Contribution of Adipose-Derived Factor D/Adipsin to Complement Alternative Pathway Activation: Lessons from Lipodystrophy

Abstract

Factor D (FD) is an essential component of the complement alternative pathway (AP). It is an attractive pharmaceutical target because it is an AP-specific protease circulating in blood. Most components of the complement activation pathways are produced by the liver, but FD is highly expressed by adipose tissue. Two critical questions are: 1) to what degree does adipose tissue contribute to circulating FD levels and 2) what quantity of FD is sufficient to maintain a functional AP? To address these issues, we studied a novel mouse strain with complete lipodystrophy (LD), the fld mouse with partial LD, an FD-deficient mouse, and samples from lipodystrophic patients. FD was undetectable in the serum of LD mice, which also showed minimal AP function. Reconstitution with purified FD, serum mixing experiments, and studies of partial LD mice all demonstrated that a low level of serum FD is sufficient for normal AP activity in the mouse system. This conclusion was further supported by experiments in which wild-type adipose precursors were transplanted into LD mice. Our results indicate that almost all FD in mouse serum is derived from adipose tissue. In contrast, FD levels were reduced ∼50% in the sera of patients with congenital generalized LD. Our studies further demonstrate that a relatively small amount of serum FD is sufficient to facilitate significant time-dependent AP activity in humans and in mice. Furthermore, this observation highlights the potential importance of obtaining nearly complete inhibition of FD in treating alternative complement activation in various autoimmune and inflammatory human diseases.

FIGURE 1

Western blot analysis of mouse FD in serum and adipose tissue. (A) Untreated serum sample from WT mouse (0.3 µl) and also of recombinant mouse FD (30 ng) were treated with PNGase F. (B) Western blot of mouse FD in various adipose tissues: EWAT (Epididymal White Adipose Tissue), IWAT (Inguinal White Adipose Tissue), BAT (Brown Adipose Tissue). (C) Western blot of 1) undifferentiated MEFS prior to differentiation (UN0, lane 1), 2) MEFs differentiated into white adipocytes with the Dexamethasone/Insulin/IBMX protocol (DIX) for 9 d (DIX9, lane 2), 3) MEFs differentiated into brown adipocytes with DIX plus the PPAR-γ agonist troglitazone (DIXγ) protocol for 9 d (DIXγ9, lane 3) and 4) undifferentiated MEFs at the conclusion of the differentiation period (UN9, lane 4). (D) Western blot of mouse FD in various fat tissues (EWAT, IWAT, and BAT) before and after treatment with PNGaseF. These results demonstrate that mouse FD both in serum and adipose tissues is highly and variably N-linked glycosylated.

FIGURE 2

Lipodystrophic mice. (A) WT (left) and LD (right) littermates. Note interscapular defect in LD mouse. (B) Upper, Western blot for adiponectin; lower, loading control (heavy chain). Pinned view of WT (C) and LD (D) mouse showing absence of inguinal fat in LD. (E, F) skinned WT and LD P8 mice showing complete lack of adipose tissue in LD mice. Hematoxylin and Eosin (H&E) staining of newborn postnatal day 0 (P0) WT (G) and LD (H) demonstrating absence of BAT in LD. H&E of liver in WT (I) and LD (J) mice. Oil Red O staining of WT (K) and LD (L) mice. Note the hepatic steatosis in LD mice. White arrowheads, adipose depots in WT mice (missing in LD mice). Black arrowheads, brown adipose in WT mice. In LD mice, black arrowheads point to where BAT should be, but there is only a tissue plane between skeletal muscles.

FIGURE 3

Absence of FD in LD mice. (A)Western blot analysis of complement components (FD, C3, FB and P) in sera of WT and LD mice. FD was undetectable while FB was increased approximately two-fold in the serum of LD mice. Serum C3 and P levels were similar between WT and LD mice. NS, non-specific band. A small amount of C3 alpha 2, a degradation fragment of complement activation produced during sample collection, was observed in some samples. (B) Quantification of (A) by densitometry. *, p<0.05.

FIGURE 4

Deficiency of AP activity in sera of LD mice. (A) LPS-based in vitro binding assay in sera of WT (Black, n=8), LD (Grey, n=7), and FB^−/− (White, n=4) mice. Left, assay performed with 20% serum; Right, assay performed with 10% serum. Samples normalized to WT 20% sera result equals 100% activity. (B) Rabbit RBC based-hemolysis assay on sera from WT (n=14), LD (n=9), FD^−/− (n=8) or FB^−/− (n=3) mice. Data normalized to WT equals 100% activity. These functional assays demonstrate a severe defect in AP activity of LD mice.

