siRNA sequences were designed based on the available bioinformatics information to target the MASP-1 splice variant of the mouse MASP1 gene (22 siRNAs) or the MASP2 gene (46 siRNAs). GalNAc-siRNA conjugates were synthesized using solid phase synthesis methods as previously described (36, 58).

The sequence of interest (MASP-1 or MASP-2) was cloned into the multiple cloning region (XhoI-NotI sites) located downstream of the Renilla STOP codon in the 3'UTR. The psiCHECK2-Dual-Glo system enables detection of siRNA-mediated silencing of target sequences fused to a Renilla luciferase reporter gene. RNAi-mediated cleavage and degradation of the fusion mRNA can be measured by a loss in Renilla signal following siRNA treatment. The psiCHECK2 vector also contains a second reporter gene, Firefly luciferase, which is driven by a different promoter and allows for normalization of Renilla expression. An African Green monkey Cos-7 cell line was used in vitro to examine the effect of MASP-1 and MASP-2 siRNA duplexes on the expression of MASP-1 and MASP-2 in the psiCHECK2 system. All transfections were repeated for a total of three times. Inhibitory Concentration (IC50) of each MASP-1 or MASP-2 duplexes were calculated from the expression curve and is included on the plots.

The effect of MASP-1 and MASP-2 duplexes on MASP1 and MASP2 expression was initially characterized in 8 weeks old C57BL/6J mice (Charles River Laboratories). Single doses of 10 mg/kg of MASP-1 and of MASP-2 duplexes were administered via a subcutaneous injection. Livers were harvested on Day 7 post s.q. injection. mRNA was isolated using RNeasy 96 Universal tissue kit (Qiagen). High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems) was used to convert the mRNA samples into cDNA. Levels of MASP-1 and MASP-2 expression were analyzed using qRT-PCR reagents compatible with Roche Lightcycler 480 based on previously published studies (36). The MASP-1 and MASP-2 mRNAs were measured using a customized TaqMan assay (ThermoFisher) and normalized to the endogenous GAPDH expression (ThermoFisher).

WT mice were injected s.q. with 10 mg/kg on days −10, −5, and on day 0 either with GalNAc-Luciferase-siRNA (n = 10) or GalNac-MASP-1 (n = 10) or with GalNAc-MASP-2-siRNA (n = 10). Then CAIA was induced in these WT mice by using a mixture of 5 mAb to bovine collagen type II (CII) (Arthrogen-CIA, Chondrex) according to published studies (59–61). All WT mice received i.p., injections of 8 mg of Arthrogen on day 0 and 50 μg of LPS from E. coli strain 0111B4 on day 3. All mice were sacrificed at day 10 post anti-collagen antibodies injection. The total duration of the CAIA experiment was day 20 post the first siRNA injection. The clinical disease activity (CDA) in all groups of WT mice was determined every day by two trained laboratory personnel acting independently and blinded as to treatment according to our previously published methods (36, 40, 61, 62).

Mouse liver from first cohort of WT mice (n = 30) injected (s.q.) with GalNac-Luciferase (n = 10), GalNAc-MASP-1 (n = 10) and GalNAc-MASP-2 (n = 10) siRNAs with disease, sacrificed at day 20, were homogenized using a bullet blender (36). The expression of MASP-1 and MASP-2 were measured from the liver by qRT-PCR in a blinded fashion. Similarly, in a second cohort, liver from WT mice (n = 34) treated with GalNAc-conjugated Luciferase siRNA (n = 8) or with MASP-1 siRNA (n = 8) or with MASP-2 siRNA (n = 9) or with MASP-1 plus MASP-2 siRNAs (n = 9) without disease were processed for RNA extraction. In this study, out of 34 mice, LPS (50 μg/mouse/i.p.) was injected, at day 10, in mice treated with GalNAc-conjugated Luciferase siRNA (n = 4) or with MASP-1 siRNA (n = 4) or with MASP-2 siRNA (n = 5) or with MASP-1 plus MASP-2 siRNAs (n = 5) and these mice were sacrificed at day 14. The remaining 16 mice, i.e., 4 mice per treatment group, were not injected with LPS and sacrificed at day 25. Total RNA from the liver was isolated from the liver homogenates using RNAeasy Mini kit (Qiagen Inc., Germantown, MD). 18S ribosomal RNA (18S rRNA) was used as an internal control in each experiment and all data were expressed in pg/ng mRNA/18S RNA. All mRNA samples were analyzed in duplicate. All sample were analyzed by amplifying at 40 cycles according to the methods published previously in Nature Protocols (14, 63). All qRT-PCR data were analyzed by using cDNA based on the standard curve made by using liver RNA from a normal WT mice.

