Dense Deposit Disease and C3 Glomerulopathy

Histologic Features

Electron-dense deposits are seen within the glomerulus in all forms of C3 glomerulopathy. DDD is defined by the intramembranous location of these deposits, their intensely osmiophilic appearance forming ribbons, and the associated transformation of the GBM (Fig. 1). Confirmation of DDD requires EM, although the diagnosis can be suspected with a high degree of confidence if the typical LM and IF/IHC features are present. The distinction between DDD and C3GN is sometimes difficult even using EM,^{17, 18, 19} with debates about the extent of intramembranous deposits required for EM diagnosis of DDD.²⁰ Furthermore, because the ribbons of DDD typically are discontinuous and may be more prominent in some capillary loops than others, they may be missed as a result of biopsy sampling error, leading to a diagnosis of C3GN. Such discrepancies may be important in the context of prognostic or therapeutic studies comparing outcomes in DDD and C3GN.

The ultrastructural characteristics of C3GN are less clearly defined, other than by the absence of the hallmarks of DDD (Fig. 2). In the original French C3GN series by Servais et al,¹⁰ two main histologic groups were described. The first group had typical features of MPGN type 1, with mesangial proliferation, double contours, and subendothelial, mesangial, and (less commonly) subepithelial deposits. The second group lacked mesangial proliferation, a membranoproliferative pattern, or subendothelial deposits. EM was performed in only two patients (confirming the absence of dense intramembranous deposits).¹⁰ This series recently was updated to provide data on 29 DDD patients, 56 C3GN patients, and 49 MPGN type 1 patients (both adults and children).²¹ Occasionally, the EM appearance of C3GN closely resembles that previously described as a subtype of MPGN type 3 by Strife et al.²² In CFHR5 nephropathy, histologic appearances (Fig. 2E and F) include mesangial and/or capillary wall C3 deposits on IF; a normal or mesangioproliferative pattern on LM; and mesangial, subendothelial, and occasional subepithelial deposits of only moderate density on EM. Subepithelial deposits may in fact be present in all types of C3 glomerulopathy, sometimes with so-called humps identical to those characteristically seen in postinfectious GN. In one DDD series, paramesangial subepithelial deposits correlated with low serum C3 owing to C3 nephritic factor (C3NeF) activity²³ and were composed of C3c.²⁴ Laser microdissection and mass spectrometry of glomerular tissue has shown a similar proteomic profile for DDD²⁵ and C3GN¹⁷ that includes alternative and terminal pathway complement components.

Clinical Features

DDD has an estimated prevalence of 2 to 3 per million population²⁶ and traditionally is viewed as a diagnosis of childhood and young adulthood.²⁷ Several series have shown a significantly younger mean age at diagnosis among DDD patients compared with MPGN types 1/3^{28, 29, 30} or C3GN.²¹ However, some DDD cohorts have comprised numerous older patients,³¹ including one recent US series of 18 adults and 14 children in which 39% of adults were older than age 60 at diagnosis.⁴ Although this study and several others^{30, 32, 33, 34} found that DDD was almost twice as common in females as males, some have found the opposite,^{35, 36} whilst the majority of studies have found no differences between males and females with DDD^{3, 5, 21, 27, 37, 38, 39, 40, 41} or C3GN.^{21, 42} Native renal manifestations of C3 glomerulopathy are diverse and nonspecific, with hypertension, hematuria, proteinuria (with or without nephrotic syndrome), and renal failure all reported. Most studies have found no difference in presenting features between DDD and MPGN type 1.^{28, 29, 32, 38, 40} However, nephrotic syndrome was approximately twice as common in MPGN type 1 as either DDD or C3GN in the recently updated French C3 glomerulopathy series.²¹

In CFHR5 nephropathy, persistent microscopic hematuria with or without episodes of synpharyngitic macroscopic hematuria (clinically identical to those seen in IgA nephropathy⁴³) is the usual presentation. Identical features occasionally have been reported in DDD patients,^{34, 44} in whom the clinical and histologic overlap with acute postinfectious GN suggests that intercurrent infection is important in renal injury when AP dysregulation is present. Thus, symptoms of respiratory tract infection preceding renal symptoms have been described in children (not usually adults) with DDD,^{4, 5, 17, 35, 38, 39, 40, 41, 45, 46} sometimes in association with increased anti–streptolysin O titers.^{17, 38, 41, 46}

