- Email: [email protected]
Patterns of Complement Activation in Idiopathic Membranoproliferative Glomerulonephritis, Types I, II, and III
Patterns of Complement Activation in Idiopathic Membranoproliferative Glomerulonephritis, Types I, II, and III William S. Varade, MD, Judith Forristal, and Clark D. West, MD • Complement profiles on 22 hypocomplementemic patients with membranoproliferative glomerulonephritis (MPGN) type I, on 11 with MPGN II, and on 16 with MPGN III, gave evidence that the nephritic factor of the amplification loop (NFa) is responsible for the hypocomplementemia in MPGN II and the nephritic factor of the terminal pathway (NFl) for the hypocomplementemia in MPGN III. In contrast, in MPGN I, there was evidence for three complement-activating modalities, NFa, NFl' and immune complexes. As a result, four different patterns of complement activation were seen. NFa, found in MPGN II, produces a complement profile characterized mainly by C3 depression. In addition, four of seven (57%) severely hypocomplementemic MPGN II patients (C3 < 30 mg/dL) had slightly depressed levels of factor B, and one of seven (14%) of properdin, but in all the C5 concentration was normal. In contrast, all eight severely hypocomplementemic patients with MPGN III had depressed C5 and properdin levels, and six of eight (75%) depressed levels of C6, C7, and/or C9. Of eight MPGN III patients with moderate hypocomplementemia, 50% had depressed C5 and properdin levels and the remainder, depressed C3 only. This spectrum of profiles is most likely produced by varying concentrations of NFl' In MPGN I, nine of 23 (39%) had a profile indicating only classical pathway activation; seven of 23 (39%), a pattern compatible with NFl alone; four of 23 (9%), evidence for both classical pathway activation and NFl; and three of 23 (13%), a pattern compatible with NFa. The unique multifactorial origin of the hypocomplementemia in MPGN I, often giving evidence of classical pathway activation, together with previously reported differences in glomerular morphology and clinical features at onset, makes it distinct from MPGN III. Depressed C8 levels were found to some extent in all hypocomplementemic states. The levels were uncommonly depressed in patients with NFa, most markedly depressed with NFl' and moderately reduced with classical pathway activation. The cause is not known. Diagnostically, profiles showing classical pathway activation and low levels of C6, C7, and/or C9 are specific for MPGN I. Those showing only classical activation are likewise diagnostic of MPGN I if systemic lupus erythematosus (SLE) and chronic bacteremia are ruled out. © 1990 by the National Kidney Foundation, Inc. INDEX WORDS: Membranoproliferative glomerulonephritis; complement; immune complexes; nephritic factors.
F
OLLOWING ITS original description, membranoproliferative glomerulonephritis (MPGN) was divided into three types based on glomerular ultrastructure. Type II, or dense-deposit disease, has a glomerular ultrastructure differing markedly from the other two types and is considered a separate disease. Types I and III are not as easily distinguished. They differ ultrastructurally in that the deposits in type I are subendothelial and the basement membrane has a normal appearance, whereas the deposits in type III are on both sides of the basement membrane and, when impregnated with silver, the membrane appears disrupted and fragmented. 12 By uranyl-lead staining, dense deFrom the Children's Hospital Research Foundation. and the Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, OH. Address reprint requests to Clark D. West, MD, Children's Hmpital Medical Center, Eiland and Bethesda Ave, Cincinnati, OH 45229. © 1990 by the National Kidney Foundation, Inc. 0272-6386/90/1603-0003$3.00/0 196
posits are seen within the membrane in a patchy distribution. Additional differences between MPGN I and III have been noted in earlier studies from this laboratory.3 First, judging from serum C4 levels and from the composition of glomerular deposits with respect to IgG and C4, classical pathway activation is occurring only in MPGN I. Second, MPGN III tends to be clinically silent, often detected by chance discovery of proteinuria and hematuria in the healthy subject as compared with onset with signs of systemic disease in patients with MPGN I. Finally, mesangial cell proliferation is significantly less in MPGN III than in MPGN I. The present study is concerned with details of the complement profile in all three types of MPGN. Whereas the renal glomerulus in this disease becomes an alternative pathway activator, 4 there is no evidence that complement activation in situ in the glomerulus materially affects the serum levels of components of the system. Instead, the hypocomplementemia can be ascribed to three cir-
American Journal of Kidney Diseases, Vol XVI, No 3 (September). 1990: pp 196-206
197
COMPLEMENT PROFILES IN MPGN I, II, AND III
culating complement-reactive modalities. Two of these are nephritic factors. One is the nephritic factor of the amplification loop (NFa), and the other the nephritic factor of the terminal pathway (NF t).56 In addition, classical pathway activation is occurring in MPGN I, presumably produced by circulating immune complexes. These three modalities acting singly or in combination produce several different complement profiles. MATERIALS AND METHODS
values were determined in serum samples from 163 normal hospital personnel. 8
Immunofluorescence IgG, Clq, C4, and C5 were detected in glomerular deposits by fluorescein-labeled monospecific antibody prepared in this laboratory. The data reported are those for the first biopsy obtained at this institution. The biopsies were from both normocomplementemic and hypocomplementemic patients who had glomeruli that showed diffuse proliferation; those with focal proliferation were excluded. 9 In all cases, data for immunofluorescence were recorded on the day of the biopsy before the type of MPGN had been determined.
Patients
Statistical Methods
Between 1957 and 1988, the diagnosis of MPGN has been made on 77 patients as the result of a renal biopsy performed at this institution. The type of MPGN was determined by characteristics of the ultrastructure, as described previously, 12 or, for some cases of MPGN II, by light microscopy of sections stained by the Jones methenamine silver method. Of the 77 patients, 35 had MPGN I; 17, MPGN II; and 25, MPGN III.
The means of complement component levels were compared using Student's t test. Immunofluorescence data for MPGN I and III were compared by the chi-square test.
Complement Profiles Blood for complement profiles was obtained by venipuncture, allowed to clot at room temperature, and the serum either used immediately for measurement of complement components or stored at - 70°C. It has been shown that prolonged storage of serum at - 70 °C has no effect on the measurement of serum complement component levels. 7 Two or more complement profiles were obtained on all patients except one. The profiles selected for the study were those with the lowest C3 level for which complete data were available. Serum complement protein levels were determined by radial immunodiffusion using monospecific goat antisera. 7 Reference Table 1.
RESULTS
Among the 77 patients with MPGN, 18 were found to be consistently normocomplementemic and 59 were hypocomplementemic as defined by a serum C3 level below the lower limit of the normal range (86 mg/dL). Of the 59 hypocomplementemic patients, data for 49 are reported. The frequency of normocomplementemic patients and the reasons for exclusion of data are given for each type of MPGN in Table 1. Of the 10 hypocomplementemic patients excluded from the study, the data for seven were incomplete and in three the complement profile was, or could have been, al-
Reasons for Exclusion of Patient Data From Analysis of Complement Profiles No. of Patients
MPGN I (n 11
=
Reasons for Exclusion
35) Normocomplementemic Severely hypocomplementemic* with homozygous C6 and C7 deficiency Moderately hypocomplementemic* with factor I deficiency
MPGN II (n 3 3
17)
MPGN III (n 4
25)
3
Normocomplementemic Moderately hypocomplementemic; terminal component levels missing Normocomplementemic Severely hypocomplementemic; terminal component levels missing Severely hypocomplementemic with Marder's syndrome 1213 Moderately hypocomplementemic; terminal component levels missing
*Severe hypocomplementemia, C3 <30 mg/dL; moderate hypocomplementemia, C3 30 to 86 mg/dL; lower limit of normal range, 86 mg/dL.
198
tered by other abnormalities. One had a persistent, severe factor I deficiency (serum level, 0 to 14 % of normal mean), which would have contributed to the hypocomplementemia by affecting the turnover of the amplification 100p.1O Data were not available to determine whether the deficiency was inherited. Another had homozygous deficiency of C6 and C7, which would compromise terminal pathway activation, and a third had Marder's syndrome. 11 . 12 The timing of the acquisition of the serum specimens used in the study varied. In 33 of the 49 patients, the serum used was that obtained at the time of, or shortly after, presentation at this institution. In six , the serum was obtained within 2 years of presentation, in six, from 2 to 5 years of presentation , and in four, after 5 years. Thirteen of the 16 profiles not performed on initial specimens were from patients seen before 1969, most of whom either were not receiving prednisone or were receiving it daily for short periods. In such patients, the hypocomplementemia may persist for long periods or wax and wane.13 After 1969, most patients were receiving an alternate-day prednisone regimen . Most of these rapidly and permanently lost their hypocomplementemia. '4 Inherited complement component deficiencies were commonly observed among the patients with MPGN I and III who were included in the study. 8 Among those with MPGN I, patient P0210 had heterozygous Clq deficiency; patient D16I9 had heterozygous C6 deficiency; patient HI207, heterozygous deficiency of determinants of the alphagamma chain of C8 1s ; and patient Rl901 had hereditary angioneurotic edema (HANE). Those with MPGN III with inherited deficiencies were moderately hypocomplementemic and their individual profiles are not graphed. One was heterozygous deficiency in C2 and the other in C7. Levels of complement components affected by these deficiencies have been omitted from the graphs and are not included in the statistical analysis. It is assumed, in the absence of evidence to the contrary, that heterozygous deficiencies of Clq , C2 , C6, C7, and C8 do not affect the levels of other components. The level of C4 has been omitted from the data for patient Rl901 with HANE; its level was 13% of the normal mean. The level of C2 was within the normal range .
VARADE, FORRISTAL, AND WEST
ClqC2C4
P B C3C5
C6C7C8C9
•••••••••••
G 1000
........... ---.... ---- -_ ....
50~ ---0:';;;:=
5°R
~ - 5~t'::
~ - 100
:: ~
....,. .... -- ............_:::: ................ --_ ...
___._-•..
""'
----
a: w
Cl. I
~ W ~ ...J
«
~
a:
o
0 ._n_ - 50 ••••• - •• 5 o~
-
:.::.:
W 1923
-
---.. --- ... --- ... :.::: :.::: ::.:.: .....- ..
