Validation of dot blot immunoassay for measurement of complement opsonization of nanoparticles

Abstract

Complement plays a critical role in the immune response toward nanomaterials. The complement attack on a foreign surface results in the deposition of C3, assembly of C3 convertases, the release of anaphylatoxins C3a and C5a, and finally, the formation of membrane attack complex C5b-9. Various technologies can measure complement activation markers in the fluid phase, but measurements of surface C3 deposition are less common. Previously, we developed an ultracentrifugation-based dot blot immunoassay (DBI) to measure the deposition of C3 and other protein corona components on nanoparticles. Here, we validate the repeatability of the DBI and its correlation with pathway-specific and common fluid phase markers. Moreover, we discuss the advantages of DBI, such as cost-effectiveness and versatility, while addressing potential limitations. This study provides insights into complement activation at the nanosurface level, offering a valuable tool for nanomedicine researchers in the field.

Graphical abstract

INTRODUCTION

Complement, a crucial component of humoral innate immunity, plays a pivotal role in opsonizing nanopharmaceuticals and pathogens through the third complement protein (C3) while concurrently releasing proinflammatory factors such as C3a and C5a.^{1, 2} Several approaches exist for measuring complement activation, including by hemolysis of sensitized sheep red blood cells (CH50) and measurements of pathway-specific and terminal activation markers such as Bb, C3a, iC3b, C3bc, C4a, C5a, and sC5b-9 by enzyme-linked immunosorbent assay (ELISA).³ At the same time, several groups, including ours, assessed the level of C3 on various surfaces using flow cytometry and fluorescent microscopy, for example, on bacteria⁴, polymer beads⁵, erythrocytes⁶ , and nanoparticles⁷. The disadvantage of these assays is that they are not entirely quantitative, are labor-intensive, and are particularly challenging to apply to nanosized particles. However, there is a clear need to measure the level of deposited C3 in multiple donors and nanoparticle types, as C3b/iC3b is the main opsonin aiding nanoparticle ingestion by phagocytic cells, and its disposition could also be correlated with other components of the protein corona. Dot blot immunoassays (DBI) are commonly employed to rapidly screen samples, including proteins and nucleic acids⁸, due to their simplicity and efficiency. DBI is especially useful when dealing with arrays of samples or when the quantity of sample material is limited⁹. We developed a DBI to measure C3 deposition on nanoparticles^{7, 10, 11}, capitalizing on the sedimentation behavior of nanoparticles and liposomes under high g-forces. We successfully quantified C3 deposition among human donors and animal strains, measured the impact of complement regulators (IC50 value) on diverse nanomaterials, and determined the number of deposited proteins per nanoparticle^{7, 10, 12}. Here, the repeatability of DBI is validated, and C3 deposition is correlated with other complement activation markers measured in the fluid phase.

MATERIALS

Materials:

The nitrocellulose membrane (0.45 μm pore) was from Bio-Rad. Instant non-fat milk (Cat# A614–1005) was from Foothold, USA (Landover, MD). 2x Laemmli Sample Buffer and precast Mini-PROTEAN TGX Gels were from Bio-Rad. Secondary IgG antibodies labeled with IRDye 800CW were from Li-COR Biosciences (Lincoln, NE). Goat IgG fraction against human C3 (Cat#55033) was from MP Biomedicals (Solon, OH, USA). This antibody was previously validated to recognize main C3 fragments ⁴⁸. Recombinant hirudin (lepirudin) was from Aniara Diagnostica, LLC (West Chester Township, OH, USA). Purified C3 was from Complement Technology Inc (Tyler, TX, USA). Complement factors were measured by Excera labs (University of Colorado) using commercially available ELISA.

Nanoparticles:

Chemicals for SPIO NW synthesis, including iron salts, epichlorohydrin, and ammonia, were from Sigma-Aldrich (Saint Louis, MO, USA). Pharmaceutical grade dextran (20kDa molecular weight, T-20) was from Pharmacosmos (Holbæk, Denmark). Large SPIO NWs were prepared by precipitation of FeCl₂ and FeCl₃ salts with T-20 dextran by ammonia^{13, 14}. NW size was confirmed with ZetaSizer Nano (Malvern, UK). Nanoparticles were stored at 4°C at 10 mg Fe /mL in double distilled water and diluted with sterile 1x phosphate-buffered saline (PBS) before use.

