- Email: [email protected]
Glomerular paramesangial deposits: Association with hypocomplementemia in membranoproliferative glomerulonephritis types I and III
Glomerular Paramesangial Deposits: Association With Hypocomplementemia in Membranoproliferative Glomerulonephritis Types I and III Clark D. West, MD, and A. James McAdams, MD ● Of 22 subjects previously reported with some form of factor H dysfunction, 12 had a glomerulonephritis that appeared to not be of immune complex origin. Factor H dysfunction results in elevated circulating levels of the C3b-dependent C3 convertase, C3b,Bb. Of the 12 cases with glomerulonephritis, the glomerular deposits in the six whose biopsy specimens were studied were predominately subepithelial on the paramesangial portion of the glomerular basement membrane. In a subsequent study, similar deposits were found in patients with membranoproliferative glomerulonephritis (MPGN) type II, also a nephritis that is probably not of immune complex origin. Paramesangial deposits were found in these patients only in biopsy specimens obtained when the C3 level was low, at which time convertase stabilized by nephritic factor would be present in the circulation. This association of paramesangial deposits with circulating convertase was further tested by correlating these deposits with the level of C3 at the time of biopsy in MPGN types I and III. The results in type III MPGN were similar to those in type II; paramesangial deposits were frequently present when the C3 level was low as a result of circulating nephritic factor of the terminal pathway, NFt, and were usually absent when the C3 level was in the upper two thirds of the normal range. Deposits persisted in those patients with C3 levels that had been low but that had increased during the year before biopsy to within the lower one third of the normal range. The persistence of paramesangial deposits in MPGN type III, as compared with MPGN type II, may be related to the differences in composition and function of the two NF stabilized convertases (C3bn,Bb,P,NFt and C3b,Bb,NFa, respectively) that circulate in these two disorders. In contrast to MPGN type III, the hypocomplementemia in MPGN type I is thought to be, for the most part, the result of classical pathway activation, which is not associated with elevated circulating convertase levels. In agreement with this, paramesangial deposits were found in only two of 34 biopsy specimens. At the time of those two biopsies, both patients had a complement profile indicating that the NFt was circulating, as in MPGN type III. In three other cases with profiles compatible with circulating NFt, paramesangial deposits were not found. In all patients with type I MPGN, electron microscopy and immunofluorescence of the glomeruli gave results typical of an immune complex nephritis. Thus, even though the complement profile in MPGN type I may at times indicate the presence of a nephritic factor, circulating immune complexes appear to be basic to pathogenesis. The observations support the hypothesis that elevated levels of the C3b-dependent convertase, as found in the ‘‘experiments of nature’’ with factor H dysfunction and in MPGN types II and III, are associated with paramesangial deposits. The nature of this association and the role of these deposits in producing the nephritis is not clear. r 1998 by the National Kidney Foundation, Inc. INDEX WORDS: Nephritic factors, MPGN; paramesangial glomerular deposits; hypocomplementemia; C3; factor H.
W
E HAVE PREVIOUSLY noted that subjects with dysfunction of factor H of the complement system are predisposed to glomerulonephritis that does not appear to be of immune complex origin.1 Three conditions were cited: homozygous factor H deficiency in humans and piglets, a circulating factor H inhibitor, and heterozygous absence of a factor H binding site on C3b (Marder disease). Among 22 reported patients with these forms of factor H dysfunction, 12 have had evidence of glomerulonephritis, and in each of these, C3 was present in the glomeruli but immunoglobulin G (IgG) was absent or present in only trace amounts. Because factor H serves as a cofactor for factor I and mediates extrinsic decay of the C3b-dependent convertase, C3b,Bb, factor H dysfunction allows this convertase to circulate in excess. As a result of this excess convertase, serum C3 levels are low.
A common feature of the glomerular morphology with factor H dysfunction in humans is subepithelial deposits on the paramesangial portion of the glomerular basement membrane. These were prominent in the renal biopsy specimens from four patients with nephritis with Marder disease2,3 and from the two patients with homozygous factor H deficiency we have had the opportunity to examine. To further test the association between factor H dysfunction and From the Children’s Hospital Research Foundation and the Department of Pediatrics, College of Medicine, University of Cincinnati, Cincinnati, OH. Received February 19, 1997; accepted in revised form August 29, 1997. Address reprint requests to Clark D. West, MD, Children’s Hospital Medical Center, Division of Nephrology, 3333 Burnet Ave, Cincinnati, OH 45229. E-mail: [email protected]
r 1998 by the National Kidney Foundation, Inc. 0272-6386/98/3103-0004$3.00/0
American Journal of Kidney Diseases, Vol 31, No 3 (March), 1998: pp 427-434
427
428
paramesangial deposits, biopsy specimens from patients with membranoproliferative glomerulonephritis (MPGN) type II, another nephritis that probably is not of immune complex origin, were examined in a previous study.4 The nephritic factor of the amplification loop (NFa) found in this disease greatly slows the factor H-mediated decay of the convertase5,6 by binding to the activated factor B, forming in the circulation C3b,Bb,NFa.7 It was reasoned that if factor H dysfunction is associated with paramesangial deposits, then deposits should be present only in biopsy specimens obtained when this nephritic factor is circulating and the C3 level is low. In agreement with this hypothesis, paramesangial deposits were found in 12 of 14 biopsy specimens obtained when the C3 level was low and in none of 11 biopsy specimens obtained when the C3 level was normal (P ⬍ 0.001).4 Subepithelial deposits on the capillary loop and the so-called intramembranous dense deposits, also found in the glomeruli in this disease, did not correlate with the C3 level. Since Marder disease, factor H deficiency, and MPGN type II are widely variable with respect to the cause of the factor H dysfunction, it was postulated that the paramesangial deposits were associated with the elevated levels of circulating convertase, which is present in all such cases. The nature of the association of circulating convertase with nephritis is not clear. It should be noted that other investigators have hypothesized that MPGN type II is a consequence of a defect in immune complex clearance because of the acquired nephritic factor-induced deficiency of C3.8 This would result in enhanced deposition of immune complexes in the glomeruli. However, the absence in this disease of IgG from the glomerular deposits and of signs of classical pathway activation9-11 argue against this hypothesis. The above observations have been extended in the present report to patients with MPGN type III. As in the previous study of patients with MPGN type II, a correlation is made between paramesangial deposits and the C3 level at the time of biopsy. In this type of MPGN, the circulating nephritic factor is known as the nephritic factor of the terminal pathway (NFt). It differs from NFa in that it requires both properdin and prolonged incubation of the serum mixture
WEST AND MCADAMS
