Small vessel vasculitis caused by hepatitis El virus; immune complexes Small vessel vasculitis and HBsAG Richard G. Gower, M.D., William F. Sausker, M.D., Peter F. Kohler, M.D., George E. Thorne, M.D., and Rawle M. McIntosh, M.D. Denver, Cola.
In a comprehensive study of 80 patients with vasculitis, 4 had concurrent hepatitis B virus (HBV) infection. Polyarteritis nodosa was present in 2 and in the other 2, cutaneous vasculitis, presenting clinically as palpable or Henoch-Schiinlein purpura. In one of these patients skin biopsies demonstrated granular deposits of IgM, C3, C4, and the hepatitis B surface antigen (HBsAg) and electron-dense deposits of aggregated 20-nm particles resembling HBsAg in postcapillary venules. Evidence for circulating HBsAg-immune complexes included increased serum Clq binding activity, decreased serum complement, and a cryoprecipitate containing both HBsAg and IgM anti-HBs. Aggregated 20-nm particles resembling intact HBsAg were also seen by negative staining electron microscopy of the serum cryoprecipitate. This patient fulfills all the criteria for a specific immune complex vasculitis caused by his immune response to a chronic HBV infection. These findings emphasize that HBV infection may be associated with small vessel vasculitis as well as polyarteritis nodosa, mixed cryoglobulinemia, and glomerulonephritis. A similar immune response to other viral infections may be expressed as palpable (Henoch-Schiinlein) purpura also.
Persistent virus infection may cause chronic immupe complex diseaseas documentedin an increasing number of spontaneousand experimental animal models,l In man chronic hepatitis B virus (HBV) infection has been directly, albeit infrequently, implicated in the immunopathogeneds of glomerulonephritis,2* a and indirectly associatedwith necrotizing vasculitis of muscular arteries in 30% to 50% of patients with polyarteritis nodosa(PAN).4*5 What is not well recognizedis the etiologic role of HBV infection in the more common syndrome of cutaneous necrotizing venuliti9 manifested clinically as palpable From the Vasculitis Study Group, Departmentsof Medicine, Dermatology, and Pediatrics, University of Colorado Medical Center; and the Section of Dermatology, Fitzsimons Army Medical Center. Supported by National Institutes of Health Asthma and Allergic Disease Center Grant AI 12028and General Clinical Research Program Grant 5 MO1 RR5I. Presentedin part at the American Congress of Allergy and Immunology, New York, N. Y., March, 1977. Received for publication Jan. IO, 1978. Accepted for publication June 27, 1978. Reprint requeststo: Peter F. Kohler, M.D., 4200 E. Ninth Ave., Denver, CO 80262. Vol. 62, No. 4, pp. 222-228
(Henoch-Schonlein) purpura which is much more common than PAN. In an ongoing, prospective study of 80 consecutive patients with vasculitis, PAN occurred in 4 patients compared to cutaneous necrotizing venulitis, clinically manifested as palpable or Henoch-Schonlein purpura, in 67. In 4 of these 80 patients HBV infection was diagnosed by the presence of hepatitis B surface antigen (HBsAg) concurrent with the vasculitis. Two of these patients had classical PAN, and 2 had cutaneous necrotizing venulitis. This report describes the extensive immunologic and immunohistologic studies done in one of the patients with Henoch-Schiinlein purpura which confirm that HBsAg * anti-HBs immune complexes were directly involved in the immunopathogenesisof his recurrent cutaneousvenulitis. CASE REPORT The patient is a 21-year-oldSamoanmalewith a 2-year history of diffuse,palpablepurpuraexacerbated by standing and accompanied by swelling of the hands and feet, arthralgias, and myalgias. There was no history of clinical hepatitis or likely exposure to hepatitis B virus. On physical examination he had a blood pressure of 95/125, erythema
VOLUME NUMBER
Vasculitis
62 4
FIG. 1. A, Macular spontaneous skin vessel and nuclear same skin biopsy tioned horizontally. ing finely granular the upper dermis
caused
by hepai :itis B
223
and papular purpuric lesions on the patient’s lower legs. B, Biopsy of a lesion demonstrating intensive neutrophilic cellular infiltrate in a high dermal dust. (Hematoxylin and eosin; x200.) C, Direct immunofluorescence of the demonstrating finely granular deposits of IgM in a high dermal vessel sec(x320.) 0, Direct immunofluorescence of the same skin biopsy demonstratdeposits of HBsAg in two high dermal vessels cut vertically. The location in indicates postcapillary involvement. (x320; see text.)
