Complotype affects the extent of down-regulation by Factor I of the C3b feedback cycle in vitro

Abstract

Sera from a large panel of normal subjects were typed for three common polymorphisms, one in C3 (R102G) and two in Factor H (V62I and Y402H), that influence predisposition to age-related macular degeneration and to some forms of kidney disease. Three groups of sera were tested; those that were homozygous for the three risk alleles; those that were heterozygous for all three; and those homozygous for the low-risk alleles. These groups vary in their response to the addition of exogenous Factor I when the alternative complement pathway is activated by zymosan. Both the reduction in the maximum amount of iC3b formed and the rate at which the iC3b is converted to C3dg are affected. For both reactions the at-risk complotype requires higher doses of Factor I to produce similar down-regulation. Because iC3b reacting with the complement receptor CR3 is a major mechanism by which complement activation gives rise to inflammation, the breakdown of iC3b to C3dg can be seen to have major significance for reducing complement-induced inflammation. These findings demonstrate for the first time that sera from subjects with different complement alleles behave as predicted in an in-vitro assay of the down-regulation of the alternative complement pathway by increasing the concentration of Factor I. These results support the hypothesis that exogenous Factor I may be a valuable therapeutic aid for down-regulating hyperactivity of the C3b feedback cycle, thereby providing a treatment for age-related macular degeneration and other inflammatory diseases of later life.

Introduction

During recent years it has become apparent that there are multiple polymorphisms among complement components which affect their function. The first of these to be described 1 was the C3 F/S polymorphism (R102G), where the F allele (102G) was shown to have increased functional activity 2. It was shown much later that this increase in function is due to the decreased affinity of C3 convertases, including C3F for Factor H 3. Neither allele is hypomorphic, so there is no difference in C3 levels due to the polymorphism 4 and data below.

In the last decade there has been a veritable explosion in the discovery of complement allotypes and their association with disease. These are reviewed by Harris et al. 5, who coined the term complotype ‘to represent the pattern of genetic variants in complement genes inherited by an individual which alters risk for both inflammatory disorders and infectious diseases involving complement’. Certain complotypes predispose very significantly to a group of inflammatory diseases, of which age-related macular degeneration is much the most common. Some of the risk alleles are also associated with atypical haemolytic uraemic syndrome and a group of other kidney diseases, now known collectively as C3 glomerulopathies 6, and quite possibly also to other inflammatory diseases of later life, including atherosclerosis 7 and systemic vasculitis 8.

The significant variant alleles that predispose to disease all share the property of increasing the activity of the C3b feedback cycle, which amplifies all complement pathways irrespective of triggering stimulus 9. Significant alleles that occur at reasonable frequency include C3F, loss of function alleles in Factor H (62I and 402H). They also include a gain of function allele in Factor B (R32 W/Q), but because of the ascertainment difficulties described below it was not possible to include this polymorphism in the present study.

These hyperactive complotypes are likely to be advantageous in conferring resistance against infection in early life but in later life, once immunoglobulin (Ig)G antibodies have been formed to the infecting bacteria, they become disadvantageous in predisposing to inflammatory diseases. There is therefore considerable interest in finding methods for reversing this hyperactivity of the C3b feedback cycle.

The C3b feedback cycle is discussed in detail by Lachmann 9. Its rate is essentially controlled by the balance of two reactions: the positive feedback reaction, where C3b reacts with Factor B, which forms the alternative pathway C3 convertase after the Factor B is cleaved by Factor D; and the C3b breakdown reaction, where C3b reacts with Factor H [or some other proteins that have co-factor activity for Factor I (CD46 and CR1)], which allows cleavage of C3b to iC3b by Factor I. In the total absence of Factor I the breakdown reaction cannot take place, and all C3 is converted to C3b by the positive feedback reaction. However, raising the level of Factor I above the physiological level then has the reverse effect of reducing complement activation by accelerating the cleavage of C3b to iC3b. In the presence of CR1 there is also further cleavage of iC3b to C3c and C3dg. This final Factor I cleavage is important, as it destroys the proinflammatory effects of iC3b.

