Alternative Pathway Diagnostics

Summary :

Uncontrolled alternative pathway activation is the primary driver of several diseases, and it contributes to the pathogenesis of many others. Consequently, diagnostic tests to monitor this arm of the complement system are increasingly important. Defects in alternative pathway regulation are strong risk factors for disease, and drugs that specifically block the alternative pathway are entering clinical use. A range of diagnostic tests have been developed to evaluate and monitor the alternative pathway, including assays to measure its function, expression of alternative pathway constituents, and activation fragments. Genetic studies have also revealed many disease-associated variants in alternative pathway genes that predict the risk of disease and prognosis. Newer imaging modalities offer the promise of non-invasively detecting and localizing pathologic complement activation. Together, these various tests help in the diagnosis of disease, provide important prognostic information, and can help guide therapy with complement inhibitory drugs.

1. INTRODUCTION

Although the alternative pathway (AP) was first described by Pillemer in 1954,¹ the last 20 years have seen an explosion in our understanding of this arm of the complement cascade.^2,3 It is now clear that uncontrolled AP activation is the primary driver of several diseases, and it contributes to the pathogenesis of many others. These discoveries have spurred the development of new therapies that specifically target this activation pathway.^4-7 Moreover, several complement proteins are unique to the AP, providing biomarkers of AP activation. In addition, multiple different defects in AP proteins (mutations, autoantibodies, endogenous antagonists) are risk factors for disease, and identification of these molecular causes of AP dysregulation provides important prognostic and mechanistic information.

2. THE IMPORTANCE OF AP DIAGNOSTICS

The complement system provides a useful source of biomarkers that can aid clinicians in diagnosing AP-mediated diseases and monitoring target coverage in patients treated with complement inhibitory drugs. In clinical trials, complement biomarkers can help in patient stratification and in monitoring target engagement. Complement activation fragments are soluble and can be easily measured in plasma or other body fluids. Cleavage of C3 and C4 also results in covalent fixation of fragments (C3b/iC3b/C3dg and C4b/C4d, respectively) to cell surfaces, providing tissue biomarkers of activation. Complement genetics identify the molecular mechanisms of complement deficiency and can help in classifying a patient’s risk of developing disease, the prognosis, the risk of recurrence, and can help in guiding therapy. These biomarkers of complement activation are clinically informative in many different autoimmune and inflammatory diseases, and they are particularly useful in conjunction with complement therapeutics. They provide biologic plausibility for using these drugs, and they can be used to monitor drug pharmacodynamics. The following examples highlight the critical role that complement diagnostics can play in various clinical settings:

• C3 glomerulopathy (C3G) is a disease of uncontrolled AP activation in the kidney. Diagnostic tests of AP activation can reveal whether this arm of the complement cascade is active and can reveal the underlying mechanism(s) of disease. Specific examples of this are discussed in section 3, below.

• Atypical hemolytic uremic syndrome (aHUS) is an AP-mediated disease, and eculizumab is currently the treatment of choice for this disease. Clinically, aHUS can be very hard to distinguish from other forms of thrombotic microangiopathy (TMA). AP diagnostics may provide evidence of a link between complement dysregulation and the disease (although it remains a diagnosis of exclusion), and these tests can help in selecting patients for prolonged treatment with complement inhibition (potentially life-long).

• aHUS frequently recurs after kidney transplantation. The risk of recurrence is largely determined by whether the underlying defect in AP regulation involves a plasma protein or a cell surface complement regulator. Identification of a genetic cause of AP dysregulation can help determine the risk of recurrence, and whether it is safe for a relative to serve as the kidney donor.

2.1. Standardization of complement biomarkers

Because of the complexity of the complement cascade, there are multiple technical challenges when interrogating the complement system. To improve the accuracy of these tests, there have been efforts to standardize clinical complement assays in laboratories around the world, and to rigorously assess the quality of these assays. The Committee for the Standardization and Quality Assessment in Complement Measurements is a subcommittee of the International Union of Immunological Societies and the International Complement Society that was created that was created to help standardize the assays and develop calibration reagents. The committee has standardized the methods for 20 assays, including measurement of complement proteins, functional assessment of complement proteins, measurement of auto-antibodies to complement proteins, and measurement of activation fragments.⁸

3. MEASUREMENT OF AP PROTEINS AND ACTIVATION FRAGMENTS

3.1. Plasma levels of complement proteins

AP activation consumes plasma C3 and factor B (Figure 1). Plasma levels of C3 and C4 have long been used to monitor activity of several autoimmune conditions,⁹ particularly immune-complex mediated autoimmune diseases.^10-12 Decreased levels of C3 in the setting of normal C4 levels suggests AP activation, and this pattern provided early clues that some diseases are primarily mediated by activation of the AP.¹³ C3 and C4 are routinely followed to help monitor complement-mediated diseases, although they have limited sensitivity and specificity due to the wide range of normal levels.

Figure 1. Protein biomarkers of alternative pathway activation.

C3b can initially be generated by through the classical pathway (CP), alternative pathway (AP), or lectin pathway (LP). C3b then combines with factor B (FB) to form the alternative pathway C3 convertase (C3bBb). Cleavage of C3 and factor B generates C3a and Ba, respectively. Consequently, alternative pathway activation is associated with a reduction of C3 and factor B levels, and a simultaneous increase in C3a and Ba levels. The alternative pathway convertase is stabilized by properdin (P), and it is negatively regulated by factor H (FH). The factor H related proteins (FHRs) can block factor H. Expression levels of these positive and negative regulators (P, FH, and FHRs) can affect the risk of some diseases and the prognosis.