and periarticular swelling of the hands, and palpable purpura on his legs (Fig. 1, A). Liver function tests were only moderately abnormal with serum glutamic oxaloacetic transaminase (SGOT), 86 III/L (Normal, 9-30), alkaline phosphatase, 86 IU/L (Normal, 56-232), and total bilirubin, I .2 mg/dl (normal, 0.1-I .3). Repeated urinalyses showing l-3 RBC/hpf, O-2+ protein, and a creatinine clearance of 84 ml/min were suggestive of renal involvement. Short courses of antihistamines and prednisone did not alter his skin lesions.
MATERIALS AND METHODS Serologic tests for circulating
immune complexes in-
cluded isolation and characterization of cryoproteins,r. 8 a
Clq binding assay,a and negative staining electron microscopy of the cryoprecipitate.7 Total hemolytic complement,
C3, C4, CS, Clq, pmperdin, and rheumatoid factor (RF) were also determined. I0
Controls Four subjects had similar serologic studies and electron microscopy and immunofluorescent studies of skin biopsies from histamine-injected sites as described below. Three of 4 (G. L., C. M., D. P.) were chronic carriers of HBsAg and had renal disease but no clinical signs of cutaneous vasculitis in contrast to the patient. The fourth subject (E. M.) was normal. The HBsAg-positive controls were selected to ensure that the HBV carder state by itself was not associated with the immunohistologic findings observed in the patient. Informed consent was obtained from all five subjects. Biopsies of lesions and normal skin from 4 additional patients with palpable putpura who were HBsAg.negative
224 Gower et al.
J. ALLERGY
TABLE I. Serologic
data on patient
and controls
Serum HBsAg
Pt.
+
SL
+
CM RP
t +
CLIN. IMMUNOL. OCTOBER 1978
Cryoprotein Anti-HBs
-t
Qual*
HBsAg
Anti-HBs
+
+
+
f +
+ +
+ +
Clq assayt (40%)
(nlrZIo,l)
13.7 ND 3.9 16.2
34 ND 45 38
*Qualitativetestfor cryoprateins. SClqbindingassayresultsareexpressed as% ‘zsClqpresentin the 2.5%polyethyleneglycol-inducedprecipitatesof serum.
TABLEII.
lmmunofluorescent patient and controls
results from One hour after histamine
Spontaneous lesion
HBsAg HBcAg IgM W 16 C3 c4 Clq CS Properdin Factor B Fibrin
3-t Neg
1+ Neis N% 1+ 2+ Neg Neg Neg Nets ND
Patient
3+ Neg 1+ 1+ Neg 2+ 3+ Neg Neg Neg Neg 1+
Controls
Neg(3/3) Neg (3/3) 1+ (l/4) Neg Neg Neg ND ND ND ND ND Nets
were also studied to ensure that cutaneousvenulitis lesions were not associatedwith false-positive immunofluorescent localization of HBsAg , Skin biopsies A spontaneouslesion from the patient’s leg was biopsied using circular block technique with 1% lidocaine (without epinephrine). The specimenwas bisected, and one-half was placedimmediately into liquid nitrogen and storedat ~70” C for immunofluorescent study. The other half was further divided and placed in glutaraldehyde for electron microscopy (EM) studies and in formaldehyde for light microscopy. I1 Normal skin was processedsimilarly. In an attempt to induce skin lesions, 0.55 pg histamine phosphate(Eli Lilly & Co., Indianapolis) in 0.02 ml sterile water was injected intradermally into the lower leg of the patient and controls to induce vasodilation and promoteegress of circulating immune complexes and inflammatory cells. Punch biopsies of the histamine-injected sites were done one hour after injection. The specimenswere processed as above for immunofluorescentstudy and electron microscopy. Biopsies were done on two occasions,4 months apart, with similar findings,
lmmunofluorescence
(IF)
