- Email: [email protected]
Binding of synthetic double-stranded DNA by serum from patients with systemic lupus erythematosus
CLINICAL STUDIES
Binding of Synthetic Double-Stranded DNA by Serum from Patients with Systemic Lupus Erythematosus Correlation with Renal Histology
CHARLES R. STEINMAN,
M.D.
EDITH GRISHMAN, M.D. HARRY SPIERA, M.D. UTIS DEESOMCHOK,
M.D.
New York, New York
From the Mount Sinai School of Medicine, New York, New York. This study was supported in part by Grants AM 15544 and AM 00918 from the National Institutes of Health, National Institute of Arthritis, Metabolic and Digestive Diseases, and by grants from the New York Lupus Erythematosus Foundation and the New York Chapter of the Arthritis Foundation. Requests for reprints should be addressed to Dr. Charles R. Steinman, Mount Sinai School of Medicine, 1 East 100th Street, New York, New York 10029. Manuscript accepted June 2, 1978.
Detection of antibody to double-stranded DNA by direct binding assays has proved useful in clinical management of patients with systemic lupus erythematosus (SLE). Recent confusion regarding specificity of these antibodies for SLE appedrs to be due, at least in part, to contamination of natural DNA preparations with nondouble-stranded DNA antigens. Measurement of Mnding of a synthetic, self-complementary DNA copolymer (dAT) rather than of natural DNA (KB) has been shown to obviate some of these difficulties, apparently because of freedom of dAT from nondouble-stranded DNA antigens. Among the advantages found in this way was a higher degree of specificity of antibodies to double-stranded DNA for clinically-judged active lupus nephritis than had been suspected. Since activity of nephritis is difficult to assess clinically, histologic data were sought to confirm these observations. Thirty-two kidney specimens were examined by light and/or electron microscopy. The degree of histologic activity and the amount and location of glomerular electron-dense deposits were semiquantitated blindly. The binding of both dAT and KB DNA was measured by the ammonium sulfate method. Correlation with the amount of elestron-dense deposits was highly significant for dAT binding and somewhat less so for KB DNA binding as determined by both parametric and nonparametric statistical methods. Significant correlation with histologic activity was found for dAT but not KB DNA binding. These results are consistent with previous data and suggest that dAT binding may provide a useful, noninvasive means of clinically assessing both nephritis activity and the intensity of glomerular immune-complex deposition as reflected by the amount of electron-dense deposits. If it can be confirmed that the latter provides long-term prognostic information, then dAT binding (and perhaps its reponse to therapy) may also prove of value in this regard. Antibodies specific for double-stranded DNA have long been thought to occur exclusively in systemic lupus erythematosus (SLE) either in humans or in animal models [ 11. Extensive clinical and investigative work has indicated that such antibodies, along with complement and DNA, are deposited in the glomerular basement membrane in patients with lupus nephritis, where
March 1977
The American Journalof Medicine
Volume 62
319
BINDING OF DNA IN SYSTEMIC
LUPUS ERYTHEMATOSUS-STEINMAN
ET AL
they are thought to act as at least one pathogenetic factor in this disorder [ 21. In addition, it has been shown that the amount of such antibody in serum from patients with SLE is a useful clinical parameter of disease activity [3,4]. Recognition [5,6] that the naturallyderived “native” DNA preparations generally used in assaying these antibodies are virtually always contaminated with small but immunologically significant amounts of singlestranded DNA has resulted in ambiguous interpretation of recent studies that had appeared to challenge these concepts by suggesting the presence of antidoublestranded DNA in people with diseases other than SLE [ 7-91 and in normal subjects [ lo]. Such contamination of complex natural double-stranded DNA may be virtually unavoidable during routine isolation and purification procedures [ 61. Although special procedures for removal of single-stranded DNA may be utilized, their reliability is uncertain and, even if successful, do not obviate the risk of reintroducing single-stranded regions during storage and handling [6]. Accordingly, an alternate approach was proposed [6] in which a radiolabelled, synthetic, self-complementary, alternating copolymer of deoxyribonucleotides would be substituted for the natural antigen in binding assays. Such a polymer is, on thermodynamic grounds, unstable in the singlestranded state at conditions under which the ammonium sulfate binding assay is performed [ 11,121. Hence contamination with single-stranded DNA would not be anticipated. Since antidouble-stranded DNA is thought to recognize only antigenic sites on the deoxyribosephosphate backbone, and not individual bases [ 11, it was expected that such a polymer would be an appropriate model antigen for detection of antibody to double-stranded DNA. In previous work [6] the alternating copolymer of deoxyadenytate and deoxythymidylate (dAT) was shown to be free of detectable single-stranded DNA by several sensitive assays. Immunologic analysis along with examination of a large number of serums from patients with SLE or with nonSLE rheumatologic disorders as well as normal control subjects for antidouble-stranded DNA utilizing C i4-labeled dAT and DNA from human cells (KB DNA) led to the following conclusions: (1) dAT contained most, if not all, of the antigens intrinsic to natural double-stranded DNA; (2) as compared with KB DNA, binding of dAT was more nearly specific for SLE in that increased dAT binding by serum from normal or nonSLE control subjects was not found; (3) binding by serum from patients with SLE, as detected by the ammonium sulfate assay, appeared to result from interaction with immunoglobulins, presumed to be antibody; (4) increased dAT binding correlated more closely than did that of KB DNA with active lupus nephritis, as judged by clinical criteria; (5) in contrast to KB binding, no