~
o
a: u.
oz -
~ > W o
50~ 0 ::::nn -~. . - 50
........ -••••
_ _ uu
W 0112
o ••• - ••••
z
-
...... -_ .......... _- -..... ...... _.. _-
-100
5l . .-
H 1011 0'\.. . . . . . . . . . . ..... .. ...
.... -
....
_- -......... ::'::--_._-_. .....
H 1207
......... . -~
- 100
5~~._ ...... - 50
- 100
50
-----
n.
-
....
f•••---.
o ::::;:.:_
-~ ~~
Clq C2 C4
..... _,. ---- ----
-
............. ---- ----
-•••--•
...'" .."',. ....... --.-......_- .... P B C3 C5
M 1005
..... _::;:---91.. ..... - - ., ... - - _ ...
N 0319
::.:.: ....... --... -
••• •
C6 C7 C8 C9
Fig 1. Complement profiles of patients with MPGN I with terminal pathway activation predominating. The shaded areas bounded by the dashed lines delineate ± 2 SO from the mean of 163 normal subjects. Compo· nent concentrations in percent deviation from the normal mean are indicated by the heavy horizontal lines. Patients are designated by code. Patient H1207 had heterozygous C8 deficiency.
Complement Profiles-MPGN I Inspection of the complement profiles of the patients with MPGN I indicated considerable heterogeneity. The first step in analyzing the data was to separate a group in which one or more of the terminal components, C6, C7, and C9, were depressed. This group, designated the predominantly terminal activation group, contained seven patients (Fig 1). In addition to terminal component
COMPLEMENT PROFILES IN MPGN I, II, AND III
199
depression, some of these patients also had signs of classical pathway activation; four of the seven had depressed levels of at least one early reacting component. The depressed C6, C7, and C9 levels are assumed to indicate the presence of NF t , 57 and the depressed classical pathway components, of circulating immune complexes. From the patients remaining, two additional groups could be separated. One, shown in Fig 2, was designated the classical activation group and consisted of 8 patients. Each patient in this group had depressed levels of at least two early reacting components while levels of C6, C7, and C9 were well within the normal range.
Clq C2C4
•••
P B C3 C5
••• •
C6C7C8C9
••• •
H 1304
T:T:TIT H:F ':":~ :: :: ::::::: ~=.:.:.:.:.:.~=
-
I 2012
~ :'!:;:: ~:T~~ ~:~.~.-: ':'::.~.~
.... i;i=,;,=j,. .;:~',;:.:. ~
_
R 1901
,......"'.'" '!::: ::: 7.7.'":.-'. .............
D 1619 Clq C2 C4
P
B C3 C5
• • • •
• ••
C6 C7C8 C9
••••
R 0416
......
-
:::::::::::::: -
!='!:'!':'!
- .........
M 1619 :::::::
......... ::::::: ..... .
:,..:,..:....;, ';':.i,:.i.:';' ---- ----
......
S 1812 T................... :T:Tr ....... :::::'...
!'!'!'!~!'!!'!'!!'!
...... ............. ....... :"' :"'' ''-
:::::::
S 0301
C 0104 i,;;i,i"':..i, ::....
. ......
D 1217
-5
.... ...................
.:.:.:.:,;,:~ ~ ~.~.~-.:.
R 1010
50~
: ';"';"':"
""."."
~ ~= ;~;~~ ITfu
-1 00
•• •
ClqC2 C4
:::::::
-
~
.i.:.i.:.;,~
:::::::
Fig 3. Complement profiles of patients with MPGN I in whom C3 activation predominated. Data presentation as in Fig 1. Patient R1901 had hereditary angioneurotic edema and patient 01619, heterozygous C6 deficiency.
...... .
,.... ----
B 1305
W 1812 ............................
--
i.i.i.:';':';'
~:i:i.:~
••• • •••• P B C3 C5 C6C7C8C9
.::.:.:.;.:.;, ,;,:.::";':':'
• C7 •• • •P • • • C6 B C3 C5 C8 C9
Fig 2. Complement profiles of patients with MPGN I showing classical pathway activation. Data presentation as in Fig 1. Patient 0210 had partial C1 q deficiency.
The third group, designated the predominantly C3 activation group, consisted of seven patients (Fig 3). Three of these had profiles characterized only by C3 depression. The other four had, in addition, depressed levels of C5 and/or properdin. All levels of C lq, C2, C6, C7, and C9 were within the normal range and only patient R1901 with HANE had, as would be expected, a depressed level of C4. The three groups can be compared in Fig 4. C6, C7, and C9 were significantly depressed in the predominantly terminal activation group only, but
VARADE, FORRISTAL, AND WEST
200
IPredominately terminal activation, n- 7!
individual patients, C8 was depressed in all in the predominantly terminal activation group and in approximately half of those in the other two groups.
Complement Profiles-MPGN III
Clq C2 C4
P
B C3 C5
C6 C7 CB C9
IClassical activation. n- BI 50
Complement profiles in the patients with MPGN III gave considerable evidence of terminal pathway activation with scant evidence of classical pathway activation. Profiles for a group of eight with severe hypocomplementemia (serum C3 level < 30 mg/dL) are shown in Fig 5. Levels of early reacting components were rarely depressed; one
.::::
Clq C2 C4
• • •
Clq C2 C4
P
B C3 C5
IPredominately C3 activation .
C6 C7 CB C9
-~ ~~
P B C3 C5
• • • •
C6 C7 C8 C9
••••
E 0805
.. ..... ::.::.: ......
----
..-.. .
- -'';'
P 0113 nz 7!
...... .
. .
.... ~---- ~-,,~:l
P 1019
--P
B C3 C5
R 0112
C6 C7 CB C9
Fig 4. Mean values for profiles of patients with MPGN I shown in Figs 1 through 3. The mean ± 2 SD are indicated by bars on the vertical lines. P values indicate means significantly lower than those for 163 normal subjects.
- 5a - 100
---- ----
.......
-_ .... ----
.. ;::: ..... ..... : ... . . ~:.:.~:,;.
S 0413
.. ....
....... .. .... .. ....
~.i.:~:.:.=..::..;;.:.;.:.:;,,::.:.:..
in this group mean levels of Clq, C2, and C4 were also significantly less than normal. However, the latter components were more profoundly depressed in the classical activation group but mean levels of properdin and C5 were higher. The higher levels reflect the presence of the three severely hypocomplementemic patients (RI901, D1619, and S0301) who had normal C5 and properdin levels. As will be noted below, this profile is typical of that seen in MPGN II produced by NF a . The remaining four patients in the predominantly C3 activation group had low levels of C5 and/or properdin, probably as the result of low plasma levels of NF t • Depressed levels of properdin, factor Band C8 were seen in all three groups, with properdin and C8 being most consistently depressed. In terms of
S 0301 ~ ~l.:.:~
-- .. .:.- :'::.::.:':"
W 0401
5~~~~ JI2 ----
-50 - 100
-.. ::.:.:
................. - ..
~:~:.:.:.:.~
.......
W 0412
4O4O
4O4O
C6 C7 C8 C9 Fig 5. Complement profiles of severely hypocomplementemic patients with MPGN III. Data presentation as in Fig 1.
201
COMPLEMENT PROFILES IN MPGN I, II, AND III
! Severe hypoc omplementemia, nos !
50
I-
Z
~
O~~~~~~~~~~--~~~~~
ffi
- 50
:'i
- 100
Cl. I
w
Clq C2 C4
P
B C3 C5
C6 C7 CS C9
~
! Moderate hypocomplementemia , n- s !
Fig 6. Comparison of mean component concentrations for severely and moderately hypocomplementemic patients with MPGN III. Data presentation as in Fig 4.
patient had a level of Clq slightly below the lower limit of the normal range, but all levels of C2 and C4 were within the normal range. However, six of eight had depressed levels of at least one of the three terminal components, C6, C7, and C9, and in all patients levels of C5 were markedly depressed. The pooled data for the above severely hypocomplementemic patients can be compared with that for eight moderately hypocomplementemic
.
+100
;J.
+50
~
(serum C3 level, 30 to 86 mg/dL) patients in Fig 6. With severe hypocomplementemia, mean levels of C6, C7, and C9 were significantly depressed, but were normal with moderate hypocomplementemia. Whereas both groups had a mean C2 level significantly less than that in normal subjects, it seems doubtful that this indicates classical pathway activation since (1) the mean levels were approximately equal in the two groups despite the marked differences in C3 concentration, (2) depression of C4 was not seen in any patient and mean levels were normal, and (3) the significance of the deviation of the mean of C2 from that of normal subjects was borderline (P < 0.05). In both groups there was significant depression of properdin, factor B, and C8 levels similar to that observed in the MPGN I patients.
Comparison of Profiles Characterized by Severe Hypocomplementemia in MPGN I and III To pinpoint the difference between profiles in MPGN I and III, means for the severely hypocomplementemic patients in the two groups are compared in Fig 7. The most significant difference is a depressed level of C4 in MPGN 1. C2 also tends to be lower in MPGN 1. C5 and C7 were lower in MPGN III, while C6, C8, and C9 were comparably depressed in the two types.
Complement Profiles-MPGN II Complement profiles of seven severely hypocomplementemic patients with MPGN II are summarized in Fig 8. With the severely depressed C3, the only other definite abnormality was a de-
W
ISevere hypocomplementemia, n"71
::;;
::;;
a: >oZ ZW
::;;(j
oa: W a: u.a.
5
. I
0
Z
::; ...J
«
- 50
::; a:>OZ z~
~ :; W
o
50
W
< .05 <. 001 n 16 8 16 8 15 8
M~~ IN IN IN
Clq C2 C4
< .02
<. 0 1
16 8 16 8 16 8 16 8
15 8
IN 1][ 1][ IN
IN 1][ IN 1][
P
B
C3 C5
16 8 15 8 16 8
C6 C7 C8 C9
Fig 7. Comparison of mean concentrations of complement components for severely hypocomplementemic patients with MPGN I and III. P values indicate significance of differences between means for MPGN I and III. The shaded areas indicate ± 2 SO from the mean for normal subjects.