FIGURE 5

Rescue of AP defect in LD and FD^−/− mice by purified human FD. (A) Rabbit RBC-based hemolysis assay was performed to assess AP activity in WT (White), LD (Grey), and FD^−/− (Black) sera (n=3 for all groups). Various quantities of FD were added to the reaction mixture. Human FD (100 ng/ml) rescued AP defect in the LD and FD^−/− mice. All values were normalized to WT equals 100%. (B) RBC hemolytic assays were with sera from WT, C3^+/−, FB^+/− and FD^+/− mice (n=3). All values were normalized to WT equals 100%.*, p<0.05.

FIGURE 6

Serum mixing experiments to assess the relative requirements of FD (Red triangle), C3 (Black inverted triangle), FB (Black square) or P (Black circle) in AP activation. Hemolytic assays were performed at a final concentration of 20% serum with different percentages of mixtures between WT and KO sera. For sake of clarity, 20% WT sera in the assay then correlates to 100% WT. (A) High-range WT sera (WT sera is 12.5–50% of total sera), (B) Mid-range (2.5–10%), (C) Low-range, (0.05–5%). Properdin levels affect AP modestly in the 20% serum assay system. Whereas FD, FB and C3 are all required for AP activity, the quantity of FD required is much lower than FB and C3. Note that 100% WT and 0% WT (100% KO) were performed with each experiment. Data are plotted with logarithmic x-axis to show spread of values with break to include 0% WT data.

FIGURE 7

FD level and AP activity in fld mice. (A) Western blot analysis for complement components in fld mice.(B) Quantification of (A). (C) Titration of FD level in fld mice. (D) Rabbit RBC-based hemolytic assay to measure AP activity in fld mice. Although there was ~30% of FD remaining in fld mice, fld sera (white) had ~80% of AP activity compared to WT mice (black), non-significant, n=3 per group. NS, non-specific band. *, p<.0.05.

FIGURE 8

FD concentrations and AP activity profile in LD mice following an adipose tissue transplant. (A) Western blot analysis of FB and FD in mice. In the titration experiments, FD levels in LD mice receiving a fat transplant were estimated to be about 1–5% of the WT mice; FB served as a loading control. (B) Rabbit RBC-based hemolytic assay to measure AP activity in WT, LD, and LD mice (n=3–5) that had received pre-adipocyte transplant (LD+Fat), *, p<0.05. FD^−/− and FB^−/− served as controls.

FIGURE 9

Cobra venom factor (CVF) leads to C3 cleavage in sera of LD mice. (A) CVF induces cleavage in vitro of C3 in sera of WT and LD mice, but not FD^−/− or FB^−/− mice. (B) CVF injection results in C3 cleavage in WT, LD and FD^−/− mice, but not in FB^−/− mice. All gels were run under reducing conditions. All experiments were performed at least twice with representative results shown. NS, non-specific band.

FIGURE 10

Reduction in circulating FD levels in patients with CGL. (A) FD levels were measured by ELISA for the following groups: 1) control (Black, n=4), 2) CGL due to AGPAT2 or BSCL2 mutations (Light Gray, n=5), 3) AGL (Dark Gray, n=5) and 4) FPL due to mutations in LMNA or PPARG (White, n=8). *, p<0.05. (B) Reduced FD in CGL patients observed in Western blot. P<0.05 CGL vs. controls. FD lane is 10 ng of purified human FD. HC, heavy chain of IgG. Gel was run under reducing condition. (C) Scatter plot of adipose mass (kg) vs FD levels (ng/ml). R²=0.388. AGL, Black circle; FPL, unknown type, Black square; CGL-atypical Dunnigan’s, Black triangle; CGL-BSCL1, inverted Black triangle; CGL-BSCL2, Black diamond; FPL-LMNA, White circle; FPL, PPAR-γ, White square. (D) Mixing of normal human sera control with either FD or C3 depleted sera. Experiment was repeated three times.

FIGURE 11

Kinetics and dose response curve for human Factor D in the rabbit RBC hemolysis and LPS microplate assays for AP. A-D were performed in FD-depleted human serum reconstituted with hFD. A) Dose response curve for hFD with a 5 min incubation. B) Kinetics of hemolysis when using 400 ng/ml hFD. C) Dose response curve of hFD when used with a 30 min incubation. RBC hemolysis data are expressed as percent hemolysis. D) Dose response curve for hFD added to FD-depleted human serum in the LPS microplate assay. E) Dose response curve for FD mixing experiment using WT and FD^−/− mouse serum in the LPS microplate assay.