3.2. Positive and negative complement regulators

Positive and negative regulators of the AP influence the probability that amplified activation will occur. Measuring the level of these proteins, therefore, provides insight into the degree to which the AP is controlled/dysregulated and can provide diagnostic or prognostic information. Factor H plays an integral role in regulating the AP, and low factor H expression is a strong risk factor for some diseases.^14-16 Complete absence of factor H is rare, but genetic studies have linked common and rare genetic variants in the complement factor H gene to several kidney diseases.¹⁷ Patients with functional variants frequently have normal antigenic levels of factor H. Measuring plasma levels of factor H, therefore, does not always reveal impairments in factor H function. It is believed that the factor H related proteins (FHRs) can serve as factor H antagonists. There is experimental evidence, for example, that they can bind to the same ligands as factor H, including glycosaminoglycans and C3b/C3d.^18-20 They may, therefore, competitively inhibit binding of factor H. Consistent with this hypothesis, high concentrations of specific FHRs are associated with faster progression of several diseases.^21-24 High plasma levels of the FHRs may have the same physiologic significance as low levels of factor H, although the FHRs are currently only measured in the research setting.

3.3. Complement activation fragments

Complement activation within tissues can cause significant injury without causing a reduction in plasma levels of the pathway components. Consequently, measurement of the activation fragments is more sensitive for detecting tissue-limited activation. AP activation generates C3a, C3b, Ba, Bb, C5a, and sC5b-9, all of which can be measured by enzyme linked immunosorbent assay (ELISA). Unfortunately, several technical factors limit the accuracy of these tests. It is important, for example, that the plasma is collected carefully and that complement fragments are not generated or degraded during sample handling and storage. Complement is activated during the clotting process. Therefore, blood should be collected directly into ethylenediaminetetraacetic acid (EDTA) to prevent activation ex vivo, and the plasma should be quickly separated and frozen.²⁵

Plasma AP fragments have been used to monitor the activity of several diseases. Ba and Bb levels are elevated in glomerular diseases, including C3 nephropathy, ANCA vasculitis, IgA nephropathy, lupus, and chronic kidney disease.^26-29 Plasma Ba levels seem to be increased in all patients who have decreased kidney function, possibly as a result of increased factor D levels.³⁰ This association complicates the interpretation of elevations of plasma Ba in patients with kidney disease. AP activation may be a general consequence of reduced kidney function and accumulation of factor D. Thus, Ba levels may increase in all diseases that affect kidney function, regardless of whether the primary disease process causes generation of Ba. Plasma Ba levels are also increased in age related macular degeneration (AMD), another AP-mediated disease.^31,32 Elevated levels of sC5b-9 is a biomarker of terminal pathway activation. It may be an indicator of prognosis in some AP-mediated diseases, although this is controversial.^33,34 Increased levels of complement fragments are clearly not specific to any one disease, however, and the tests are not widely standardized. Even in aHUS, where sC5b-9 is commonly measured, there is conflicting data regarding the sensitivity and specificity of sC5b-9 as a marker of disease activity.³⁵ Nevertheless, it can provide evidence that complement activation is ongoing, and it is increasingly used to monitor the response of patients treated with complement inhibitory drugs.

4. URINE AND OTHER BODY FLUIDS

Although complement proteins are highly abundant in plasma, their expression in other extravascular compartments is more restricted.³⁶ Detection of AP fragments in body fluids, therefore, can help localize activation to affected organs.

4.1. Urine

Given the central role that complement activation plays in many types of kidney disease, studies have examined whether complement proteins in the urine can serve as biomarkers of kidney inflammation. Complement proteins are not detected in urine from healthy individuals but are elevated in various kidney disease states.^37-42 However, several factors complicate the interpretation of these studies (Figure 2).

Figure 2. Sources of complement activation fragments in urine.

Complement activation fragments in urine can come from several sources. A) Fragments that are generated systemically can filter into the urine, particularly small fragments such as C3a, C3d, C4a, Ba, and C5a. The concentration of systemically generated proteins may also increase if reabsorption of the fragments decreases due to tubular damage. B) Complement fragments generated by activation within the glomerular capillaries can freely enter the urine. C) Intact C3 and factor B that enter the urinary space can activate within the tubular lumen. D) Fragments generated within the tubulointerstitium can also leak into the urine.

First, complement activation fragments generated outside of the kidney can filter into the urine, so their detection does not necessarily reflect intra-renal complement activation. Another possible confounder is that many glomerular diseases are associated with increased permeability of the glomerulus to plasma proteins. Once intact C3 and factor B enter the urinary space they are exposed to an acidic environment, high concentrations of ammonia, and they come in contact with the apical surface of tubular epithelial cells – all of which promote AP activation.^43-46 Therefore, AP fragments in urine of proteinuric patients may be a general consequence of leakage of C3 and factor B across the capillary wall, and are not necessarily generated by the glomerular disease process. Urine C5b-9 (uC5b-9; ~1000 kD) is probably too large to filter into the urine from plasma, even in proteinuric patients.³⁹ However, uC5b-9 may still reflect non-specific activation of component proteins (C5, C6, C7, C8, C9) that have passed through a damaged glomerulus and been activated within the urinary system. It is challenging, therefore, to distinguish whether urine complement fragments are generated systemically, by inflammation in the kidney, or simply reflect filtration of the precursor proteins into the urinary space because of glomerular damage.