Skin. Direct IF staining of skin biopsy specimenswas performed using monospecific fluorescein-conjugatedantisera to IgG, IgM, IgA, C3, and fibrinogen obtained from Hyland Laboratories, Costa Mesa, Calif. Molar ratios of fluorescein to protein for these conjugated antisera were 2.4 (IgG), 3.3 (IgM), 3.0 (IgA), 2.3 (C3), 2.4 (fibrinogen), 1.4 (HBsAg), and 2.5 (HBcAg). Specimenswere stained for HBsAg and HBcAg with fluorescein isothiocyanate (FITC)-labeled reagents kindly donated by Dr. Adam Nowoslawski.* Immunofluorescencewas read on a scaleof O-3+ . Specificity of all fluorescein-labeledantibody, except for the anti-HBc, was confirmed by immunoelectrophoresis, Ouchterlony analysis, blocking with unlabeled antibody, and absorption with specific antigens. The staining with anti-HBs and anti-HBc’ reagents was also compared to horseradish peroxidase (I-JRPO)-labeled Fab,,reagents on sections of liver from HBsAg-positive and negative patients. Indirect immunofluorescentstudies employed goat antiClq, C4, and C5; horse anti-properdin; and rabbit antifactor B followed by fluoroscein-labeled rabbit anti-goat IgG and horse IgG and goat anti-rabbit IgG. The 4-p cryostat sectionsof skin were air-dried and then overlaid for 30 min with each antiserum in a humidified chamber at room temperature.Sections were washed with phosphate-bufferedsaline (PBS), pH 7.4, for 10 min after the application of each antiserum before viewing under epi-illumination with a Zeiss microscope equipped with a quartz halogen bulb and appropriate interference filters. Liver. Cryostat sections (6-8 CL) were taken from a specimenfixed in periodate-lysine-parformaldehyde.I* Direct IF staining was performed as above using antisera to IgG, C3, HBsAg, and HBcAg.
Light and electron
microscopy
(EM)
Skin. Specimenswere fixed in 4% glutaraldehyde and processed as previously described.” Plastic-embedded “thick” sections (500 nm) were stained for light microscopy with azure II, methylene blue, and basic fuchsin.13 “Thin” sections (70 nm) were double-contrasted with
*Departmentof Immunopathology,StateInstitute of Hygiene, Warsaw,Poland.
VOLUME NUMBER
Vasculitis caused by hepatitis B 225
62 4
FIG. 2.A, Electron microscopic view of an involved postcapillary venule showing electron-dense deposits (arrowhead) within the multilaminated basement membrane (BMI; E, endothelial cell; R, red cell; C, collagen fibers. (x30,500 original magnification.) Lumen of venule is to the left. Red cells are within the wall of the venule. B, Higher magnification (~102,000) of the electrondense deposit showing 20-27 nm discrete particles within the deposit. uranyl acetate and lead citrate and viewed with a Phillips 300 electron microscope. Particular attention was paid to blood vessel morphology, cellular identification, and a search for electron-dense deposits. Liver. Specimens were prepared and evaluated by light microscopy in two ways: (1) tissue was fixed in formaldehyde and stained with hematoxylin and eosin; and (2) tissue was fixed in periodate-lysine-paraformaldehyde for three hours, washed, embedded in Ames OCT compound, and quick-frozen in a mixture of ethanol and solid carbon dioxide.” Frozen sections (6-8 /L) were placed on albumincoated microscope glass slides and HBs and HBc antigens were localized using HRPO-labeled Fab’ of human IgG anti-HBs and anti-HBc.*
RESULTS Serologic studies Patient. Pertinent serologic data are listed here (normal values in parentheses) and in Table I: total hemolytic complement was 34 (33-61 CHSOp/ml); C3, 910 pg/ml (1,160-1,840); C4, 274 pg/ml (287-675); C5, 44 pug/ml (54-98); Clq, 170 pgiml *Courtesy of Drs. Gotaro Yamade and Paul K. Nakane, Department of Pathology, University of Colorado School of Medicine, Denver, Co10
TABLE III. Association of immune complex manifestations with hepatitis B virus infection incidence
Prodromal serum sickness and acute infection Polyarteritis nodosa Cutaneous venulitis (Henoch-Schanlein pwura) Essential mixed cryoglobulinemia Chronic glomemlonephritis
20%-30% (19) 30%-50% (4,5) 5% (present SW9 50% (18)
