Partial complement factor H deficiency is associated with both C3

glomerulopathy and thrombotic microangiopathy

Katherine A. Vernon1, Marieta M. Ruseva1; H. Terence Cook1, Marina Botto1, Talat H.

Malik1, Matthew C. Pickering1

1 Centre for Complement and Inflammation Research, Imperial College, London

Running title: Partial FH deficiency and renal disease

Abstract: 199 words

Text: 2376 words

CORRESPONDING AUTHOR

Matthew Pickering

matthew.pickering@imperial.ac.uk

Centre for Complement and Inflammation Research

Division of Immunology and Inflammation

Imperial College London

Hammersmith Campus

Du Cane Road

London W12 0NN

Tel: +44 (0) 20 8383 2398

Fax: +44 (0) 20 8383 2379

ABSTRACT

The complement-mediated renal diseases, C3 glomerulopathy (C3G) and atypical

haemolytic uraemic syndrome (aHUS), are strongly associated with inherited and

acquired abnormalities in the regulation of the complement alternative pathway (AP).

The major negative AP regulator is the plasma protein complement factor H (FH).

Abnormalities in FH result in uncontrolled activation of complement C3 through the

AP and are associated with susceptibility to both C3G and aHUS. FH is

predominantly synthesized in the liver. We generated mice with hepatocyte-specific

FH deficiency and demonstrated that these animals have reduced plasma FH levels

with secondary reduction in plasma C3. Unlike mice with complete FH deficiency,

hepatocyte-specific FH-deficient animals did not develop either plasma C5 depletion

or accumulation of C3 along the glomerular basement membrane. In contrast, the

sub-total FH deficiency was associated with mesangial C3 accumulation consistent

with C3G. Although, there was no evidence of spontaneous thrombotic

microangiopathy (TMA), the hepatocyte-specific FH-deficient animals developed

severe C5-dependent TMA following induction of complement activation within the

kidney using accelerated serum nephrotoxic nephritis. Taken together, our data

indicate that sub-total FH deficiency can give rise to either spontaneous C3G or

aHUS following a complement-activating trigger within the kidney and that the latter

is C5-dependent.

INTRODUCTION

Both atypical haemolytic uraemic syndrome (aHUS) and C3 glomerulopathy (C3G)

can develop in the setting of abnormal complement regulation 1-3. C3G includes the

entities dense deposit disease and C3 glomerulonephritis, and is characterized by

the abnormal accumulation of C3 within the glomerulus which may or may not be

associated with inflammation, such as membranoproliferative glomerulonephritis 4. In

practice, it is considered when there is C3-dominant glomerulonephritis, with

dominance defined as C3 reactivity at least two orders of magnitude more intense

than any other immune reactant 5. In aHUS, there is complement-mediated damage

to the renal endothelium with consequent development of thrombotic

microangiopathy (TMA) 2.

C3G and aHUS are typically associated with genetic and acquired abnormalities that

result in uncontrolled activation of the complement alternative pathway (AP).

Complement factor H (FH) is the major negative regulator of the AP and

abnormalities in FH have been associated with aHUS and C3G 3. Genetic changes in

complement components in aHUS include loss of function in regulators (FH,

complement factor I [FI], CD46) and gain of function in activation proteins

(complement factor B and C3) 2. In C3G genetic abnormalities are less frequent and

include loss of function changes in FH, gain of function change in C3 and structural

changes within the factor H-like gene family 1. Acquired factors include

autoantibodies that enhance AP activation, most commonly C3 nephritic factors in

the case of C3G and anti-FH autoantibodies in aHUS 1,2.

Intriguingly, case reports have shown that complete FH deficiency can manifest as

aHUS or C3G 3. The factors that determine which pathology develops in the setting

of FH deficiency are not fully understood but mechanistic insights into C3G and

aHUS have derived from functional characterization of the genetic defects together

with the study of animal models that mimic some of these defects 3,6. Pigs 7 and mice 8 with complete FH deficiency develop spontaneous C3G but not aHUS. In both

species, the deficiency results in uncontrolled AP activation and secondary

consumption of C3 3 and C5 7,9, which results in absence of plasma complement

haemolytic activity. In FH-deficient mice (Cfh-/-) spontaneous C3G depended on AP

activation 8 but not on C5 10 and developed in wild-type kidneys transplanted into

Cfh-/- recipients 11. Notably, spontaneous C3G did not develop in the heterozygous

FH-deficient pig or mouse strain 7,8. FH consists of 20 short consensus repeat (SCR)

domains with SCR1-4 essential for C3 regulation and SCR19-20 important in

interacting with surface polyanions and C3b. FH mutations in aHUS predominantly

affect the surface recognition sites within SCR19-20 6. C5-dependent spontaneous

aHUS did occur in Cfh-/- mice transgenic for a FH protein (FHΔ16-20) that lacked

surface recognition domains 9,12. This led to the paradigm that aHUS occurs when

there is a combination of defective regulation of complement activation along the

glomerular endothelium (e.g. in the presence of FH mutations that impair surface

recognition) and sufficient circulating intact C3 and C5 to enable complement-

mediated endothelial injury to occur 9,12.