uC5b-9 can be detected in most proteinuric diseases, but membranous nephropathy (MN) seems to be associated with higher levels than other diseases with comparable levels of proteinuria.^42,47,48 This makes sense physiologically, as immune-complexes deposit between the glomerular basement membrane and the podocyte in MN, so complement activation fragments can easily enter the urinary space. Studies in a rat model of MN showed that uC5b-9 reflected ongoing complement activation within in the glomerular capillaries.⁴⁹ Subsequent studies in MN patients reported that uC5b-9 levels were still elevated after correction for the level of proteinuria, and they predicted a faster decline of kidney function.⁵⁰ A similar study demonstrated that uC5b-9 decreased in treated patients, even those with persistent proteinuria, indicating that uC5b-9 was more reflective of immunologic activity than proteinuria.⁵¹ Less is known about the significance of uC5b-9 in other glomerular diseases, but there is evidence that it is a useful biomarker in IgA nephropathy^42,52 and focal segmental glomerulosclerosis.⁵³ Nevertheless, many questions remain regarding the significance of uC5b-9 levels, the optimal methods of collecting and processing the urine, and whether the results should be corrected using either urine creatinine or an index of glomerular permeability to large proteins (e.g., albumin).

In tubulointerstitial kidney diseases, complement fragments may enter the urine after the AP is activated on injured tubules. For example, the AP is activated in the tubulointerstitium of patients with acute kidney injury (AKI),⁵⁴ and urine Ba levels quickly increased in patients with AKI after cardiac surgery.³⁷ Levels were particularly high in patients who subsequently needed hemodialysis. Similar increases in urine Ba were seen in critically ill children who developed AKI, supporting its values as a biomarker of AKI across different populations and types of kidney insults.⁵⁵

Interestingly, one study found that urine C3d (~35 kD) is increased in patients who have impaired tubular function, but not in patients who have proteinuria with normal tubular function.³⁸ By calculating the fractional excretion of C3d the authors concluded that increased levels of urine C3d were simply due to decreased tubular reabsorption of C3d filtered from the circulation, not due to activation of C3 within the kidney. Thus, tubular injury may also affect the levels of some complement fragments in the urine by reducing the constitutive reabsorption of these molecules, further complicating interpretation of urine studies. Along the same lines, urine properdin may reflect infiltration of the kidney by leukocytes, and is not necessarily caused by intra-renal AP activation per se. Increased urine properdin may still be informative as a marker of kidney inflammation, though, and it has been investigated as a candidate biomarker in lupus nephritis⁵⁶ and kidney transplant failure.⁵⁷

4.2. Cerebrospinal fluid

As cerebrospinal fluid (CSF) is obtained by lumbar puncture, an invasive procedure, most samples come from patients with a suspected disease. It is difficult, therefore, to determine the normal range of complement fragments present in CSF from healthy subjects. Nevertheless, studies have reported increased levels of sC5b-9 in CSF from patients with meningitis, traumatic brain injury, demyelinating diseases, and subarachnoid hemorrhage.^58-60 CSF sC5b-9 is currently being investigated as a biomarker to rapidly distinguish between bacterial and viral meningitis. All diseases that cause disruption of the blood brain barrier likely increase the passage of plasma proteins into the CSF. Thus, a challenge in interpreting CSF sC5b-9 levels is to determine the relative contribution of blood brain barrier breakdown, production of complement proteins within the central nervous system, and the correlation of sC5b-9 levels with disease activity.

4.3. Aqueous humor

Aqueous humor is difficult to sample, and it is typically only obtained in patients who are undergoing ophthalmologic surgery. Although this limits its utility as a source of biomarkers, studies of complement fragments in aqueous humor have provided evidence that the AP is involved in some eye diseases. One study measured C3a and Ba in samples collected from 31 age-related macular degeneration (AMD) patients and 30 control subjects (patients without AMD who were undergoing cataract surgery).⁶¹ Ba and C3a were present in the control samples, but levels of both complement fragments were significantly higher in patients with AMD. The investigators did not detect elevated levels of these analytes in plasma, and they concluded that the fragments were generated by AP activation in the eye. A similar study reported that C3a and Ba were elevated in vitreous humor from patients with proliferative diabetic retinopathy.⁶² The results were normalized using the level of albumin in the vitreous humor to correct for vascular leakage of plasma proteins into the eye, so the results indicate that there is AP activation within the retinas of patients with diabetic eye disease.

5. CELL-BASED COMPLEMENT ASSAYS

The alternative pathway hemolytic assay (AH50) is a widely available clinical test. In this assay, patient serum is incubated with rabbit, guinea pig, or chicken erythrocytes (Figure 3). Human factor H does not bind erythrocytes from these species, so serum lyses the cells if there is an intact AP. The AH50 assay is used to screen for deficiency of AP proteins, as low levels of any AP component will reduce the degree of erythrocyte lysis. Plate-based assays are now commercially available that can similarly screen patient samples for normal activity of the alternative, classical, and lectin pathways.

Figure 3. Cell-based complement assays.