The absolute levels of MASP-1 protein were measured by using Time-resolved immunofluorometric assay (TRIFMA), a sandwich-type immunoassay using europium-labeled detecting agents as described (64). In this method, the concentrations of MASP-1 protein in the sera from WT mice injected with GalNAc-Lucifease-siRNA or GalNAc-MASP-1-siRNA or GalNAc-MASP-2-siRNA were determined by using microtiter wells coated with monoclonal anti-mouse MASP-1 antibody, followed by incubation with samples and detection using biotinylated anti-MASP-1 antibody and subsequently by adding Eu3+-conjugated streptavidin and reading of signal by time-resolved flourometri (Perkin Elmer, Hvidovre, Denmark) (64).

To assess the functionality of the LP, a 96-well ELISA Costar plates was pre-coated with 9 ug/ml of mannan diluted in 0.05 M Sodium carbonate buffer, pH 9.5 for overnight at 4°C. The ELISA plates were flipped gently with no wash followed by blocking with 1% BSA 1xPBS (with Ca²⁺ Mg²⁺) for 1 h at room temperature. After washing 3x with 1xPBS 0.05% Tween 20, 100 μl of 10% serum diluted in MBL binding buffer [20 mM Tris, 1 M NaCl, 0.05 % (v/v) Triton X-100, 10 mM CaCl₂, 15 mM NaN₃, 1 mg/ml HSA, pH 7. 4] was added to each well in duplicate, and the ELISA plates were left at 4°C overnight. Serum was subsequently removed by washing, and recombinant human C4 (rHuC4) protein (Complement Technology Inc., Tyler, Texas) diluted in sodium barbital buffer was added to the wells. One percent sera from WT and MASP-2^−/− mice were used as a positive and negative controls, respectively. No target control contained no rHuC4. ELISA plates were incubated for 1.5 h at 37°C followed by 3x washings with 1xPBS 0.05% T20 (with Ca²⁺ Mg2⁺). After 1 h anti-C4C biotinylated antibody (1 μg/ml) diluted in 1xPBS 0.05% T20 (with Ca²⁺ Mg²⁺) and incubated at room temperature for 2 h. One hundred microliter of Streptavidin (dilution 1:1,000) was added to each well after washing three times (with Ca²⁺ Mg²⁺). The ELISA plates were developed using 100 μl of 1:1 diluted Super Signal™ West Pico Plus Chemiluminescent Substrate (Thermo Scientific, Rockford, IL). The reaction was stopped using 50 μl with 2N H₂SO₄ and absorbance read at 450 nm, correcting for background at 550 nm.

To detect the presence or absence of MASP-2 protein, before, during and after MASP-2 silencing, by Western blot analysis, a total of 40 μL of D-Mannose-Agarose beads (resin) (Sigma) were equilibrated in wash buffer containing TBS, 5 mM CaCl₂ + 1 mM Pefabloc SC (Sigma, 76307). Ten μL of serum was mixed with 20 μL of wash buffer, added to the resin, and incubated for 1.5 h. The beads were washed three times, re-suspended in 50 μL of wash buffer, and transferred to Pierce Microspin column (Thermo Scientific). The spin columns were centrifuged for 2 min at 4,000 × g, until the resin become dry. Elution of MBL-MASP-2 complexes was performed by adding 30 μL of electrophoresis sample buffer to each tube [7.5 μL of NuPAGE LDS sample buffer (Life technologies, NP0007) + 3 μL NuPAGE reducing agent (Life technologies, NP0009) + 19.5 μL of ddH₂O] and centrifuging the resin until it is dry. Then 30 μL of the eluate was heated at 80°C for 5 min and loaded on the 10% SDS gel using 1xMOPS (Invitrogen) and 1xNupage (Invitrogen) running buffer. Proteins were transferred to a PVDF membrane using 1xNupage tran (Invitrogen) transfer buffer, and the membrane was blocked in PBS + 5% milk for 2 h. The blot was incubated with a primary antibody, biotinylated monoclonal Rat anti-human MASP-2/MAp19 (1 μg/ml) (6G12, cross reacting with mouse MASP-2) in PBS + 5% milk and then incubated at 4°C overnight (65). HRP-conjugated Streptavidin (diluted 1:1,000) was used as a secondary antibody. Finally, the blot was developed by incubating with 12 mL of SuperSignal West Pico PLUS Chemiluminescent Substrate (ThermoFisher) for 10 min. The blot was scanned using Genegnome XRQ Chemiluminescence Imaging System and Gene Tool analysis software from Syngene (Frederick, MD), and the density of each band was quantified using the Quantity One® software (Bio-Rad, Hercules, CA).