320
March 1977
The American Journal of Medicine
Volume 62
correlation with the presence of active nonrenal lupus (including cerebritis) was observed; and (6) when increased, the amount of dAT binding fluctuated with activity of disease. Hence, it appeared possible that there might be a closer and more selective association between the presence of antidouble-stranded DNA antibody, as measured by dAT binding and lupus nephritis, than had previously been suspected from studies based on use of the natural antigen [3,4]. Because of the difficulty of assessing activity of lupus nephritis by clinical criteria, renal biopsy data are essential for confirmation of these conclusions. The present report is of such data and allows correlation of histologic findings with serum binding of both dAT and naturallyderived DNA (KB). MATERIALS
AND
METHODS
A total of 32 kidney specimens from 31 patients was available for examination. Three were obtained at autopsy. Six of the patients, aged between 14 and 47, were male, and 25 were female. Biopsies were performed only when thought to be clinically indicated, generally in the presence of clinical evidence of renal disease. Twenty-five patients were receiving a cytostatic or corticosteroid drug at the time tissue was obtained. Of these, eight were receiving more than 30 mg/day of prednisona or the equivalent, and/or a cytostatic agent, 17 were receiving less than this dose of corticosteroid and five were receiving no corticosteroids or cytostatic agents. Information about therapy was unavailable for two patients. The histologic specimens were all examined by Histology. a single observer (E.G.). Precautions were taken to be certain that her interpretation was performed blindly. In all but three cases, at least 6 glomeruli were examined by light microscopy. Exclusion of these three cases did not significantly altar the results of the study. Under light microscopy, the glomerular changes were classified, and the degree of activity, as judged by previously defined criteria [ 13,141, was quantitated on an arbitrary scale of 0 to 3. Based on examination of electron micrographs of multiple areas, of 1 or, in 14 cases, 2 glomeruli, electron-dense deposits were classified by location and their amount was semiquantitated on a 0 to 3 scale [ 141. Binding assays were performed by DNA Binding Assays. a modified ammonium sulfate precipitation technic [3,6,15]. C14-labeled KB DNA was obtained from Electronucleonics, Bethesda, Maryland. As judged by susceptibility to a singlestranded-DNA-specific nuclease, similar preparations have been found to be contaminated with between 9 and 37 per cent single-stranded DNA [6]. No data are available on possible contamination with other potential antigens. No attempt was made to further purify the KB DNA preparations prior to use in binding assays. Previous experience has suggested that at least partial elimination of single-stranded DNA contaminants by nuclease pretreatment or, less effectively, by chromatography on methylated albumin kieselguhr resulted in an antigen whose binding was more nearly similar to that of dAT [6]. C14-labelled dAT was purchased from Biopolymers (Chagrin Falls, Ohio) or from Grand Island Patients.
ET AL.
BINDING OF DNA IN SYSTEMIC LUPUS ERYTHEMATOSUS-STEINMAN
TABLE
I
100
Histologic Findings Location of ElectronDense Deposits
Type of Glomerular Lesion by LM ______ Minimal
4
Mesangial Membranous (pure) Focal proliferative Diffuse proliferative Not examined
8 1 4 14 1
Total
32
Mesangiai, and/or subepithelial alone Intramembranous” Subendothelial’ No deposits
r
.
I-
’
oj
r=0.582
-
I
/
16 5 5 4 2
Not examined
32
NOTE: LM = light microscopy. *Specimens with intramembranous or subendothelial deposits also frequently contained mesangial and/or subepithelial deposits but are classified here only under the former two categories.
EDD
Biological Co. (Grand Island, New York). Material obtained from the latter source required sonication prior to use [6]. Serums were obtained within two weeks of biopsy or autopsy except in five cases in which two to five weeks intervened. Exclusion of these specimens did not alter the conclusions reached. RESULTS
6
I are summarized the light and electron microscopic findings. The small number of specimens showing minimal disease by light microscopy is consistent with the clinical indications generally required for performance of biopsies. Similarly, the distribution of electron-dense deposits does not appear exceptional
0
In Figure 1 is shown the correlation between binding of dAT (top) or of KB DNA (bottom), with the amount of electron-dense deposits. The solid lines are those calculated for regression of binding upon electron-dense deposits [ 171. Binding of both dAT and KB DNA correlated significantly with electron-dense deposits although the correlation coefficient, r, was higher for dAT than for KB DNA binding. Use of the nonparametric rank correlation method of Spearman also led to a significant correlation between these variables but with smaller correlation coefficients [ 171.