Here we report the phenotype of hepatocyte-specific FH-deficient mice (hepatocyte-

Cfh-/-) and show that this strain has plasma FH levels of approximately 20% of

normal. This sub-total FH deficiency resulted in plasma C3 dysregulation and

accumulation of C3 within mesangial areas. However, unlike mice with total FH

deficiency, hepatocyte-Cfh-/- mice did not develop plasma C5 deficiency or glomerular

basement membrane (GBM) C3 deposition. The hepatocyte-Cfh-/- animals did not

develop spontaneous aHUS but, following experimentally triggered renal injury using

accelerated serum nephrotoxic nephritis (NTN), they developed a severe C5-

dependent TMA.

RESULTS

Hepatocyte-Cfh-/- mice have reduced plasma FH and C3 but normal plasma C5 The conditional FH-deficient (CfhloxP/loxP) strain was developed using recombineering

techniques 13 to introduce loxP sites either side of exons 2 and 3 of the Cfh gene in

embryonic stem cells (Figure 1A). CfhloxP/loxP mice were viable and back-crossed onto

the C57BL/6 genetic background. To generate hepatocyte-specific FH-deficient mice

(hepatocyte-Cfh-/-), the CfhloxP/loxP strain was inter-crossed with mice selectively

expressing Cre recombinase in hepatocytes using the albumin promoter/enhancer

sequence (Alb-Cre)14. Hepatocyte Cre recombinase-mediated excision of exons 2

and 3 was confirmed by analysis of hepatic RNA (supplemental Figure 1). Consistent

with hepatocytes being the predominant source of plasma FH synthesis, circulating

FH levels were significantly reduced in hepatocyte-Cfh-/- mice (Figure 1B). This sub-

total FH deficiency was associated with marked reduction in plasma C3 levels

although not as low as that seen in mice with complete FH deficiency (Figure 2A).

However, unlike Cfh-/- mice, plasma C5 levels appeared normal when assessed

semi-quantitatively by western blotting (Figure 2B). Unexpectedly CfhloxP/loxP animals

had significantly lower plasma FH and C3 compared to Alb-Cre mice (Figure 1B and

2A). This indicated that the production of FH from the targeted cfh allele was

impaired.

Hepatocyte-Cfh-/- mice develop spontaneous mesangial but not GBM C3 deposition Cfh-/- mice spontaneously develop C3 accumulation along the GBM 8. To determine if

sub-total FH deficiency was associated with the same renal phenotype we examined

renal histology in 9 month old hepatocyte-Cfh-/- and Cfh-/- mice. We included CfhloxP/loxP

and Alb-Cre strains as control groups (supplemental table 1). We found that serum

urea levels were <20mmol/l with the exception of one animal in the hepatocyte-Cfh-/-

cohort (26.5mmol/l) and two animals in the Cfh-/- cohort (20.5 and 21.7mmol/l).

Albuminuria was only detectable in one of the hepatocyte-Cfh-/- (2.1mg/ml) mice.

Haematocrits and peripheral blood smears were normal (supplemental Figure 2).

Renal light microscopy showed minimal glomerular hypercellularity in hepatocyte-

Cfh-/- and Cfh-/- mice and no evidence of glomerular thrombosis (Figure 2C and D). As

expected all Cfh-/- mice had linear glomerular capillary wall C3 staining8. In contrast,

glomerular C3 deposition was granular and mesangial in distribution in age-matched

hepatocyte-Cfh-/- mice (figure 2E). The normal tubulo-interstitial C3 staining pattern

(seen in the control strains, Figure 2E) was reduced in intensity in the hepatocyte-

Cfh-/- strain and, as previously shown 8, absent in Cfh-/- animals. These data indicated

that sub-total FH deficiency was sufficient to prevent C3 accumulation along the

GBM.

Hepatocyte-Cfh-/- and CfhloxP/loxP mice develop TMA during NTNTo test the hypothesis that the impaired complement regulation associated with sub-

total FH deficiency predisposed to enhanced renal damage during complement

activation within the glomerulus, we examined the response of hepatocyte-Cfh-/- mice

to NTN. Hepatocyte-Cfh-/- mice rapidly developed renal failure and animals were

humanely culled on day 3 after administration of nephrotoxic serum (NTS, Figure 3).