We measured, by qRT PCR, the expression of complement C3, factor B (FB), factor D (FD), Properdin, C5, MASP-1, MASP-2, and MASP-3 from the liver of mice treated with GalNAc-Luciferase siRNA, GalNAc-MASP-1 siRNA, GalNAc-MASP-2 siRNA, and GalNAc-MASP-1 plus MASP-2 siRNAs. The expression of the above-mentioned complement components were measured according to our published methods (14, 36).

To determine the increase or decrease in FD protein in the circulation of WT mice, with and without LPS, after silencing with GalNAc-MASP-1/MASP-2 duplexes simultaneously, Western blots for FD were performed according to our published methods (36, 40).

Recombinant Human TLR4 (R&D Systems) was re-suspended in a buffer containing 20 mM Tris-HCl pH 7.5, 150 mM NaCl, 5 mM CaCl₂, and 1 mM BME. TLR4 was fluorescently labeled using the NanoTemper Monolith His-Tag Labeling Kit RED-tris-NTA 2nd Generation kit (MO-L018) per the manufacturer's protocol. TLR4 was titrated with recombinant MBL (clinical grade) with concentrations varying from 20.8 μM to 635 pM, while keeping the labeled TLR4 at 5 nM. Samples were loaded into standard capillaries, and Microscale thermophoresis (MST) experiments were carried out using a Monolith NT.115 pico (NanoTemper Technologies GmbH, Munich, Germany). The thermophoresis was monitored at 20% LED power and high MST power. Data were analyzed and fit to the Kd model using MO. Affinity software (version 2.3) (NanoTempet Technologies, CA) in duplicate.

To detect TLR4 on the cell surface of cells by flow cytometry, PE-conjugated mouse anti-TLR4 antibody (CD284) (Biolegend) was used along with matched isotype control, rat IgG2a k PE-conjugated (BioLegend). Primary adipocytes derived from adipose tissue of C57BL6 mice were also examined by flow cytometry. The Mouse 3T3 L1 embryonic cell line was used to examine the expression of TLR4 receptors and FD by qRT PCR as mentioned above. 3T3L1 cells were cultured based on our previously published study (66). 3T3L1 cell line was used as analogous to adipocytes as these cells differentiate into adipocyte-like morphology under certain conditions (66). Differentiated 3T3L1 cells were also stimulated with various doses of LPS (E. coli) (1.25 or 2.50 or 5 μg/ml) or with recombinant human MBL (10 or 20 μg/ml) for 48 h to determine their effect on the expression of FD and TLR4.

Some of the TRIFMA data were analyzed by using the Excel program. Analysis of variance (ANOVA) was used to comparing more than two independent groups. Two-tailed Student's t-test was used to find out the differences between two means. DATA were expressed as standard error of the mean (SEM). The level of significance was defined as p < 0.05.

In vitro screening led to the selection of one duplex each of GalNAc-MASP-1 (IC50 = 0.0027 nM) and MASP-2 (IC50 = 0.1042 nM) (Figures 1A,B) siRNA. GalNAc-MASP-1-siRNA or GalNAc-MASP-2-siRNA were both 23 bp duplexes. The duplex of GalNAc-MASP-2-siRNA have no off-target effect on MASP-1, and vice versa, as shown below (Figure 4).