for its demonstration and, in addition, it activates the terminal pathway.12-14 For comparison, the results of a similar study with MPGN type I are also reported. As noted previously,15,16 patients with this type are heterogeneous with respect to their complement profiles. Some have evidence only of classical pathway activation, some have circulating NFt with signs of terminal pathway activation, and some have activation of both pathways. As a group, however, their level of C4 is significantly lower than that of patients with MPGN type III,17 indicating classical pathway activation, and IgG and C4 can be seen in the glomeruli by immunofluorescence.17 All have subendothelial and mesangial deposits, as typically seen in an immune complex nephritis such as lupus. Thus, despite evidence that NFt is present is some cases, the basis of the disease appears to be the deposition of circulating immune complexes. Classical pathway activation produced by the immune complexes would not cause the C3-dependent convertase to circulate in excess; hence, according to our hypothesis, paramesangial deposits should not be found. MATERIALS AND METHODS Concurrent serum C3 levels were available for 34 renal biopsy specimens from 19 patients with MPGN type I and for 44 biopsy specimens from 21 patients with MPGN type III, all performed over a 20-year period from 1970 to 1990. The type of MPGN was determined by characteristics of the glomerular ultrastructure as described previously.17-19 The biopsy specimens were obtained with a Menghini needle; they were immediately snap-frozen or fixed and processed for light microscopy, electron microscopy, and immunofluorescence. Tissue for light microscopy was routinely fixed in 10% formalin and stained with hematoxylineosin, periodic acid-Schiff, and silver methenamine as described previously.20 Immunohistology was performed on snap-frozen tissue from all 78 biopsies using antisera to IgG and C3. In addition, antiserum to C4 was used on 13 of the 17 biopsy specimens obtained from patients with type I at a time when they were hypocomplementemic. The antisera were prepared in this laboratory. Tissue for electron microscopy was fixed in 2% osmium tetroxide, embedded in Epon, and stained with uranyl lead or silver by impregnation.19 Paramesangial deposits were most easily identified in silver-impregnated electron micrographs. Because these deposits are typically not numerous, visualizing them in the low-contrast uranyl-lead–stained sections was difficult, especially in the face of the structural distortion and the abundance of other deposits in MPGN type III. Hence, advantage was taken of the silver-impregnated electron micrographs, with their inherent high contrast between structure and deposits. In the detection of these deposits,
PARAMESANGIAL DEPOSITS IN MPGN I AND III
advantage was also taken of the irregularity of involvement in MPGN type III by processing less altered glomeruli. The number of glomeruli searched for paramesangial deposits in the 78 biopsy specimens varied from one (the maximum available in three biopsy specimens) to 13 (mean, 5.1 glomeruli per biopsy specimen). Deposits were sought by electron microscopy without knowledge of the C3 levels. All serum C3 and C4 levels were measured by radial immunodiffusion.
RESULTS
MPGN Type III C3 levels at the time of biopsy as well as all levels available within 80 months before the biopsy are plotted in Fig 1. To avoid confusing overlap of the lines, data for patients with C3 levels at the time of biopsy in the upper two thirds of the normal range, in the lower one third of the normal range, and below the normal range are shown in Fig 1A, Fig 1B, and Fig 1C, respectively. Previously published data4 for MPGN type II, similarly graphed, are shown for comparison in Fig 2. It is apparent that the results in these two types of MPGN are similar in that paramesangial deposits were present in virtually all biopsy specimens obtained when the C3 level
429
was low (Figs 1C and 2). The results differed for biopsy specimens obtained when the C3 level was in the normal range. Although we had found4 that in MPGN type II, paramesangial deposits were absent in all patients when the C3 level was in the normal range (Fig 2), in type III, these deposits tended to be absent only when the level was in the upper two thirds of the normal range (Fig 1A) and to be present at levels in the lower one third of the normal range (Fig 1B). Their presence at low normal levels may reflect the fact that in MPGN type III, these deposits are more persistent than they are in type II. Thus, as shown in Fig 1B, before these biopsies, the C3 level had been low, and paramesangial deposits had presumably formed and had persisted during the year before the biopsy, when the level was slowly increasing into the normal range. With one exception, the biopsies that showed lownormal levels of C3 had been preceded by approximately 2 years of the alternate-day prednisone regimen. Presumably in response to this regimen, the C3 level had slowly increased into the normal range.
Fig 1. MPGN type III. Serum levels of C3 before and at the time of renal biopsy, obtained when the C3 level was (A) in the upper two thirds of the normal range, (B) in the lower one-third of the normal range, and (C) below the normal range. The time of biopsy is indicated by the vertical line. Closed circles at the time of biopsy indicate that paramesangial deposits were present; open circles indicate that paramesangial deposits were absent. (A) Data for 11 biopsy specimens from nine patients, (B) data for seven biopsy specimens from seven patients, and (C) data for 26 biopsy specimens from 16 patients.
430
WEST AND MCADAMS
paramesangial deposits persist for very long periods after the NFt stabilized convertase disappears from the circulation. By immunofluorescence, the paramesangial deposits that were readily detectable in type II4 could not be distinguished in type III because of the abundance of other deposits. Thus, their composition with respect to immunoglobulins and complement components could not be determined. From the total fluorescence pattern, however, there was no evidence that nephritic factor stabilized convertase was present. For this to be present according to our hypothesis, IgG and C3 should be found co-deposited only in biopsy specimens obtained during hypocomplementemia and not when the C3 level was in the upper two thirds of the normal range. Apparent codeposition of IgG and C3 was found in only six of 26 biopsy specimens obtained during hypocomplementemia and in three of 18 obtained when the C3 level was in the upper two thirds of the normal range (P ⫽ NS). All patients in this study, as well as all patients in the previous study of MPGN type II,4 received Fig 2. MPGN type II. Serum levels of C3 before and at the time of renal biopsy. The time of biopsy is indicated by the vertical line. Closed circles at the time of biopsy indicate that paramesangial deposits were present; open circles indicate that paramesangial deposits were absent. The data are for 25 biopsy specimens from 12 patients.4
By electron microscopy, there were minor differences in the appearance of the paramesangial deposits in MPGN types II and III. Paramesangial deposits that were contiguous across the basement membrane with submesangial deposits, the so-called waist basement membrane deposits,4 were less frequent (seven of 33) in type III than in type II (five of 11), but the difference was not significant. In addition, the paramesangial deposits in two biopsy specimens from patients with type III had an appearance suggesting that they had been present for long periods of time. In one, most of the deposits were at least partially covered; in the other, they contained curvilinear striate bodies (Fig 3).21 These were seen in the two biopsy specimens obtained when the C3 level was in the upper two thirds of the normal range (Fig 1A). No deposits had this appearance in patients with type II. It is possible that in some patients with type III,
Fig 3. Silver-impregnated electron micrograph from a patient with MPGN type III showing a typical, welldefined, large, subepithelial, paramesangial deposit (arrow). Note the curved membrane fragments within the deposit, which show at higher resolution in other patients in this series features characterized as striate bodies.13 Most paramesangial subepithelial deposits observed in this study did not show this feature. To the right, the complex deposits characteristic of the type III lesion can be seen.