All Specimens Number of specimens Correlation coefficient t P
31
Subendothelial and Intramembranous 10
TABLE III
0.956
Subendothelial
I 2.0
I 2.5
I 30
Correlation of Binding with Histologic Activity of Nephritis
5
Number of specimens (N j Correlation coefficient (r) t
13
0.977
0.759
P 10.33
I 1.5
dAT
Diffuse Proliferative
(r) 6.35
I I.0
Antigen Usedin Binding Assay
(N) 0.752
I 13.5
The open circles indicate those specimens with electron-dense deposits in intramembranous and/or subendothelial locations (i.e., excluding specimens with no deposition or with deposits in mesangial and/or subepithelial locations). As seen in Table II, this subgroup exhibited somewhat higher correlation coefficients than did the entire group but with the pre-
Correlation of dAT Binding with Electron-Dense Deposits
II
0
EDD Figure 1. Correlation of dA T binding (top) and KS DNA binding (bottom) with amount of electronxfense deposits (EDD). Open circles represent specimens with subendothelial or intramembranous deposits. Solid lines represent those calculated for the regression of binding on electron-dense deposits. Correlation coefficients are indicated by r.
[161.
TABLE
l
0
In Table
10.23
4.20 <0.005
NOTE: dAT oxyadenylate
March 1977
31
KB DNA 29
0.600
0.275
4.18
1.525 >o. 1
= self-complimentary alternating copolymer of deand deoxythymidylate. KB DNA = human DNA.
The American Journal of Medicine
Volume 62
321
BINDING OF DNA IN SYSTEMIC LUPUS ERYTHEMATOSUS-STEINMAN ET AL
viously observed difference between dAT and KB DNA binding being maintained. The five specimens exhibiting subendothelial localization of electron-dense deposits similarly exhibited a higher correlation coefficient than did the entire group although, despite the significant p value, the small number of specimens in this group makes interpretation of this finding somewhat uncertain. The specimens with diffuse proliferative disease on light microscopy did not differ in their correlation from that of the entire group. Binding of dAT (Table III) also correlated significantly with the degree of activity of glomerulitis as judged by light microscopy, although at a lower level than with electron-dense deposits. KB binding did not exhibit such a correlation. COMMENTS
Although several reports have demonstrated a correlation between naturallyderived “native” DNA binding levels and clinical activity of SLE [3,4], no data have appeared on such a correlation with histologic parameters of renal disease. The binding of natural native DNA has been said to correlate with the amount of glomerular immune-complex deposition as identified by immunofluorescence, although no data are given [ 181. Since it has been shown that the usually measured clinical parameters of renal disease are not reliable guides to histology [ 191, absence of biopsy information has been an obstacle to clear interpretation of DNA binding data. The present data demonstrate that dAT binding correlates closely with glomerular electron-dense deposits (which have been shown to contain immunoglobulin and complement [20] and hence are presumed to be immune complexes) as well as with activity of nephritis on light microscopy so that the previously observed correlation with clinically-judged activity of lupus nephritis appears confirmed. That dAT binding correlates more closely with histologic findings than does KB DNA binding is consistent with a large body of data suggesting that antidoublestranded DNA antibodies are of pathogenetic importance in renal lupus [2] and that routinely-prepared, naturally-derived DNA antigens may detect antisinglestranded DNA as well as antidouble-stranded DNA antibodies [5,6]. The possibility that they can also detect antibody to other, non-nucleic acid antigenic contaminants (such as nucleoprotein) also must be considered.
Hence, binding of naturally-derived antigens might not be expected to exhibit as much specificity for either antidouble-stranded DNA or activity of nephritis as would a homogeneous double-stranded DNA antigen. In general the observed correlations were quite close, particularly in view of the semiquantitative nature of the histologic measurements as well as the likelihood of biopsy sampling error. However, because of these limitations, it is unclear whether the few aberrant values that were observed resulted from biologic or methodologic factors. Hence it is difficult to form firm conclusions about the certainty with which dAT binding can predict those histologic variables with which it correlates. Several recent studies [ 14,211 appear to agree, on the basis of limited data, that the histologic finding most useful for predicting prognosis is the amount of electron-dense deposits, particularly if located subendothelially. To the extent that dAT binding accurately reflects the amount of deposition, it too may provide similar information by a noninvasive procedure that can be easily monitored clinically. No clear correlation of dAT binding with location of deposits was apparent, although it remains possible that the previously noted methodologic imprecision and the small number of samples with subendothelial deposits could have obscured such an association. Finally, it must be emphasized that the present results do not yet allow conclusions regarding possible correlations between therapeutically-induced changes in dAT binding with changes in renal morphology for which serially-obtained data would be required. However, a recent report [22] has suggested that, among patients whose biopsy specimens initially showed diffuse proliferative disease with subendothelial deposits, failure of normalization of naturally derived “native” DNA binding during treatment implies a poor prognosis. These data appear to be consistant with results of the present study and suggest that presumed therapeutically-induced changes (or lack thereof) in dAT binding also may have prognostic importance. ACKNOWLEDGMENT
The advice of Dr. Jacob Churg is gratefully acknowledged as is the technical assistance of Ms. Ping Tong and the help of Mrs. Anita Sardo in preparation of the manuscript.