To determine the in vivo efficacy of GalNAc-MASP-1-siRNA and GalNAc-MASP-2-siRNA duplexes on the expression of MASP-1 and MASP-2, healthy C57Bl/6 mice were used (Figure 2). In this experiment, mice were injected with a single dose (10 mg/kg) of either PBS or GalNAc-MASP-1-siRNA or GalNac-MASP-2-siRNA duplexes on day 0 and sacrificed at day 7. A robust silencing in the siRNA-treated groups was observed, i.e., 93% of MASP-1 (Figure 2A) and 69% of MASP-2 (Figure 2B) expression was silenced after a single subcutaneous administration. Initially 58% decrease in MASP-3 expression was also noticed with silencing of MASP-1 but none with MASP-2 silencing.

To examine the systemic effect of GalNAc-MASP-1-siRNA and GalNAc-MASP-2-siRNA on arthritis, mice were injected (s.q.) three times with GalNAc-Luciferase, GalNAc-MASP-1-siRNA, or GalNAc-MASP-2 siRNA before the induction of disease. All mice injected with anti-CII mAb and LPS developed disease after day 3. At day 10, the CDA in mice injected with GalNAc-Luciferase, GalNAc-MASP-1, or GalNAc-MASP-2 duplexes were 10.7 ± 0.597, 7.7 ± 1.57, and 6.4 ± 1.44, respectively (Figure 3A). In mice treated with GalNAc-MASP-1-siRNA, there was a significant 35% decrease (p < 0.030) in the CDA at day 7, whereas there was no difference at day 10 (Figure 3A). However, in mice treated with GalNAc-MASP-2 siRNA there was from day 5 to 10 a consistent decrease in the CDA. At day 10, there was a significant (p < 0.013) 40% decrease in the CDA of mice treated with GalNAc-MASP-2-siRNA compared with GalNAc-luciferase, and at day 7 there was a significant (p < 0.0003) 60% decrease in the CDA in mice treated with GalNAc-MASP-2 siRNA compared with the mice treated with GalNAc-Luciferase-siRNA (Figure 3A).

The prevalence of disease at day 10 in WT mice injected with GalNAc-Luciferase, GalNAc-MASP-1 and GalNAc-MASP-2 duplexes was 100, 100, and 90%, respectively (Figure 3B). There was no significant (p < 0.68) and (p < 0.88) effect on the weights of mice treated with GaLNAc-MASP-1 and GalNAc-MASP-2 siRNA throughout the study (Figure 3C). No toxicity was seen, and no mouse died during the course of this experiment. Overall these data suggest that GalNAc-MASP-2, but not GalNAc-MASP-1 partially attenuated CAIA in mice.

A sandwich type immunoassay, i.e., a Time-resolved immunofluorometric assay (TRIFMA) to measure the absolute levels of MASP-1 was used in the sera from mice from the CAIA model (Figure 4). MASP-1 protein levels were measured at baseline (−day 10 relative to disease induction with anti-CII mAbs), after three injections of siRNA (day 0) and after the induction of disease (at day 10) (Figures 4A–C). The levels of MASP-1 were significantly (p < 0.0001) decreased by 95% in mice after three injections with GalNAc-MASP-1 siRNA, relative to the mice injected with the GalNAc-Luciferase siRNA, before the development of disease (day 0) and at the end of the study (day 10). (Figures 4A–C). In mice injected with control GalNAc-Luciferase-siRNA or GalNAc-MASP-2-siRNA there was no effect on MASP-1 levels (Figures 4A,C). These data show that MASP-1 duplex specifically and significantly silenced MASP-1 levels systemically for a long time not only before the development of arthritis but even after the development of arthritis. Despite the robust suppression of MASP-1 level, MASP-1 does not play a significant role in the pathogenesis of CAIA, because mice treated with GalNAc-MASP-1-siRNA were not protected as compared to mice treated with control GalNac-Luciferase siRNA (Figure 3A).