~ffi
[a.
z
- 50
o
~
t
>
w
o
- 100
Clq C2 C4
P
B C3 C5
C6 C7 C8 C9
Fig 8. Mean values for components in severely hypocomplementemic patients with MPGN II. Data presentation as in Fig 4.
202
VARADE, FORRISTAL, AND WEST
Table 2. Comparison of MPGN I and III With Respect to Positive Immunofluorescence for Glomerular IgG, C1q, C4, and C5 in Renal Biopsies MPGN Type
III P
IgG
C1q
C4
C5
24/25* (96)t 10/20 (50)
17/22 (77) 7/19 (37)
11/15 (73) 1/16 (6)
15/15 (100) 16/16 (100)
<0.001
<0.01
<0.001
NS
* Numerator indicates number of biopsies with fluorescence positive for the protein and the denominator, the number of biopsies tested. tPercent positive.
pressed level of factor B, present in four patients (57 %). The depression of C5 is of marginal significance; no patient actually had a C5 level below the lower limit of the normal range. The level of properdin was slightly depressed in one patient and of C4 in another, but mean levels did not differ significantly from normal. The long error bars for C4 reflect a markedly elevated level in one patient. Complement profiles were also available for four patients with MPGN II who were moderately hypocomplementemic. They had principally low levels ofC3, but one had a low properdin level and another, a low level of C4. The patients with MPGN II had a much higher frequency of uremia than those with types I and III. Of the 14 hypocomplementemic patients, seven had compromised renal function. Whether uremia in this setting would influence the complement profile is not clear, but it should be noted that severely compromised renal function was present in all patients with low levels of factor B, in one with a low C4 level, and in another with a low properdin level. Immunofluorescence of Renal Biopsies in MPGN I and III The results of labeled antibody studies to detect glomerular IgG, C 1q, C4, and C5 in renal biopsies of patients with MPGN I and III are shown in Table 2. It is apparent that IgG and C4 were found less often in MPGN III than in MPGN I and that the difference was highly significant. C 1q was also less frequently found in MPGN III, but the difference was less significant. C4 was observed in only trace amounts in one patient with MPGN III. The numbers of MPGN I patients in the classical, predominantly terminal, and predominantly C3 activation groups are too small to relate immunofluorescence results to the complement profile.
The Complement Profile in Determining the Type of MPGN The study has shown that severe hypocomplementemia (C3 < 30 mg/dL) in the patient with MPGN may be accompanied by the following categories of abnormalities: (1) signs of classical pathway activation; (2) signs of C6, C7, and C9 activation, (4) depressed levels of C5 and properdin only, or (5) normal levels of all other components. Experience with serial complement profiles in this laboratory has indicated that if the hypocomplementemia remains severe, the complement profile will remain in the same category. As the C3 rises above 30 mg/dL, the accompanying abnormalities may disappear, but usually return with the same pattern should severe hypocomplementemia again develop. The glomerulonephritides in which these profiles may be encountered are shown in Table 3. For the patient with histologic and clinical evidence of MPGN, signs of classical pathway activation, with or without signs of C6, C7, and C9 activation, indicate the presence of MPGN I. However, it should be noted that classical pathway activation without signs of C6, C7, and C9 activation could also be seen in systemic lupus erythematosus (SLE), with chronic bacteremia and early in the course of acute post-streptococcal glomerulonephritis (AGN). None of the other profiles are diagnostic of type; they could be seen in either MPGN II or III as well as in MPGN I, and they also overlap with other hypocomplementemic glomerulonephritides. Correlation of Complement Profiles and Extended Haplotypes A previous study from this laboratory16 indicated that 28 % of patients with MPGN I and III had, in the major histocompatibility complex on the short arm of chromosome 6, the extended hap-
COMPLEMENT PROFILES IN MPGN I, II, AND III
lotype, AI, B8, DR3, SCOl, GL02. Of 10 patients with this extended haplotype, the hypocomplementemia in three could be ascribed to NFl' two had classical pathway activation only with no evidence of NFl' and hypocomplementemia was not documented in our laboratory in five. DISCUSSION
The overall frequency of hypocomplementemia in the children and young adults of the present study is 59 of 77 (77 %). This is slightly greater than the frequency of 28 of 44 (64 %) observed by Cameron et al 17 in patients presenting under the age of 20. However, the difference is not significant. The frequency of hypocomplementemia in the MPGN I patients was 68 % and in the MPGN III patients, 84 %. Again, the difference is not significant. Depression of the levels of C6, C7, and C9, found frequently in the patients of the present study, is not seen in the hypocomplementemia produced by NFa , the nephritic factor found in MPGN II and partiallipodystrophy71s (Fig 8), nor in that produced by circulating immune complexes in SLE and chronic bacteremia. IS Also, in hypocomplementemic patients with acute post-streptococcal glomerulonephritis, C6, C7, and C9 levels are not depressed.7.IS.19 Although this would suggest that terminal components are not activated in these disTable 3.
Nephritides
203
eases, it should be noted that in SLE and AGN, C5b-9 (the membrane attack complex) may be found in the circulation,5.20 indicating that the terminal pathway is being activated. However, the activation is apparently not sufficient to depress the serum levels of the three terminal components. The observations of Mollnes et al 5 and Clardy et al 6 strongly suggest that the effector of the terminal component reduction in MPGN is a hitherto unrecognized "slow" nephritic factor. This nephritic factor differs from that first described in that to be effective, properdin must be present, 6 and in mixtures with normal serum, it activates C5 and C9 and probably the other terminal components 6 and forms C5b-9 complexes. 5 This nephritic factor has been designated NFIIlh 6 or the nephritic factor of the terminal pathway, NFl' IS The nephritic factor originally described and found in MPGN II and partial lipodystrophy has been designated NF n,6 or NF a , the nephritic factor of the amplification loop. IS When present in high concentration, the two nephritic factors can be distinguished by rates of C3 conversion in vitro. With NFl' C3 conversion continues for up to 4 hours,5.6 whereas C3 conversion produced by high concentrations of NF a is rapid but ceases after 20 minutes. 21 At lower concentrations, differences in the rate and duration of C3 conversion are not distinguishing features;
Pathways of Complement Activation in Hypocomplementemic Glomerulonephritis Classical* Pathway Activation
Severe hypocomplementemia (C3, < 30 mg/dL) MPGN I + MPGN I, SLE, AGN§, chronic bacteremia + MPGN I, MPGN III MPGN I, MPGN III, AGN MPGN It MPGN II, partial lipodystrophy Moderate hypocomplementemia (C3, 30-86 mg/dL) MPGN I, SLE, chronic bacteremia + MPGN I, MPGN III, AGN MPGN It MPGN II 'Depression of at least one early reacting component. tDepression of C6, C7, and/or Cg. :j:Properdin usually also in normal concentration. §Observed only occasionally early in course of AGN. ,Found in three of 23 (12%) patients with MPGN I.
Terminalt Pathway Activation
Depressed C5 and Properdin
Normal C5t
+ +
+ +
+
+
204
with low concentrations of both, C3 conversion can continue for several hours. The two could be distinguished more accurately by assessing their properdin dependence, or by measuring activation of components of the terminal pathway or the production of C5b-9, and relating the results to the extent of C3 conversion under standard conditions. Unfortunately, the serum remaining after measurement of profiles in the patients in the present study was not sufficient to allow identificaiton and assay of nephritic factors. In lieu of this, hypocomplementemia characterized by low levels of C6, C7 and/or C9 is, as previouslY,7.18 assumed to indicate the presence of NF t . Depressed levels of C6, C7, and/or C9, indicating that NF t was present, were found in seven of 22 (32 %) hypcomplementemic patients with MPGN I and six of 16 (37%) with MPGN III. With a more sensitive method of detection, the frequency of NF t would be much greater. C6, C7, and C9 levels were depressed only in severely hypocomplementemic patients (none had a C3 level > 21 % of the normal mean) in whom the concentration of NF t was presumably high. Lower concentrations of NF t would presumably depress only C3, C5, and properdin levels and very low concentrations, only C3. Of the 16 hypocomplementemic MPGN III patients, six had depressed levels of C6, C7, and/or C9 combined with depressed C5 and properdin, seven had depression of C5 and properdin only, and three had, as the only abnormality, a mild to moderate C3 depression. These profiles would appear to represent the full range of those produced by varying concentrations of NF t • From the data for patients with MPGN II, it is readily apparent that NFa has a different mode of action. Severely hypocomplementemic patients lacked signs of classical and terminal pathway activation, and levels of C5 and, usually, properdin were normal. In four of seven, levels of factor B were low. We l8 and others22 have seen the same profile in patients with partial lipodystrophy and nephritis. In a patient with MPGN II, others 23 have found C5 catabolism to be normal despite markedly increased C3 catabolism. The above observations indicate that the hypocomplementemia in both MPGN II and III is unifactorial in origin, in MPGN II due to NFa and in MPGN III, to NFl' In MPGN I, the hypocomplementemia is multifactorial in origin. In eight of 22 patients, there
VARADE, FORRISTAL, AND WEST