In a representative experiment, 2 out of 14 hepatocyte-Cfh-/- mice were culled 3 days

after receiving NTS, whereas there were no deaths in the wild-type (n=9) and

CfhloxP/loxP (n=12) groups. Maximal haematuria developed in all hepatocyte-Cfh-/- mice

1 day post-NTS (supplemental Figure 3). In comparison to wild-type mice, at day 3

post-NTS hepatocyte-Cfh-/- animals had greater haematuria, higher urea levels, lower

haematocrits, marked red cell fragmentation and glomerular thrombosis (Figure 3

and 4). Glomerular C3 was highest in the hepatocyte-Cfh-/- mice (Figure 3F).

Differences in glomerular IgG and sheep IgG were minimal and the immune

response to sheep IgG in adjuvant was equivalent between the experimental groups

(supplemental Figures 4-6). CfhloxP/loxP mice, in which we had demonstrated slight

reduction in plasma FH levels (Figure 1B), developed a phenotype intermediate in

severity between wild-type and hepatocyte-Cfh-/- mice. Together these data indicated

that sub-total FH deficiency was associated with a hypersensitive response during

NTN and that the severity of TMA was related to the degree of FH deficiency.

The hypersensitive response of hepatocyte-Cfh-/- mice during NTN is dependent on C5.Since hepatocyte-Cfh-/- animals had normal C5 levels we hypothesized that the

enhanced renal damage during nephrotoxic nephritis could be influenced by C5

activation. To test this hypothesis we administered either a monoclonal anti-mouse

C5 antibody (BB5.1) or an isotype-matched control antibody to hepatocyte-Cfh-/-

mice. Administration of BB5.1 ameliorated the renal damage with reduced

haematuria, urea, red cell fragmentation and glomerular thrombosis and neutrophil

scores (Figure 5). Notably the thrombosis was lower in the isotype-matched antibody

treated group than that typically seen in untreated hepatocyte-Cfh-/- mice during NTN

(Figure 3E). These data indicated that the hypersensitive response of hepatocyte-

Cfh-/- mice during NTN was dependent on C5.

DISCUSSION

Mice with hepatocyte-specific FH deficiency were established by inter-crossing

conditional FH-deficient mice with mice expressing Cre recombinase under the

control of a hepatocyte-specific promoter. The hepatocyte-Cfh-/- mice had low plasma

FH and C3 levels and spontaneous development of mesangial C3 deposition. This

type of lesion has been reported as one morphological type of human C3G 15.

Heterozygous FH-deficient animals have normal glomerular C3 staining pattern with

slight but significantly lower plasma C3 levels than wild-type mice 8. The FH levels in

the hepatocyte-Cfh-/- mice (~20% of normal) were lower than those in heterozygous

FH-deficient mice 12, but still sufficient to prevent glomerular C3 accumulation along

the GBM, a characteristic feature of mice with complete FH-deficiency 8. However,

the sub-total FH level in the hepatocyte-Cfh-/- mice resulted in spontaneous

mesangial C3 deposition. Paramesangial deposits composed solely of C3c have

been reported in DDD specifically during episodes of hypocomplementaemia 16. It

was hypothesized that these derived from plasma C3c that had been released

following the FI-mediated cleavage of iC3b to C3dg. This mechanism could be

operating in the hepatocyte-Cfh-/- mice continuously due to the low FH levels. The

normal tubulo-interstitial C3 staining seen in wild-type mice was reduced in intensity

in hepatocyte-Cfh-/- mice. This is most likely due to the reduced plasma C3 levels

since tubulo-interstitial C3 staining in mice is dependent on circulating C3 11,17,18.

Renal disease in C3G may become clinically evident or be exacerbated by an

environmental trigger, such as infection. For example, upper respiratory tract

infection is an exacerbating factor in IgA nephropathy (synpharyngitic macroscopic

haematuria) and C3G 19,20. The hypothesis is that the inability to regulate the AP not

only increases susceptibility to spontaneous C3G but also increases the likelihood of

renal injury during an inter-current event that triggers complement activation within

the kidney. To model this we examined the response of hepatocyte-Cfh-/- animals to

NTN. Our observations showed that partial FH deficiency enhanced renal injury

during experimental NTN and that this was C5-dependent. Hepatocyte-Cfh-/- mice

were extremely sensitive to renal injury in this model and developed a renal

phenotype consistent with aHUS. Our interpretation is that hepatocyte-Cfh-/- mice

were susceptible during NTN due to a combination of the inability of the partial FH

deficiency to adequately regulate complement within the kidney and the ability of the

partial FH deficiency to preserve sufficient C5 in plasma to mediate injury. The

hepatocyte-Cfh-/- mice were housed in specific pathogen-free conditions so we do not

know how the renal phenotype might be influenced by infection. The C5-dependence

of the TMA in our experimental model suggests that the therapeutic utility of C5

inhibition in C3G may be particularly useful in patients with C3G and defective

complement regulation who present with features of intercurrent TMA.