Plasma samples:

Whole blood (3–5 mL) was collected in Vacutainer^® Z (no anticoagulant for sera) from consented donors at the University of Colorado Blood Donor Center under the Colorado Multiple Institutional Review Board (COMIRB) generic protocols for anonymous collection. Only age and gender were made available to the investigators. Whole lepirudin blood (3–5 ml, (10 μg recombinant lepirudin/mL blood) was collected between 2021 and 2023 at the University of Colorado Hospital Emergency Department under COMIRB protocol 17–1642 for the Emergency Medicine Specimen Bank¹⁵. Serum or plasma was collected by separation according to the protocol described previously¹⁶. Aliquots were stored at −80°C and were subjected to no more than three freeze-thaw cycles.

METHODOLOGY AND RESULTS

C3 opsonization is measured in non-diluted, lepirudin-anticoagulated plasma or complementsufficient sera^{17, 18}. We did not observe significant differences between serum and plasma in terms of complement activation¹⁶. DBI was described in our previous publications^{10, 12, 16, 19–21}. Nanoparticles in PBS are added to freshly thawed serum or plasma in a polypropylene tube at a 1:3 volume ratio. For iron oxide nanoparticles, the final concentration in plasma or serum is 0.25 mg Fe/mL but can go as low as 0.05 mg/mL. For liposomal drugs (e.g., Onivyde^®, Doxil^®), the final concentration is 0.25 mg drug/mL. After a brief mix, the sample is incubated in a 37°C water bath for 30 minutes.

According to Fig. 1, the activation occurs through three pathways, resulting in the deposition of C3 isoforms from C3(H₂O) to C3 cleavage products (C3b/iC3b/C3d). The activation also leads to the release of fluid phase markers, e.g., C3a, C5a, and sC5b-9, along with activation markers C4a, C4b, Bb, etc. Post-incubation, the sample is immediately diluted 30–50 times in 1x PBS without Ca²⁺ and Mg²⁺. The dilution is to halt complement activation. If necessary, 10 mM EDTA can be added to serum/plasma before dilution if the supernatant is to be assayed for fluid phase markers. An additional function of the dilution step is to decrease serum density, which is crucial for subsequent ultracentrifugation of low-density nanoparticles (e.g., liposomes) but is less critical for dense materials like gold or iron oxide. We use Beckman TLA-100.3 or TLA-45 rotors with 6 or 12 positions for 2 mL tubes, respectively. Centrifugation forces of 180,000g and above can pellet down most nanoparticle types, including ultrasmall iron oxides (ferumoxytol) and liposomes⁷. The pelleted particles are resuspended in PBS, and the centrifugation step is repeated 2 more times. After two washes, there is no immunodetectable complement C3 in the supernatant. Indeed, considering the average serum C3 concentration to be 1.5 mg/mL and two washes resulting in a dilution factor of ~1:2500, the residual C3 concentration should be 0.6 μg/mL, or below the detection concentration of ~0.78 μg/mL (standard curve in Fig. 2 below).

Fig. 1.

Pathways of C3 deposition and general workflow of the dot blot assay.

Fig. 2. Repeatability of dot blot assay.

A) Standard dilutions of purified C3 are dotted on the nitrocellulose membrane (in triplicates). The detection limit is ~1.56 ng C3/dot; B) NWs were incubated in non-diluted plasma or in plasma diluted 32 times with PBS without Ca^2+/Mg2⁺. DBI was repeated for 5 consecutive days. The membrane (non-diluted plasma) shows similar C3 deposition between repetitions (in triplicates). Only non-diluted plasma shows complement activation. Integrated density shows a 12.75% coefficient of variation; C) NWs were added to serum at different volume ratios to change the level of complement activation. There is a linear correlation between NP/serum ratio and C3 deposition; D) plasma from 8 donors was collected in 2021 and then in 2023. Plasma was stored at −80°C. DBI was performed in 2023. There was a significant 21% decrease in the C3 deposition for all donors, suggesting deterioration of plasma during storage (p-value of paired, two-sided t-test). All data are the means of technical triplicates.

The nanoparticle pellets are resuspended at 0.5 mg/mL and dotted on a nitrocellulose membrane (usually 0.45μm but can be 0.22 μm pore size) in 2 μL, 3–4 replicates per condition (Fig. 1). After drying the spots, the membrane is blocked in 5% w/v non-fat dry milk in PBS-Tween-20 (0.1%) and probed with anti-C3 antibodies. We prefer polyclonal antibodies that recognize intact C3 and major cleavage fragments (Fig. 1). Some monoclonals can be used, for example, anti-C3/C3b/iC3b antibody (e.g., commonly used 6C9 clone). Fragment-specific antibodies that recognize neoepitopes rather than whole C3 (e.g., anti-C3d) can be used. Besides C3, other antibodies against components of the protein corona, including immunoglobulins, can be used (discussed elsewhere¹²). The membranes are finally probed with secondary IRDye800- or IRDye700-labeled antibodies and scanned with a Li-COR Odyssey near-infrared scanner.