was evidence of classical pathway activation only; in seven, evidence of NFl only; in three, evidence of NF a and in four, there was some evidence for both classical activation and NFl' The three patients with evidence for NFa were markedly hypocomplementemic, yet had normal levels of C5 and properdin, a profile not seen with NFl' All of the four with combined classical-terminal activation had definite evidence of terminal activation, but in some, the evidence for classical activation was less convincing. Patients GIOOO and WOl12 (Fig 1) had depressed C4 levels with or without depressed Clq levels, but patients HlOll and HI923 had only slightly depressed levels of Clq in the absence of a nephrotic syndrome. The multifactorial nature of the hypocomplementemia in MPGN I may be even greater than the present study indicates. Two patients have been described 24 who had NFc, the nephritic factor of the classical pathway. Both also had a C3b,Bb stabilizing factor, which was presumably NFl (both had low C5Ievels). NFc has been previously found in patients with AGN25 and SLE.26 The basis for the depression of determinants of the alpha-gamma chain of C8, found in most patients with MPGN I and III, independent of the levels of other terminal components, is not known. Although the data indicate that NFt augments the depression of C8, additional factor(s) must be responsible, since C8 levels also tend to be depressed in SLE,18 in the nephritis of chronic bacteremia l8 and, to a minimal extent, in AGN.718 In patients with SLE, the levels tend to parallel closely those of factor B (unpublished observations by authors). C8 levels were infrequently depressed in hypocomplementemic patients with MPGN II or partial lipodystrophy. 18 Hypocomplementemia and the pathogenesis of MPGN can be correlated only for profiles indicating classical pathway activation and here the correlation is not straightforward. Although it is well known that measurements of circulating immune complexes by different methods give results that correlate poorly with each other 2728 and with disease activity,2829 there is nevertheless strong circumstantial evidence that circulating complexes can produce a glomerular morphology very similar to that of idiopathic MPGN 13031 and that at least those found in SLE are sufficiently reactive with complement to produce the hypocomplementemia. 32 On this basis, the classical pathway ac-
COMPLEMENT PROFILES IN MPGN I, II, AND III
tivation observed in MPGN I must be related to pathogenesis but its precise relationship depends on whether the complexes reacting with complement are promptly cleared or persist with ultimate deposition in the glomeruli. 32 It has to be assumed that the normocomplementemic patients with MPGN I (Table 1), as well as those without evidence in their profile of classical pathway activation either had complexes that were nephritogenic but minimally reactive with complement or had complexes that had disappeared before the patient presented. The relationship between nephritic factors and pathogenesis is less clear. The present study indicates that they do not relate to the abnormalities in glomerular ultrastructure insofar as both NFl and NFa can be present in MPGN I yet the ultrastructure has no resemblance to that of MPGN II or MPGN III. The nephritic factors would be effective in generating disease only by partially depleting the plasma of complement and in this nonspecific way compromising immune complex disposal. However, our previous observations have shown no well-defined correlation between the presence of hypocomplementemia 14 or of nephritic factors 33 and the course of the disease. The evidence provided by both the complement profiles and the immunofluorescence data that circulating immune complexes are frequently present in MPGN I but not in MPGN III, reinforces evidence given previously3 that these two types are distinct entities. Whereas in the previous publication,3 MPGN I and III were considered variants of the same disease, more recent observations do not support this conclusion. Studies from this laboratory16 have shown that some patients with MPGN I and III have a common genetic basis for their disease in that 29 % have, on the short arm of chromosome 6, the extended haplotype, AI, B8, DR3, SCOl, GL02. This extended haplotype is found in only 3 % of control subjects. Because this extended haplotype has also been found with high frequency in insulin-dependent diabetes mellitus,34 glutensensitive enteropathy,35 and SLE,36 its high frequency in MPGN I and III indicates only that both types have their origin in autoimmune phenomena and not that they are necessarily the same disease or variants of the same disease. Parenthetically, the previous publication 16 indicated that this haplotype was associated with a poor prognosis and the question arises as to whether it might be associated
205
with other manifestations of disease. The fact that the 10 patients in the present study with this haplotype had widely differing complement profiles indicates that it does not determine which nephritic factor is produced or whether complement reactive immune complexes are circulating. In summary, differences in disease presentation, glomerular morphology and in mechanisms of complement perturbation give evidence that MPGN I and III are of differing pathogenesis. The pathogenesis of MPGN I, like that of SLE and the nephritis of chronic bacteremia, is based on the glomerular deposition of immune complexes, while that of MPGN III is independent of antibody and related to intraglomerular activation of the alternative pathway with deposits composed of C3, properdin, and C5. These proteins, with factor B, are the constituents of the alternative pathway C3/ C5 convertase. The general absence of factor B3738 could be attributed to the relatively low affinity of its association with C3b. 39 The factor or factors that cause the glomerulus to activate the alternative pathway should be the subject of further investigation.
REFERENCES 1. Strife CF, McEnery PT, McAdams AJ, et al: Membranoproliferative glomerulonephritis with disruption of the glomerular basement membrane. Clin Nephrol 7:65-72, 1977 2. West CD, McAdams AJ: The chronic glomerulonephritides of childhood. Part II. J Pediatr 93: 167-176, 1978 3. Jackson EC, McAdams AJ, Strife CF, et al: Differences between membranoproliferative glomerulonephritis types I and III in clinical presentation, glomerular morphology, and complement perturbation. Am J Kidney Dis 9:115-120, 1987 4. Pan C, Strife CF, West CD: The glomerular capillary as an activator of the alternative pathway of complement in nephritis. Kidney Int 35:372, 1989 (abstr) 5. Mollnes TE, Ng YC, Peters DK, et al: 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 6. Clardy Cw, Forristal J, Strife CF, et al: 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 7. Clardy Cw, Forristal J, Strife CF, et al: Serum terminal complement component levels in hypocomplementemic glomerulonephritides. Clin Immunol Immunopathol 50:307-310, 1989 8. Coleman TH, Forristal J, Kosaka T, et al: Inherited complement component deficiencies in membranoproliferative glomerulonephritis. Kidney Int 24:681-690, 1983 9. Strife CF, McAdams AJ, West CD: Membranoproliferative glomerulonephritis characterized by focal, segmental proliferative lesions. Clin Nephrol 18:9-16, 1982
206 10. Wyatt RJ, Forristal J, Davis CA, et al: Control of serum C3 levels by {3lH and C3b inactivator. J Lab Clin Med 95:905917, 1980 II. Marder HK, Coleman TH, Forristal J, et al: An inherited defect in the C3 convertase, C3b,Bb, associated with glomerulonephritis. Kidney Int 23:749-758, 1983 12, Linshaw MA, Stapleton BF, Cuppage FE, et al: Hypocomplementemic glomerulonephritis in an infant and mother: Evidence for an abnormal C3. Am J Nephrol 7:470-477, 1987 13. West CD, McAdams AJ: Serum {3IC globulin levels in persistent glomerulonephritis with low serum complement: Variability unrelated to clinical course. Nephron 7: 193-202, 1970 14. McEnery PT, McAdams AJ, West CD: The effect of a high dose alternate day prednisone regimen on the natural history of membranoproliferative glomerulonephritis. Medicine (Baltimore) 64:401-424, 1985 15. Tedesco F, Densen P, Villa MA, et al: Two types of dysfunctional eighth component of complement (C8) molecules in C8 deficiency in man. J Clin Invest 71:183-191, 1983 16. Welch TR, Beischel L, Balakrishnan K, et al: Majorhistocompatibility-complex extended haplotypes in membranoproliferative glomerulonephritis. N Engl J Med 314: 14761481,1986 17. Cameron JS, Turner DR, Heaton J, et al: Idiopathic mesangio-capillary glomerulonephritis: Comparison of types I and II in children and adults and long term prognosis. Am J Med 74:175-192,1983 18. West CD: The complement profile in clinical medicine: Inherited and acquired conditions lowering the serum concentrations of complement and control proteins. Complement Inflamm 6:49-64, 1989 19. Wyatt RJ, Forristal J, West CD, et al: Complement profiles in acute post-streptococcal glomerulonephritis. Pediatr Nephrol 2:219-223,1988 20. Falk RJ, Dalmasso Ap, Kim Y, et al: Radioimmunoassay of the attack complex of complement in serum from patients with systemic lupus erythematosus. N Engl J Med 312: 15941599, 1985 21. Vallota EH, Forristal J, Spitzer RE, et al: Characteristics of a non-complement dependent C3-reactive complex formed from factors in nephritic .and normal serum. J Exp Med 131: 1306-1334, 1970 22. Sissons JGp, West RJ, Fallows J, et al: The complement abnormalities of lipodystrophy. N Engl J Med 294:461-465, 1976 23. Sissons JGp, Liebowitch D, Amos N, et al: 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 24. Tanuma Y, Ohi H, Watanabe S, et al: C3 nephritic factor and C4 nephritic factor in the serum of two patients with hypocomplementemic membranoproliferative glomerulonephritis. Clin Exp Immunol 76:82-85, 1989
VARADE, FORRISTAL, AND WEST 25. Halbwachs L, Leveille p, Lesavre P, et al: Nephritic factor of the classical pathway of complement. Immunoglobulin G autoantibody directed against the classical C3 pathway con vcrtase enzyme. J Clin Invest 65: 1249-1256, 1980 26. Daha MR, Hazevoet HM, van Es LA, et al: Stabilization of the classical pathway convertase, C 42, by a factor (F-42) isolated from sera of patients with systemic lupus erythematosus (SLE). Immunology 40:417-424, 1980 27. Lambert PH, Dixon FJ, Zubler RH, et al: A WHO collaborative study for the evaluation of eighteen methods for detecting immune complexes in serum. J Lab Clin Immunol I: 115, 1978 28. Fust G, Kavai M, Szegedi GY, et al: Evaluation of different methods for detecting circulating immune complexes. An interlaboratory study. J Immunol Methods 38:281-289, 1980 29. Davis CA, Marder H, West CD: Circulating immune complexes in membranoproliferative glomerulonephritis. Kidney Int 20:728-732, 1981 30. Michael AF, Herdman RC, Fish AJ, et al: Chronic membranoproliferative glomerulonephritis with hypocomplementemia. Transplant Proc 1:925-932, 1969 31. Strife CF, McDonald BM, Ruley EJ, et al: Shunt nephritis: The nature of the serum cryoglobulins and their relation to the complement profile. J Pediatr 88:403-413, 1976 32. Edberg JC, Tosic L, Wright EL, et al: Quantitative analysis of the relationship between C3 consumption, C3b capture, and immune adherence of complement-fixing antibody/DNA immune complexes. J Immunol 141 :4258-4265, 1988 33. Vallota EH, Forristal J, Davis NC, et al: The C3 nephritic factor and membranoproliferative nephritis: Correlation of serum levels of the nephritic factor with C3 levels, with therapy and with progression of the disease. J Pediatr 80:947959, 1972 34. Raum D, Awdeh Z, Yunis EJ, et al: Extended major histocompatibility complex haplotypes in type I diabetes mellitus. J Clin Invest 74:449-454, 1984 35. Alper CA, Fleischnich E, Awdeh Z, et al: Extended major histocompatibility complex haplotypes in patients with gluten-sensitive enteropathy. J Clin Invest 79:251-256, 1987 36. Welch TR, Beischel LS, Balakrishnan K, et al: Major histocompatibility complex extended haplotypes in systemic lupus erythematosus. Dis Markers 6:247-255, 1988 37. Michael AF, McLean RH: Evidence for activation of the alternative pathway in glomerulonephritis, in Hamburger J, Crosnier J, Maxwell MH (eds): Advances in Nephrology, vol 4. Chicago, IL, Yearbok Medical, 1968, pp 49-66 38. Wyatt RJ, McAdams AJ, Forristal J, et al: Glomerular deposition of complement-control proteins in acute and chronic glomerulonephritis. Kidney Int 16:505-512, 1979 39. Kazatchkine MD, Fearon DT, Austen KF: Human alternative complement pathway: Membrane associated sialic acid regulates the competition between Band {3-1H for cell bound C3b. J Immunol 122:75-81, 1979