REFERENCES 1. 2.
3.
322
Levine L, Stellar B: Nucleic acid immune systems. Prog Allergy 12: 161, 1968. Agnello V, Koffler D. Kunkel HG: Immune complex systems in the nephritis of systemic lupus erythematosus. Kidney Int 3: 90, 1973. Pincus T, Schur P, Rose J, et al.: Measurement of serum
March 1977
The
American Journal of M8dlCfne
Volume 82
4. 5.
DNA-binding activity in systemic lupus erythematosus. N Engl J Med 281: 701, 1969. Hughes G, Cohen S, Christian C: Anti-DNA activity in systemic lupus erythematosus. Ann Rheum Dis 30: 259, 197 1. Samaha R, Irvin W: Deoxyribonucleic acid (DNA) strandedness: partial characterization of the antigenic regions
BINDING OF DNA IN SYSTEMIC: LUPUS ERYTHEMATOSUS--STEINMAN
6.
7.
8.
9. 10.
11.
12. 13.
binding antibodies in lupus erythematosus serum. J Clin Invest 56: 446, 1975. Steinman CR, Deesomchok U, Spiera H: Detection of antibody using synthetic antigens. Characterization and clinical significance of binding of poly(deoxyadenylatedeoxythymidylate) by serum. J Clin Invest 57: 1330, 1976. Epstein W, Tan M, Esterbrook M: Serum antibody to doublestranded RNA and DNA in patients with idiopathic and secondary uveites. N Engl J Med 285: 1502, 1971. Hahn B, Sharp G, Irvin W, et al.: Immune responses to hydralazine and nuclear antigens in hydralazine induced lupus erythematosus. Ann Intern Med 76: 365, 1972. Rodinis, P. Palefsky H, Beeker M, et al.: Native DNA binding in rheumatoid arthritis. Ann Rheum Dis 33: 357, 1974. Hasselbacher P, LeRoy E: Serum DNA binding activity in healthy subjects and in rheumatic disease. Arthritis Rheum 17: 63, 1974. Schachman H, Adler J, Radding C, et al.: Enzymatic synthesis of deoxyribonucleic acid:Vlll. Synthesis of a polymer of deoxyadenylate and deoxythymidylate. J Biol Chem 235: 3242, 1960. Geiduschek EP: On the factors controlling the reversibility of DNA denaturation. J Molec Biol 4: 467, 1962. Pollak VE, Pirani CL, Schwartz FD: Clinical and experimental. The natural history of the renal manifestations of systemic lupus erythematosus. J Lab Clin Med 63: 537, 1964.
14.
15.
16.
17. 18.
19. 20.
21.
22.
March 1977
ET AL.
Grishman E, Porush JG, Lee SL, Churg J: Renal biopsies in lupus nephritis. Correlation of electron microscopic findings with clinical course. Nephron 10: 25, 1973. Wold RT, Young FE, Tan EM, Fan RS: Deoxyribonucleic acid antibody. A method to detect its primary interaction with deoxyribonucleic acid. Science 161: 806, 1968. Dujovne I, Pollak VE, Pirani CL, Dillard MG: The distribution and character of glomerular deposits in systemic lupus erythematosus. Kidney Int 2: 33, 1972. Snedecor G, Cochran W: Statistical Methods, 6th ed, Ames Iowa, Iowa State University Press, 1967, p 135. Hill GS, Hinglais N, Tron F, Bach JF: Systemic lupus erythematosus. Correlation of renal biopsies with immunologic data at the time of biopsy (abstract). Kidney Int 8: 449, 1975. Dubois EL: Lupus Erythematosus, 2nd ed., Los Angeles, University of Southern California Press, 1974, p 87. Comerford FR, Cohen AS: The nephropathy of systemic lupus erythematosus. Medicine (Baltimore) 46: 425, 1967. Dillard MG, Tillman LR, Sampson CC: Lupus nephritis. Correlations between the clinical course and presence of electrondense deposits. Lab Invest 32: 261, 1975. Adler MK, Baumgarten A, Hecht B, et al.: Prognostic significance of DNA-binding capacity patterns in patients with lupus nephritis. Ann Rheum Dis 34: 444, 1975.