PARAMESANGIAL DEPOSITS IN MPGN I AND III
431
an alternate-day prednisone regimen, a regimen that, as noted above, is usually accompanied by a slow increase in the serum C3 level. Although the data in Fig 1 can be interpreted as indicating that paramesangial deposits slowly disappear as the C3 level increases into the normal range, the possibility should be considered that their disappearance was instead a direct effect of the prednisone. However, analysis of the data does not support this interpretation. There were 19 biopsy specimens obtained after the patients had been on the alternate-day regimen for ⱖ23 months; of these, paramesangial deposits were still present in 10 (53%). Thus, no direct correlation between the deposits and prolonged use of the alternate-day regimen was demonstrable. MPGN Type I A graph relating paramesangial deposits to the C3 level at and before biopsy in 19 patients with MPGN type I is shown in Fig 4. In contrast to the hypocomplementemic patients with MPGN type III, the C3 level in these patients came into the normal range very quickly after the alternate-day regimen was started.22 Mild hypocomplementemia transiently recurred in some patients. All patients in this group with a low C3 level had either a low serum level of C4 (10 of 17) or C4 in their glomerular deposits. As further evidence of classical pathway activation, all had IgG in their glomerular deposits. The results with respect to frequency of paramesangial deposits differed markedly from those in MPGN types II and III. No paramesangial deposits were present when the C3 level was normal, but when it was low they were present in only two of 17 biopsy specimens (12%) compared with 12 of 14 (86%) in MPGN type II (P ⬍ 0.001) and 25 of 26 (96%) in MPGN type III (P ⬍ 0.001). It is of interest that at the time of both of the MPGN type I biopsies in which paramesangial deposits were present, the complement profile gave evidence that NFt was circulating (patients H1011 and N031915) similar to that found in MPGN type III. However, the overall glomerular morphology in these two biopsies was typical of MPGN type I, and in both the glomerular deposits contained IgG and C4. It should be noted that three other MPGN type I patients with a complement profile compatible with the presence of NFt did not have paramesangial deposits.
Fig 4. MPGN type I. Serum levels of C3 before and at the time of renal biopsy. The time of biopsy is indicated by the vertical line. Closed circles at the time of biopsy indicate that paramesangial deposits were present; open circles indicate that paramesangial deposits were absent. Data are for 34 renal biopsy specimens from 19 patients.
DISCUSSION
In the three types of MPGN, C3 is activated by three modalities: the nephritic factor of the amplification loop (NFa), the nephritic factor of the terminal pathway (NFt),12,13 and circulating immune complexes. The two nephritic factors are very different. NFa in a serum mixture activates C3 relatively rapidly,12,23 does not activate terminal components, and does not require properdin.12,14 NFt, on the other hand, activates C3 slowly, activates terminal components, and requires properdin.12-14 Both activate C3 by making the C3b-dependent C3 convertase, C3b,Bb, resistant to both intrinsic and extrinsic decay, the latter mediated by factor H.5,6 The composition of the NFa stabilized convertase is simply C3b,Bb,NFa,7 whereas that stabilized with NFt is presumably C3bn,Bb,P,NFt. NFa produces a complement profile characterized mainly by a low level of C315,24; a metabolic
432
study gave no evidence that it activates C5 in vivo.25 With NFt, in addition to a depressed C3 level, there is a reduced level of properdin, a markedly depressed level of C5, usually depressed levels of one or more of the other four terminal components,15,16 and elevated levels of terminal complement complexes.13 NFa is frequently found in patients with MPGN type II, as well as in those with partial lipodystrophy.24 NFt appears to be solely responsible for the hypocomplementemia in MPGN type III.12,15,16 Judging from the complement profile, NFt also may be present in some patients with MPGN type I. However, most patients with MPGN type I have a low C4 level, which indicates classical pathway activation, presumably by circulating immune complexes.15,17 It should be noted that the hypocomplementemia of the patients of the present report was always mediated by one or more of the above modalities, even though diminished C3 synthesis may have contributed heavily to the severity of the C3 depression. Diminished C3 synthesis is dependent on a negative feedback induced by circulating C3dg,26 and for C3dg to circulate, complement activation by nephritic factor or immune complexes is necessary. It was shown in a previous report that in patients with MPGN type II, paramesangial deposits in the glomeruli are frequently present when NFa is circulating and are not found when it is absent.4 The present results in patients with MPGN type III can be interpreted as indicating that when NFt is circulating, paramesangial deposits are present but that they remain for a time after NFt disappears. Data for prebiopsy C3 levels in patients who had C3 levels in the low-normal range at the time of biopsy gave evidence that the deposits could persist for up to 1 year after NFt disappears. Because paramesangial deposits are present in patients with homozygous factor H deficiency and with Marder disease,2,3 conditions in which the C3b-dependent C3 convertase is circulating, we hypothesize that the association of paramesangial deposits with nephritic factor in MPGN types II and III is because nephritic factor also allows this convertase to circulate. With paramesangial deposits in some way related to circulating convertase, it would not be surprising that the persistence of the deposits after the stabilized convertase left
WEST AND MCADAMS
the circulation would differ for stabilized convertases as different in composition and complement reactivity as are those in types II and III. If nephritic factors, circulating convertase, paramesangial deposits, and glomerulonephritis interrelate, there should be an association between the severity and/or progression of MPGN and the hypocomplementemia. Possible reasons for the lack of an association were given previously.1 Briefly, these are (1) that the infrequency of nephritis with Marder disease and homozygous factor H deficiency raises the possibility that a combination with one or more other factors is needed before circulating convertase is nephritogenic; (2) that when observing patients, measurements of C3 and of renal function may be too infrequent to detect a correlation of diminishing function with hypocomplementemia; and (3), most importantly, that glomerular injury may lead to processes, such as hyperfiltration, that aggravate the injury at a time when the C3 level is in the normal range and nephritic factor is absent. Assuming paramesangial deposits are part of the nephritogenic process in type III, their tendency to persist after NFt disappears might also be a cause for the nephritis to progress after the hypocomplementemia disappears. As noted above, the origin of the hypocomplementemia in MPGN type I is complex. There is evidence for classical pathway activation and in some patients for the presence of NFt.15 Why this nephritic factor is present in this type is not clear. It appears, however, to have little influence on the manifestations of the disease. Thus, by electron microscopy and immunofluorescence, the disease has the hallmarks of an origin from the deposition of circulating immune complexes. The glomerular deposits are overwhelmingly subendothelial and mesangial, and they contain IgG and C4; in addition, serum C4 levels are often low.15,17 Patients with chronic antigenemia, known to cause immune complexes to circulate, may have a glomerular morphology identical to that of MPGN type I.27,28 The fact that patients with type I often have symptoms of systemic disease, not seen with types II and III, also suggests an immune complex origin.17 In addition, although the complement profile of five patients at the time of biopsy could have been produced by NFt, paramesangial deposits were present in only two, and neither had any other