It has long been known 10,11 that raising the serum concentration of Factor I and/or of Factor H can down-regulate the feedback cycle, but from a therapeutic perspective elevating Factor I is more likely to be effective. There are a number of reasons for this.

First, the human serum concentration of Factor I is much lower than that of Factor H (reported around 35 μg/ml and 250 μg/ml, respectively). Less Factor I protein would therefore be required to produce an effect.

Secondly, in some diseases and tissue sites, Factor H function is blocked by binding of dimers of Factor H-related proteins to cell surfaces for which they have a much enhanced avidity 12. This evidence indicates that this is a dominant pathogenic effect and increasing Factor H concentration is unlikely to overcome the effect on complement amplification.

The third and most important reason is that the main inflammatory mediator produced by complement activation, iC3b, is produced by Factor I in the breakdown cycle. iC3b reacts with the complement receptor CR3 on neutrophils triggering cell activation and phagocytosis. Augmenting concentrations of Factor H will lead to the formation of more iC3b in the first instance, and will not accelerate the breakdown of iC3b that is already present. Conversely, raising Factor I levels will both remove C3b from the feedback cycle and enhance the conversion of iC3b to C3dg (9 and data in this paper). We are therefore pursuing the possibility of developing Factor I as a therapeutic aid, and this is occurring against the background of the continuing therapeutic development (by King’s College London) of Mirococept – a truncated CR1 derivative 13 – as a co-factor for the conversion of C3b to iC3b and of C3b to C3c and C3dg 9.

CR1 is the only co-factor that allows Factor I cleavage of iC3b to C3dg under physiological conditions 9, because the affinity of Factor H for iC3b is too low for it to act as co-factor. At reduced ionic strength, however, Factor H can act efficiently as co-factor for the cleavage of iC3b. When the breakdown of iC3b in serum, carefully cleared of cell fragments, is looked at after many hours of incubation it is not possible to distinguish between minute contamination with soluble CR1 and extremely inefficient co-factor activity of Factor H – or both. Where iC3b has been fixed on red blood cell CR1 in vivo [for example, in immune complex diseases such as systemic lupus erythematosus (SLE) or in cold autoantibody haemolytic anaemia] the CR1 on the cells is removed by proteolysis in the reticulo–endothelial system 14, and it is feasible that traces of this could appear in the serum.

As a preliminary investigation, in-vitro complement activation studies have been carried out on serum samples from subjects of three different complotypes, using the canonical alternative pathway activator – yeast cell walls (zymosan). The effect of Factor I supplementation on iC3b formation and its subsequent breakdown to C3dg was measured. Most experiments were performed without adding a source of CR1. This is normally present on human erythrocytes, but the combination of a particulate activator and erythrocytes were problematic in the assay used. In some later experiments a soluble CR1 derivative (Mirococept) was used to provide an efficient co-factor for the cleavage of iC3b to C3dg by Factor I.

Materials and methods

Sera studied

Sera were sourced through the Cambridge BioResource (REC reference 04/Q0108/44, http://www.cambridgebioresource.org.uk/). DNA samples from the database were genotyped for three single nucleotide polymorphisms (SNPs), one in C3 and two in Factor H, where there are known associations with susceptibility to age-related macular degeneration and/or atypical haemolytic uraemic syndrome and/or C3 glomerulopathies 3–6. Factor B polymorphism is characterized by HindIII restriction fragment length polymorphism (RFLP). SNPs in linkage disequilibrium with the HindIII sites can be used to tag this site. We used rs12757487 within the linkage disequilibrium (LD) region to tag the HindIII site. However, these SNPs could not be investigated by this method, as they failed to genotype on a number of different genotyping platforms. Table 1 shows the rs numbers and the nucleotide and amino acid variation that made up the final selection criteria, as well as the number of subjects tested and the numbers in each of the selected complotype groups. Volunteers from the Cambridge BioResource were subsequently assigned to three groups according to their genotype at these three SNPs. These are shown in detail in Table 1. Briefly, group A were subjects who are heterozygotes at all three SNPs (and make up 5·6% of the population studied); group B were subjects who are susceptible homozygotes at all three SNPs (and make up 0·6% of the population studied); and group C were subjects who are protective homozygotes at all three SNPs (and made up 3·3% of the population studied).