The American Journal of Medklne
Volume 62
323
Binding of Synthetic Double-Stranded DNA by Serum from Patients with Systemic Lupus Erythematosus Correlation with Renal Histology
CHARLES R. STEINMAN,
M.D.
EDITH GRISHMAN, M.D. HARRY SPIERA, M.D. UTIS DEESOMCHOK,
M.D.
New York, New York
From the Mount Sinai School of Medicine, New York, New York. This study was supported in part by Grants AM 15544 and AM 00918 from the National Institutes of Health, National Institute of Arthritis, Metabolic and Digestive Diseases, and by grants from the New York Lupus Erythematosus Foundation and the New York Chapter of the Arthritis Foundation. Requests for reprints should be addressed to Dr. Charles R. Steinman, Mount Sinai School of Medicine, 1 East 100th Street, New York, New York 10029. Manuscript accepted June 2, 1978.
Detection of antibody to double-stranded DNA by direct binding assays has proved useful in clinical management of patients with systemic lupus erythematosus (SLE). Recent confusion regarding specificity of these antibodies for SLE appedrs to be due, at least in part, to contamination of natural DNA preparations with nondouble-stranded DNA antigens. Measurement of Mnding of a synthetic, self-complementary DNA copolymer (dAT) rather than of natural DNA (KB) has been shown to obviate some of these difficulties, apparently because of freedom of dAT from nondouble-stranded DNA antigens. Among the advantages found in this way was a higher degree of specificity of antibodies to double-stranded DNA for clinically-judged active lupus nephritis than had been suspected. Since activity of nephritis is difficult to assess clinically, histologic data were sought to confirm these observations. Thirty-two kidney specimens were examined by light and/or electron microscopy. The degree of histologic activity and the amount and location of glomerular electron-dense deposits were semiquantitated blindly. The binding of both dAT and KB DNA was measured by the ammonium sulfate method. Correlation with the amount of elestron-dense deposits was highly significant for dAT binding and somewhat less so for KB DNA binding as determined by both parametric and nonparametric statistical methods. Significant correlation with histologic activity was found for dAT but not KB DNA binding. These results are consistent with previous data and suggest that dAT binding may provide a useful, noninvasive means of clinically assessing both nephritis activity and the intensity of glomerular immune-complex deposition as reflected by the amount of electron-dense deposits. If it can be confirmed that the latter provides long-term prognostic information, then dAT binding (and perhaps its reponse to therapy) may also prove of value in this regard. Antibodies specific for double-stranded DNA have long been thought to occur exclusively in systemic lupus erythematosus (SLE) either in humans or in animal models [ 11. Extensive clinical and investigative work has indicated that such antibodies, along with complement and DNA, are deposited in the glomerular basement membrane in patients with lupus nephritis, where
March 1977
The American Journalof Medicine
Volume 62
319
BINDING OF DNA IN SYSTEMIC
LUPUS ERYTHEMATOSUS-STEINMAN
ET AL
they are thought to act as at least one pathogenetic factor in this disorder [ 21. In addition, it has been shown that the amount of such antibody in serum from patients with SLE is a useful clinical parameter of disease activity [3,4]. Recognition [5,6] that the naturallyderived “native” DNA preparations generally used in assaying these antibodies are virtually always contaminated with small but immunologically significant amounts of singlestranded DNA has resulted in ambiguous interpretation of recent studies that had appeared to challenge these concepts by suggesting the presence of antidoublestranded DNA in people with diseases other than SLE [ 7-91 and in normal subjects [ lo]. Such contamination of complex natural double-stranded DNA may be virtually unavoidable during routine isolation and purification procedures [ 61. Although special procedures for removal of single-stranded DNA may be utilized, their reliability is uncertain and, even if successful, do not obviate the risk of reintroducing single-stranded regions during storage and handling [6]. Accordingly, an alternate approach was proposed [6] in which a radiolabelled, synthetic, self-complementary, alternating copolymer of deoxyribonucleotides would be substituted for the natural antigen in binding assays. Such a polymer is, on thermodynamic grounds, unstable in the singlestranded state at conditions under which the ammonium sulfate binding assay is performed [ 11,121. Hence contamination with single-stranded DNA would not be anticipated. Since antidouble-stranded DNA is thought to recognize only antigenic sites on the deoxyribosephosphate backbone, and not individual bases [ 11, it was expected that such a polymer would be an appropriate model antigen for detection of antibody to double-stranded DNA. In previous work [6] the alternating copolymer of deoxyadenytate and deoxythymidylate (dAT) was shown to be free of detectable single-stranded DNA by several sensitive assays. Immunologic analysis along with examination of a large number of serums from patients with SLE or with nonSLE rheumatologic disorders as well as normal control subjects for antidouble-stranded DNA utilizing C i4-labeled dAT and DNA from human cells (KB DNA) led to the following conclusions: (1) dAT contained most, if not all, of the antigens intrinsic to natural double-stranded DNA; (2) as compared with KB DNA, binding of dAT was more nearly specific for SLE in that increased dAT binding by serum from normal or nonSLE control subjects was not found; (3) binding by serum from patients with SLE, as detected by the ammonium sulfate assay, appeared to result from interaction with immunoglobulins, presumed to be antibody; (4) increased dAT binding correlated more closely than did that of KB DNA with active lupus nephritis, as judged by clinical criteria; (5) in contrast to KB binding, no