F
OLLOWING ITS original description, membranoproliferative glomerulonephritis (MPGN) was divided into three types based on glomerular ultrastructure. Type II, or dense-deposit disease, has a glomerular ultrastructure differing markedly from the other two types and is considered a separate disease. Types I and III are not as easily distinguished. They differ ultrastructurally in that the deposits in type I are subendothelial and the basement membrane has a normal appearance, whereas the deposits in type III are on both sides of the basement membrane and, when impregnated with silver, the membrane appears disrupted and fragmented. 12 By uranyl-lead staining, dense deFrom the Children's Hospital Research Foundation. and the Department of Pediatrics, University of Cincinnati College of Medicine, Cincinnati, OH. Address reprint requests to Clark D. West, MD, Children's Hmpital Medical Center, Eiland and Bethesda Ave, Cincinnati, OH 45229. © 1990 by the National Kidney Foundation, Inc. 0272-6386/90/1603-0003$3.00/0 196
posits are seen within the membrane in a patchy distribution. Additional differences between MPGN I and III have been noted in earlier studies from this laboratory.3 First, judging from serum C4 levels and from the composition of glomerular deposits with respect to IgG and C4, classical pathway activation is occurring only in MPGN I. Second, MPGN III tends to be clinically silent, often detected by chance discovery of proteinuria and hematuria in the healthy subject as compared with onset with signs of systemic disease in patients with MPGN I. Finally, mesangial cell proliferation is significantly less in MPGN III than in MPGN I. The present study is concerned with details of the complement profile in all three types of MPGN. Whereas the renal glomerulus in this disease becomes an alternative pathway activator, 4 there is no evidence that complement activation in situ in the glomerulus materially affects the serum levels of components of the system. Instead, the hypocomplementemia can be ascribed to three cir-
American Journal of Kidney Diseases, Vol XVI, No 3 (September). 1990: pp 196-206
197
COMPLEMENT PROFILES IN MPGN I, II, AND III
culating complement-reactive modalities. Two of these are nephritic factors. One is the nephritic factor of the amplification loop (NFa), and the other the nephritic factor of the terminal pathway (NF t).56 In addition, classical pathway activation is occurring in MPGN I, presumably produced by circulating immune complexes. These three modalities acting singly or in combination produce several different complement profiles. MATERIALS AND METHODS
values were determined in serum samples from 163 normal hospital personnel. 8
Immunofluorescence IgG, Clq, C4, and C5 were detected in glomerular deposits by fluorescein-labeled monospecific antibody prepared in this laboratory. The data reported are those for the first biopsy obtained at this institution. The biopsies were from both normocomplementemic and hypocomplementemic patients who had glomeruli that showed diffuse proliferation; those with focal proliferation were excluded. 9 In all cases, data for immunofluorescence were recorded on the day of the biopsy before the type of MPGN had been determined.
Patients
Statistical Methods
Between 1957 and 1988, the diagnosis of MPGN has been made on 77 patients as the result of a renal biopsy performed at this institution. The type of MPGN was determined by characteristics of the ultrastructure, as described previously, 12 or, for some cases of MPGN II, by light microscopy of sections stained by the Jones methenamine silver method. Of the 77 patients, 35 had MPGN I; 17, MPGN II; and 25, MPGN III.
The means of complement component levels were compared using Student's t test. Immunofluorescence data for MPGN I and III were compared by the chi-square test.
Complement Profiles Blood for complement profiles was obtained by venipuncture, allowed to clot at room temperature, and the serum either used immediately for measurement of complement components or stored at - 70°C. It has been shown that prolonged storage of serum at - 70 °C has no effect on the measurement of serum complement component levels. 7 Two or more complement profiles were obtained on all patients except one. The profiles selected for the study were those with the lowest C3 level for which complete data were available. Serum complement protein levels were determined by radial immunodiffusion using monospecific goat antisera. 7 Reference Table 1.
RESULTS
Among the 77 patients with MPGN, 18 were found to be consistently normocomplementemic and 59 were hypocomplementemic as defined by a serum C3 level below the lower limit of the normal range (86 mg/dL). Of the 59 hypocomplementemic patients, data for 49 are reported. The frequency of normocomplementemic patients and the reasons for exclusion of data are given for each type of MPGN in Table 1. Of the 10 hypocomplementemic patients excluded from the study, the data for seven were incomplete and in three the complement profile was, or could have been, al-
Reasons for Exclusion of Patient Data From Analysis of Complement Profiles No. of Patients
MPGN I (n 11
=
Reasons for Exclusion
35) Normocomplementemic Severely hypocomplementemic* with homozygous C6 and C7 deficiency Moderately hypocomplementemic* with factor I deficiency
MPGN II (n 3 3
17)
MPGN III (n 4
25)
3
Normocomplementemic Moderately hypocomplementemic; terminal component levels missing Normocomplementemic Severely hypocomplementemic; terminal component levels missing Severely hypocomplementemic with Marder's syndrome 1213 Moderately hypocomplementemic; terminal component levels missing
*Severe hypocomplementemia, C3 <30 mg/dL; moderate hypocomplementemia, C3 30 to 86 mg/dL; lower limit of normal range, 86 mg/dL.
198
tered by other abnormalities. One had a persistent, severe factor I deficiency (serum level, 0 to 14 % of normal mean), which would have contributed to the hypocomplementemia by affecting the turnover of the amplification 100p.1O Data were not available to determine whether the deficiency was inherited. Another had homozygous deficiency of C6 and C7, which would compromise terminal pathway activation, and a third had Marder's syndrome. 11 . 12 The timing of the acquisition of the serum specimens used in the study varied. In 33 of the 49 patients, the serum used was that obtained at the time of, or shortly after, presentation at this institution. In six , the serum was obtained within 2 years of presentation, in six, from 2 to 5 years of presentation , and in four, after 5 years. Thirteen of the 16 profiles not performed on initial specimens were from patients seen before 1969, most of whom either were not receiving prednisone or were receiving it daily for short periods. In such patients, the hypocomplementemia may persist for long periods or wax and wane.13 After 1969, most patients were receiving an alternate-day prednisone regimen . Most of these rapidly and permanently lost their hypocomplementemia. '4 Inherited complement component deficiencies were commonly observed among the patients with MPGN I and III who were included in the study. 8 Among those with MPGN I, patient P0210 had heterozygous Clq deficiency; patient D16I9 had heterozygous C6 deficiency; patient HI207, heterozygous deficiency of determinants of the alphagamma chain of C8 1s ; and patient Rl901 had hereditary angioneurotic edema (HANE). Those with MPGN III with inherited deficiencies were moderately hypocomplementemic and their individual profiles are not graphed. One was heterozygous deficiency in C2 and the other in C7. Levels of complement components affected by these deficiencies have been omitted from the graphs and are not included in the statistical analysis. It is assumed, in the absence of evidence to the contrary, that heterozygous deficiencies of Clq , C2 , C6, C7, and C8 do not affect the levels of other components. The level of C4 has been omitted from the data for patient Rl901 with HANE; its level was 13% of the normal mean. The level of C2 was within the normal range .
VARADE, FORRISTAL, AND WEST
ClqC2C4
P B C3C5
C6C7C8C9
•••••••••••
G 1000
........... ---.... ---- -_ ....
50~ ---0:';;;:=
5°R
~ - 5~t'::
~ - 100
:: ~
....,. .... -- ............_:::: ................ --_ ...
___._-•..
""'
----
a: w
Cl. I
~ W ~ ...J
«
~
a:
o
0 ._n_ - 50 ••••• - •• 5 o~
-
:.::.:
W 1923
-
---.. --- ... --- ... :.::: :.::: ::.:.: .....- ..
~
o
a: u.
oz -
~ > W o
50~ 0 ::::nn -~. . - 50
........ -••••
_ _ uu
W 0112
o ••• - ••••
z
-
...... -_ .......... _- -..... ...... _.. _-
-100
5l . .-
H 1011 0'\.. . . . . . . . . . . ..... .. ...
.... -
....
_- -......... ::'::--_._-_. .....
H 1207
......... . -~
- 100
5~~._ ...... - 50
- 100
50
-----
n.
-
....
f•••---.
o ::::;:.:_
-~ ~~
Clq C2 C4
..... _,. ---- ----
-
............. ---- ----
-•••--•
...'" .."',. ....... --.-......_- .... P B C3 C5
M 1005
..... _::;:---91.. ..... - - ., ... - - _ ...
N 0319
::.:.: ....... --... -
••• •
C6 C7 C8 C9
Fig 1. Complement profiles of patients with MPGN I with terminal pathway activation predominating. The shaded areas bounded by the dashed lines delineate ± 2 SO from the mean of 163 normal subjects. Compo· nent concentrations in percent deviation from the normal mean are indicated by the heavy horizontal lines. Patients are designated by code. Patient H1207 had heterozygous C8 deficiency.
