Case History

A 72-year-old man (AG) presented with recurrent hemoptysis and abnormal chest radiograph. His work-up had previously been thought to be negative with the exception of an elevated titer for antiglomerular basement membrane antibodies. His medical history was remarkable for obstructive pulmonary disease, previously treated tuberculosis, and a 90 pack-year history of cigarette smoking. Physical examination on admission showed mild respiratory distress. Laboratory findings included a hemoglobin level of 107 g/L, a serum creatinine level of 70.72 µmol/L, and an erythrocyte sedimentation rate of 20 mm/h. The urinalysis result was normal. Test results for antinuclear antibodies and antineutrophil cytoplasmic antibodies were negative. The patient's chest radiograph showed right middle and right upper lobe infiltrates.

Fiberoptic bronchoscopy and thoracoscopy showed diffuse hemorrhage from the right upper and right middle lobes. Wedge biopsy specimens from both lobes were consistent with acute and chronic pulmonary hemorrhage; centrolobular emphysema was seen, and vasculitis was not. Cytologic studies, bacterial cultures, acid-fast bacilli stains, and immunofluorescent studies for tissue-bound antibodies were negative; no malignancy was noted.

Because a commercial reference laboratory reported an antiglomerular basement membrane antibody level of 44 (normal less than 10), limited Goodpasture syndrome was suspected. The patient was treated with steroids, cyclophosphamide, and several courses of plasmapheresis. He improved clinically until hospital day 26, when a right pleural effusion developed. Thoracentesis yielded 1 L of bloody fluid. Cytologic examination of the fluid showed an adenocarcinoma; the patient died soon thereafter.

Methods

Top Methods Results Discussion Author & Article Info References

Bovine type IV collagens, their hexamers, and recombinant noncollagenous domain 1 (NC1) domains for human 1- 6 type IV collagens were purified before use [3]. Immunoblotting, enzyme-linked immunosorbent assay (ELISA), radioimmunoassay, and immunofluorescence were done using standard protocols [2, 4, 5].

Results

Top Methods Results Discussion Author & Article Info References

Our patient (AG) was initially treated for the Goodpasture syndrome because the commercial serologic data and chest radiograph suggested a life-threatening autoimmune disease, despite the absence of nephritis [1]. Several cases of pulmonary hemorrhage due to antiglomerular basement membrane antibodies in the absence of renal involvement have been reported [6, 7].

To address the possibility that AG may have had such a case, we assayed plasmapheresis fluid from the first treatment for anti- 3(IV) NC1 antibodies using both native bovine and recombinant human antigens. The results of ELISA and immunoblotting showed minimal or no binding to the 3(IV) NC1 (Figure 1; top and bottom, respectively); instead, intense reactivity with both bovine 1(IV) NC1 and human recombinant 1(IV) NC1 was seen. We also noted weak binding to the human recombinant 3(IV) NC1 and 6(IV) NC1 (Figure 1, middle). Serial serum samples were further studied in retrospect using an antiglomerular basement membrane antibody radioimmunoassay [5] and were found to be nonreactive. The lung biopsy specimen from AG showed no endogenous IgG bound to alveolar basement membrane (Figure 2, left). Indirect immunofluorescence on renal tissue showed that AG's antibodies were bound to the mesangial matrix, tubular basement membrane, and glomerular basement membrane. This anti- 1(IV) NC1 antibody binding was most intense in the mesangial matrix and glomerular basement membrane (Figure 2, middle). The reference Goodpasture sera containing 3(IV) NC1 antibodies (LL) bound predominantly to the glomerular basement membrane in a linear manner [2] (Figure 2, right). The binding of AG's antibody to the mesangial matrix is characteristic of anti- 1(IV) antibodies.

View larger version (28K): [in this window] [in a new window]   Figure 1. Identification of the type IV collagen chains that bind to the patient's (AG) autoantibodies. Top. Enzyme-linked immunosorbent assay (ELISA) with bovine antigens. The results show that AG's antibody binds to the 1(IV) noncollagenous domain 1 (NC1) domain with minimal binding to the 3(IV) NC1 domain. The antibody of the patient with the Goodpasture syndrome (LL) shows specific binding to the 3(IV) NC1 domain and minimal binding to the 1(IV) NC1 domain. The control antibodies show no specific binding to type IV collagen. The dimers represent the 3(IV) NC1 domain in dimeric form. Middle. Results of ELISA with recombinant human type IV collagen NC1 domains. AG's antibody binds strongly to the recombinant 1(IV) NC1 domain and binds much more weakly to the 3(IV) NC1 and 6(IV) NC1 domains. The Goodpasture antibodies (LL) bind predominantly to the 3(IV) NC1 domain. The control antibodies do not specifically bind to any of the recombinants. Bottom. Immunoblot analysis of the antibodies from AG and LL using the recombinant human type IV collagen NC1 domains. AG's antibody binds predominantly to the 1(IV) NC1 domain, and LL's antibodies bind predominantly to the 3(IV) NC1 domain. The dilution of the antibodies used in all assays was 1:200. The 405-nm absorbance values shown in the graphs are each a mean of two separate readings.

View larger version (66K): [in this window] [in a new window]   Figure 2. Tissue immunohistochemistry. Left. Lung biopsy specimen. Middle. Binding of the patient (AG) antibodies to the glomerular basement membrane and mesangial matrix of the normal human kidney section. Right. Binding of the Goodpasture (LL) antibodies to the glomerular basement membrane of the normal human kidney section predominantly in a linear fashion. The control serum specimen did not show any binding (data not shown). The dilution of the antibodies used was 1:50. The control serum specimen did not bind to any of the recombinant or bovine noncollagenous domain 1 domains.

Discussion

Top Methods Results Discussion Author & Article Info References

Although the patient's antibodies were directed at antigens present in normal tissue, no inflammation was seen; therefore, we doubt that the hemoptysis was caused by immune reaction. Recent research provides some confirmation of this. Rabbits injected with bovine 1(IV) NC1 dimers produce circulating antibodies to 1(IV) NC1 domains but do not develop nephritis [4]. In contrast, when they are injected with 3(IV) NC1 dimers, rabbits develop a fulminant Goodpasture-like syndrome. These results suggest that the target glomerular basement membrane antigen may determine the likelihood of tissue damage.

Our patient received three courses of five plasmaphereses. Despite the elimination of anti- 1(IV) NC1 reactivity, the patient continued to have hemoptysis, perhaps because of an erosive endobronchial tumor that was not apparent on bronchoscopy. Several reports have shown basement membrane degradation in metastatic cancer [8, 9]. Increased amounts of type IV collagen have also been seen in the sera of patients with primary or metastatic bone and soft-tissue tumors [10]. These antibodies may reflect paraneoplastic immune responses against the 1(IV) collagen chain, as seen in our patient.

We believe that a definitive diagnosis of the Goodpasture syndrome requires the strong presence of anti- 3(IV) NC1 antibodies in patient serum. Although antibodies against 1(IV) NC1 are occasionally present in small amounts in a few patients with the Goodpasture syndrome, we have found that they are always associated with antibodies directed to 3(IV) NC1 domains [2]. From a clinical perspective, sensitive and specific assays using immunoabsorbed antigens [5], bovine 3(IV) NC1 chains, or recombinant human 3(IV) NC1 chains [3] are needed to accurately diagnose the Goodpasture syndrome.

Drs. Palevsky and Grippi: Pulmonary Division, Department of Medicine, Hospital of the University of Pennsylvania, 422 Curie Boulevard, Philadelphia, PA 19104.