The spontaneous phenotype (mesangial C3 glomerulopathy with no evidence of

TMA) and the phenotype observed during NTN (acute severe TMA) were distinct.

There is clear evidence of phenotypic variation within these diseases among

individuals with the same complement mutation. For example, in CFHR5

nephropathy the spectrum of renal disease varies from microscopic haematuria to

end-stage renal failure 21. Similarly, in aHUS incomplete penetrance in the setting of

complement mutations is common and clinical onset is typically preceded by an

environmental change e.g. infection, drugs and pregnancy 22. However, variation

between C3G and aHUS phenotypes within the same individual is less well

understood. Case reports of C3G as the presenting feature with aHUS occurring later

in the clinical course include: an adult male with complete FH deficiency who

presented with biopsy-proven C3G (membranoproliferative glomerulonephritis,

[MPGN]) but approximately one year later developed aHUS 23; an individual with anti-

FH autoantibodies who presented with C3G (MPGN) and developed recurrent C3G

in the first renal transplant and TMA in the second transplant 24 ; a young male with

mutations in both C3 and CD46 and a family history of aHUS who presented initially

with C3G (MPGN) before developing clinical features of aHUS 25; a young male with

heterozygous FH mutation who presented with C3G (MPGN) and later developed

aHUS 26,27 ; an adult female with C3G (MPGN) with no complement mutations who

later developed aHUS 27 and an adult male with C3G (mesangial C3 deposits) with

no complement mutations who later developed concurrent aHUS 27. Case reports of

aHUS as the presenting feature with C3G occurring later in the clinical course appear

less numerous but include: a young male with combined heterozygous FH and FI

mutations who presented with aHUS and subsequently developed glomerular C3

deposits in the transplant kidney 28 and a young male with homozygous FH

deficiency who presented with aHUS and then developed isolated glomerular C3

deposits in the absence of TMA within the transplant kidney 28. In this context the

hepatocyte-Cfh-/- mice recapitulate the clinical observations and provide definitive

experimental evidence that the same genetic defect can predispose to C3G and

aHUS.

The mice carrying the targeted Cfh alleles had significantly reduced basal plasma FH

and C3 levels indicating that expression of FH by the targeted locus was impaired.

These animals were also hypersensitive to renal injury during NTN which, based on

the hepatocyte-Cfh-/- NTN phenotype, was likely a consequence of the reduced FH

levels. Impaired expression of targeted genes due to the introduction of loxP sites

and/or selectable markers is well recognized and may be rectified by removal of the

selectable marker 29. In our construct FRT sites were placed either side of the

neomycin resistance gene to enable the neomycin resistance gene to be excised by

Flp recombinase and this is currently in progress. With our present dataset, the

impaired FH expression from the targeted allele prevented us from deriving any firm

conclusions with respect to the contribution of extra-hepatocyte FH to the total FH

pool.

In summary, our experimental data demonstrated that two distinct renal phenotypes,

C3G and aHUS, can develop in the setting of sub-total FH deficiency. Like

spontaneous aHUS in a murine model of FH-associated aHUS 9, experimentally

triggered aHUS in hepatocyte-Cfh-/- animals was C5-dependent, further supporting

the key role of C5 in complement-mediated glomerular TMA. Our data illustrate the

complex relationship between FH insufficiency and complement-mediated renal

injury and that the therapeutic utility of C5 inhibition in this setting depends on the

renal phenotype.

CONCISE METHODS

MiceMice were housed in specific pathogen-free conditions and all animal procedures

were performed in accordance with institutional guidelines and approved by the UK

government. Recombineering reagents, EL350 cells, PL451 plasmid and PL452

plasmid were obtained from NCI at Frederick (www.ncifrederick.cancer.gov). Mice

used in models of experimental renal disease were matched for age and sex, and

aged between 10 to 12 weeks at the time of the experiment. Mice expressing Cre

recombinase under the control of the murine albumin promoter (Alb-Cre) were

purchased from the Jackson laboratories (B6.Cg-Tg[Alb-cre]21Mgn/J, strain number

003574, http://jaxmice.jax.org/strain/003574.html). Genotyping was performed by

PCR on digested tissue using primers pairs. Alb-Cre primer pairs: 5’-

GCGGTCTGGCAGTAAAAACTATC-3’ and 5’-GTGAAACAGCATTGCTGTCACTT-3’

generated a 100bp product in presence of Alb-Cre allele. Targeted Cfh allele primer

pairs: 5’-AGGGGGAAAGAGAGAGATG-3 and 5’-TCCTAGTAGCAAAGCTAAACAG-

3’ generated a 450bp or 527bp amplicon in wild-type and targeted Cfh alleles

respectively. Floxed Cfh allele primer pairs: 5’-CTGTCAGCAAGAATTATTTGGC-3’

and 5’-ACACATCGTGGCTTTTCATTGC-3’ generated a 610bp or 318bp amplicon in

wild-type and floxed Cfh alleles respectively. The 9 month old cohorts comprised:

hepatocyte-Cfh-/- (n=25; 12M and 13F), Cfh-/- (n=23; 8M and 15F), CfhloxP/loxP (n=9; 4M

and 5F) and Alb-Cre (n=8; 2M and 6F) mice. Two unexplained deaths occurred: one

hepatocyte-Cfh-/- mouse aged 5 weeks and one Alb-Cre mouse aged 17 weeks. In

both cases renal histology was normal by light microscopy.