320
March 1977
The American Journal of Medicine
Volume 62
correlation with the presence of active nonrenal lupus (including cerebritis) was observed; and (6) when increased, the amount of dAT binding fluctuated with activity of disease. Hence, it appeared possible that there might be a closer and more selective association between the presence of antidouble-stranded DNA antibody, as measured by dAT binding and lupus nephritis, than had previously been suspected from studies based on use of the natural antigen [3,4]. Because of the difficulty of assessing activity of lupus nephritis by clinical criteria, renal biopsy data are essential for confirmation of these conclusions. The present report is of such data and allows correlation of histologic findings with serum binding of both dAT and naturallyderived DNA (KB). MATERIALS
AND
METHODS
A total of 32 kidney specimens from 31 patients was available for examination. Three were obtained at autopsy. Six of the patients, aged between 14 and 47, were male, and 25 were female. Biopsies were performed only when thought to be clinically indicated, generally in the presence of clinical evidence of renal disease. Twenty-five patients were receiving a cytostatic or corticosteroid drug at the time tissue was obtained. Of these, eight were receiving more than 30 mg/day of prednisona or the equivalent, and/or a cytostatic agent, 17 were receiving less than this dose of corticosteroid and five were receiving no corticosteroids or cytostatic agents. Information about therapy was unavailable for two patients. The histologic specimens were all examined by Histology. a single observer (E.G.). Precautions were taken to be certain that her interpretation was performed blindly. In all but three cases, at least 6 glomeruli were examined by light microscopy. Exclusion of these three cases did not significantly altar the results of the study. Under light microscopy, the glomerular changes were classified, and the degree of activity, as judged by previously defined criteria [ 13,141, was quantitated on an arbitrary scale of 0 to 3. Based on examination of electron micrographs of multiple areas, of 1 or, in 14 cases, 2 glomeruli, electron-dense deposits were classified by location and their amount was semiquantitated on a 0 to 3 scale [ 141. Binding assays were performed by DNA Binding Assays. a modified ammonium sulfate precipitation technic [3,6,15]. C14-labeled KB DNA was obtained from Electronucleonics, Bethesda, Maryland. As judged by susceptibility to a singlestranded-DNA-specific nuclease, similar preparations have been found to be contaminated with between 9 and 37 per cent single-stranded DNA [6]. No data are available on possible contamination with other potential antigens. No attempt was made to further purify the KB DNA preparations prior to use in binding assays. Previous experience has suggested that at least partial elimination of single-stranded DNA contaminants by nuclease pretreatment or, less effectively, by chromatography on methylated albumin kieselguhr resulted in an antigen whose binding was more nearly similar to that of dAT [6]. C14-labelled dAT was purchased from Biopolymers (Chagrin Falls, Ohio) or from Grand Island Patients.
ET AL.
BINDING OF DNA IN SYSTEMIC LUPUS ERYTHEMATOSUS-STEINMAN
TABLE
I
100
Histologic Findings Location of ElectronDense Deposits
Type of Glomerular Lesion by LM ______ Minimal
4
Mesangial Membranous (pure) Focal proliferative Diffuse proliferative Not examined
8 1 4 14 1
Total
32
Mesangiai, and/or subepithelial alone Intramembranous” Subendothelial’ No deposits
r
.
I-
’
oj
r=0.582
-
I
/
16 5 5 4 2
Not examined
32
NOTE: LM = light microscopy. *Specimens with intramembranous or subendothelial deposits also frequently contained mesangial and/or subepithelial deposits but are classified here only under the former two categories.
EDD
Biological Co. (Grand Island, New York). Material obtained from the latter source required sonication prior to use [6]. Serums were obtained within two weeks of biopsy or autopsy except in five cases in which two to five weeks intervened. Exclusion of these specimens did not alter the conclusions reached. RESULTS
6
I are summarized the light and electron microscopic findings. The small number of specimens showing minimal disease by light microscopy is consistent with the clinical indications generally required for performance of biopsies. Similarly, the distribution of electron-dense deposits does not appear exceptional
0
In Figure 1 is shown the correlation between binding of dAT (top) or of KB DNA (bottom), with the amount of electron-dense deposits. The solid lines are those calculated for regression of binding upon electron-dense deposits [ 171. Binding of both dAT and KB DNA correlated significantly with electron-dense deposits although the correlation coefficient, r, was higher for dAT than for KB DNA binding. Use of the nonparametric rank correlation method of Spearman also led to a significant correlation between these variables but with smaller correlation coefficients [ 171.