Complement Profiles-MPGN I Inspection of the complement profiles of the patients with MPGN I indicated considerable heterogeneity. The first step in analyzing the data was to separate a group in which one or more of the terminal components, C6, C7, and C9, were depressed. This group, designated the predominantly terminal activation group, contained seven patients (Fig 1). In addition to terminal component
COMPLEMENT PROFILES IN MPGN I, II, AND III
199
depression, some of these patients also had signs of classical pathway activation; four of the seven had depressed levels of at least one early reacting component. The depressed C6, C7, and C9 levels are assumed to indicate the presence of NF t , 57 and the depressed classical pathway components, of circulating immune complexes. From the patients remaining, two additional groups could be separated. One, shown in Fig 2, was designated the classical activation group and consisted of 8 patients. Each patient in this group had depressed levels of at least two early reacting components while levels of C6, C7, and C9 were well within the normal range.
Clq C2C4
•••
P B C3 C5
••• •
C6C7C8C9
••• •
H 1304
T:T:TIT H:F ':":~ :: :: ::::::: ~=.:.:.:.:.:.~=
-
I 2012
~ :'!:;:: ~:T~~ ~:~.~.-: ':'::.~.~
.... i;i=,;,=j,. .;:~',;:.:. ~
_
R 1901
,......"'.'" '!::: ::: 7.7.'":.-'. .............
D 1619 Clq C2 C4
P
B C3 C5
• • • •
• ••
C6 C7C8 C9
••••
R 0416
......
-
:::::::::::::: -
!='!:'!':'!
- .........
M 1619 :::::::
......... ::::::: ..... .
:,..:,..:....;, ';':.i,:.i.:';' ---- ----
......
S 1812 T................... :T:Tr ....... :::::'...
!'!'!'!~!'!!'!'!!'!
...... ............. ....... :"' :"'' ''-
:::::::
S 0301
C 0104 i,;;i,i"':..i, ::....
. ......
D 1217
-5
.... ...................
.:.:.:.:,;,:~ ~ ~.~.~-.:.
R 1010
50~
: ';"';"':"
""."."
~ ~= ;~;~~ ITfu
-1 00
•• •
ClqC2 C4
:::::::
-
~
.i.:.i.:.;,~
:::::::
Fig 3. Complement profiles of patients with MPGN I in whom C3 activation predominated. Data presentation as in Fig 1. Patient R1901 had hereditary angioneurotic edema and patient 01619, heterozygous C6 deficiency.
...... .
,.... ----
B 1305
W 1812 ............................
--
i.i.i.:';':';'
~:i:i.:~
••• • •••• P B C3 C5 C6C7C8C9
.::.:.:.;.:.;, ,;,:.::";':':'
• C7 •• • •P • • • C6 B C3 C5 C8 C9
Fig 2. Complement profiles of patients with MPGN I showing classical pathway activation. Data presentation as in Fig 1. Patient 0210 had partial C1 q deficiency.
The third group, designated the predominantly C3 activation group, consisted of seven patients (Fig 3). Three of these had profiles characterized only by C3 depression. The other four had, in addition, depressed levels of C5 and/or properdin. All levels of C lq, C2, C6, C7, and C9 were within the normal range and only patient R1901 with HANE had, as would be expected, a depressed level of C4. The three groups can be compared in Fig 4. C6, C7, and C9 were significantly depressed in the predominantly terminal activation group only, but
VARADE, FORRISTAL, AND WEST
200
IPredominately terminal activation, n- 7!
individual patients, C8 was depressed in all in the predominantly terminal activation group and in approximately half of those in the other two groups.
Complement Profiles-MPGN III
Clq C2 C4
P
B C3 C5
C6 C7 CB C9
IClassical activation. n- BI 50
Complement profiles in the patients with MPGN III gave considerable evidence of terminal pathway activation with scant evidence of classical pathway activation. Profiles for a group of eight with severe hypocomplementemia (serum C3 level < 30 mg/dL) are shown in Fig 5. Levels of early reacting components were rarely depressed; one
.::::
Clq C2 C4
• • •
Clq C2 C4
P
B C3 C5
IPredominately C3 activation .
C6 C7 CB C9
-~ ~~
P B C3 C5
• • • •
C6 C7 C8 C9
••••
E 0805
.. ..... ::.::.: ......
----
..-.. .
- -'';'
P 0113 nz 7!
...... .
. .
.... ~---- ~-,,~:l
P 1019
--P
B C3 C5
R 0112
C6 C7 CB C9
Fig 4. Mean values for profiles of patients with MPGN I shown in Figs 1 through 3. The mean ± 2 SD are indicated by bars on the vertical lines. P values indicate means significantly lower than those for 163 normal subjects.
- 5a - 100
---- ----
.......
-_ .... ----
.. ;::: ..... ..... : ... . . ~:.:.~:,;.
S 0413
.. ....
....... .. .... .. ....
~.i.:~:.:.=..::..;;.:.;.:.:;,,::.:.:..
in this group mean levels of Clq, C2, and C4 were also significantly less than normal. However, the latter components were more profoundly depressed in the classical activation group but mean levels of properdin and C5 were higher. The higher levels reflect the presence of the three severely hypocomplementemic patients (RI901, D1619, and S0301) who had normal C5 and properdin levels. As will be noted below, this profile is typical of that seen in MPGN II produced by NF a . The remaining four patients in the predominantly C3 activation group had low levels of C5 and/or properdin, probably as the result of low plasma levels of NF t • Depressed levels of properdin, factor Band C8 were seen in all three groups, with properdin and C8 being most consistently depressed. In terms of
S 0301 ~ ~l.:.:~
-- .. .:.- :'::.::.:':"
W 0401
5~~~~ JI2 ----
-50 - 100
-.. ::.:.:
................. - ..
~:~:.:.:.:.~
.......
W 0412
4O4O
4O4O
C6 C7 C8 C9 Fig 5. Complement profiles of severely hypocomplementemic patients with MPGN III. Data presentation as in Fig 1.
201
COMPLEMENT PROFILES IN MPGN I, II, AND III
! Severe hypoc omplementemia, nos !
50
I-
Z
~
O~~~~~~~~~~--~~~~~
ffi
- 50
:'i
- 100
Cl. I
w
Clq C2 C4
P
B C3 C5
C6 C7 CS C9
~
! Moderate hypocomplementemia , n- s !
Fig 6. Comparison of mean component concentrations for severely and moderately hypocomplementemic patients with MPGN III. Data presentation as in Fig 4.
patient had a level of Clq slightly below the lower limit of the normal range, but all levels of C2 and C4 were within the normal range. However, six of eight had depressed levels of at least one of the three terminal components, C6, C7, and C9, and in all patients levels of C5 were markedly depressed. The pooled data for the above severely hypocomplementemic patients can be compared with that for eight moderately hypocomplementemic
.
+100
;J.
+50
~
(serum C3 level, 30 to 86 mg/dL) patients in Fig 6. With severe hypocomplementemia, mean levels of C6, C7, and C9 were significantly depressed, but were normal with moderate hypocomplementemia. Whereas both groups had a mean C2 level significantly less than that in normal subjects, it seems doubtful that this indicates classical pathway activation since (1) the mean levels were approximately equal in the two groups despite the marked differences in C3 concentration, (2) depression of C4 was not seen in any patient and mean levels were normal, and (3) the significance of the deviation of the mean of C2 from that of normal subjects was borderline (P < 0.05). In both groups there was significant depression of properdin, factor B, and C8 levels similar to that observed in the MPGN I patients.
Comparison of Profiles Characterized by Severe Hypocomplementemia in MPGN I and III To pinpoint the difference between profiles in MPGN I and III, means for the severely hypocomplementemic patients in the two groups are compared in Fig 7. The most significant difference is a depressed level of C4 in MPGN 1. C2 also tends to be lower in MPGN 1. C5 and C7 were lower in MPGN III, while C6, C8, and C9 were comparably depressed in the two types.
Complement Profiles-MPGN II Complement profiles of seven severely hypocomplementemic patients with MPGN II are summarized in Fig 8. With the severely depressed C3, the only other definite abnormality was a de-
W
ISevere hypocomplementemia, n"71
::;;
::;;
a: >oZ ZW
::;;(j
oa: W a: u.a.
5
. I
0
Z
::; ...J
«
- 50
::; a:>OZ z~
~ :; W
o
50
W
< .05 <. 001 n 16 8 16 8 15 8
M~~ IN IN IN
Clq C2 C4
< .02
<. 0 1
16 8 16 8 16 8 16 8
15 8
IN 1][ 1][ IN
IN 1][ IN 1][
P
B
C3 C5
16 8 15 8 16 8
C6 C7 C8 C9
Fig 7. Comparison of mean concentrations of complement components for severely hypocomplementemic patients with MPGN I and III. P values indicate significance of differences between means for MPGN I and III. The shaded areas indicate ± 2 SO from the mean for normal subjects.
~ffi
[a.
z
- 50
o
~
t
>
w
o
- 100
Clq C2 C4
P
B C3 C5
C6 C7 C8 C9
Fig 8. Mean values for components in severely hypocomplementemic patients with MPGN II. Data presentation as in Fig 4.