Measurement and detection of mouse FH, C3 and C5Mouse blood samples were collected directly into EDTA-containing Eppendorf tubes

and kept on ice until centrifugation and aliquots stored at -80°C. Repeated

freeze/thaw cycles were avoided. Plasma FH levels were measured by enzyme-

linked immunosorbent assay (ELISA) using a polyclonal sheep anti-human FH

antibody (antibodies-online.com) and a biotinylated goat anti-human FH antibody

(Quidel, San Diego, USA). FH levels in the samples were quantified relative to

normal wild-type mouse plasma used in the standard curve. Measurement of plasma

C3 was performed by ELISA as previously described 30 using a unconjugated and

horse radish peroxidase-conjugated polyclonal goat anti-mouse C3 antibody (MP

Biomedical, Cambridge, UK) with a standard curve generated using acute phase

serum containing a known quantity of C3 (Calbiochem, Hertfordshire, UK). Western

blotting for mouse C5 was performed as previously described 31.

Assessment of renal and haematological function Haematuria was detected using Hema-Combistix urine reagent strips (Siemens,

Frimley, UK) using the following scale: negative, trace, 1+, 2+ and 3+. Plasma urea

was measured using an ultraviolet method according to the manufacturer’s

instructions (R-Biopharm, Glasgow, UK). Plasma samples were diluted 1:40 in PBS

and a standard preparation of urea of known concentration was included as an

internal control. Albuminuria was measured using radial immunodiffusion as

previously described 10 with a rabbit anti-mouse albumin antibody (AbD Serotec,

Oxford, UK) and purified mouse albumin standard (Sigma-Aldrich, Dorset, UK).

Haematocrit was calculated after centrifuging EDTA-treated blood in ammonium-

heparin capillary tubes (VWR International Ltd, Lutterworth, UK) at 13,000rpm for 5

minutes and measuring the packed cell height to total height ratio. Peripheral blood

smears were prepared on glass slides and stained using the Diff Quick staining kit

(Polysciences, Warrington, USA). Red cell fragments were quantified as the number

per 200 red blood cells counted at high power (x40).

Assessment of renal histology Kidneys were fixed in Bouin’s solution (Sigma-Aldrich, Dorset, UK) for 2-4 hours and

then transferred to 70% ethanol. Sections were embedded in paraffin and stained

with periodic acid-Schiff (PAS) reagent and examined by light microscopy. Sections

were scored for glomerular thrombosis and glomerular neutrophil infiltration.

Analyses were all done blinded to the sample identity with 20 glomeruli scored per

section. Glomerular thrombosis was scored according to the amount of PAS-positive

material per glomerular cross-section as follows: grade 0, no PAS-positive material;

grade 1, 0-25%; grade 2, 25-50%; grade 3, 50-75%; grade 4, 75-100%. Glomerular

neutrophils were scored according to the number of neutrophils per glomerular cross-

section. Glomerular histology was graded as follows: grade 0, normal; grade 1,

segmental hypercellularity in 10–25% of the glomeruli; grade 2, hypercellularity

involving >50% of the glomerular tuft in 25–50% of glomeruli; grade 3,

hypercellularity involving >50% of the glomerular tuft in 50–75% of glomeruli; grade

4, glomerular hypercellularity in >75% or crescents in >25% of glomeruli.

Immunostaining for C3, IgG and sheep IgG was performed on acetone-dried 5µm

frozen sections as previously described 32. Antibodies used were: fluorescein-

isothiocyanate (FITC)-labeled polyclonal goat anti-mouse C3 (MP Biomedical,

Cambridge, UK), FITC-conjugated polyclonal goat anti-mouse IgG Fc γ-chain specific

antibody (Sigma-Aldrich, Dorset, UK) and a FITC-conjugated monoclonal mouse

anti-goat/sheep IgG (Sigma-Aldrich, Dorset, UK) Quantitative immunofluorescence

analysis was performed using an Olympus fluorescent microscope with digital

camera (Olympus, Southend, UK), Image-Pro Plus software (Media Cybernetics,

Silver Spring, USA) and a minimum of 10 glomeruli were examined per section as

previously described 32. Mean fluorescence intensity was expressed in arbitrary

fluorescence units (AFU). Glomerular C3 from NTN was scored according to the

number of glomerular cross-section quadrants that stained for C3. Grade 0, 1, 2, 3

and 4 representing staining of 0, 1, 1-2, 2-3, or all 4 quadrants respectively.