Western blot analysis for MASP-2 protein was done to examine the effect on MASP-2 on MASP-1 silencing by GalNAc-MASP-2-siRNA on circulating protein levels (Figure 5). Sera from a subset of mice evaluated in the CAIA model were examined by western blot (Figure 5A). We found that there was no effect on MASP-2 levels in the sera from each mouse injected with GalNAc-Luciferase-siRNA (Figure 5A, lanes 2, 3, 4) or GalNAc-MASP-1-siRNA (Figure 5A, lanes 5, 6, 7). A distinct band of ~75 kDa of MASP-2 was present along with ~19kDa MAp19 (Figure 5A, lanes 2–8 and 11, lane 1, marker). These MASP-2 or MAp19 bands were present only in the sera from mice before treatment GalNAc-MASP-2-siRNA (day −10; Figure 5A, lane 8) but completely absent or barely visible after treatment with GalNAc-MASP-2-siRNA (day 0; Figure 5A, lane 9) and after disease development (day 10; Figure 5A, lane 10). Identical results were seen using serum from another mouse (Figure 5A, lanes 11, 12, 13). Sera from FD^−/− and MASP-2^−/− mice were used as positive and negative controls, respectively, to show the specificity of Western blot analysis for MASP-2 protein (Figure 5B, lanes 1, 2). MASP-2 protein is absent in serum from MASP-2^−/− mouse (Figure 5B, lane 1) as expected but present in the serum from FD^−/− mouse (Figure 5B, lane 2). These data show that liver targeting of MASP-2 by GalNAc-MASP-2-siRNA was highly effective, and this effect lasted systemically for at least 10 days after the last siRNA dose. Our data once again confirmed that most of the MASP-2 in the circulation is generated by the liver and it can be inhibited below threshold detection levels even during acute inflammation (Figure 5).

To examine the functionality of MASP-2 silencing in liver by GalNAc-MASP-2-siRNA, C4b deposition via MBL pathway of the complement on mannancoated microtiter well surfaces was measured by ELISA. This procedure specifically excluded the activation from the CP due to the buffer composition, i.e., no C1q binding occurs to the surface. There is sufficient MASP-1 to activate MASP-2 in GalNac-MASP-1-siRNA treated mice. Sera from mice treated with a single dose of GalNAc-Luciferase-siRNA, GalNAc-MASP-1-siRNA, or GalNAc-MASP-2-siRNA were added to ELISA plates pre-coated with mannan (Figures 6A–C). There was a complete and significant inhibition of C4b deposition both before and after the induction of disease (p < 0.05) in the sera from mice injected with GalNAc-MASP-2-siRNA (Figure 6C). These data were remarkably consistent in all mice (10 out of 10) treated with GalNAc-MASP-2-siRNA (Figure 6C). On the other hand, there was no significant effect on C4b deposition from mice treated identically with GalNAc-Luciferase-siRNA or GalNAc-MASP-1-siRNA neither before nor after induction of disease (Figures 6A,B). Overall, these data show that liver-targeted MASP-2 silencing by GalNAc-MASP-2-siRNA completely abrogated C4b deposition (10 out of 10 mice) in the circulation and indicated that most of the MASP-2 is generated by the liver (Figure 6C). Furthermore, MASP-2 duplexes functionally impaired the lectin pathway of the complement. There was no significant off-target effect on C4b deposition of GalNAc-Luciferase-siRNA or GalNAc-MASP-1-siRNA (Figure 6).

Liver expression of MASP-1 and MASP-2 mRNA was examined in mice from the CAIA model. qRT PCR data from the liver of these mice at day 10 confirmed that there was a significant (p < 0.05) decrease in the expression of MASP-1 (70–95%) and MASP-2 (90%) with no off target effect on each other (Figure 7). However, in there was a decrease (58%) in expression of MASP-3 in the liver from normal mice with arthritis treated with GalNAc-MASP-1-siRNA (data not shown) but not with GalNAc-MASP-2-siRNA (data not shown). Later on we have not seen any effect on MASP-3 under normal physiological or mild pro-inflammatory conditions (Figure 8C). These data were analyzed in a blinded fashion using two different probes and primers by qRT PCR and results were identical.