Nonrenal manifestations of DDD include ocular drusen, acquired partial lipodystrophy (PLD), type 1 diabetes mellitus, and monoclonal gammopathy of undetermined significance (MGUS). Ocular drusen are lipoproteinaceous deposits of complement-containing debris between the basal surface of the retinal pigment endothelium and Bruch’s membrane. The clinical association of ocular drusen with DDD was first reported in 1989,⁴⁷ and appears to be related to structural and functional similarities between Bruch’s membrane, a multilaminar extracellular matrix that separates the retina from the blood vessels in the choroid, and the GBM. Ocular drusen are also a manifestation of age-related macular degeneration (AMD),⁴⁸ the leading cause of blindness in the industrialized world, for which a number of shared genetic risk factors with DDD are discussed later. However, unlike AMD, drusen in patients with DDD are said to occur at a young age and to be associated with only a modest risk of progressive visual loss.⁴⁹ In the recent therapeutic trial of eculizumab in DDD and C3GN, among several patients identified as having ocular drusen, one patient had C3GN.⁵⁰

PLD involves loss of subcutaneous fat from the face and upper body. An association with GN was first noted in 1958.⁵¹ Early reports of PLD in patients with DDD⁵² and C3NeF-associated MPGN (without EM)⁵³ were followed by several MPGN cohorts in which PLD was associated specifically with DDD.^{29, 31, 38} In the updated French C3 glomerulopathy series, PLD was noted only in the DDD group,²¹ affecting 17% of DDD patients (versus <5% in other recent DDD cohorts^{26, 54}). However, in an early PLD series comprising six patients with MPGN (all having C3NeF and low serum C3 levels), one patient had only subendothelial deposits on EM.⁵⁵ Moreover, the Necker hospital group reported two patients with PLD and MPGN in whom IF was positive for both C3 and IgG.⁵⁶ In a 2004 literature review of PLD, MPGN was the most common associated renal lesion, but only slightly more than half of MPGN cases were shown to be DDD, with MPGN types 1/3 also represented.⁵⁷ In the majority of cases of PLD, including those associated with low serum C3 levels, renal disease was not found.⁵⁷ A significantly increased risk of diabetes mellitus type 1 was reported in patients and families with DDD in a US questionnaire-based study of 98 patients (although ascertainment bias cannot be excluded).³⁹ It has been suggested that MGUS is associated with DDD^{4, 58, 59} and C3GN,^{18, 60} and that it may confer a poor renal prognosis.

Prognosis

Progression to end-stage kidney disease (ESKD), despite treatment, is said to occur in approximately half of patients with a diagnosis of DDD of 10 or more years.^{31, 34, 49, 61, 62} In the recently updated French C3 glomerulopathy series, overall 10-year progression to ESKD was 36.5% (with a mean follow-up period of 11 years).²¹ No significant difference was seen between the three diagnostic categories (DDD, C3GN, and MPGN type 1). This is consistent with the finding in earlier MPGN cohort studies that DDD was not more likely than other forms of MPGN to progress to ESKD.^{7, 29, 31, 40, 63, 64, 65} Other series found that an apparent increase in ESKD rates probably was confounded by a higher incidence of crescents^{30, 38} or reduced GFR⁶¹ at diagnosis in the DDD versus non-DDD patient groups.

The impact of clinical or histologic variables on outcomes of C3 glomerulopathy has been assessed in numerous retrospective cohort studies. In the two most recent DDD cohorts, reduced GFR at diagnosis was associated significantly on multivariate analysis with progression to ESKD.^{4, 21} In CFHR5 nephropathy, in which progressive disease is characterized by hypertension, proteinuria, and chronic renal failure, ESKD has been reported in 18 of 91 carriers (or 20%) of the heterozygous CFHR5 gene mutation. Of these, 14 were men and 4 were women, a difference that remains unexplained. No other features have been shown consistently to correlate with disease outcomes in DDD or other forms of C3 glomerulopathy. This is partly because of the limited number of end points that can be assessed in mostly small cohorts. In the updated French C3 glomerulopathy cohort,²¹ adults with DDD had significantly reduced renal survival compared with children with DDD, adults with C3GN, and adults with MPGN type 1. However, the DDD adult subgroup comprised only 11 patients, and because other differences between these subgroups were not assessed it is difficult to determine if this was a robust conclusion. For some variables, including the presence of crescents on the diagnostic biopsy specimen (found to predict a rapid progression to ESKD in some DDD cohorts^{38, 66} but not others^{4, 29, 35}), it now seems especially clear that prognostic value cannot be assessed adequately through retrospective cohort analyses. This is exemplified by the US DDD cohort study in which it was suggested that the reason crescents did not correlate with outcomes may have been the increased use of immunosuppressants in such patients.⁴