A) In hemolytic assays, patient serum is added to erythrocytes, and complement activity is assessed by the degree of lysis. In classical pathway assays, antibodies reactive to the cells are used to activate complement. In alternative pathway assays, erythrocytes are used to which factor H does not bind, so the cells lyse even without the addition of antibodies. B) A modified hemolytic assay can also be used to test the function of factor H. Erythrocytes to which factor H ordinarily binds are used. If patient serum lyses the cells, it may indicate defective binding of factor H to the cell surface. C) Endothelial cell assays test the ability of patient serum to deposit C5b-9 on the surface of activated cells. Formation of C5b-9 on the cell surface may indicate alternative pathway dysregulation. D) Flow cytometry assays are used to measure deposition of complement proteins on erythrocytes, platelets, or extracellular vesicles. Deposition of complement fragments on these cells and cell fragments may indicate endovascular complement activation.

5.1. Factor H functional assay

Approximately 30% of aHUS patients have factor H (CFH) mutations. In most cases, the patients have normal antigenic levels of factor H, but the mutation reduces binding of the protein to anionic molecules on cell surfaces. To detect these impairments, Sanchez-Correl et al. modified the AH50 assay.⁶³ They used sheep erythrocytes in the assay, as factor H ordinarily binds to sialic acid on the outer surface of these cells. In patients with mutations in the sialic acid-binding region of factor H (SCR 20), however, factor H fails to bind to the cell surface and AP-mediated lysis of the cells increases. A limitation of the assay is that 30-50% of aHUS patients have low C3 levels,⁶⁴ even during remission.⁶⁵ If C3 is consumed by the disease process, then there may not be sufficient C3 in the sample to support complement-mediated lysis of the erythrocyte, even in the setting of impaired factor H function. This may decrease the sensitivity of the assay.

5.2. Endothelial cell assays

A similar cell-based assay utilizes human endothelial cells (ECs) as the target surface to study dysregulated AP activation.⁶⁵ Using this method, ECs are exposed to serum from patients and C5b-9 deposition on the cells is used as a readout of AP activation on the cell surface. An advantage of the EC assay compared to the hemolysis assay described above is that it detects functional defects in all of the soluble AP components: factor H, FHRs, factor I, C3, and factor B. Interestingly, serum from patients with active aHUS causes complement deposition on quiescent ECs, whereas serum from aHUS patients in remission – but who still have functional defects in one of the AP proteins – only causes increased activation on cells that have been activated with adenosine diphosphate. This difference suggests that the assay is sensitive to some trigger of AP activation present in the serum of patients with active disease but absent in patients who are in remission. On the other hand, exposure of activated ECs to patient serum detects impaired AP regulation, even in patients who do not have active disease. Similarly, serum unaffected individuals who carry mutations in the complement regulatory proteins also causes greater complement deposition on the activated ECs. This assay is useful for detecting AP dysregulation and can also be used to monitor the efficacy of complement inhibitory drugs.

5.3. Extracellular vesicles

Extracellular vesicles (EVs) are membrane bound particles (<1 μm) that are secreted or released from cells. EVs carry cytosolic molecules as well as plasma membrane associated proteins from their cell of origin. Complement proteins on the surface of EVs can serve as markers of complement activation on the parent cell, or the EVs themselves activate the AP after release from the cell.⁶⁶ The classical pathway is activated on the endothelium of patients with antibody mediated transplant rejection, and a study showed that C4d+ plasma EVs can serve as a biomarker of this disease.⁶⁷ Similarly, in vitro studies showed that platelets exposed to serum with dysfunctional factor H release EVs, although this approach has not yet been applied to clinical samples from patients with aHUS.⁶⁸

6. COMPLEMENT GENETICS

6.1. Complement gene mutations as drivers of disease

Genetic analysis has provided invaluable evidence linking complement deficiencies and impairments in AP regulation with several diseases (Table 1).^69-71 It has long been known that aHUS is associated with perturbations in complement proteins, but the genetic studies revealed that variants in Factor H are causative factors in the development of the disease.² Over the past 25 years, several cohort studies convincingly identified variants in five biologically plausible candidate genes, all of which affect the AP C3 convertase.^2,17 Loss-of-function variants and gain-of-function variants were identified in genes for three AP regulators [CFH, CD46 or membrane-cofactor protein, and Factor I (CFI)] and in C3 and Factor B (CFB), the main components of the AP C3 convertase, respectively. The frequency of rare variants in cohorts of patients with aHUS varies from 26% to 62%, depending on the inclusion criteria,^2,72 In contrast, the frequency of rare variants in these complement genes is only moderately increased in secondary forms of HUS, including Shiga toxin producing E. coli-associated HUS and de novo TMA after renal transplantation,⁷³ arguing against constitutive AP dysregulation as a major driver of disease in these settings. Rare variants have also been reported in non-complement genes, including thrombomodulin (THBD), plasminogen (PLG), cobalamin C (CBL) and clusterin (CLU). The role of these genes in aHUS remains uncertain, as their frequency in aHUS may not be significantly more common than in healthy controls.^74,75

Table 1.

Susceptibility to disease in patients with rare variants or deletions of alternative pathway genes.