Induction of serum nephrotoxic nephritisNephrotoxic nephritis (NTN) was induced by the intravenous injection of 200μl of

sheep nephrotoxic serum (NTS; a sheep IgG serum fraction containing anti-mouse

GBM antibodies), in to mice that had been pre-immunised with an intra-peritoneal

injection of 200µg of sheep IgG (Sigma-Aldrich, Dorset, UK) in complete Freund’s

adjuvant (CFA) (Sigma-Aldrich, Dorset, UK) as previously described 32. For anti-C5

antibody treatment mice received either 2mg of BB5.1 or isotype-matched antibody

subcutaneously 24 hours before NTS. The animals received two more doses of

either BB5.1 or isotype-matched antibody on day 1 and 3 post-NTS administration.

To assess the immune response to sheep IgG hepatocyte-Cfh-/- (n=8), CfhloxP/loxP

(n=7), and wild-type (n=10) mice were injected with 200 µg of sheep IgG in CFA.

Plasma samples were collected at day 2, 4 and 9 post-immunization. IgG response

to sheep IgG was measured by ELISA on microtitre plates coated with 3.5 µg/ml of

sheep IgG. The plates were incubated with mouse plasma serially diluted in PBS

containing 1% BSA. Bound mouse IgG was detected with a peroxidase-conjugated

sheep anti-mouse IgG antibody (product number: 515-035-062; Jackson Immune

Research, US). The antibody titre was expressed as the inverse of the greatest

serum dilution that gave two-fold the background reading.

Statistical analysis Non-parametric data was represented as the median with the range of values given

in parentheses. Two groups were analysed using the Mann-Whitney test whilst one-

way analysis of variance with Bonferroni multiple comparison test was used to

analyse three or more groups. Data were analyzed using GraphPad Prism version

3.0 for Windows (GraphPad Software, San Diego, USA).

ACKNOWLEDGEMENTS

KAV was a funded by Kidney Research UK Clinical Fellowship (TF8/2009). MCP is a

Wellcome Trust Senior Fellow in Clinical Science (WT082291MA) and MMR is

funded by this fellowship.

STATEMENT OF COMPETING FINANCIAL INTERESTS

MCP has received fees from Alexion Pharmaceuticals for invited lectures and for pre-

clinical testing of complement therapeutics and consultancy fee from Achillion

Pharmaceuticals.

REFERENCES

1. Barbour TD, Ruseva MM, Pickering MC. Update on C3 glomerulopathy.

Nephrology, dialysis, transplantation : official publication of the European Dialysis

and Transplant Association - European Renal Association 2014.

2. Le Quintrec M, Roumenina L, Noris M, Fremeaux-Bacchi V. Atypical

hemolytic uremic syndrome associated with mutations in complement regulator

genes. Seminars in thrombosis and hemostasis 2010;36:641-52.

3. Pickering MC, Cook HT. Translational mini-review series on complement

factor H: renal diseases associated with complement factor H: novel insights from

humans and animals. Clin Exp Immunol 2008;151:210-30.

4. Pickering MC, D'Agati VD, Nester CM, et al. C3 glomerulopathy: consensus

report. Kidney Int 2013;84:1079-89.

5. Hou J, Markowitz GS, Bomback AS, et al. Toward a working definition of C3

glomerulopathy by immunofluorescence. Kidney Int 2014;85:450-6.

6. de Cordoba SR, de Jorge EG. Translational mini-review series on

complement factor H: genetics and disease associations of human complement

factor H. Clin Exp Immunol 2008;151:1-13.

7. Hogasen K, Jansen JH, Mollnes TE, Hovdenes J, Harboe M. Hereditary

porcine membranoproliferative glomerulonephritis type II is caused by factor H

deficiency. J Clin Invest 1995;95:1054-61.

8. Pickering MC, Cook HT, Warren J, et al. Uncontrolled C3 activation causes

membranoproliferative glomerulonephritis in mice deficient in complement factor H.

Nat Genet 2002;31:424-8.

9. de Jorge EG, Macor P, Paixao-Cavalcante D, et al. The development of

atypical hemolytic uremic syndrome depends on complement C5. J Am Soc Nephrol

2011;22:137-45.

10. Pickering MC, Warren J, Rose KL, et al. Prevention of C5 activation

ameliorates spontaneous and experimental glomerulonephritis in factor H-deficient

mice. Proc Natl Acad Sci U S A 2006;103:9649-54.

11. Alexander JJ, Wang Y, Chang A, et al. Mouse podocyte complement factor H:

the functional analog to human complement receptor 1. J Am Soc Nephrol

2007;18:1157-66.