PARAMESANGIAL DEPOSITS IN MPGN I AND III
hallmarks of MPGN type III. This finding suggests that deposition of circulating immune complex predominates as the nephritogenic agent in MPGN type I, even though nephritic factors may be present in the circulation. In summary, there is strong evidence for a ‘‘convertase’’ nephritis. Nephritis was present in approximately half the ‘‘experiments of nature’’ in which native convertase circulates, it is present in a high proportion of those with convertase stabilized by NFa, and, to our knowledge, it is present in all those with convertase stabilized by NFt. There is circumstantial evidence that paramesangial deposits are in some way associated with convertase circulating in excess; they are found when both native convertase circulates in excess and when nephritic factor-stabilized convertase circulates. Furthermore, the deposits disappear when nephritic factor-stabilized convertase leaves the circulation, albeit after a delay in MPGN type III, and they are rarely found in MPGN type I, a nephritis of immune complex origin. The nature of the paramesangial deposit-convertase association and the role of the deposits in producing the nephritis is not clear. ACKNOWLEDGMENT Jean Snyder prepared the renal biopsy sections for immunohistology and electron microscopy, and Judith Forristal measured the levels of the complement components. Their expertise, essential to the study, is gratefully acknowledged.
REFERENCES 1. West CD: Nephritic factors predispose to chronic glomerulonephritis. Am J Kidney Dis 24:956-963, 1994 2. Marder HK, Coleman TH, Forristal J, Beischel L, West CD: An inherited defect in the C3 convertase, C3b, Bb, associated with glomerulonephritis. Kidney Int 23:749-758, 1983 3. Linshaw MA, Stapleton FB, Cuppage FE, Forristal J, West CD, Schreiber RD, Wilson CB: Hypocomplementemic glomerulonephritis in an infant and mother. Evidence for an abnormal form of C3. Am J Nephrol 7:470-477, 1987 4. West CD, McAdams AJ: Paramesangial glomerular deposits in membranoproliferative glomerulonephritis type II correlate with hypocomplementemia. Am J Kidney Dis 25:853-861, 1995 5. Weiler JM, Daha MR, Austen KF, Fearon DT: Control of the amplification convertase of complement by the plasma protein beta-1H. Proc Natl Acad Sci U S A 73:32683272, 1976 6. Daha MR, Fearon DT, Austen KF: C3 nephritic factor (C3 NeF): Stabilization of fluid phase and cell bound alternative pathway convertase. J Immunol 116:1-7, 1976
433
7. Daha MR, Austen KF, Fearon DT: The incorporation of C3 nephritic factor (C3 NeF) into a stabilized C3 convertase, C3bBb (C3 NeF), and its release after decay of convertase function. J Immunol 119:812-817, 1977 8. Peters DK, Gwyn Williams D: Complement and mesangiocapillary glomerulonephritis. Role of complement deficiency in the pathogenesis of nephritis. Nephron 13:187197, 1974 9. Habib R, Gubler M-C, Loirat C, Ben Maiz H, Levy M: Dense deposit disease: A variant of membranoproliferative glomerulonephritis. Kidney Int 7:204-215, 1975 10. Vargas R, Thomson KJ, Wilson D, Cameron JS, Turner DR, Gill D, Chantler C, Ogg CS: Mesangiocapillary glomerulonephritis with dense ‘‘deposits’’ in the basement membranes of the kidney. Clin Nephrol 5:73-82, 1976 11. Lamb V, Tisher CC, McCoy RC, Robinson RR: Membranoproliferative glomerulonephritis with dense intramembranous alterations. A clinicopathologic study. Lab Invest 36:607-616, 1977 12. Clardy CW, Forristal J, Strife CF, West CD: A properdin dependent nephritic factor slowly activating C3, C5, and C9 in membranoproliferative glomerulonephritis, types I and III. Clin Immunol Immunopathol 50:333-347, 1989 13. Mollnes TE, Ng YC, Peters DK, Lea T, Tschopp J, Harboe M: The effect of nephritic factor on C3 and on the terminal pathway of complement in vivo and in vitro. Clin Exp Immunol 65:73-79, 1986 14. Tanuma Y, Ohi H, Hatano M: Two types of C3 nephritic factor: Properdin-dependent C3 NeF and properdinindependent C3 NeF. Clin Immunol Immunopathol 56:226238, 1990 15. Varade WS, Forristal J, West CD: Patterns of complement activation in idiopathic membranoproliferative glomerulonephritis, types I, II, and III. Am J Kidney Dis 16:196-206, 1990 16. Clardy CW, Forristal J, Strife CF, West CD: Serum terminal complement component levels in hypocomplementemic glomerulonephritides. Clin Immunol Immunopathol 50:307-310, 1989 17. Jackson EC, McAdams AJ, Strife CF, Forristal J, Welch TR, West CD: Differences between membranoproliferative glomerulonephritis types I and III in clinical presentation, glomerular morphology, and complement perturbation. Am J Kidney Dis 9:115-120, 1987 18. West CD, McAdams AJ: The chronic glomerulonephritides of childhood. Part II. J Pediatr 93:167-176, 1978 19. Strife CF, McEnery PT, McAdams AJ, West CD: Membranoproliferative glomerulonephritis with disruption of the glomerular basement membrane. Clin Nephrol 7:6572, 1977 20. McAdams AJ, McEnery PT, Bove KE, West CD: Childhood nephritis, in Rosenberg AH, Bolandi RP (eds): Perspectives in Pediatric Pathology. Chicago, IL, Yearbook Medical, 1973, pp 189-225 21. Bariety J, Callard P: Striated membranous structures in renal glomerular tufts. An electron microscopy study of 340 human renal biopsies. Lab Invest 32:636-641, 1975 22. McEnery PT, McAdams AJ, West CD: The effect of prednisone in a high-dose, alternate-day regimen on the
434
natural history of idiopathic membranoproliferative glomerulonephritis. Medicine (Baltimore) 64:401-424, 1985 23. Vallota EH, Forristal J, Spitzer RE, Davis NC, West CD: Characteristics of a non-complement dependent C3reactive complex formed from factors in nephritic and normal serum. J Exp Med 131:1306-1334, 1970 24. Sissons JGP, West RJ, Fallows J, Williams DG, Boucher BJ, Amos N, Peters DK: The complement abnormalities of lipodystrophy. N Engl J Med 194:461-465, 1976 25. Sissons JGP, Liebowitch D, Amos N, Peters DK: Metabolism of the fifth component of complement and its relation to the metabolism of the third component in patients with complement activation. J Clin Invest 59:704-715, 1977
WEST AND MCADAMS
26. Charlesworth JA, Gwyn William D, Sherington E, Lachmann PJ, Peters DK: Metabolic studies of the third component of complement and the glycine-rich beta glycoprotein in patients with hypocomplementemia. J Clin Invest 53:1578-1587, 1974 27. Michael AF, Herdman RC, Fish AJ, Pickering RJ, Vernier RL: Chronic membranoproliferative glomerulonephritis with hypocomplementemia. Transplant Proc 1:925932, 1969 28. Strife CF, McDonald BM, Ruley EJ, McAdams AJ, West CD: Shunt nephritis: The nature of the serum cryoglobulins and their relation to the complement profile. J Pediatr 88:403-413, 1976
W