In a comprehensive study of 80 patients with vasculitis, 4 had concurrent hepatitis B virus (HBV) infection. Polyarteritis nodosa was present in 2 and in the other 2, cutaneous vasculitis, presenting clinically as palpable or Henoch-Schiinlein purpura. In one of these patients skin biopsies demonstrated granular deposits of IgM, C3, C4, and the hepatitis B surface antigen (HBsAg) and electron-dense deposits of aggregated 20-nm particles resembling HBsAg in postcapillary venules. Evidence for circulating HBsAg-immune complexes included increased serum Clq binding activity, decreased serum complement, and a cryoprecipitate containing both HBsAg and IgM anti-HBs. Aggregated 20-nm particles resembling intact HBsAg were also seen by negative staining electron microscopy of the serum cryoprecipitate. This patient fulfills all the criteria for a specific immune complex vasculitis caused by his immune response to a chronic HBV infection. These findings emphasize that HBV infection may be associated with small vessel vasculitis as well as polyarteritis nodosa, mixed cryoglobulinemia, and glomerulonephritis. A similar immune response to other viral infections may be expressed as palpable (Henoch-Schiinlein) purpura also.
Persistent virus infection may cause chronic immupe complex diseaseas documentedin an increasing number of spontaneousand experimental animal models,l In man chronic hepatitis B virus (HBV) infection has been directly, albeit infrequently, implicated in the immunopathogeneds of glomerulonephritis,2* a and indirectly associatedwith necrotizing vasculitis of muscular arteries in 30% to 50% of patients with polyarteritis nodosa(PAN).4*5 What is not well recognizedis the etiologic role of HBV infection in the more common syndrome of cutaneous necrotizing venuliti9 manifested clinically as palpable From the Vasculitis Study Group, Departmentsof Medicine, Dermatology, and Pediatrics, University of Colorado Medical Center; and the Section of Dermatology, Fitzsimons Army Medical Center. Supported by National Institutes of Health Asthma and Allergic Disease Center Grant AI 12028and General Clinical Research Program Grant 5 MO1 RR5I. Presentedin part at the American Congress of Allergy and Immunology, New York, N. Y., March, 1977. Received for publication Jan. IO, 1978. Accepted for publication June 27, 1978. Reprint requeststo: Peter F. Kohler, M.D., 4200 E. Ninth Ave., Denver, CO 80262. Vol. 62, No. 4, pp. 222-228
(Henoch-Schonlein) purpura which is much more common than PAN. In an ongoing, prospective study of 80 consecutive patients with vasculitis, PAN occurred in 4 patients compared to cutaneous necrotizing venulitis, clinically manifested as palpable or Henoch-Schonlein purpura, in 67. In 4 of these 80 patients HBV infection was diagnosed by the presence of hepatitis B surface antigen (HBsAg) concurrent with the vasculitis. Two of these patients had classical PAN, and 2 had cutaneous necrotizing venulitis. This report describes the extensive immunologic and immunohistologic studies done in one of the patients with Henoch-Schiinlein purpura which confirm that HBsAg * anti-HBs immune complexes were directly involved in the immunopathogenesisof his recurrent cutaneousvenulitis. CASE REPORT The patient is a 21-year-oldSamoanmalewith a 2-year history of diffuse,palpablepurpuraexacerbated by standing and accompanied by swelling of the hands and feet, arthralgias, and myalgias. There was no history of clinical hepatitis or likely exposure to hepatitis B virus. On physical examination he had a blood pressure of 95/125, erythema
VOLUME NUMBER
Vasculitis
62 4
FIG. 1. A, Macular spontaneous skin vessel and nuclear same skin biopsy tioned horizontally. ing finely granular the upper dermis
caused
by hepai :itis B
223
and papular purpuric lesions on the patient’s lower legs. B, Biopsy of a lesion demonstrating intensive neutrophilic cellular infiltrate in a high dermal dust. (Hematoxylin and eosin; x200.) C, Direct immunofluorescence of the demonstrating finely granular deposits of IgM in a high dermal vessel sec(x320.) 0, Direct immunofluorescence of the same skin biopsy demonstratdeposits of HBsAg in two high dermal vessels cut vertically. The location in indicates postcapillary involvement. (x320; see text.)