Disease	Genes	Variants	Transmission mode	Age at onset	Frequency
Infection with gram-positive bacteria (e.g., S. pneumoniae and S. pyogenes) and Gram-negative bacteria (e.g., N. meningitidis).	C3, CFB, CFD Properdin	Homozygous, two heterozygous variants	Recessive; X-linked complement deficiency (Properdin)	Childhood, Teenager and young adults (properdin)	Rare
aHUS	CFH, MCP DGKe	Homozygous or two heterozygous variants	Recessive	Childhood	Few cases
C3G/ Ig-MPGN	CFH	Homozygous or two heterozygous variants	Recessive	Teenager and young adults	1% of aHUS
aHUS	CFH, MCP, CFI, C3, CFB	Heterozygous	Sporadic/familial cases (20%)	Over the life-span	50% of aHUS
aHUS	CFHR-CFHR1 hybrids	Heterozygous	Familial/sporadic	over the life-span	5% of aHUS
C3G/ Ig-MPGN	CFH, CFI, C3	Heterozygous	Sporadic/familial cases	Teenager and adults	15-20% of C3G
C3G/ Ig-MPGN	CFH-CFHR hybrids	Heterozygous	Sporadic/familial cases	Teenager and adults	Reported cases

aHUS, atypical hemolytic uremic syndrome; C3G, C3 glomerulopathy; CFB, complement factor B; CFD, complement factor D; CFH, complement factor H; DGKe, diacylglycerol kinase epsilon; MCP, membrane cofactor protein; CFI, complement factor I.

Inherited causes of complement overactivity can also predispose patients to other AP-mediated diseases, including C3 glomerulopathy (C3G) and immunoglobulin-mediated membranoproliferative glomerulonephritis (Ig-MPGN), (Table 1). In approximately 10 to 20% of cases, patients with C3G/Ig-MPGN carry pathogenic variants in CFH, CFI, or C3.⁷⁶ Many of the variants involved in these diseases involve the human CFHR-CFH gene cluster. This region of chromosome 1 comprises the complement genes CFHR1, CFHR2, CFHR3, CFHR4, CFHR5, and CFH and is susceptible to copy number variations, translocations, deletions, or duplications. More than 20 different hybrid genes in this region have been reported in aHUS and C3G. Most of the reported hybrid genes are private (found in a single individual or family), but CFHR5 nephropathy is a unique subtype of C3G caused by duplication of exons 2 and 3 of the CFHR5 gene.⁷⁷ This duplication causes altered function of the CFHR5 protein and is endemic in Cyprus.

Although less is known about the genetic risk of other AP-mediated diseases, polymorphisms in the CFH, CFB, and C3 genes appear to exhibit strong associations with the risk of developing AMD. Pathogenic Factor I variants have also been associated with an increased risk of AMD and may predict a response to treatment with complement inhibition.^78,79 Large-scale genome-wide association studies have also linked variants in the gene for complement receptor-1 (CR1) with Alzheimer’s disease,⁸⁰ and deletion of the CFHR1 and CFHR3 genes is associated with decreased risk of developing IgA nephropathy.⁸¹

6.2. Genetic screening in patients with AP-mediated disease

Genetic screening has entered mainstream clinical practice and provides important prognostic information that guides clinical decision making for individual patients with aHUS. Identification of molecular defects in complement genes helps to confirm the diagnosis and is invaluable for family planning. Specific AP gene variants also predict aHUS disease severity, recurrence after renal transplantation, and recurrence after anti-C5 treatment discontinuation.^73,82-84 For patients suspected of having aHUS, the screening genetic tests should include next generation and or Sanger sequencing of the five complement genes (CFH, C3, CFI, CFB, and CD46) and multiplex ligation-dependent probe amplification (MLPA) to identify CFHR1/CFH hybrid genes (caused by non-allelic homologous recombination).

Genetic screening is commonly performed in patients with C3G. This analysis is not necessary for the diagnosis of this disease, however, which is based on immunostaining and electron microscopy of the kidney biopsy. In addition, there are not currently any approved treatments for the C3G, so genetic testing does inform treatment decisions. Similarly, genetic screening is not necessary in patients with AMD, as it does not inform the diagnosis or management of these patients.

6.3. The difficulty in classifying rare variants

In the near future, extensive genetic testing will likely become the standard of care for patients with kidney disease and complement genes may be implicated in the complex architecture of several more diseases. A clear positive result proving the link of the variant with a dysregulation of the AP can lead to changes in clinical management (e.g., discontinuation of steroids), can help inform prognosis (e.g., post-transplantation outcomes), and may reduce the need for invasive procedures (e.g., kidney biopsy).^73,85 However, interpreting complement gene sequencing can be quite challenging. As a general rule, the variants underlying aHUS or C3G are rare, with a minor allele frequency (MAF) below 0.1% (Figure 4). They typically impact the formation (C3 and CFB gain-of-function variants) or the regulation (CFH, CFI and MCP loss-of-function variants) of the AP C3 convertase. The detection of a rare variant does not necessarily indicate complement dysregulation, however, and not all detected gene variants are clinically relevant. Thus, gene variants are classified along a gradient ranging from an almost certainly pathogenic role to those that are very likely benign. As genetic analysis is performed on more and more patients, an increasing number of variants of uncertain (or unknown) significance (VUS) will be identified but whose relation to the function or health of an organism is not known. The clinical relevance of individual variants should be regularly re-evaluated as new information is obtained.

Figure 4. Contribution of alternative pathway (AP) gene variants to disease.