Table 1.

Sera studied

rs number	Observed	Variant	Type	Description		Minor allele frequency
				Minor allele	Number tested	Observed	Expected
rs800292	c/t SNP	Factor H	V62I	t (I) protective	10 658	0·232	0·213
rs1061170	c/t SNP	Factor H	Y402H	c (H) susceptible	10 551	0·383	0·42
rs2230199	c/g SNP	C3 SF	R102G	g (G) susceptible	10 586	0·206	0·19
Number with 3 genotypes		9989
Genotype C	Homozygous protective at all three loci	331	3·3%
Genotype A	Heterozygous at all three loci	558	5·6%
Genotype B	Homozygous susceptible at all three loci	55	0·6%

SNP = single nucleotide polymorphism.

Equal numbers of volunteers from each group were invited to participate in the study and donate a fresh sample of blood for sera isolation. Written informed consent was received from all volunteers who participated in the study under approved ethics application (REC reference: HBREC 2012·06). Consent was received in all cases by Cambridge BioResource research nurses. The blood was taken and the sera were separated as described 15 and stored at −80°C in small aliquots until they were used.

The experimenters remained blind to genotype throughout the study until all the sera had been collected and results had been obtained.

Quantitation of Factor I in the serum samples

The concentration of Factor I was measured in 61 sera using commercially available immunodiffusion plates (Bindarid plates from The Binding Site, Edgbaston, Birmingham, UK). The assay was recalibrated using as standards the Factor I purified from plasma and supplied by CompTech (Complement Technology, Inc., Tyler, TX, USA) and the recombinant Factor I prepared by GlaxoSmithKline (Stevenage, UK). It should be noted that there are differences in the way that the concentrations of Factor I can be expressed. The extinction coefficient at 280 nm can be calculated from the amino acid composition to be 1·53 for a solution of 1 mg/ml. This gives the concentration of the ‘core protein’, but ignores the 13·7% carbohydrate in the molecule. For the intact protein, the optical density (OD) used for calibration is 1·35 for 1 mg/ml.

Quantitation of C3, Factor H and Factor B in the serum samples

Antigenic quantitation of C3, Factor H and Factor B in the sera was performed by Dr Sanja Ugrinovic at the Clinical Immunology Laboratories at Cambridge University Hospitals NHS Trust using standard nephelometric techniques.

Factor I

Two sources of Factor I were tested: Factor I purified from plasma prepared by CompTech, and recombinant Factor I prepared by GlaxoSmithKline. The two preparations gave identical activity to protein ratios as measured by conglutination assays (data not shown). All the clone 9 capture assays reported in this paper were performed using the recombinant Factor I.

Typically, a serum from each genotype (A, B and C, 18 sera in each group) was tested alongside each other on any one experimental day. The clone 9 antibody 16,17 recognizes C3g, which appears as a neo-antigen in iC3b and in C3dg but is not exposed in native C3 or in C3b. Therefore, these plates will capture iC3b or C3dg from the reaction mixture. The clone 4 detection antibody (a monoclonal anti-C3c) 17 reacts with a conformational epitope in C3c and reacts with C3, C3b, iC3b and C3c. Therefore, only bound iC3b is detected by this antibody on the clone 9-coated plate. When iC3b is converted to C3dg the C3c is released and with it the ability to be detected by the biotinylated-clone 4 antibody.

Using this assay, we were able to investigate the effect of adding Factor I to sera of the three different genotypes and also additionally adding a co-factor for the third Factor I cleavage. For this purpose we used Mirococept (previously known as APT070), which contains the first three short consensus repeats (SCRs) of CR1 bound to a membrane-interactive amphiphilic peptide ‘tail’ 13. CR1 and this product derived from it are essential co-factors that allow Factor I to cleave iC3b to C3dg and C3c. They also share with other complement control proteins (CD46 and Factor H) co-factor activity for the Factor I conversion of C3b to iC3b. There is very little, if any, soluble CR1 present in sera, as is known from the long half-life (more than 18 h) of EAC43bi cells incubated in human serum before they lose their (iC3b requiring) reaction with conglutinin 16. This is unlike the situation in whole blood, where there is CR1 present on the red cells.