and periarticular swelling of the hands, and palpable purpura on his legs (Fig. 1, A). Liver function tests were only moderately abnormal with serum glutamic oxaloacetic transaminase (SGOT), 86 III/L (Normal, 9-30), alkaline phosphatase, 86 IU/L (Normal, 56-232), and total bilirubin, I .2 mg/dl (normal, 0.1-I .3). Repeated urinalyses showing l-3 RBC/hpf, O-2+ protein, and a creatinine clearance of 84 ml/min were suggestive of renal involvement. Short courses of antihistamines and prednisone did not alter his skin lesions.
MATERIALS AND METHODS Serologic tests for circulating
immune complexes in-
cluded isolation and characterization of cryoproteins,r. 8 a
Clq binding assay,a and negative staining electron microscopy of the cryoprecipitate.7 Total hemolytic complement,
C3, C4, CS, Clq, pmperdin, and rheumatoid factor (RF) were also determined. I0
Controls Four subjects had similar serologic studies and electron microscopy and immunofluorescent studies of skin biopsies from histamine-injected sites as described below. Three of 4 (G. L., C. M., D. P.) were chronic carriers of HBsAg and had renal disease but no clinical signs of cutaneous vasculitis in contrast to the patient. The fourth subject (E. M.) was normal. The HBsAg-positive controls were selected to ensure that the HBV carder state by itself was not associated with the immunohistologic findings observed in the patient. Informed consent was obtained from all five subjects. Biopsies of lesions and normal skin from 4 additional patients with palpable putpura who were HBsAg.negative
224 Gower et al.
J. ALLERGY
TABLE I. Serologic
data on patient
and controls
Serum HBsAg
Pt.
+
SL
+
CM RP
t +
CLIN. IMMUNOL. OCTOBER 1978
Cryoprotein Anti-HBs
-t
Qual*
HBsAg
Anti-HBs
+
+
+
f +
+ +
+ +
Clq assayt (40%)
(nlrZIo,l)
13.7 ND 3.9 16.2
34 ND 45 38
*Qualitativetestfor cryoprateins. SClqbindingassayresultsareexpressed as% ‘zsClqpresentin the 2.5%polyethyleneglycol-inducedprecipitatesof serum.
TABLEII.
lmmunofluorescent patient and controls
results from One hour after histamine
Spontaneous lesion
HBsAg HBcAg IgM W 16 C3 c4 Clq CS Properdin Factor B Fibrin
3-t Neg
1+ Neis N% 1+ 2+ Neg Neg Neg Nets ND
Patient
3+ Neg 1+ 1+ Neg 2+ 3+ Neg Neg Neg Neg 1+
Controls
Neg(3/3) Neg (3/3) 1+ (l/4) Neg Neg Neg ND ND ND ND ND Nets
were also studied to ensure that cutaneousvenulitis lesions were not associatedwith false-positive immunofluorescent localization of HBsAg , Skin biopsies A spontaneouslesion from the patient’s leg was biopsied using circular block technique with 1% lidocaine (without epinephrine). The specimenwas bisected, and one-half was placedimmediately into liquid nitrogen and storedat ~70” C for immunofluorescent study. The other half was further divided and placed in glutaraldehyde for electron microscopy (EM) studies and in formaldehyde for light microscopy. I1 Normal skin was processedsimilarly. In an attempt to induce skin lesions, 0.55 pg histamine phosphate(Eli Lilly & Co., Indianapolis) in 0.02 ml sterile water was injected intradermally into the lower leg of the patient and controls to induce vasodilation and promoteegress of circulating immune complexes and inflammatory cells. Punch biopsies of the histamine-injected sites were done one hour after injection. The specimenswere processed as above for immunofluorescentstudy and electron microscopy. Biopsies were done on two occasions,4 months apart, with similar findings,
lmmunofluorescence
(IF)