Pathogenesis

Family Studies

Renal biopsy data for a number of family studies in C3 glomerulopathy are shown in Table 1. These studies indicate a genetic basis for a small number of C3 glomerulopathy cases, although the genetic defect has been characterized and an underlying pathogenic mechanism proposed in only a handful of families. These include cases associated with homozygous Cfh deficiency or homozygous loss-of-function Cfh mutation, heterozygous gain-of-function C3 mutation, and CFHR mutations leading to enhanced Cfh deregulation (summarized later and in Figure 3). Two Algerian brothers of consanguineous parents were reported with severe Cfh deficiency and autosomal-recessive “atypical DDD” as shown on EM by discontinuous intramembranous deposits.⁴⁴ Renal disease presented during infancy with intermittent macroscopic hematuria and infections (including pharyngitis) together with failure to thrive and developmental delay in the younger sibling. Both brothers were negative for C3NeF but had very low serum levels of C3, Cfb, properdin, and C5 with normal classic pathway components. A homozygous missense mutation in SCR 7 of the Cfh gene, causing a substitution of a cysteine with a serine residue, subsequently was identified in the elder brother.⁸⁰ Both parents and two other children without renal disease had half-normal Cfh levels, consistent with heterozygosity. A young girl presenting with recurrent macroscopic hematuria and undetectable serum C3, Cfb, and Cfh was found to have C3GN with a mesangioproliferative pattern and predominant mesangial deposits on EM.⁸¹ A homozygous missense mutation was detected on SCR 16 of Cfh that resulted in a cysteine to serine change. Both unaffected parents were heterozygous for the Cfh mutation and displayed low-normal Cfh levels, while an unaffected sister had two normal Cfh alleles and high-normal Cfh levels. Intriguingly, an earlier report concerned a patient with severe Cfh deficiency in association with type 3 collagen glomerulopathy⁸² (more likely representing chronic thrombotic microangiopathy¹⁶) in whom a different homozygous missense mutation was identified at the same site on SCR 16.⁸³ The substitution of a cysteine residue (in this case with a tyrosine) altered the secondary structure of the Cfh protein, resulting in its nonsecretion.⁸⁴ A young girl was reported with macroscopic hematuria, very low C3 and Cfh levels, and endocapillary GN with predominant glomerular C3.⁸⁵ A homozygous missense mutation on SCR 2, causing an exchange of proline to serine, was apparently the result of paternal isodisomy. The patient also was homozygous for a risk haplotype for the membrane cofactor protein (MCP) gene (discussed later).⁸⁶

Table 1.

Renal Biopsy Data in Family Studies of C3 Glomerulopathy

Open in a new tab

Figure 3.

Open in a new tab

Two young daughters of consanguineous Turkish parents had C3GN with low C3, positive C3NeF, and only slightly reduced serum Cfh levels.^{87, 88} A loss-of-function homozygous mutation in Cfh involving deletion of a single codon in SCR 4 was found to produce a circulating mutant Cfh protein with impaired binding to C3b, defective decay accelerating activity, and reduced factor I (Cfi) cofactor activity. Both patients also were homozygous for the DDD/AMD-associated Y402H Cfh polymorphism (discussed later). A mother and her twin adult sons had DDD in association with half-normal C3 and Cfb levels and normal Cfh levels.⁸⁹ A gain-of-function heterozygous mutation of the C3 gene involving deletion of two codons was found to produce a hyperfunctional C3 protein. Mutant C3 was unable to be activated by the C3 convertase of the AP. However, after its conversion by proteases to C3b or by spontaneous hydrolysis of the internal thioester (C3 tickover) to C3(H₂O), it formed a C3 convertase that was resistant to regulation by Cfh. This resulted in accelerated breakdown of wild-type C3 that appeared to be restricted to the fluid phase, because regulation in vitro by decay accelerating factor (DAF, a surface-expressed complement regulatory protein) was intact. Similarly, mutant C3b and C3(H₂O) did not undergo Cfi-mediated proteolysis in the fluid phase (being dependent on Cfh cofactor activity), whereas proteolysis proceeded efficiently in the presence of MCP.