Common variants in complement genes contribute to several common, polygenic diseases. Lower frequency variants contribute to several rare diseases. In general, the rarer the variant, the greater penetrance of disease in affected individuals. aHUS, atypical hemolytic uremic syndrome; C3G, C3 glomerulopathy; AMD, age related macular degeneration; AD, Alzheimer’s disease; IgAN, IgA nephropathy.

Numerous lines of evidence are used to assess the probability that a genetic variant is relevant to the disease, as outlined in guidance from the American College of Medical Genetics and Genomics (ACMG).⁸⁶ Analysis of functional alterations in complement proteins takes into account the level of expression of the encoded protein (in plasma for CFH and CFI and at the granulocyte surface for CD46), the impact of the variant on the function of the encoded protein (assessed using in vitro assays performed in specialized laboratories) and prediction of the pathogenicity of a variant based on involvement of functional domains and in silico analyses. Certain types of variants (e.g., nonsense, frameshift, canonical +/− 1 or 2 splice sites, initiation codon deletion, single or multi-exon deletion) can be assumed to disrupt gene function by causing a complete lack of transcription or by inducing nonsense-mediated decay of an altered transcript.⁸⁷

In aHUS and C3G, the majority of unique rare variants are point genetic variants leading to missense changes. For example, of the 216 rare unique CFH variants reported in one study, 64% were missense.⁷⁴ Functional studies can be a powerful tool in support of pathogenicity; however, not all functional assays accurately predicting the impact on protein function in vivo.^88,89 In silico prediction of the consequences of amino acid changes can also be attempted, but in cases of CFH and CFB variants these have proved unreliable.⁷³ Thus, a significant proportion of variants identified in these patients are of uncertain biologic and clinical relevance. Conversely, incomplete penetrance in carriers of pathogenic variants is common. Genetic data are now available for >140,000 individuals from diverse populations in the Genome Aggregation Database (gnomAD) (Table 2). Rare variants with MAF <0.1% were identified in 30 of the 503 European controls, five of which were pathogenic variants with demonstrated functional deficiency.⁹⁰ Thus, identification of a rare variant is not equivalent to finding a disease-causing association.

Table 2.

Frequency of complement rare variants in the European-American population.

	1000 Genomes Project* (N=503)	%
Rare Variant (MAF<0.1%)	30	6.0
Pathogenic variant (P)	5	1.0
VUS	25	5.0
CFH	8	1.6
Homozygous	0	0.0
Heterozygous	8	1.6
Pathogenic	2	0.4
VUS	6	1.2
CFI	7	1.4
Homozygous	0	0.0
Heterozygous	0	0.0
Pathogenic	2	0.4
VUS	5	1.0
MCP	2	0.4
Homozygous	0	0.0
Heterozygous	2	0.4
Pathogenic	1	0.2
VUS	1	0.2
C3	13	2.6
Pathogenic	0	0.0
VUS	13	2.6

MAF, minor allele frequency; VUS, variant of uncertain significance; CFH, complement factor H; CFI, complement factor I; MCP, membrane cofactor protein.

^*These data derive from European cohort in the Genome Aggregation Database (gnomAD). The frequency of variants might vary markedly between populations.

Misinterpretation of the results can have important consequences for the patients and their relatives, causing incorrect estimation of the risk of disease transmission. Some rare genetic variants can markedly increase the risk of AP-mediated diseases. More commonly, however, the overall genetic risk is determined by the cumulative effects of many common genetic variants, each of which has a small effect by itself. It is now established that the overall genetic predisposition of an individual to AP-mediated diseases results from the combination of rare and common variants. Complement gene haplotypes, mainly homozygous CFH-H3 or/and MCP ggaac (present in ~5% of heathy individuals), along with triggers (infections, pregnancy) also contribute to the development of aHUS.^72,91,92

7. COMPLEMENT AUTOANTIBODIES AND “NEPHRITIC FACTORS”

Acquired complement abnormalities due to autoantibodies against specific components of the AP are associated with C3G, MPGN and aHUS. Pathways leading to autoantibody-induced pathology differ among diseases, and autoantibodies directed against the same protein can have diverse effects depending on the targeted epitope. In clinical practice, screening for autoantibodies is a useful diagnostic tool in these diseases.^76,93

7.1. Nephritic Factors

In more than 50% of cases C3G is associated with autoantibodies to the AP C3 convertase, the so-called C3 Nephritic Factors (C3NeF).⁹⁴ C3NeF are IgG that recognize a neo-epitope on the C3bBb complex but are not able, by definition, to bind C3b or Bb separately. C3NeF impair the ability of Factor H, CR1 or DAF to accelerate the decay of fluid-phase and cell-bound alternative pathway C3 convertases. In the fluid-phase this increases activation of the AP and thus more frequently decreases levels of plasma C3. The effect of C3NeF on the terminal pathway (i.e., increased levels of sC5b9) is variable.

