-1 in crescentic glomerulo-nephritis. Kidney mt 48:576—586, 1995