E HAVE PREVIOUSLY noted that subjects with dysfunction of factor H of the complement system are predisposed to glomerulonephritis that does not appear to be of immune complex origin.1 Three conditions were cited: homozygous factor H deficiency in humans and piglets, a circulating factor H inhibitor, and heterozygous absence of a factor H binding site on C3b (Marder disease). Among 22 reported patients with these forms of factor H dysfunction, 12 have had evidence of glomerulonephritis, and in each of these, C3 was present in the glomeruli but immunoglobulin G (IgG) was absent or present in only trace amounts. Because factor H serves as a cofactor for factor I and mediates extrinsic decay of the C3b-dependent convertase, C3b,Bb, factor H dysfunction allows this convertase to circulate in excess. As a result of this excess convertase, serum C3 levels are low.
A common feature of the glomerular morphology with factor H dysfunction in humans is subepithelial deposits on the paramesangial portion of the glomerular basement membrane. These were prominent in the renal biopsy specimens from four patients with nephritis with Marder disease2,3 and from the two patients with homozygous factor H deficiency we have had the opportunity to examine. To further test the association between factor H dysfunction and From the Children’s Hospital Research Foundation and the Department of Pediatrics, College of Medicine, University of Cincinnati, Cincinnati, OH. Received February 19, 1997; accepted in revised form August 29, 1997. Address reprint requests to Clark D. West, MD, Children’s Hospital Medical Center, Division of Nephrology, 3333 Burnet Ave, Cincinnati, OH 45229. E-mail: [email protected]
r 1998 by the National Kidney Foundation, Inc. 0272-6386/98/3103-0004$3.00/0
American Journal of Kidney Diseases, Vol 31, No 3 (March), 1998: pp 427-434
427
428
paramesangial deposits, biopsy specimens from patients with membranoproliferative glomerulonephritis (MPGN) type II, another nephritis that probably is not of immune complex origin, were examined in a previous study.4 The nephritic factor of the amplification loop (NFa) found in this disease greatly slows the factor H-mediated decay of the convertase5,6 by binding to the activated factor B, forming in the circulation C3b,Bb,NFa.7 It was reasoned that if factor H dysfunction is associated with paramesangial deposits, then deposits should be present only in biopsy specimens obtained when this nephritic factor is circulating and the C3 level is low. In agreement with this hypothesis, paramesangial deposits were found in 12 of 14 biopsy specimens obtained when the C3 level was low and in none of 11 biopsy specimens obtained when the C3 level was normal (P ⬍ 0.001).4 Subepithelial deposits on the capillary loop and the so-called intramembranous dense deposits, also found in the glomeruli in this disease, did not correlate with the C3 level. Since Marder disease, factor H deficiency, and MPGN type II are widely variable with respect to the cause of the factor H dysfunction, it was postulated that the paramesangial deposits were associated with the elevated levels of circulating convertase, which is present in all such cases. The nature of the association of circulating convertase with nephritis is not clear. It should be noted that other investigators have hypothesized that MPGN type II is a consequence of a defect in immune complex clearance because of the acquired nephritic factor-induced deficiency of C3.8 This would result in enhanced deposition of immune complexes in the glomeruli. However, the absence in this disease of IgG from the glomerular deposits and of signs of classical pathway activation9-11 argue against this hypothesis. The above observations have been extended in the present report to patients with MPGN type III. As in the previous study of patients with MPGN type II, a correlation is made between paramesangial deposits and the C3 level at the time of biopsy. In this type of MPGN, the circulating nephritic factor is known as the nephritic factor of the terminal pathway (NFt). It differs from NFa in that it requires both properdin and prolonged incubation of the serum mixture
WEST AND MCADAMS
for its demonstration and, in addition, it activates the terminal pathway.12-14 For comparison, the results of a similar study with MPGN type I are also reported. As noted previously,15,16 patients with this type are heterogeneous with respect to their complement profiles. Some have evidence only of classical pathway activation, some have circulating NFt with signs of terminal pathway activation, and some have activation of both pathways. As a group, however, their level of C4 is significantly lower than that of patients with MPGN type III,17 indicating classical pathway activation, and IgG and C4 can be seen in the glomeruli by immunofluorescence.17 All have subendothelial and mesangial deposits, as typically seen in an immune complex nephritis such as lupus. Thus, despite evidence that NFt is present is some cases, the basis of the disease appears to be the deposition of circulating immune complexes. Classical pathway activation produced by the immune complexes would not cause the C3-dependent convertase to circulate in excess; hence, according to our hypothesis, paramesangial deposits should not be found. MATERIALS AND METHODS Concurrent serum C3 levels were available for 34 renal biopsy specimens from 19 patients with MPGN type I and for 44 biopsy specimens from 21 patients with MPGN type III, all performed over a 20-year period from 1970 to 1990. The type of MPGN was determined by characteristics of the glomerular ultrastructure as described previously.17-19 The biopsy specimens were obtained with a Menghini needle; they were immediately snap-frozen or fixed and processed for light microscopy, electron microscopy, and immunofluorescence. Tissue for light microscopy was routinely fixed in 10% formalin and stained with hematoxylineosin, periodic acid-Schiff, and silver methenamine as described previously.20 Immunohistology was performed on snap-frozen tissue from all 78 biopsies using antisera to IgG and C3. In addition, antiserum to C4 was used on 13 of the 17 biopsy specimens obtained from patients with type I at a time when they were hypocomplementemic. The antisera were prepared in this laboratory. Tissue for electron microscopy was fixed in 2% osmium tetroxide, embedded in Epon, and stained with uranyl lead or silver by impregnation.19 Paramesangial deposits were most easily identified in silver-impregnated electron micrographs. Because these deposits are typically not numerous, visualizing them in the low-contrast uranyl-lead–stained sections was difficult, especially in the face of the structural distortion and the abundance of other deposits in MPGN type III. Hence, advantage was taken of the silver-impregnated electron micrographs, with their inherent high contrast between structure and deposits. In the detection of these deposits,
PARAMESANGIAL DEPOSITS IN MPGN I AND III
advantage was also taken of the irregularity of involvement in MPGN type III by processing less altered glomeruli. The number of glomeruli searched for paramesangial deposits in the 78 biopsy specimens varied from one (the maximum available in three biopsy specimens) to 13 (mean, 5.1 glomeruli per biopsy specimen). Deposits were sought by electron microscopy without knowledge of the C3 levels. All serum C3 and C4 levels were measured by radial immunodiffusion.