All Specimens Number of specimens Correlation coefficient t P
31
Subendothelial and Intramembranous 10
TABLE III
0.956
Subendothelial
I 2.0
I 2.5
I 30
Correlation of Binding with Histologic Activity of Nephritis
5
Number of specimens (N j Correlation coefficient (r) t
13
0.977
0.759
P 10.33
I 1.5
dAT
Diffuse Proliferative
(r) 6.35
I I.0
Antigen Usedin Binding Assay
(N) 0.752
I 13.5
The open circles indicate those specimens with electron-dense deposits in intramembranous and/or subendothelial locations (i.e., excluding specimens with no deposition or with deposits in mesangial and/or subepithelial locations). As seen in Table II, this subgroup exhibited somewhat higher correlation coefficients than did the entire group but with the pre-
Correlation of dAT Binding with Electron-Dense Deposits
II
0
EDD Figure 1. Correlation of dA T binding (top) and KS DNA binding (bottom) with amount of electronxfense deposits (EDD). Open circles represent specimens with subendothelial or intramembranous deposits. Solid lines represent those calculated for the regression of binding on electron-dense deposits. Correlation coefficients are indicated by r.
[161.
TABLE
l
0
In Table
10.23
4.20 <0.005
NOTE: dAT oxyadenylate
March 1977
31
KB DNA 29
0.600
0.275
4.18
1.525 >o. 1
= self-complimentary alternating copolymer of deand deoxythymidylate. KB DNA = human DNA.
The American Journal of Medicine
Volume 62
321
BINDING OF DNA IN SYSTEMIC LUPUS ERYTHEMATOSUS-STEINMAN ET AL
viously observed difference between dAT and KB DNA binding being maintained. The five specimens exhibiting subendothelial localization of electron-dense deposits similarly exhibited a higher correlation coefficient than did the entire group although, despite the significant p value, the small number of specimens in this group makes interpretation of this finding somewhat uncertain. The specimens with diffuse proliferative disease on light microscopy did not differ in their correlation from that of the entire group. Binding of dAT (Table III) also correlated significantly with the degree of activity of glomerulitis as judged by light microscopy, although at a lower level than with electron-dense deposits. KB binding did not exhibit such a correlation. COMMENTS
Although several reports have demonstrated a correlation between naturallyderived “native” DNA binding levels and clinical activity of SLE [3,4], no data have appeared on such a correlation with histologic parameters of renal disease. The binding of natural native DNA has been said to correlate with the amount of glomerular immune-complex deposition as identified by immunofluorescence, although no data are given [ 181. Since it has been shown that the usually measured clinical parameters of renal disease are not reliable guides to histology [ 191, absence of biopsy information has been an obstacle to clear interpretation of DNA binding data. The present data demonstrate that dAT binding correlates closely with glomerular electron-dense deposits (which have been shown to contain immunoglobulin and complement [20] and hence are presumed to be immune complexes) as well as with activity of nephritis on light microscopy so that the previously observed correlation with clinically-judged activity of lupus nephritis appears confirmed. That dAT binding correlates more closely with histologic findings than does KB DNA binding is consistent with a large body of data suggesting that antidoublestranded DNA antibodies are of pathogenetic importance in renal lupus [2] and that routinely-prepared, naturally-derived DNA antigens may detect antisinglestranded DNA as well as antidouble-stranded DNA antibodies [5,6]. The possibility that they can also detect antibody to other, non-nucleic acid antigenic contaminants (such as nucleoprotein) also must be considered.
Hence, binding of naturally-derived antigens might not be expected to exhibit as much specificity for either antidouble-stranded DNA or activity of nephritis as would a homogeneous double-stranded DNA antigen. In general the observed correlations were quite close, particularly in view of the semiquantitative nature of the histologic measurements as well as the likelihood of biopsy sampling error. However, because of these limitations, it is unclear whether the few aberrant values that were observed resulted from biologic or methodologic factors. Hence it is difficult to form firm conclusions about the certainty with which dAT binding can predict those histologic variables with which it correlates. Several recent studies [ 14,211 appear to agree, on the basis of limited data, that the histologic finding most useful for predicting prognosis is the amount of electron-dense deposits, particularly if located subendothelially. To the extent that dAT binding accurately reflects the amount of deposition, it too may provide similar information by a noninvasive procedure that can be easily monitored clinically. No clear correlation of dAT binding with location of deposits was apparent, although it remains possible that the previously noted methodologic imprecision and the small number of samples with subendothelial deposits could have obscured such an association. Finally, it must be emphasized that the present results do not yet allow conclusions regarding possible correlations between therapeutically-induced changes in dAT binding with changes in renal morphology for which serially-obtained data would be required. However, a recent report [22] has suggested that, among patients whose biopsy specimens initially showed diffuse proliferative disease with subendothelial deposits, failure of normalization of naturally derived “native” DNA binding during treatment implies a poor prognosis. These data appear to be consistant with results of the present study and suggest that presumed therapeutically-induced changes (or lack thereof) in dAT binding also may have prognostic importance. ACKNOWLEDGMENT
The advice of Dr. Jacob Churg is gratefully acknowledged as is the technical assistance of Ms. Ping Tong and the help of Mrs. Anita Sardo in preparation of the manuscript.