Experimental protocol

Stage 1

The sera were thawed rapidly at 37°C, vortexed briefly, and then placed on ice. The sera were then treated with zymosan with the addition of recombinant Factor I (to 25–50 μg/ml) and/or Mirococept (APT070) (to 10, 20, 40 or 80 μg/ml). Zymosan (Sigma, Gillingham, Dorset, UK) was diluted in ‘alternative pathway buffer’ [veronal-buffered saline pH 7·2 containing 10 mM ethylene glycol tetraacetic acid (EGTA) and 2 mM Mg]. The soluble additives were added to 500 μl aliquots of serum prior to the addition of zymosan to a final concentration of 5% and rotated in a 37°C incubator. At selected time-points, 40 μl of each sample was removed into 10 μl of 50 mM ethylenediamine tetraacetic acid (EDTA). The sampling times were: 0, 30, 60, 120, 180, 240, 480 min. At each time-point the samples were microfuged to remove the zymosan and frozen at −80°C until tested by ELISA.

Stage 2: coating

Nunc Maxisorb microtitre plates were coated with a clone 9 anti-C3g antibody (50 μl/well clone 9 at 25 μg/ml in 0·1 M carbonate/bicarbonate coating buffer pH 9·5), incubated overnight at 4°C, followed by five washes in phosphate-buffered saline (PBS)/0·05%Tween 20. The plates were blocked with PBS/0·05%Tween 20/4% dried skimmed milk (2 h at room temperature), then washed again as above.

Stage 3: capture

For capture, 50-μl supernatants of the incubation mixtures as described above were loaded onto the plates and incubated for 1 h at room temperature.

Stage 4: detection

Captured antigen (iC3b) was measured using biotinylated clone 4 anti-C3c monoclonal antibody 17. After washing the plates (five times in PBS/0·05%Tween 20), 50 μl biotinylated-clone 4 at 5 μg/ml diluted in PBS/gelatin was added to each well and the plate incubated for 1 h at room temperature. The plates were washed (five times in PBS/0·05%Tween 20) before incubation with the extravidin–peroxidase (Sigma cat. no. E2886) diluted 1/1000 in PBS/gelatin for 1 h at room temperature. The plates were washed (five times in PBS/0·05%Tween 20) and then incubated with 100 μl per well 3,3′,5,5′-Tetramethylbenzidine (TMB substrate; Life Technologies, Paisley, UK) for 15–30 min. The reaction was stopped by the addition of 1 M H₂SO₄ (50 μl per well) and the plates read immediately at 450 nm on a plate reader. The maximum OD achieved in the first 2 h of incubation is taken as a measure of the extent of activation of the alternative pathway and the percentage reduction of this level to its lowest level (usually at 8 h) as a measure of the breakdown of iC3b to C3c and C3dg.

Statistical analyses

All the comparisons made were of ratios of amounts of iC3b generated in a particular serum in a single experiment. Ratios (which have variance in both numerator and denominator) have to be analysed using non-parametric methods, and the appropriate test (the Kruskal–Wallis test) was performed using GraphPad Prism programs. Because of the influence of acute-phase reactions, the level of the individual components was distributed normally and the same non-parametric statistics were used.

Results

Antigenic concentrations of Factor I, C3, Factor H and Factor B

The concentration of Factor I in 61 serum samples was measured using immunodiffusion plates. Figure 1a shows the distribution of serum Factor I levels across the different genotypes. Although there is a trend in the mean levels of Factor I suggesting that genotype C > A > B, these differences are not significant on statistical analysis across all groups; nor when comparing only genotype C with genotype B. The mean level for the whole population was 45 μg/ml (10th–90th percentile, 34–52 μg/ml). This compares well with the mean level of 40 μg/ml (10th–90th percentile: 27–84 μg/ml) found by Reynolds et al. 4, using a similar assay and performed in a complement laboratory of high repute, and is not greatly different from the long-standing estimate of 35 μg/ml 18. These results show that the effect of complotype does not depend upon systematic differences in the pre-existing levels of Factor I.