Analysis of C3NeF is an important part of the diagnostic workup of C3G. Their identification is difficult, however, likely due to considerable heterogeneity among patients. There is also no standardization for the detection of C3NeF and the results can be influenced by the methods of screening. The gold standard for detecting C3NeF is a hemolytic assay using C3b-opsonized sheep erythrocytes, to which factor B and factor D are added to create C3 convertase on the cell surface.⁹⁵ C3NeF stabilize the convertase, and the degree of lysis is used as a functional readout. However, this sensitive test is only performed in a limited number of specialized immunology laboratories. A more commonly used assay examines the breakdown of C3 after it is mixed with patient serum. C3NeF in the serum increases the cleavage of C3, which is then analyzed by two-dimensional immune-electrophoresis, immunofixation electrophoresis, or Western blotting. Finally, some labs use a method in which a solid-phase C3bBb is generated on the bottom of an ELISA plate using NiCl₂, which stabilizes the C3b and C3bBb complexes.⁹⁶ Patient samples are added to the wells, and binding of IgG to the convertase is examined by ELISA. This assay is simpler to perform than the hemolytic assay, but it does not provide functional proof that the antibody stabilizes the convertase.

Another type of nephritic factor stabilizes the C5 convertase of the AP (C5NeF).⁹⁷ The assay for C5Nef uses sheep erythrocytes coated with C3bBb. Properdin is then added to samples and mediates conversion of the C3bBb to a C5 convertase (C3bBbP). Stabilization of the convertase is then detected by increased hemolysis of the cells after terminal complement components are added to the reaction. C5NeF has been detected as a unique nephritic factor in some patients (10% of cases) or as coexisting with C3NeF (39% of cases). Patients with C5NeF may also be more likely to develop C3 glomerulonephritis than they are to develop dense deposit disease, indicating that it may have a physiologic role that is distinct from C3NeF.⁹⁷

7.2. FH autoantibodies

Auto-antibodies directed against FH are found in 5% to 56% of aHUS patients and less frequently in C3G. These autoantibodies are more common in children and may be more common in certain ethnic groups.⁹⁸ In aHUS, anti-factor H antibodies primarily recognize the C-terminal region of the protein and interfere with its ability to bind cell surfaces.^99,100 Anti-factor H antibodies are strongly associated with the CFHR1-CFHR3 homozygous deletion in aHUS patients.¹⁰¹ In C3G, on the other hand, the anti-factor H antibodies bind to the N-terminal domains of factor H, impairing the regulatory functions of the protein.¹⁰² The screening for anti-factor H antibodies is performed by ELISA and the results are expressed in arbitrary units. It is best practice to screen all aHUS and C3G patients for factor H autoantibodies, and efforts ha e been made to standardize these assays.¹⁰³ However, there is no consensus for the reporting of factor H autoantibody titers, nor does detection of these antibodies currently guide treatment of these diseases.

7.3. Anti-factor B antibodies

Antibodies directed against factor B were first reported in a DDD patient in 2010.¹⁰⁴ The patient had a low plasma C3 level and was negative for C3NeF activity measured by hemolytic assay. The anti-factor B antibody in this patient bound to the Bb fragment by ELISA, and it stabilized the C3 convertase. Anti-factor B antibodies were subsequently identified in additional C3G patients, and investigators proposed that they may have C3NeF-like activity.¹⁰⁵ More recently, Chauvet et al. reported that 91% of children with acute post-infectious glomerulonephritis (APIGN) had anti-Factor B auto-antibodies versus 14% of children with hypocomplementemic-C3 glomerulopathy.¹⁰⁶ In APIGN, anti-factor B autoantibodies were transient. The antibodies correlated with depressed levels of plasma C3, and with increased levels of sC5b-9. The anti-factor B antibodies enhance the activity of AP convertases in vitro, demonstrating that they can cause AP dysregulation. Further studies are needed to determine the prognostic value of screening for anti-factor B antibodies in APIGN.

7.4. Anti-C3b autoantibodies

Autoantibodies against C3b (IgG) are detected in up to 30% of the patients with lupus nephritis and more rarely in cases with C3G.^107,108 These antibodies recognize epitopes shared between C3(H₂O), C3b, iC3b, and C3c. They enhance the formation of the C3 convertase and prevent the inactivation of C3b by Factor H. It has not clearly been established that these antibodies are pathogenic or correlate with disease severity, and further studies are warranted to explore the link between anti-C3b autoantibodies and tissue injury.

7.5. FI autoantibodies

Auto-antibodies directed against FI have been reported in three aHUS patients. Their significance in disease pathogenesis is unclear.¹⁰⁹

8. TISSUE IMMUNOSTAINING

Biopsy samples are frequently stained for complement proteins, including C3c (representing both C3b and iC3b), C1q, and C4d. The detection of these proteins is interpreted as evidence of complement activation within the tissue that has been biopsied. Furthermore, complement activaiton leads to rapid deposition of millions of complement molecules. Because C3 and C4 fragments are covalently attached to target surfaces, they are also durable markers of complement activation. For example, the diagnosis of antibody mediated transplant rejection is based on detection of C4d deposits, not IgG, even though the C4d is interpreted as evidence of antibody-mediated injury.¹¹⁰ Investigators have utilized both immunostaining and mass spectrometry to detect AP proteins within biopsy samples. These methods have revealed factor B fragments, properdin, factor H, and factor H related proteins within affected tissues.^111-113

Although standard clinical analysis of biopsies does not include stains for these AP proteins, investigators have used the presence or absence of C4d deposits to draw conclusions as to whether C3 deposits are the result of AP activation.⁵⁴ C4d staining has been proposed as a means of distinguishing immune-complex mediated glomerulonephritis (in which complement activation is presumably initiated through the classical pathway) from C3 glomerulopathy (in which activaiton may be initiated through the alternative pathway).¹¹⁴ However, not all studies have supported the use of C4d to make this distinction.^115-117 AP-mediated diseases can initially be triggered by activation of the classical pathway. C3G, for example, can become activate after infections cause deposition of immune-complexes in the kidney. C4d deposits are also frequently seen in normal glomeruli, complicating the interpretation of C4d staining in diseased kidneys.¹¹⁸