RESULTS
MPGN Type III C3 levels at the time of biopsy as well as all levels available within 80 months before the biopsy are plotted in Fig 1. To avoid confusing overlap of the lines, data for patients with C3 levels at the time of biopsy in the upper two thirds of the normal range, in the lower one third of the normal range, and below the normal range are shown in Fig 1A, Fig 1B, and Fig 1C, respectively. Previously published data4 for MPGN type II, similarly graphed, are shown for comparison in Fig 2. It is apparent that the results in these two types of MPGN are similar in that paramesangial deposits were present in virtually all biopsy specimens obtained when the C3 level
429
was low (Figs 1C and 2). The results differed for biopsy specimens obtained when the C3 level was in the normal range. Although we had found4 that in MPGN type II, paramesangial deposits were absent in all patients when the C3 level was in the normal range (Fig 2), in type III, these deposits tended to be absent only when the level was in the upper two thirds of the normal range (Fig 1A) and to be present at levels in the lower one third of the normal range (Fig 1B). Their presence at low normal levels may reflect the fact that in MPGN type III, these deposits are more persistent than they are in type II. Thus, as shown in Fig 1B, before these biopsies, the C3 level had been low, and paramesangial deposits had presumably formed and had persisted during the year before the biopsy, when the level was slowly increasing into the normal range. With one exception, the biopsies that showed lownormal levels of C3 had been preceded by approximately 2 years of the alternate-day prednisone regimen. Presumably in response to this regimen, the C3 level had slowly increased into the normal range.
Fig 1. MPGN type III. Serum levels of C3 before and at the time of renal biopsy, obtained when the C3 level was (A) in the upper two thirds of the normal range, (B) in the lower one-third of the normal range, and (C) below the normal range. The time of biopsy is indicated by the vertical line. Closed circles at the time of biopsy indicate that paramesangial deposits were present; open circles indicate that paramesangial deposits were absent. (A) Data for 11 biopsy specimens from nine patients, (B) data for seven biopsy specimens from seven patients, and (C) data for 26 biopsy specimens from 16 patients.
430
WEST AND MCADAMS
paramesangial deposits persist for very long periods after the NFt stabilized convertase disappears from the circulation. By immunofluorescence, the paramesangial deposits that were readily detectable in type II4 could not be distinguished in type III because of the abundance of other deposits. Thus, their composition with respect to immunoglobulins and complement components could not be determined. From the total fluorescence pattern, however, there was no evidence that nephritic factor stabilized convertase was present. For this to be present according to our hypothesis, IgG and C3 should be found co-deposited only in biopsy specimens obtained during hypocomplementemia and not when the C3 level was in the upper two thirds of the normal range. Apparent codeposition of IgG and C3 was found in only six of 26 biopsy specimens obtained during hypocomplementemia and in three of 18 obtained when the C3 level was in the upper two thirds of the normal range (P ⫽ NS). All patients in this study, as well as all patients in the previous study of MPGN type II,4 received Fig 2. MPGN type II. Serum levels of C3 before and at the time of renal biopsy. The time of biopsy is indicated by the vertical line. Closed circles at the time of biopsy indicate that paramesangial deposits were present; open circles indicate that paramesangial deposits were absent. The data are for 25 biopsy specimens from 12 patients.4
By electron microscopy, there were minor differences in the appearance of the paramesangial deposits in MPGN types II and III. Paramesangial deposits that were contiguous across the basement membrane with submesangial deposits, the so-called waist basement membrane deposits,4 were less frequent (seven of 33) in type III than in type II (five of 11), but the difference was not significant. In addition, the paramesangial deposits in two biopsy specimens from patients with type III had an appearance suggesting that they had been present for long periods of time. In one, most of the deposits were at least partially covered; in the other, they contained curvilinear striate bodies (Fig 3).21 These were seen in the two biopsy specimens obtained when the C3 level was in the upper two thirds of the normal range (Fig 1A). No deposits had this appearance in patients with type II. It is possible that in some patients with type III,
Fig 3. Silver-impregnated electron micrograph from a patient with MPGN type III showing a typical, welldefined, large, subepithelial, paramesangial deposit (arrow). Note the curved membrane fragments within the deposit, which show at higher resolution in other patients in this series features characterized as striate bodies.13 Most paramesangial subepithelial deposits observed in this study did not show this feature. To the right, the complex deposits characteristic of the type III lesion can be seen.