Skin. Direct IF staining of skin biopsy specimenswas performed using monospecific fluorescein-conjugatedantisera to IgG, IgM, IgA, C3, and fibrinogen obtained from Hyland Laboratories, Costa Mesa, Calif. Molar ratios of fluorescein to protein for these conjugated antisera were 2.4 (IgG), 3.3 (IgM), 3.0 (IgA), 2.3 (C3), 2.4 (fibrinogen), 1.4 (HBsAg), and 2.5 (HBcAg). Specimenswere stained for HBsAg and HBcAg with fluorescein isothiocyanate (FITC)-labeled reagents kindly donated by Dr. Adam Nowoslawski.* Immunofluorescencewas read on a scaleof O-3+ . Specificity of all fluorescein-labeledantibody, except for the anti-HBc, was confirmed by immunoelectrophoresis, Ouchterlony analysis, blocking with unlabeled antibody, and absorption with specific antigens. The staining with anti-HBs and anti-HBc’ reagents was also compared to horseradish peroxidase (I-JRPO)-labeled Fab,,reagents on sections of liver from HBsAg-positive and negative patients. Indirect immunofluorescentstudies employed goat antiClq, C4, and C5; horse anti-properdin; and rabbit antifactor B followed by fluoroscein-labeled rabbit anti-goat IgG and horse IgG and goat anti-rabbit IgG. The 4-p cryostat sectionsof skin were air-dried and then overlaid for 30 min with each antiserum in a humidified chamber at room temperature.Sections were washed with phosphate-bufferedsaline (PBS), pH 7.4, for 10 min after the application of each antiserum before viewing under epi-illumination with a Zeiss microscope equipped with a quartz halogen bulb and appropriate interference filters. Liver. Cryostat sections (6-8 CL) were taken from a specimenfixed in periodate-lysine-parformaldehyde.I* Direct IF staining was performed as above using antisera to IgG, C3, HBsAg, and HBcAg.
Light and electron
microscopy
(EM)
Skin. Specimenswere fixed in 4% glutaraldehyde and processed as previously described.” Plastic-embedded “thick” sections (500 nm) were stained for light microscopy with azure II, methylene blue, and basic fuchsin.13 “Thin” sections (70 nm) were double-contrasted with
*Departmentof Immunopathology,StateInstitute of Hygiene, Warsaw,Poland.
VOLUME NUMBER
Vasculitis caused by hepatitis B 225
62 4
FIG. 2.A, Electron microscopic view of an involved postcapillary venule showing electron-dense deposits (arrowhead) within the multilaminated basement membrane (BMI; E, endothelial cell; R, red cell; C, collagen fibers. (x30,500 original magnification.) Lumen of venule is to the left. Red cells are within the wall of the venule. B, Higher magnification (~102,000) of the electrondense deposit showing 20-27 nm discrete particles within the deposit. uranyl acetate and lead citrate and viewed with a Phillips 300 electron microscope. Particular attention was paid to blood vessel morphology, cellular identification, and a search for electron-dense deposits. Liver. Specimens were prepared and evaluated by light microscopy in two ways: (1) tissue was fixed in formaldehyde and stained with hematoxylin and eosin; and (2) tissue was fixed in periodate-lysine-paraformaldehyde for three hours, washed, embedded in Ames OCT compound, and quick-frozen in a mixture of ethanol and solid carbon dioxide.” Frozen sections (6-8 /L) were placed on albumincoated microscope glass slides and HBs and HBc antigens were localized using HRPO-labeled Fab’ of human IgG anti-HBs and anti-HBc.*
RESULTS Serologic studies Patient. Pertinent serologic data are listed here (normal values in parentheses) and in Table I: total hemolytic complement was 34 (33-61 CHSOp/ml); C3, 910 pg/ml (1,160-1,840); C4, 274 pg/ml (287-675); C5, 44 pug/ml (54-98); Clq, 170 pgiml *Courtesy of Drs. Gotaro Yamade and Paul K. Nakane, Department of Pathology, University of Colorado School of Medicine, Denver, Co10
TABLE III. Association of immune complex manifestations with hepatitis B virus infection incidence
Prodromal serum sickness and acute infection Polyarteritis nodosa Cutaneous venulitis (Henoch-Schanlein pwura) Essential mixed cryoglobulinemia Chronic glomemlonephritis
20%-30% (19) 30%-50% (4,5) 5% (present SW9 50% (18)