12. Pickering MC, de Jorge EG, Martinez-Barricarte R, et al. Spontaneous

hemolytic uremic syndrome triggered by complement factor H lacking surface

recognition domains. J Exp Med 2007;204:1249-56.

13. Liu P, Jenkins NA, Copeland NG. A highly efficient recombineering-based

method for generating conditional knockout mutations. Genome Res 2003;13:476-84.

14. Postic C, Shiota M, Niswender KD, et al. Dual roles for glucokinase in glucose

homeostasis as determined by liver and pancreatic beta cell-specific gene knock-

outs using Cre recombinase. J Biol Chem 1999;274:305-15.

15. Servais A, Fremeaux-Bacchi V, Lequintrec M, et al. Primary

glomerulonephritis with isolated C3 deposits: a new entity which shares common

genetic risk factors with haemolytic uraemic syndrome. J Med Genet 2007;44:193-9.

16. West CD, Witte DP, McAdams AJ. Composition of nephritic factor-generated

glomerular deposits in membranoproliferative glomerulonephritis type 2. Am J Kidney

Dis 2001;37:1120-30.

17. Paixao-Cavalcante D, Hanson S, Botto M, Cook HT, Pickering MC. Factor H

facilitates the clearance of GBM bound iC3b by controlling C3 activation in fluid

phase. Mol Immunol 2009;46:1942-50.

18. Lenderink AM, Liegel K, Ljubanovic D, et al. The alternative pathway of

complement is activated in the glomeruli and tubulointerstitium of mice with

adriamycin nephropathy. Am J Physiol Renal Physiol 2007;293:F555-64.

19. Gale DP, de Jorge EG, Cook HT, et al. Identification of a mutation in

complement factor H-related protein 5 in patients of Cypriot origin with

glomerulonephritis. Lancet 2010;376:794-801.

20. Vernon KA, Goicoechea de Jorge E, Hall AE, et al. Acute presentation and

persistent glomerulonephritis following streptococcal infection in a patient with

heterozygous complement factor H-related protein 5 deficiency. Am J Kidney Dis

2012;60:121-5.

21. Athanasiou Y, Voskarides K, Gale DP, et al. Familial C3 glomerulopathy

associated with CFHR5 mutations: clinical characteristics of 91 patients in 16

pedigrees. Clin J Am Soc Nephrol 2011;6:1436-46.

22. Caprioli J, Castelletti F, Bucchioni S, et al. Complement factor H mutations

and gene polymorphisms in haemolytic uraemic syndrome: the C-257T, the A2089G

and the G2881T polymorphisms are strongly associated with the disease. Hum Mol

Genet 2003;12:3385-95.

23. Vaziri-Sani F, Holmberg L, Sjoholm AG, et al. Phenotypic expression of factor

H mutations in patients with atypical hemolytic uremic syndrome. Kidney Int

2006;69:981-8.

24. Lorcy N, Rioux-Leclercq N, Lombard ML, Le Pogamp P, Vigneau C. Three

kidneys, two diseases, one antibody? Nephrology, dialysis, transplantation : official

publication of the European Dialysis and Transplant Association - European Renal

Association 2011;26:3811-3.

25. Brackman D, Sartz L, Leh S, et al. Thrombotic microangiopathy mimicking

membranoproliferative glomerulonephritis. Nephrology, dialysis, transplantation :

official publication of the European Dialysis and Transplant Association - European

Renal Association 2011;26:3399-403.

26. Gnappi E, Allinovi M, Vaglio A, et al. Membrano-proliferative

glomerulonephritis, atypical hemolytic uremic syndrome, and a new complement

factor H mutation: report of a case. Pediatric nephrology 2012;27:1995-9.

27. Manenti L, Gnappi E, Vaglio A, et al. Atypical haemolytic uraemic syndrome

with underlying glomerulopathies. A case series and a review of the literature.

Nephrology, dialysis, transplantation : official publication of the European Dialysis

and Transplant Association - European Renal Association 2013;28:2246-59.

28. Boyer O, Noel LH, Balzamo E, et al. Complement factor H deficiency and

posttransplantation glomerulonephritis with isolated C3 deposits. Am J Kidney Dis

2008;51:671-7.

29. Fiering S, Epner E, Robinson K, et al. Targeted deletion of 5'HS2 of the

murine beta-globin LCR reveals that it is not essential for proper regulation of the

beta-globin locus. Genes Dev 1995;9:2203-13.

30. Rose KL, Paixao-Cavalcante D, Fish J, et al. Factor I is required for the

development of membranoproliferative glomerulonephritis in factor H-deficient mice.

J Clin Invest 2008;118:608-18.

31. Ruseva MM, Vernon KA, Lesher AM, et al. Loss of properdin exacerbates C3

glomerulopathy resulting from factor H deficiency. J Am Soc Nephrol 2013;24:43-52.