PARAMESANGIAL DEPOSITS IN MPGN I AND III
431
an alternate-day prednisone regimen, a regimen that, as noted above, is usually accompanied by a slow increase in the serum C3 level. Although the data in Fig 1 can be interpreted as indicating that paramesangial deposits slowly disappear as the C3 level increases into the normal range, the possibility should be considered that their disappearance was instead a direct effect of the prednisone. However, analysis of the data does not support this interpretation. There were 19 biopsy specimens obtained after the patients had been on the alternate-day regimen for ⱖ23 months; of these, paramesangial deposits were still present in 10 (53%). Thus, no direct correlation between the deposits and prolonged use of the alternate-day regimen was demonstrable. MPGN Type I A graph relating paramesangial deposits to the C3 level at and before biopsy in 19 patients with MPGN type I is shown in Fig 4. In contrast to the hypocomplementemic patients with MPGN type III, the C3 level in these patients came into the normal range very quickly after the alternate-day regimen was started.22 Mild hypocomplementemia transiently recurred in some patients. All patients in this group with a low C3 level had either a low serum level of C4 (10 of 17) or C4 in their glomerular deposits. As further evidence of classical pathway activation, all had IgG in their glomerular deposits. The results with respect to frequency of paramesangial deposits differed markedly from those in MPGN types II and III. No paramesangial deposits were present when the C3 level was normal, but when it was low they were present in only two of 17 biopsy specimens (12%) compared with 12 of 14 (86%) in MPGN type II (P ⬍ 0.001) and 25 of 26 (96%) in MPGN type III (P ⬍ 0.001). It is of interest that at the time of both of the MPGN type I biopsies in which paramesangial deposits were present, the complement profile gave evidence that NFt was circulating (patients H1011 and N031915) similar to that found in MPGN type III. However, the overall glomerular morphology in these two biopsies was typical of MPGN type I, and in both the glomerular deposits contained IgG and C4. It should be noted that three other MPGN type I patients with a complement profile compatible with the presence of NFt did not have paramesangial deposits.
Fig 4. MPGN type I. Serum levels of C3 before and at the time of renal biopsy. The time of biopsy is indicated by the vertical line. Closed circles at the time of biopsy indicate that paramesangial deposits were present; open circles indicate that paramesangial deposits were absent. Data are for 34 renal biopsy specimens from 19 patients.
DISCUSSION
In the three types of MPGN, C3 is activated by three modalities: the nephritic factor of the amplification loop (NFa), the nephritic factor of the terminal pathway (NFt),12,13 and circulating immune complexes. The two nephritic factors are very different. NFa in a serum mixture activates C3 relatively rapidly,12,23 does not activate terminal components, and does not require properdin.12,14 NFt, on the other hand, activates C3 slowly, activates terminal components, and requires properdin.12-14 Both activate C3 by making the C3b-dependent C3 convertase, C3b,Bb, resistant to both intrinsic and extrinsic decay, the latter mediated by factor H.5,6 The composition of the NFa stabilized convertase is simply C3b,Bb,NFa,7 whereas that stabilized with NFt is presumably C3bn,Bb,P,NFt. NFa produces a complement profile characterized mainly by a low level of C315,24; a metabolic
432
study gave no evidence that it activates C5 in vivo.25 With NFt, in addition to a depressed C3 level, there is a reduced level of properdin, a markedly depressed level of C5, usually depressed levels of one or more of the other four terminal components,15,16 and elevated levels of terminal complement complexes.13 NFa is frequently found in patients with MPGN type II, as well as in those with partial lipodystrophy.24 NFt appears to be solely responsible for the hypocomplementemia in MPGN type III.12,15,16 Judging from the complement profile, NFt also may be present in some patients with MPGN type I. However, most patients with MPGN type I have a low C4 level, which indicates classical pathway activation, presumably by circulating immune complexes.15,17 It should be noted that the hypocomplementemia of the patients of the present report was always mediated by one or more of the above modalities, even though diminished C3 synthesis may have contributed heavily to the severity of the C3 depression. Diminished C3 synthesis is dependent on a negative feedback induced by circulating C3dg,26 and for C3dg to circulate, complement activation by nephritic factor or immune complexes is necessary. It was shown in a previous report that in patients with MPGN type II, paramesangial deposits in the glomeruli are frequently present when NFa is circulating and are not found when it is absent.4 The present results in patients with MPGN type III can be interpreted as indicating that when NFt is circulating, paramesangial deposits are present but that they remain for a time after NFt disappears. Data for prebiopsy C3 levels in patients who had C3 levels in the low-normal range at the time of biopsy gave evidence that the deposits could persist for up to 1 year after NFt disappears. Because paramesangial deposits are present in patients with homozygous factor H deficiency and with Marder disease,2,3 conditions in which the C3b-dependent C3 convertase is circulating, we hypothesize that the association of paramesangial deposits with nephritic factor in MPGN types II and III is because nephritic factor also allows this convertase to circulate. With paramesangial deposits in some way related to circulating convertase, it would not be surprising that the persistence of the deposits after the stabilized convertase left
WEST AND MCADAMS
the circulation would differ for stabilized convertases as different in composition and complement reactivity as are those in types II and III. If nephritic factors, circulating convertase, paramesangial deposits, and glomerulonephritis interrelate, there should be an association between the severity and/or progression of MPGN and the hypocomplementemia. Possible reasons for the lack of an association were given previously.1 Briefly, these are (1) that the infrequency of nephritis with Marder disease and homozygous factor H deficiency raises the possibility that a combination with one or more other factors is needed before circulating convertase is nephritogenic; (2) that when observing patients, measurements of C3 and of renal function may be too infrequent to detect a correlation of diminishing function with hypocomplementemia; and (3), most importantly, that glomerular injury may lead to processes, such as hyperfiltration, that aggravate the injury at a time when the C3 level is in the normal range and nephritic factor is absent. Assuming paramesangial deposits are part of the nephritogenic process in type III, their tendency to persist after NFt disappears might also be a cause for the nephritis to progress after the hypocomplementemia disappears. As noted above, the origin of the hypocomplementemia in MPGN type I is complex. There is evidence for classical pathway activation and in some patients for the presence of NFt.15 Why this nephritic factor is present in this type is not clear. It appears, however, to have little influence on the manifestations of the disease. Thus, by electron microscopy and immunofluorescence, the disease has the hallmarks of an origin from the deposition of circulating immune complexes. The glomerular deposits are overwhelmingly subendothelial and mesangial, and they contain IgG and C4; in addition, serum C4 levels are often low.15,17 Patients with chronic antigenemia, known to cause immune complexes to circulate, may have a glomerular morphology identical to that of MPGN type I.27,28 The fact that patients with type I often have symptoms of systemic disease, not seen with types II and III, also suggests an immune complex origin.17 In addition, although the complement profile of five patients at the time of biopsy could have been produced by NFt, paramesangial deposits were present in only two, and neither had any other
PARAMESANGIAL DEPOSITS IN MPGN I AND III
hallmarks of MPGN type III. This finding suggests that deposition of circulating immune complex predominates as the nephritogenic agent in MPGN type I, even though nephritic factors may be present in the circulation. In summary, there is strong evidence for a ‘‘convertase’’ nephritis. Nephritis was present in approximately half the ‘‘experiments of nature’’ in which native convertase circulates, it is present in a high proportion of those with convertase stabilized by NFa, and, to our knowledge, it is present in all those with convertase stabilized by NFt. There is circumstantial evidence that paramesangial deposits are in some way associated with convertase circulating in excess; they are found when both native convertase circulates in excess and when nephritic factor-stabilized convertase circulates. Furthermore, the deposits disappear when nephritic factor-stabilized convertase leaves the circulation, albeit after a delay in MPGN type III, and they are rarely found in MPGN type I, a nephritis of immune complex origin. The nature of the paramesangial deposit-convertase association and the role of the deposits in producing the nephritis is not clear. ACKNOWLEDGMENT Jean Snyder prepared the renal biopsy sections for immunohistology and electron microscopy, and Judith Forristal measured the levels of the complement components. Their expertise, essential to the study, is gratefully acknowledged.