202
VARADE, FORRISTAL, AND WEST
Table 2. Comparison of MPGN I and III With Respect to Positive Immunofluorescence for Glomerular IgG, C1q, C4, and C5 in Renal Biopsies MPGN Type
III P
IgG
C1q
C4
C5
24/25* (96)t 10/20 (50)
17/22 (77) 7/19 (37)
11/15 (73) 1/16 (6)
15/15 (100) 16/16 (100)
<0.001
<0.01
<0.001
NS
* Numerator indicates number of biopsies with fluorescence positive for the protein and the denominator, the number of biopsies tested. tPercent positive.
pressed level of factor B, present in four patients (57 %). The depression of C5 is of marginal significance; no patient actually had a C5 level below the lower limit of the normal range. The level of properdin was slightly depressed in one patient and of C4 in another, but mean levels did not differ significantly from normal. The long error bars for C4 reflect a markedly elevated level in one patient. Complement profiles were also available for four patients with MPGN II who were moderately hypocomplementemic. They had principally low levels ofC3, but one had a low properdin level and another, a low level of C4. The patients with MPGN II had a much higher frequency of uremia than those with types I and III. Of the 14 hypocomplementemic patients, seven had compromised renal function. Whether uremia in this setting would influence the complement profile is not clear, but it should be noted that severely compromised renal function was present in all patients with low levels of factor B, in one with a low C4 level, and in another with a low properdin level. Immunofluorescence of Renal Biopsies in MPGN I and III The results of labeled antibody studies to detect glomerular IgG, C 1q, C4, and C5 in renal biopsies of patients with MPGN I and III are shown in Table 2. It is apparent that IgG and C4 were found less often in MPGN III than in MPGN I and that the difference was highly significant. C 1q was also less frequently found in MPGN III, but the difference was less significant. C4 was observed in only trace amounts in one patient with MPGN III. The numbers of MPGN I patients in the classical, predominantly terminal, and predominantly C3 activation groups are too small to relate immunofluorescence results to the complement profile.
The Complement Profile in Determining the Type of MPGN The study has shown that severe hypocomplementemia (C3 < 30 mg/dL) in the patient with MPGN may be accompanied by the following categories of abnormalities: (1) signs of classical pathway activation; (2) signs of C6, C7, and C9 activation, (4) depressed levels of C5 and properdin only, or (5) normal levels of all other components. Experience with serial complement profiles in this laboratory has indicated that if the hypocomplementemia remains severe, the complement profile will remain in the same category. As the C3 rises above 30 mg/dL, the accompanying abnormalities may disappear, but usually return with the same pattern should severe hypocomplementemia again develop. The glomerulonephritides in which these profiles may be encountered are shown in Table 3. For the patient with histologic and clinical evidence of MPGN, signs of classical pathway activation, with or without signs of C6, C7, and C9 activation, indicate the presence of MPGN I. However, it should be noted that classical pathway activation without signs of C6, C7, and C9 activation could also be seen in systemic lupus erythematosus (SLE), with chronic bacteremia and early in the course of acute post-streptococcal glomerulonephritis (AGN). None of the other profiles are diagnostic of type; they could be seen in either MPGN II or III as well as in MPGN I, and they also overlap with other hypocomplementemic glomerulonephritides. Correlation of Complement Profiles and Extended Haplotypes A previous study from this laboratory16 indicated that 28 % of patients with MPGN I and III had, in the major histocompatibility complex on the short arm of chromosome 6, the extended hap-
COMPLEMENT PROFILES IN MPGN I, II, AND III
lotype, AI, B8, DR3, SCOl, GL02. Of 10 patients with this extended haplotype, the hypocomplementemia in three could be ascribed to NFl' two had classical pathway activation only with no evidence of NFl' and hypocomplementemia was not documented in our laboratory in five. DISCUSSION
The overall frequency of hypocomplementemia in the children and young adults of the present study is 59 of 77 (77 %). This is slightly greater than the frequency of 28 of 44 (64 %) observed by Cameron et al 17 in patients presenting under the age of 20. However, the difference is not significant. The frequency of hypocomplementemia in the MPGN I patients was 68 % and in the MPGN III patients, 84 %. Again, the difference is not significant. Depression of the levels of C6, C7, and C9, found frequently in the patients of the present study, is not seen in the hypocomplementemia produced by NFa , the nephritic factor found in MPGN II and partiallipodystrophy71s (Fig 8), nor in that produced by circulating immune complexes in SLE and chronic bacteremia. IS Also, in hypocomplementemic patients with acute post-streptococcal glomerulonephritis, C6, C7, and C9 levels are not depressed.7.IS.19 Although this would suggest that terminal components are not activated in these disTable 3.
Nephritides
203
eases, it should be noted that in SLE and AGN, C5b-9 (the membrane attack complex) may be found in the circulation,5.20 indicating that the terminal pathway is being activated. However, the activation is apparently not sufficient to depress the serum levels of the three terminal components. The observations of Mollnes et al 5 and Clardy et al 6 strongly suggest that the effector of the terminal component reduction in MPGN is a hitherto unrecognized "slow" nephritic factor. This nephritic factor differs from that first described in that to be effective, properdin must be present, 6 and in mixtures with normal serum, it activates C5 and C9 and probably the other terminal components 6 and forms C5b-9 complexes. 5 This nephritic factor has been designated NFIIlh 6 or the nephritic factor of the terminal pathway, NFl' IS The nephritic factor originally described and found in MPGN II and partial lipodystrophy has been designated NF n,6 or NF a , the nephritic factor of the amplification loop. IS When present in high concentration, the two nephritic factors can be distinguished by rates of C3 conversion in vitro. With NFl' C3 conversion continues for up to 4 hours,5.6 whereas C3 conversion produced by high concentrations of NF a is rapid but ceases after 20 minutes. 21 At lower concentrations, differences in the rate and duration of C3 conversion are not distinguishing features;
Pathways of Complement Activation in Hypocomplementemic Glomerulonephritis Classical* Pathway Activation
Severe hypocomplementemia (C3, < 30 mg/dL) MPGN I + MPGN I, SLE, AGN§, chronic bacteremia + MPGN I, MPGN III MPGN I, MPGN III, AGN MPGN It MPGN II, partial lipodystrophy Moderate hypocomplementemia (C3, 30-86 mg/dL) MPGN I, SLE, chronic bacteremia + MPGN I, MPGN III, AGN MPGN It MPGN II 'Depression of at least one early reacting component. tDepression of C6, C7, and/or Cg. :j:Properdin usually also in normal concentration. §Observed only occasionally early in course of AGN. ,Found in three of 23 (12%) patients with MPGN I.
Terminalt Pathway Activation
Depressed C5 and Properdin
Normal C5t
+ +
+ +
+
+
204
with low concentrations of both, C3 conversion can continue for several hours. The two could be distinguished more accurately by assessing their properdin dependence, or by measuring activation of components of the terminal pathway or the production of C5b-9, and relating the results to the extent of C3 conversion under standard conditions. Unfortunately, the serum remaining after measurement of profiles in the patients in the present study was not sufficient to allow identificaiton and assay of nephritic factors. In lieu of this, hypocomplementemia characterized by low levels of C6, C7 and/or C9 is, as previouslY,7.18 assumed to indicate the presence of NF t . Depressed levels of C6, C7, and/or C9, indicating that NF t was present, were found in seven of 22 (32 %) hypcomplementemic patients with MPGN I and six of 16 (37%) with MPGN III. With a more sensitive method of detection, the frequency of NF t would be much greater. C6, C7, and C9 levels were depressed only in severely hypocomplementemic patients (none had a C3 level > 21 % of the normal mean) in whom the concentration of NF t was presumably high. Lower concentrations of NF t would presumably depress only C3, C5, and properdin levels and very low concentrations, only C3. Of the 16 hypocomplementemic MPGN III patients, six had depressed levels of C6, C7, and/or C9 combined with depressed C5 and properdin, seven had depression of C5 and properdin only, and three had, as the only abnormality, a mild to moderate C3 depression. These profiles would appear to represent the full range of those produced by varying concentrations of NF t • From the data for patients with MPGN II, it is readily apparent that NFa has a different mode of action. Severely hypocomplementemic patients lacked signs of classical and terminal pathway activation, and levels of C5 and, usually, properdin were normal. In four of seven, levels of factor B were low. We l8 and others22 have seen the same profile in patients with partial lipodystrophy and nephritis. In a patient with MPGN II, others 23 have found C5 catabolism to be normal despite markedly increased C3 catabolism. The above observations indicate that the hypocomplementemia in both MPGN II and III is unifactorial in origin, in MPGN II due to NFa and in MPGN III, to NFl' In MPGN I, the hypocomplementemia is multifactorial in origin. In eight of 22 patients, there
VARADE, FORRISTAL, AND WEST
was evidence of classical pathway activation only; in seven, evidence of NFl only; in three, evidence of NF a and in four, there was some evidence for both classical activation and NFl' The three patients with evidence for NFa were markedly hypocomplementemic, yet had normal levels of C5 and properdin, a profile not seen with NFl' All of the four with combined classical-terminal activation had definite evidence of terminal activation, but in some, the evidence for classical activation was less convincing. Patients GIOOO and WOl12 (Fig 1) had depressed C4 levels with or without depressed Clq levels, but patients HlOll and HI923 had only slightly depressed levels of Clq in the absence of a nephrotic syndrome. The multifactorial nature of the hypocomplementemia in MPGN I may be even greater than the present study indicates. Two patients have been described 24 who had NFc, the nephritic factor of the classical pathway. Both also had a C3b,Bb stabilizing factor, which was presumably NFl (both had low C5Ievels). NFc has been previously found in patients with AGN25 and SLE.26 The basis for the depression of determinants of the alpha-gamma chain of C8, found in most patients with MPGN I and III, independent of the levels of other terminal components, is not known. Although the data indicate that NFt augments the depression of C8, additional factor(s) must be responsible, since C8 levels also tend to be depressed in SLE,18 in the nephritis of chronic bacteremia l8 and, to a minimal extent, in AGN.718 In patients with SLE, the levels tend to parallel closely those of factor B (unpublished observations by authors). C8 levels were infrequently depressed in hypocomplementemic patients with MPGN II or partial lipodystrophy. 18 Hypocomplementemia and the pathogenesis of MPGN can be correlated only for profiles indicating classical pathway activation and here the correlation is not straightforward. Although it is well known that measurements of circulating immune complexes by different methods give results that correlate poorly with each other 2728 and with disease activity,2829 there is nevertheless strong circumstantial evidence that circulating complexes can produce a glomerular morphology very similar to that of idiopathic MPGN 13031 and that at least those found in SLE are sufficiently reactive with complement to produce the hypocomplementemia. 32 On this basis, the classical pathway ac-
COMPLEMENT PROFILES IN MPGN I, II, AND III
tivation observed in MPGN I must be related to pathogenesis but its precise relationship depends on whether the complexes reacting with complement are promptly cleared or persist with ultimate deposition in the glomeruli. 32 It has to be assumed that the normocomplementemic patients with MPGN I (Table 1), as well as those without evidence in their profile of classical pathway activation either had complexes that were nephritogenic but minimally reactive with complement or had complexes that had disappeared before the patient presented. The relationship between nephritic factors and pathogenesis is less clear. The present study indicates that they do not relate to the abnormalities in glomerular ultrastructure insofar as both NFl and NFa can be present in MPGN I yet the ultrastructure has no resemblance to that of MPGN II or MPGN III. The nephritic factors would be effective in generating disease only by partially depleting the plasma of complement and in this nonspecific way compromising immune complex disposal. However, our previous observations have shown no well-defined correlation between the presence of hypocomplementemia 14 or of nephritic factors 33 and the course of the disease. The evidence provided by both the complement profiles and the immunofluorescence data that circulating immune complexes are frequently present in MPGN I but not in MPGN III, reinforces evidence given previously3 that these two types are distinct entities. Whereas in the previous publication,3 MPGN I and III were considered variants of the same disease, more recent observations do not support this conclusion. Studies from this laboratory16 have shown that some patients with MPGN I and III have a common genetic basis for their disease in that 29 % have, on the short arm of chromosome 6, the extended haplotype, AI, B8, DR3, SCOl, GL02. This extended haplotype is found in only 3 % of control subjects. Because this extended haplotype has also been found with high frequency in insulin-dependent diabetes mellitus,34 glutensensitive enteropathy,35 and SLE,36 its high frequency in MPGN I and III indicates only that both types have their origin in autoimmune phenomena and not that they are necessarily the same disease or variants of the same disease. Parenthetically, the previous publication 16 indicated that this haplotype was associated with a poor prognosis and the question arises as to whether it might be associated
205
with other manifestations of disease. The fact that the 10 patients in the present study with this haplotype had widely differing complement profiles indicates that it does not determine which nephritic factor is produced or whether complement reactive immune complexes are circulating. In summary, differences in disease presentation, glomerular morphology and in mechanisms of complement perturbation give evidence that MPGN I and III are of differing pathogenesis. The pathogenesis of MPGN I, like that of SLE and the nephritis of chronic bacteremia, is based on the glomerular deposition of immune complexes, while that of MPGN III is independent of antibody and related to intraglomerular activation of the alternative pathway with deposits composed of C3, properdin, and C5. These proteins, with factor B, are the constituents of the alternative pathway C3/ C5 convertase. The general absence of factor B3738 could be attributed to the relatively low affinity of its association with C3b. 39 The factor or factors that cause the glomerulus to activate the alternative pathway should be the subject of further investigation.