To examine the individual and combined effects of MASP-1 or MASP-2 duplexes on the AP components mice were injected again three times (at day −10, −5, and at day 0) with a single dose (10 mg/kg) of GalNAc-Luciferase-siRNA or GalNAc-MASP-1 siRNA or GalNAc-MASP-2 or GalNAc-MASP-1 + MASP-2 –siRNAs. No disease was induced in these mice but a single dose of LPS was injected at day 10 to generate mile pro-inflammatory conditions. No significant differences in the liver were seen in the expression of C3, FB, and C5 (Figure 8). Surprisingly a significant (p < 0.05) decrease of 72% was seen in the liver expression of FD, in mice without injection of LPS, after combined silencing of MASP-1 and MASP-2 (Figure 8F). This reduction in FD expression was reversed in mice injected with LPS (Figure 8F). MASP-2 silencing alone resulted in a reduction of FD mRNA in LPS treated mice (Figure 8F). LPS induced mild inflammatory response as confirmed by slightly enhanced expression of IL-1β and TNF-α cytokines (Figures 9A,B). The expression of Properdin was also significantly (p < 0.05) increased in the liver of mice injected with GalNAc-Luciferase-siRNA followed by an injection of LPS but not without LPS (Figure 8G). The decrease in FD expression in the liver show that MASP-1 and MASP-2 together or MASP-2 alone might be regulating the transcription of FD in adipocytes present in the liver in a steady state, but during inflammation this regulating capacity might be lost or switched to something else.

Western blot data show that there was qualitative increase in the levels of FD in the circulation of mice injected with LPS after silencing of both MASP-1/MASP-2 simultaneously, in contrast to the mice without injection of LPS (Figure 10). These data were consistent in six different mice injected with no LPS or with LPS after silencing both MASP-1/MASP-2 genes (Figure 10, lane 3 vs. 4; lane 5 vs. 6 and lane 7 vs. 8). Nonetheless there was variability (4 to 67%) regarding an increase in the levels of FD in the circulation of mice after silencing MASP-1/MASP-2 simultaneously (Figure 10). Again these FD Western blot data in the circulation confirms the mRNA expression data in the liver that combined silencing of MASP-1 and MASP-1 in the liver also effected the FD protein levels in the circulation.

MBL has been previously shown to interact with TLR4 (67, 68), we sought to determine whether the effects of silencing might be through this pathway. In order to confirm the reported binding of MBL to TLR4, MST experiments were carried out. Human TLR4 was fluorescently labeled and titrated against varying concentrations of MBL (Figure 11). All MST traces appeared normal and there was no aggregation (data not shown). The final Kd for the binding was determined to be 907.28 ± 262.77 nM (Figure 11). Thus, we confirm that MBL binds TLR4 directly.

To further understand the interaction between lectin pathway components with LPS activation, we examined the interaction between MBL and TLR4. By flow cytometry, after several repetitions, we found no distinct expression of TLR4 receptors on the surface of 3T3L1 cells as well as from the primary adipocytes derived from WT mice (data not shown). We could not rule out that a small level of TLR4 undetectable with this method is present on their surface. Nonetheless, the endogenous expression of FD and TLR4 at the mRNA level was present in 3T3L1 differentiated adipocytes (Figure 12). There was a clear significant (p < 0.006) increase in the expression of FD in differentiated 3T3L1 cells 48 h after stimulating with 10 μg/ml and 20 μg/ml of recombinant MBL but only significant (p < 0.05) decrease in TLR4 expression was seen with 10 μg/ml not with 20 μg/ml of recombinant MBL (Figures 12A,B). There was a dose-dependent increase in the expression of FD in differentiated 3T3L1 cells (p < 0.05) treated with LPS (Figure 12C). In contrast, a decreasing trend in the expression of TLR4 was seen in response to the LPS (Figure 12D). A correlation (r = 0.51) between FD and TLR4 expression in 3T3 cells in response to LPS (5 μg/ml) was noticed but it was not significant (Figures 12C,D). These data indicate although TLR4 is undetectable by flow cytometry on the surface of 3T3 cells but still it can bind to the MBL and its ligand, LPS and modulate FD expression (Figures 12A,D). These data suggest that MBL or LPS can bind to a common receptor TLR4 to regulate the transcription of FD in adipocytes.

The authors declare that this study received funding from Alnylam Pharmaceuticals Inc. as a contract to NB. The funder had the following involvement with the study: synthesizing and conjugating duplexes of MASP-1 and MASP-2 with GalNAc. NB is seeking patent protection to use these conjugates for the treatment of rheumatoid arthritis and other complement mediated diseases. The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.