Figure 1.

a. Factor I antigenic concentration in the different genotypes (61 samples). The concentrations were measured by single immunodiffusion on Binderid plates. The results for all sera are shown with the median and interquartile range. The slight trend towards lower levels from genotypes C (low risk) to A (heterozygotes) to B (high risk) is not statistically significant whether analysed non-parametrically or parametrically. Non-parametric analysis using the Kruskal–Wallis test is shown. (b–d) Antigenic concentrations of C3, Factor H and Factor B were measured nephelometrically in the Clinical Immunology Laboratory at the Cambridge University NHS Trust by Dr Sanja Ugrinovic. Results are shown with median and interquartile range. There were no significant differences between groups.

C3, Factor H and Factor B levels are shown in Fig. 1b–d. There is no significant correlation between the levels of any of these components and genotypes. It is indeed the case that the median level in the C3S homozygotes (genotype C) is slightly higher than in the C3F homozygotes (genotype B). There was one normal subject in the (protected) genotype C group who clearly had an acute-phase reaction with markedly elevated levels of C3, Factor H and Factor B and some elevation of Factor I. However, it is already known 4 that the genetic variants in C3 and Factor H that define the complotypes in this study do not show systematic differences in C3 or Factor H levels. Their differences are due to differences in activity not in levels of expression. Sera depleted of the relevant components and reconstituted with purified components of appropriate complotype (C3102G/fB32R/fh62V) have been reported to show a sixfold enhanced haemolytic titre in an alternative pathway assay compared with sera reconstituted with complotype C3102R/fB32Q/fH62I 3.

Down-regulation of alternative complement pathway by raising the concentration of Factor I

Recombinant human Factor I was added to serum from 18 subjects of each of the three genotypes and complement was activated using zymosan, as described. The amount of soluble iC3b (not bound to zymosan) was measured using the clone 9 capture assay. The effect of adding various concentrations of Factor I to a single serum sample from each genotype is shown in Fig. 2a–c, which also shows a typical time–course for the clone 9 assay. The data in Fig. 2 show that genotype B serum is substantially more resistant to the action of added Factor I when compared to genotypes A and C serum. In order to assess the effect of 50 μg/ml supplementation of Factor I on complement activation, the OD at maximum (before 2 h) and following conversion of iC3b to C3dg (8 h) was compared. The measurements chosen to analyse the test data were: (i) the percentage reduction of optical density produced by the addition of 50 μg/ml Factor I to the maximum reading achieved during the first 2 h; and (ii) the percentage reduction at 8 h.

Figure 2.

(a–c). These figures show illustrative time–course assays carried out with different concentrations of Factor I using three individual sera, one from each genotype. The optical density at 450 nm is a measure of the amount of iC3b bound to the clone 9 plate. A striking change between genotypes is seen, particularly at the highest concentration of Factor I, where the genotype B serum has maximum optical density twice that of either of the other genotypes.

The first of these readings measures the reduction in the amount of iC3b produced by the zymosan activation, and the second measures the reduction in the amount remaining after conversion to C3dg is anticipated to be substantially complete.

The percentage reductions brought about by the addition of Factor I at both time-points (Figs 3 and 4) differ highly significantly between the three groups, with genotype C > genotype A > genotype B. The differences observed between the three genotype groups are even more striking, considering that the sera taken were not controlled for age or body mass index or for the possible occurrence of any acute-phase reaction (all of which can modify complement levels). Therefore, the data in this study indicate that the genotype differences observed were considerably robust. A summary of all the results obtained with sera of the three different genotypes in their response to the addition of 50 μg of Factor I is shown on Table 2.

Figure 3.