REFERENCES 1. Strife CF, McEnery PT, McAdams AJ, et al: Membranoproliferative glomerulonephritis with disruption of the glomerular basement membrane. Clin Nephrol 7:65-72, 1977 2. West CD, McAdams AJ: The chronic glomerulonephritides of childhood. Part II. J Pediatr 93: 167-176, 1978 3. Jackson EC, McAdams AJ, Strife CF, et al: Differences between membranoproliferative glomerulonephritis types I and III in clinical presentation, glomerular morphology, and complement perturbation. Am J Kidney Dis 9:115-120, 1987 4. Pan C, Strife CF, West CD: The glomerular capillary as an activator of the alternative pathway of complement in nephritis. Kidney Int 35:372, 1989 (abstr) 5. Mollnes TE, Ng YC, Peters DK, et al: 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 6. Clardy Cw, Forristal J, Strife CF, et al: 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 7. Clardy Cw, Forristal J, Strife CF, et al: Serum terminal complement component levels in hypocomplementemic glomerulonephritides. Clin Immunol Immunopathol 50:307-310, 1989 8. Coleman TH, Forristal J, Kosaka T, et al: Inherited complement component deficiencies in membranoproliferative glomerulonephritis. Kidney Int 24:681-690, 1983 9. Strife CF, McAdams AJ, West CD: Membranoproliferative glomerulonephritis characterized by focal, segmental proliferative lesions. Clin Nephrol 18:9-16, 1982
206 10. Wyatt RJ, Forristal J, Davis CA, et al: Control of serum C3 levels by {3lH and C3b inactivator. J Lab Clin Med 95:905917, 1980 II. Marder HK, Coleman TH, Forristal J, et al: An inherited defect in the C3 convertase, C3b,Bb, associated with glomerulonephritis. Kidney Int 23:749-758, 1983 12, Linshaw MA, Stapleton BF, Cuppage FE, et al: Hypocomplementemic glomerulonephritis in an infant and mother: Evidence for an abnormal C3. Am J Nephrol 7:470-477, 1987 13. West CD, McAdams AJ: Serum {3IC globulin levels in persistent glomerulonephritis with low serum complement: Variability unrelated to clinical course. Nephron 7: 193-202, 1970 14. McEnery PT, McAdams AJ, West CD: The effect of a high dose alternate day prednisone regimen on the natural history of membranoproliferative glomerulonephritis. Medicine (Baltimore) 64:401-424, 1985 15. Tedesco F, Densen P, Villa MA, et al: Two types of dysfunctional eighth component of complement (C8) molecules in C8 deficiency in man. J Clin Invest 71:183-191, 1983 16. Welch TR, Beischel L, Balakrishnan K, et al: Majorhistocompatibility-complex extended haplotypes in membranoproliferative glomerulonephritis. N Engl J Med 314: 14761481,1986 17. Cameron JS, Turner DR, Heaton J, et al: Idiopathic mesangio-capillary glomerulonephritis: Comparison of types I and II in children and adults and long term prognosis. Am J Med 74:175-192,1983 18. West CD: The complement profile in clinical medicine: Inherited and acquired conditions lowering the serum concentrations of complement and control proteins. Complement Inflamm 6:49-64, 1989 19. Wyatt RJ, Forristal J, West CD, et al: Complement profiles in acute post-streptococcal glomerulonephritis. Pediatr Nephrol 2:219-223,1988 20. Falk RJ, Dalmasso Ap, Kim Y, et al: Radioimmunoassay of the attack complex of complement in serum from patients with systemic lupus erythematosus. N Engl J Med 312: 15941599, 1985 21. Vallota EH, Forristal J, Spitzer RE, et al: Characteristics of a non-complement dependent C3-reactive complex formed from factors in nephritic .and normal serum. J Exp Med 131: 1306-1334, 1970 22. Sissons JGp, West RJ, Fallows J, et al: The complement abnormalities of lipodystrophy. N Engl J Med 294:461-465, 1976 23. Sissons JGp, Liebowitch D, Amos N, et al: 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 24. Tanuma Y, Ohi H, Watanabe S, et al: C3 nephritic factor and C4 nephritic factor in the serum of two patients with hypocomplementemic membranoproliferative glomerulonephritis. Clin Exp Immunol 76:82-85, 1989
VARADE, FORRISTAL, AND WEST 25. Halbwachs L, Leveille p, Lesavre P, et al: Nephritic factor of the classical pathway of complement. Immunoglobulin G autoantibody directed against the classical C3 pathway con vcrtase enzyme. J Clin Invest 65: 1249-1256, 1980 26. Daha MR, Hazevoet HM, van Es LA, et al: Stabilization of the classical pathway convertase, C 42, by a factor (F-42) isolated from sera of patients with systemic lupus erythematosus (SLE). Immunology 40:417-424, 1980 27. Lambert PH, Dixon FJ, Zubler RH, et al: A WHO collaborative study for the evaluation of eighteen methods for detecting immune complexes in serum. J Lab Clin Immunol I: 115, 1978 28. Fust G, Kavai M, Szegedi GY, et al: Evaluation of different methods for detecting circulating immune complexes. An interlaboratory study. J Immunol Methods 38:281-289, 1980 29. Davis CA, Marder H, West CD: Circulating immune complexes in membranoproliferative glomerulonephritis. Kidney Int 20:728-732, 1981 30. Michael AF, Herdman RC, Fish AJ, et al: Chronic membranoproliferative glomerulonephritis with hypocomplementemia. Transplant Proc 1:925-932, 1969 31. Strife CF, McDonald BM, Ruley EJ, et al: Shunt nephritis: The nature of the serum cryoglobulins and their relation to the complement profile. J Pediatr 88:403-413, 1976 32. Edberg JC, Tosic L, Wright EL, et al: Quantitative analysis of the relationship between C3 consumption, C3b capture, and immune adherence of complement-fixing antibody/DNA immune complexes. J Immunol 141 :4258-4265, 1988 33. Vallota EH, Forristal J, Davis NC, et al: The C3 nephritic factor and membranoproliferative nephritis: Correlation of serum levels of the nephritic factor with C3 levels, with therapy and with progression of the disease. J Pediatr 80:947959, 1972 34. Raum D, Awdeh Z, Yunis EJ, et al: Extended major histocompatibility complex haplotypes in type I diabetes mellitus. J Clin Invest 74:449-454, 1984 35. Alper CA, Fleischnich E, Awdeh Z, et al: Extended major histocompatibility complex haplotypes in patients with gluten-sensitive enteropathy. J Clin Invest 79:251-256, 1987 36. Welch TR, Beischel LS, Balakrishnan K, et al: Major histocompatibility complex extended haplotypes in systemic lupus erythematosus. Dis Markers 6:247-255, 1988 37. Michael AF, McLean RH: Evidence for activation of the alternative pathway in glomerulonephritis, in Hamburger J, Crosnier J, Maxwell MH (eds): Advances in Nephrology, vol 4. Chicago, IL, Yearbok Medical, 1968, pp 49-66 38. Wyatt RJ, McAdams AJ, Forristal J, et al: Glomerular deposition of complement-control proteins in acute and chronic glomerulonephritis. Kidney Int 16:505-512, 1979 39. Kazatchkine MD, Fearon DT, Austen KF: Human alternative complement pathway: Membrane associated sialic acid regulates the competition between Band {3-1H for cell bound C3b. J Immunol 122:75-81, 1979