32. Robson MG, Cook HT, Botto M, et al. Accelerated nephrotoxic nephritis is

exacerbated in C1q-deficient mice. J Immunol 2001;166:6820-8.

FIGURE LEGENDS

Figure 1. Generation of conditional FH-deficient mouse. (A) The targeting vector

contained loxP sites (blue arrows) located either side of exons 2 and 3 of the Cfh

gene. The targeted locus contained a neomycin-resistant (Neo) positive selectable

marker flanked by flippase (Flp) recombinase recognition target (FRT) sites (black

arrows). Cre recombinase-mediated recombination between the loxP sites results in

excision of exons 2 and 3 and the positive selectable marker (floxed allele). This

causes a reading frameshift and generation of a premature stop codon in exon 4 and

therefore a null allele. Vertical numbered boxes denote exons; NcoI – Nco1

restriction enzyme; black box denotes probe location used to identify correctly

targeted cells. (B) Plasma FH levels in 12-week old Alb-Cre (median 93% of wild-

type, range 71-137%, n=6), CfhloxP/loxP (median 59.5% of wild-type levels, range 37-

103%, n=8) and hepatocyte-Cfh-/- (median 19% of wild-type levels, range 14-29%,

n=7) mice. Horizontal bars denote median values. *P=0.032, **P=0.002, P values

derived from Bonferroni multiple comparison test.

Figure 2. Plasma complement and spontaneous renal phenotype in conditional and hepatocyte-specific FH-deficient mice. (A) Plasma C3 levels in 9 month old

Alb-Cre (median 171.5µg/ml, range 158.8-273.1, n=8), CfhloxP/loxP (median

131.1µg/ml, range 118.4-272.8, n=9), hepatocyte-Cfh-/- (median 41.5µg/ml, range

28.6-82.2, n=25) and Cfh-/- (median 15.8µg/ml, range 7.9-26.6, n=23) mice.

Horizontal bars denote median values. * P<0.05, **P<0.001, ***P<0.0001, P values

derived from Bonferroni multiple comparison test. (B) Assessment of plasma C5 in

hepatocyte-Cfh-/- mice by western blotting. The intensity of the C5 protein band was

comparable between hepatocyte-Cfh-/- (lanes 1-4) and wild-type (lanes 5-7) mice. No

C5 is seen in Cfh-/- plasma (lane 8) and lane 9 contained 100ng of purified human C5

as reference. (C) Glomerular cellularity scores and (D) representative PAS-stained

glomerular images in 9-month old Alb-Cre, CfhloxP/loxP, hepatocyte-Cfh-/- and Cfh-/-

mice. Horizontal bars denote median values. Original magnification x40. (E)

Representative immunostaining images for glomerular C3 and IgG in 9-month old

Alb-Cre, CfhloxP/loxP, hepatocyte-Cfh-/- and Cfh-/- mice. The characteristic linear capillary

wall C3 deposition seen in Cfh-/- mice was replaced by a mesangial staining pattern in

hepatocyte-Cfh-/- mice. Tubulo-interstitial C3 deposition was absent in Cfh-/- mice and

normal in CfhloxP/loxP and Alb-Cre animals. It was detectable but reduced in the

hepatocyte-Cfh-/- mice. Mesangial IgG is commonly seen in older mice and was

detectable in small amounts in all of the cohorts.

Figure 3. Response of conditional and hepatocyte-specific FH-deficient mice to accelerated serum nephrotoxic nephritis. (A) Hematuria, (B) serum urea, (C)

hematocrit, (D) red cell fragments, (E) glomerular thrombosis score and (F)

glomerular C3 deposition at day 3 following administration of nephrotoxic serum to

pre-immunized wild-type, CfhloxP/loxP and hepatocyte-Cfh-/- mice. Horizontal bars

denote median values. * P<0.001, ** P<0.01, P values derived from Bonferroni

multiple comparison test. Data representative of three independent experiments.

Figure 4. Response of conditional and hepatocyte-specific FH-deficient mice to accelerated serum nephrotoxic nephritis. Representative images of peripheral

blood smears (upper panel), renal sections (middle panel) and glomerular C3

immunostaining (lower panel) in wild-type, CfhloxP/loxP and hepatocyte-Cfh-/- mice.

Figure 5. Antibody-mediated C5 inhibition in hepatocyte-specific FH-deficient mice during accelerated serum nephrotoxic nephritis. (A) Hematuria, (B) serum

urea, (C) red cell fragments, (D) glomerular thrombosis score and (E) glomerular

neutrophil numbers at day 4 following administration of nephrotoxic serum to pre-

immunized hepatocyte-Cfh-/- mice treated with either BB5.1 (n=11) or isotype-

matched antibody (n=10). Horizontal bars denote median values. P values derived

from Mann Whitney Test.