Percentage reduction in zymosan-triggered iC3b by exogenous Factor 1. The percentage reduction produced by 50 μg/ml Factor I on the maximum optical density (OD) 450 nm produced before 2 h. The graph shows the individual results with the median and the interquartile range. Results were analysed non-parametrically and gave highly significant results by the Kruskal–Wallis test. This data shows the percentage reduction in the amount of zymosan-triggered iC3b produced after supplementation of sera with Factor I. Eighteen subjects in each group.

Figure 4.

The percentage reduction in zymosan-triggered iC3b produced by the addition of 50 μg/ml Factor I on the OD 450 nm at 8 h. The graph shows the individual results with the median and the interquartile range. Results were analysed non-parametrically and gave highly significant results. This figure therefore shows the efficiency with which added Factor I allows the final clip of iC3b to C3dg which, again, shows genotype C > genotype A > genotype B. Eighteen subjects in each group.

Table 2.

Summary of results

		Genotype C (low-risk homozygotes)	Genotype A (heterozygotes)	Genotype B (high-risk homozygotes)
FI concentration (μg/ml)	Median	46·00	43·00	43·00	Kruskal–Wallis statistic	2·706
Shown in Fig. 1a	25% percentile	40·00	40·00	34·00		P = 0·2584
	75% percentile	52	52	52
	Interquartile range	12·00	12·00	18·00
C3 concentration (μg/ml)	Median	1·20	0·99	0·98	Kruskal–Wallis statistic	3·707
Shown in Fig. 1b	25% percentile	0·92	0·93	0·90		P = 0·1567
	75% percentile	1·43	1·25	1·14
	Interquartile range	0·52	0·32	0·24
Reduction by 50 μg/ml FI	Median	37·2%	29·1%	22·9%	Kruskal–Wallis statistic	19·66
of maximum reading (%)	25% percentile	29·6%	20·3%	13·4%		P = 0·0001
	75% percentile	49·1%	39·7%	26·4%
	Interquartile range	19·5%	19·4%	13·0%
Reduction by 50 μg/ml FI	Median	53·5%	44·0%	28·2%	Kruskal–Wallis statistic	18·41
of 8 h reading (%)	25% percentile	46·4%	32·6%	16·9%		P ≤ 0·0001
	75% percentile	61·0%	50·4%	43·7%
	Interquartile range	14·6%	17·8%	26·8%

Sera activated with zymosan in mg/ethylene glycol tetraacetic acid (EGTA) buffer; 18 sera/group. Each serum is from a different individual. FI = Factor I.

All these tests were carried out using serum, which does not represent the physiological situation in vivo where CR1 is present on red blood cells in the circulation, and additionally on leucocytes at sites of inflammation. Notably, CR1 is the only co-factor in humans which mediates the Factor I clip of iC3b to C3dg under physiological conditions 19. To assess the effect of added co-factor, we have used a ‘tailed’ recombinant CR1 functional protein fragment, Mirococept. This compound was tested at several dilutions using only the amount of Factor I present in the serum and also with supplementation of Factor I at 25 and 50 μg/ml. These results are shown in Table 3. It can be seen that Mirococept has major concentration-related effects on both the maximum iC3b deposition and on its breakdown, showing that CR1 is a powerful inhibitor at both stages. Nevertheless, even at high concentrations of Mirococept (80 μg/ml) (initial trials of Mirococept in perfusing kidneys before transplantation used 10 μg/ml) 20, the effect of adding further Factor I is still apparent. It is clear, therefore, in the presence of co-factor, the down-regulatory action of Factor I supplementation will be greater.

Table 3.

Effects of Factor I and mirococept combinations on iC3b formation and destruction in a normal serum

Factor I concentration μg/ml		0	0	25	25	50	50
		Max	Min	Max	Min	Max	Min
Mirococept	0	100%	76%	99%	62%	83%	42%
Concentration	10	65%	49%	74%	51%	63%	47%
μg/ml	20	63%	59%	42%	42%	33%	17%
	40	47%	37%	36%	25%	20%	16%
	80	31%	23%	16%	5%	3%	0%

Values expressed as % of maximum value with no additions of Factor I or mirococept. Max = maximum value within first 2 h; min = minimum value whenever it occurs.