In CFHR5 nephropathy, multiplex-ligation probe amplification enabled the identification of a novel heterozygous mutation in the CFHR5 gene involving internal duplication of the initial two N-terminal SCRs in two affected families originally from the Troodos region of Cyprus.⁷⁴ The finding of a common mutation in these families suggested a founder effect, and subsequently diagnostic polymerase chain reaction enabled the identification of numerous additional affected individuals and families from all over Cyprus (some with a previous diagnosis of MPGN type 1). A total of 91 carriers of the same heterozygous founder mutation from 16 apparently unrelated Greek-Cypriot families was later reported, with a disease penetrance of 90%.⁹⁰ An Irish family was first reported with autosomal-dominant MPGN type 3⁹¹ (now reclassified as C3 glomerulopathy) and linkage to the regulators of complement activation locus.⁹² A novel hybrid CFHR3-1 gene now has been identified encoding an abnormal CFHR3-1 protein in affected family members.⁹³

Treatment

No treatment has been proven to be beneficial in C3 glomerulopathy.

Antihypertensive/antiproteinuric therapy, particularly via renin-angiotensin-aldosterone system (RAAS) blockade with angiotensin-converting enzyme inhibitors or angiotensin-receptor blockers, generally is recommended in GN.¹²⁶ Evidence in support of RAAS blockade in C3 glomerulopathy specifically is limited to retrospective cohort studies (which are subject to treatment bias). In the French C3 glomerulopathy cohort, RAAS blockade, but not use of immunosuppressants, was associated with a significant increase in renal survival,²¹ whereas in the US DDD cohort a nonsignificant reduction in progression to ESKD was found with combined RAAS blockade and steroids.⁴ After early reports of children with DDD in whom alternate-day steroids were associated with histologic regression,¹²⁷ a 1992 randomized controlled trial showed clinical benefit in 70 children with MPGN, of whom 14 had DDD (without a subgroup analysis).¹²⁸ In the absence of subsequent randomized controlled trials, reports purportedly showing response to steroids in DDD are complicated by the highly variable natural history, which includes rare occurrences of sustained clinical remission even without treatment.^{129, 130} Thus, in one pediatric series of likely C3 glomerulopathy, treatment with steroids, initiated on the basis of biopsy evidence of acute glomerular inflammation including crescents, led to clinical improvement in a number of cases.¹³¹ However, an identical improvement was seen in an untreated patient. Use of steroids and immunosuppressants may appear most justified in patients with C3 glomerulopathy who develop rapidly progressive GN (because this would prompt their use for other causes of GN). However, post-transplant recurrence of C3 glomerulopathy (including with crescents on biopsy, discussed later) and aHUS¹³² suggests that immunosuppression (as used in the transplant period) may be ineffective for AP-mediated renal disease. The iatrogenic risk of infection, potentially triggering complement activation and exacerbation of nephritis, also must be weighed. Treatment with the anti-CD20 monoclonal antibody rituximab has been reported with success in a single case of DDD related to anti-Cfb/C3b autoantibodies.⁷⁶ Immunosuppressive treatment for underlying hematologic malignancies including MGUS, in patients with associated C3 glomerulopathy, leading to improvement in renal parameters in some reports⁶⁰ but not others,¹⁸ is not discussed here.

Restoration of normal AP regulation is a major goal of current research into treatment of C3 glomerulopathy. The animal models suggest that purified or recombinant Cfh replacement might be beneficial in Cfh-deficient patients. Although this therapy is not currently available, long-term plasma infusion was effective and well tolerated in the sisters with familial C3GN related to circulating mutant Cfh.⁸⁷ Elsewhere plasma infusion/exchange has been reported with only mixed success, including in the setting of recurrent post-transplant DDD.¹³³ A therapeutic role for anti-C5/C5a therapy in C3 glomerulopathy flares is suggested by experimental data in animal models and by the observation in DDD patients with rapidly progressive GN that the glomerular deposits contain C5.²³ The C5-inhibitor eculizumab has been reported with success in several cases of DDD^{134, 135} (including a patient with post-transplant recurrence associated with renal failure¹³⁶) and one of MPGN type 1.¹³⁷ However, it also has been reported with failure in another DDD patient,¹³⁴ and with mixed results in a recent prospective, uncontrolled trial in 6 adults with native or recurrent post-transplant DDD or C3GN.⁵⁰ At the conclusion of a 1-year course of eculizumab, a clinical improvement was observed in three of the trial patients (two showing reduced inflammation on biopsy¹³⁸). In the future, agents that either inhibit the C3 convertase of the AP or remove C3 (and/or C3 fragments) from the circulation may undergo evaluation (bearing in mind the key protective functions these molecules also perform in innate immunity).