REFERENCES 1. 2.
3.
322
Levine L, Stellar B: Nucleic acid immune systems. Prog Allergy 12: 161, 1968. Agnello V, Koffler D. Kunkel HG: Immune complex systems in the nephritis of systemic lupus erythematosus. Kidney Int 3: 90, 1973. Pincus T, Schur P, Rose J, et al.: Measurement of serum
March 1977
The
American Journal of M8dlCfne
Volume 82
4. 5.
DNA-binding activity in systemic lupus erythematosus. N Engl J Med 281: 701, 1969. Hughes G, Cohen S, Christian C: Anti-DNA activity in systemic lupus erythematosus. Ann Rheum Dis 30: 259, 197 1. Samaha R, Irvin W: Deoxyribonucleic acid (DNA) strandedness: partial characterization of the antigenic regions
BINDING OF DNA IN SYSTEMIC: LUPUS ERYTHEMATOSUS--STEINMAN
6.
7.
8.
9. 10.
11.
12. 13.
binding antibodies in lupus erythematosus serum. J Clin Invest 56: 446, 1975. Steinman CR, Deesomchok U, Spiera H: Detection of antibody using synthetic antigens. Characterization and clinical significance of binding of poly(deoxyadenylatedeoxythymidylate) by serum. J Clin Invest 57: 1330, 1976. Epstein W, Tan M, Esterbrook M: Serum antibody to doublestranded RNA and DNA in patients with idiopathic and secondary uveites. N Engl J Med 285: 1502, 1971. Hahn B, Sharp G, Irvin W, et al.: Immune responses to hydralazine and nuclear antigens in hydralazine induced lupus erythematosus. Ann Intern Med 76: 365, 1972. Rodinis, P. Palefsky H, Beeker M, et al.: Native DNA binding in rheumatoid arthritis. Ann Rheum Dis 33: 357, 1974. Hasselbacher P, LeRoy E: Serum DNA binding activity in healthy subjects and in rheumatic disease. Arthritis Rheum 17: 63, 1974. Schachman H, Adler J, Radding C, et al.: Enzymatic synthesis of deoxyribonucleic acid:Vlll. Synthesis of a polymer of deoxyadenylate and deoxythymidylate. J Biol Chem 235: 3242, 1960. Geiduschek EP: On the factors controlling the reversibility of DNA denaturation. J Molec Biol 4: 467, 1962. Pollak VE, Pirani CL, Schwartz FD: Clinical and experimental. The natural history of the renal manifestations of systemic lupus erythematosus. J Lab Clin Med 63: 537, 1964.
14.
15.
16.
17. 18.
19. 20.
21.
22.
March 1977
ET AL.
Grishman E, Porush JG, Lee SL, Churg J: Renal biopsies in lupus nephritis. Correlation of electron microscopic findings with clinical course. Nephron 10: 25, 1973. Wold RT, Young FE, Tan EM, Fan RS: Deoxyribonucleic acid antibody. A method to detect its primary interaction with deoxyribonucleic acid. Science 161: 806, 1968. Dujovne I, Pollak VE, Pirani CL, Dillard MG: The distribution and character of glomerular deposits in systemic lupus erythematosus. Kidney Int 2: 33, 1972. Snedecor G, Cochran W: Statistical Methods, 6th ed, Ames Iowa, Iowa State University Press, 1967, p 135. Hill GS, Hinglais N, Tron F, Bach JF: Systemic lupus erythematosus. Correlation of renal biopsies with immunologic data at the time of biopsy (abstract). Kidney Int 8: 449, 1975. Dubois EL: Lupus Erythematosus, 2nd ed., Los Angeles, University of Southern California Press, 1974, p 87. Comerford FR, Cohen AS: The nephropathy of systemic lupus erythematosus. Medicine (Baltimore) 46: 425, 1967. Dillard MG, Tillman LR, Sampson CC: Lupus nephritis. Correlations between the clinical course and presence of electrondense deposits. Lab Invest 32: 261, 1975. Adler MK, Baumgarten A, Hecht B, et al.: Prognostic significance of DNA-binding capacity patterns in patients with lupus nephritis. Ann Rheum Dis 34: 444, 1975.
The American Journal of Medklne
Volume 62
323