Discussion

It is now well established that polymorphisms in complement components that can cause acceleration of the C3b feedback cycle predispose to a group of inflammatory diseases including age related macular degeneration and haemolytic uraemic syndrome and C3 glomerulopathies. In this study we have examined a number of sera from normal donors who have been genotyped for three of these polymorphisms which are fairly common: the C3S/F polymorphism (R102G) where the rarer allele C3F (102G) with a frequency approximately 0·2 in Caucasoids and much lower in other racial groups confers susceptibility; the Factor H polymorphism (Y402H) where the rarer allele (402H) with a frequency of 0·38 confers susceptibility; the Factor H polymorphism (V62I) where the commoner allele (62V) with a frequency of 0·81 confers susceptibility.

Serum was taken from subjects who were homozygous for all three susceptibility alleles (genotype B), or were homozygous for all three protective alleles (genotype C), or were heterozygous at all three loci (genotype A). Of the 9989 individuals tested who gave results at all three loci, 55 (0·6%) were genotype B; 331 (3·3%) were genotype C; and 558 (5·6%) were genotype A. Together these make up just less than 10% of the population tested. These were normal volunteers not selected by family or clinical history for age-related macular degeneration or renal disease. The susceptible genotype B represents a very small percentage of the population.

The sera were tested in vitro for the extent of inhibition mediated by supplementation of Factor I to the activation of the alternative pathway by zymosan. The production of iC3b and its subsequent conversion to C3dg was tested in the absence or presence of a derivative of CR1 which provided co-factor activity for the split from iC3b to C3dg.

This study demonstrated that there were no significant differences in Factor I concentrations between the three genotypes as measured antigenically, although individual sera varied fairly widely. There is also quite extensive within-group variation in the rate and degree of complement activation by zymosan. This is not unexpected, as the donors will vary in age, sex and body mass index and their acute-phase state was not assessed at the time they were bled.

Conversely, there was a substantial and highly significant difference between the mean of the different genotypes on the magnitude of the lowering effect of Factor I supplementation on iC3b formation and on the conversion of iC3b to C3dg. The at-risk genotype (group B) showed a smaller degree of inhibition by Factor I compared to the protected genotype (group C) with the heterozygous genotype (group A) showing intermediate values. This was the case both at the time of maximal activation within the first 2 h, which is a measure of the formation of iC3b, and at 8 h, which is a measure of the breakdown of iC3b to C3dg.

Our results demonstrate that complement genotype at just three loci has a striking influence on the regulation of alternative pathway complement activation by Factor I – the hyperinflammatory complotype being more resistant to down-regulation by increased Factor I concentration. Furthermore, we have shown that the addition of a CR1-derived co-factor has a marked enhancing effect in the presence of physiological amounts of Factor I. This effect is enhanced further by the addition of further Factor I, and the combination allows the total inhibition of complement activation. We can therefore conclude that Factor I supplementation can reverse, in-vitro at least, the effects that the risk complotype imposes on the alternative complement pathway.

Further experiments on fewer sera using endotoxin as a soluble complement activator (which allows the effect of erythrocytes as source of CR1 to be assessed) and using a range of Factor I concentrations are to be published separately.

The results obtained encourage the further development of Factor I as a therapeutic aid to down-regulate the alternative pathway (or the C3b feedback cycle). Factor I supplementation may be particularly effective in those individuals who have a genetic predisposition to complement hyperactivity and who show any signs of the associated diseases.

Disclosure

J. E. S. and I. H. are employed by GlaxoSmithKline. The rest of the authors declare no financial or commercial conflict of interest.

Author contributions

E. L. was the major experimenter; S. N. recruited the volunteers to the study and performed the genotyping; J. S. contributed to the design of the study; I. H. prepared the recombinant Factor I; R. S. provided the Mirococept and invaluable discussions; D. S. and A. B. helped with some of the experiments; W. S. contributed to the design of the study and gave invaluable advice; P. L. designed the study and wrote the paper.