Transplantation

Histologic recurrence of DDD after renal transplantation was first reported by Galle et al¹³⁹ and is common.¹⁴⁰ Histologic recurrence of C3GN probably also was described in older studies,¹⁴¹ with recent reports^{21, 42} including cases of CFHR5 nephropathy¹⁴² and the CFHR3/1 hybrid gene (in which the related donor also later developed C3 glomerulopathy).⁹¹ A US retrospective cohort of almost 8,000 pediatric transplant recipients, including 75 children with DDD, assessed the impact of histologic recurrence of DDD on graft survival.⁶ At 5 years, 50% of pediatric DDD recipients had viable grafts (comparable with the 45% rate identified in an accompanying literature review of 78 pediatric and adult DDD recipients). This was significantly lower than the pediatric graft survival rate of 74% in all-cause ESKD and 72% in non-DDD forms of GN. Of 29 graft failures in DDD recipients during the study period, 11 were attributed to recurrent disease (9 after deceased donor transplantation). However, the contribution of recurrent disease to the overall increase in pediatric graft failure rates in DDD recipients was not assessed, except in a subgroup of 29 of the DDD recipients. Here, 4 graft losses resulting from biopsy-proven recurrent disease represented a nonsignificant reduction in median graft survival. Of note, crescents were seen on graft biopsy only in those with evidence of DDD recurrence, and were significantly associated with reduced graft survival. Biopsy-proven recurrence was always accompanied by heavy proteinuria, but no clinical or laboratory variables predicted either DDD recurrence or graft loss. The investigators concluded that there was no correlation between the isolated recurrence of dense deposits and reduced graft survival.⁶

More recent studies include an Irish cohort of 33 patients with primary idiopathic MPGN in whom graft failure rates were reported only for those with biopsy-proven recurrence.³⁰ This included 11 of a total of 18 transplants performed in recipients with (nonfamilial) DDD. The investigators concluded that 14 grafts lost to recurrent MPGN at a median of 7.5 years was likely to represent a reduction in graft half-life.³⁰ In the US questionnaire-based DDD study comprising 22 mostly pediatric transplant recipients, 10 grafts were lost within the first 5 years (7 were attributed to recurrent disease).³⁷ In one C3GN cohort, post-transplant recurrence in two individuals was heralded by increased creatinine levels, subnephrotic proteinuria, and low serum C3 levels.⁴² In CFHR nephropathy, 10 patients (mostly males) have been reported with successful transplantation (including a recipient whose female related donor was an asymptomatic carrier of the mutation), despite histologic recurrence in all three patients who underwent a biopsy.^{90, 142}

Conclusions

DDD is distinguished histologically from other forms of C3 glomerulopathy by the presence of osmiophilic, intramembranous, ribbon-like deposits. Certain clinical features, including a young age at diagnosis and (in rare cases) an association with ocular drusen, are more typical of DDD than other forms of C3 glomerulopathy. A diagnosis of DDD does not appear to confer a worse renal prognosis compared with other forms of C3 glomerulopathy. Evidence for the role of AP dysregulation in C3 glomerulopathy is drawn from linkage studies in a small number of pedigrees, cohort studies showing an association with C3NeFs and complement regulatory gene variants, and animal models. Recently, linkage was shown for copy number variation across different CFHR genes in familial C3 glomerulopathy, and a new pathophysiologic concept of Cfh deregulation has been proposed. Because progression to ESKD and post-transplant recurrence is common in all forms of C3 glomerulopathy, therapies that target underlying disease mechanisms are urgently required.^{91, 92, 93, 94, 95, 96, 97, 98, 99, 100, 101}