REFERENCES 1. West CD: Nephritic factors predispose to chronic glomerulonephritis. Am J Kidney Dis 24:956-963, 1994 2. Marder HK, Coleman TH, Forristal J, Beischel L, West CD: An inherited defect in the C3 convertase, C3b, Bb, associated with glomerulonephritis. Kidney Int 23:749-758, 1983 3. Linshaw MA, Stapleton FB, Cuppage FE, Forristal J, West CD, Schreiber RD, Wilson CB: Hypocomplementemic glomerulonephritis in an infant and mother. Evidence for an abnormal form of C3. Am J Nephrol 7:470-477, 1987 4. West CD, McAdams AJ: Paramesangial glomerular deposits in membranoproliferative glomerulonephritis type II correlate with hypocomplementemia. Am J Kidney Dis 25:853-861, 1995 5. Weiler JM, Daha MR, Austen KF, Fearon DT: Control of the amplification convertase of complement by the plasma protein beta-1H. Proc Natl Acad Sci U S A 73:32683272, 1976 6. Daha MR, Fearon DT, Austen KF: C3 nephritic factor (C3 NeF): Stabilization of fluid phase and cell bound alternative pathway convertase. J Immunol 116:1-7, 1976
433
7. Daha MR, Austen KF, Fearon DT: The incorporation of C3 nephritic factor (C3 NeF) into a stabilized C3 convertase, C3bBb (C3 NeF), and its release after decay of convertase function. J Immunol 119:812-817, 1977 8. Peters DK, Gwyn Williams D: Complement and mesangiocapillary glomerulonephritis. Role of complement deficiency in the pathogenesis of nephritis. Nephron 13:187197, 1974 9. Habib R, Gubler M-C, Loirat C, Ben Maiz H, Levy M: Dense deposit disease: A variant of membranoproliferative glomerulonephritis. Kidney Int 7:204-215, 1975 10. Vargas R, Thomson KJ, Wilson D, Cameron JS, Turner DR, Gill D, Chantler C, Ogg CS: Mesangiocapillary glomerulonephritis with dense ‘‘deposits’’ in the basement membranes of the kidney. Clin Nephrol 5:73-82, 1976 11. Lamb V, Tisher CC, McCoy RC, Robinson RR: Membranoproliferative glomerulonephritis with dense intramembranous alterations. A clinicopathologic study. Lab Invest 36:607-616, 1977 12. Clardy CW, Forristal J, Strife CF, West CD: A properdin dependent nephritic factor slowly activating C3, C5, and C9 in membranoproliferative glomerulonephritis, types I and III. Clin Immunol Immunopathol 50:333-347, 1989 13. Mollnes TE, Ng YC, Peters DK, Lea T, Tschopp J, Harboe M: The effect of nephritic factor on C3 and on the terminal pathway of complement in vivo and in vitro. Clin Exp Immunol 65:73-79, 1986 14. Tanuma Y, Ohi H, Hatano M: Two types of C3 nephritic factor: Properdin-dependent C3 NeF and properdinindependent C3 NeF. Clin Immunol Immunopathol 56:226238, 1990 15. Varade WS, Forristal J, West CD: Patterns of complement activation in idiopathic membranoproliferative glomerulonephritis, types I, II, and III. Am J Kidney Dis 16:196-206, 1990 16. Clardy CW, Forristal J, Strife CF, West CD: Serum terminal complement component levels in hypocomplementemic glomerulonephritides. Clin Immunol Immunopathol 50:307-310, 1989 17. Jackson EC, McAdams AJ, Strife CF, Forristal J, Welch TR, West CD: Differences between membranoproliferative glomerulonephritis types I and III in clinical presentation, glomerular morphology, and complement perturbation. Am J Kidney Dis 9:115-120, 1987 18. West CD, McAdams AJ: The chronic glomerulonephritides of childhood. Part II. J Pediatr 93:167-176, 1978 19. Strife CF, McEnery PT, McAdams AJ, West CD: Membranoproliferative glomerulonephritis with disruption of the glomerular basement membrane. Clin Nephrol 7:6572, 1977 20. McAdams AJ, McEnery PT, Bove KE, West CD: Childhood nephritis, in Rosenberg AH, Bolandi RP (eds): Perspectives in Pediatric Pathology. Chicago, IL, Yearbook Medical, 1973, pp 189-225 21. Bariety J, Callard P: Striated membranous structures in renal glomerular tufts. An electron microscopy study of 340 human renal biopsies. Lab Invest 32:636-641, 1975 22. McEnery PT, McAdams AJ, West CD: The effect of prednisone in a high-dose, alternate-day regimen on the
434
natural history of idiopathic membranoproliferative glomerulonephritis. Medicine (Baltimore) 64:401-424, 1985 23. Vallota EH, Forristal J, Spitzer RE, Davis NC, West CD: Characteristics of a non-complement dependent C3reactive complex formed from factors in nephritic and normal serum. J Exp Med 131:1306-1334, 1970 24. Sissons JGP, West RJ, Fallows J, Williams DG, Boucher BJ, Amos N, Peters DK: The complement abnormalities of lipodystrophy. N Engl J Med 194:461-465, 1976 25. Sissons JGP, Liebowitch D, Amos N, Peters DK: Metabolism of the fifth component of complement and its relation to the metabolism of the third component in patients with complement activation. J Clin Invest 59:704-715, 1977
WEST AND MCADAMS
26. Charlesworth JA, Gwyn William D, Sherington E, Lachmann PJ, Peters DK: Metabolic studies of the third component of complement and the glycine-rich beta glycoprotein in patients with hypocomplementemia. J Clin Invest 53:1578-1587, 1974 27. Michael AF, Herdman RC, Fish AJ, Pickering RJ, Vernier RL: Chronic membranoproliferative glomerulonephritis with hypocomplementemia. Transplant Proc 1:925932, 1969 28. Strife CF, McDonald BM, Ruley EJ, McAdams AJ, West CD: Shunt nephritis: The nature of the serum cryoglobulins and their relation to the complement profile. J Pediatr 88:403-413, 1976