9. IMAGING BIOMARKERS

As mentioned above, tissue biopsies are routinely stained for deposits of C3c, C4d and C1q, and these analyses can provide conclusive evidence that the complement system has been activated within a given tissue. Inflammatory diseases are often heterogeneous, however, and biopsies are susceptible to sample error. Furthermore, the invasive nature of biopsies limits their use, particularly in hard to reach organs or in patients at risk of bleeding. Newer imaging methods for non-invasively detecting complement fragments within tissues offer several potential advantages. First, they report on complement activation throughout an organ, or even througout the entire body. Furthermore, repeat studies can be used to monitor disease activity or the response to treatment.

C3 fragments provide an informative imaging target for several reasons. On a molar basis, C3 is probably the most abundant complement proteins deposited within inflamed tissues, and C3 fragments are covalently bound to target surfaces. Detection of tissue C3 deposits is central to the diagnosis of C3G, but it also provides information about the activity and prognosis of many other diseases.^119,120 A major challenge to developing probes for detecting C3 fragments, however, is that such a probe has to discriminate the tissue-bound fragments from precursor proteins in plasma.

Complement receptor-2 (CR2) is expressed on B cells and follicular dendritic cells, and specifically binds the C3d activation fragment. We used a recombinant form of this protein to deliver superparamagnetic iron-oxide (SPIO) nanoparticles to tissue sites of complement activation.¹²¹ Within a tissue, SPIO negatively enhance T2-weighted magnetic resonance images (i.e., it decreases the “T2 relaxation time”). In mice with lupus-like kidney disease, injection of CR2-targeted SPIO caused a significant reduction in the T2 relaxation time in the kidney, and the signal observed after injection of the probe into diseased animals was significantly greater than that seen in control mice. A follow-up study showed that the magnitude of this signal correlated with disease severity.¹²² A drawback to this method, however, is that it requires that the magnetic resonance procedure be performed before and after injection of the targeted SPIO. Investigators have also radiolabeled recombinant CR2 with ^99mTechnetium.^123,124 In a model of cardiac ischemia-reperfusion injury, they showed that the probe accumulated in the damaged heart and could be detected by single photon emission computed tomography (SPECT) imaging and only one imaging procedure was needed.

To develop higher affinity probes, we generated monoclonal antibodies specific for tissue-bound iC3b and C3d.¹²⁵ When injected systemically, the antibodies target tissue sites of inflammation. In a mouse model of laser-induced choroidal injury, fluorescently labeled antibody was detected within the retinal lesions of live mice by optical imaging. In a subsequent study the anti-C3d antibody was radiolabled with ¹²⁵I and injected into mice with Mycobacterium tuberculosis infection.¹²⁶ Strong signal was seen in granulomas of the lungs and spleens. We have also labeled the antibody with ¹²⁴I and used it as a positron emission tomograph (PET) probe. When this probe was injected into factor H-deficient mice (a model of C3G), we saw strong PET signal within the kidneys of affected animals.¹²⁷ Surpisingly, we also saw strong signal in the livers of these mice, and immunostaining confirmed that the AP was also activated in the liver sinusoids. This observation showed the value of whole body imaging for discovering unexpected targets of dysregulated AP activation. Overall, these studies show that complement activation within specific tissues can be detected using several different molecular imaging modalities. In the future, these methods may provide clinicians with the ability to monitor and quantify complement activation throughout the body.

10. CONCLUSIONS AND FUTURE DIRECTIONS

Because of the complexity of the complement cascade, complement activation perturbs the levels of many different plasma proteins. It consumes substrate proteins while simultaneously generating soluble and tissue-bound activation fragments. Measurement of these proteins provides insight into whether a patient has ongoing inflammation. Furthermore, complement genes and expression levels determine patients’ risk of many different AP-mediated diseases. Analyses of complement biomarkers have given rise to many different diagnostic assays, many of which are routinely used in clinical medicine. For some diseases, AP diagnostic tests provide fundamental diagnostic criteria, essential information about patient risk and prognosis, and they serve dynamic readouts of disease activity.

The need for accurate methods of detecting AP dysregulation and for monitoring complement activation/inhibition will likely get even greater as more and more complement therapeutics enter the clinic. However, all of the currently available AP diagnostics have technical or biologic limitations. For example, the variability in expression levels and the instability of complement proteins limit the diagnostic accuracy of plasma measurements. Furthermore, the AP is activated in a wide range of different diseases, and AP-related biomarkers are not disease specific. Nevertheless, AP diagnostics can identify subsets of patients in whom the AP is activated (e.g., complement-mediated TMA), and they can provide important information about disease activity in patients with a known diagnosis. Complement imaging may represent the “next generation” of complement diagnostics – allowing clinicians to safely and accurately monitor complement activation throughout the body or in specific organs. This may provide diagnostic information that is not available with the current tests. Even with these new imaging tools, however, comprehensive analysis of the complement system will probably require a combination of protein measurements, genetic analysis, and imaging and/or tissue analysis.

DATA AVAILABILITY STATEMENT

Data sharing is not applicable to this article as no new data were created or analyzed in this study.