In the example in Fig. 2A, we blotted dilutions of purified human C3 and confirmed linear signal (R² = 0.9989) between 0.8 ng to 50 ng (2.5×10⁹ to 1.6×10¹¹ molecules) per dot, with a limit of detection about 1.56 ng C3 per dot (0.78μg/mL). To determine the repeatability of DBI, we used core-shell superparamagnetic iron oxide nanoworms (SPIO NWs) previously described by our group¹¹. These particles showed predominantly the activation of the lectin pathway (LP) in mice and the alternative pathway (AP) in humans¹⁰. Utilizing serum from the same donor, we performed 5 consecutive DBIs of C3 deposition to account for experimental variables, such as differences in ultracentrifugation speed and immunoblotting and potential loss of particles. The image of the scanned membrane (Fig. 2B) shows the same intensity of the C3 signal between the experiments. Across experiments, the assay variability was 12.75% (Fig. 2B). Diluting serum 32-fold in PBS completely inhibited complement activation (Fig. 2B), confirming that diluting serum/plasma with PBS is sufficient to stop complement activation after the incubation.

To test the linearity of C3 deposition, we mixed NWs with serum at increasing NP/serum (v/v) ratios, from 1:6 to 1:1. Since NPs are resuspended in PBS, this results in progressive dilution of serum, which is the elegant way to fine-tune the amount of complement activation and C3 deposition. After incubating and washing the particles, the same Fe amount was dotted on the membrane. According to Fig. 2C, there was an inverse correlation between volume fraction and C3 deposition (R squared 0.967), confirming the linearity of DBI.

To assess the reproducibility of DBI as a function of storage time, we identified eight lepirudin plasma samples collected in 2021 and then recalled these donors in 2023 to collect fresh plasma. SPIO NW incubation and DBI were performed on all samples simultaneously. Fig. 2D shows a 21% reduction in C3 signal in the old sera for all donors compared with fresh sera (p-value 0.00337). Understandably, prolonged serum storage decreases complement activation and, consequently, the efficiency of C3 deposition on nanoparticles. This needs to be considered when measuring complement in old biobanked serum/plasma samples.

Using DBI, we screened sera from 35 healthy donors (Fig. 3A) and observed highly variable levels of C3 deposition (up to 6.7-fold difference). We selected three donors with high and four with low C3 deposition for subsequent studies. We routinely use 10 mM EDTA or 10 mM EGTA/Mg²⁺ to confirm the specificity of C3 deposition and to differentiate between classical/lectin and alternative pathways (Fig. 1). In EDTA-supplemented sera, the DBI signals were decreased by up to 98%. In contrast, in EGTA/Mg²⁺ supplemented sera, only one sample showed a decrease, whereas in others, the signals increased (Fig. 3B–C). Western blot analysis of C3 deposited on the particles (Fig. 3D) showed an almost complete presence of the α’2 chain, suggesting complement activation and cleavage by Factor I . EDTA effectively prevented C3 deposition, whereas EGTA/Mg²⁺ increased the deposition.

Fig. 3. Correlation between DBI, western blot, and the pathway activation markers.

A) screening of 35 healthy donors for C3 deposition; B) DBI shows complete inhibition in EDTA sera; C) only one out of 7 showed inhibition by EGTA/Mg²⁺; D) western blotting of one of the samples. Intact C3 shows a non-cleaved α-chain and β-chain (typically below 75kDa). Surface deposited C3 shows high molecular weight bands, possibly bound to other proteins and the cleaved α’2 chain of iC3b. E-F) Sera from 3 high activators and 4 from low activators were selected. AP activation marker Bb shows an increase after incubation with NWs and a significant correlation with the C3 deposition (Pearson r). Data are means of 3 technical replicates. Paired, two-sided t-test *p-value<0.05.

To correlate C3 levels with the fluid phase activation markers, sera from 7 donors (three donors with high and four with low C3 deposition) were incubated with NWs, nanoparticles were pelleted with an ultracentrifuge, and the supernatant was collected and analyzed with ELISA for activation markers C3a, C5a, terminal pathway marker sC5b-9, and the AP activation marker Bb. Figures 3E–F show a significant increase in Bb in serum after incubation with NWs and a linear correlation between C3 deposition by DBI and fluid phase Bb. Collectively, these data confirm that the AP is the main activation pathway for NWs^{10, 12}.

The levels of common fluid phase activation markers C3a, C5a, and sC5b-9 were significantly increased in serum incubated with NWs. All the markers showed a significant positive correlation with the level of NW-deposited C3 (Fig. 4A–C), suggesting that for these particles, the initial C3 deposition leads to propagation of the complement all the way to the formation of C5 convertase and assembly of the membrane attack complex.

Fig. 4. Change in fluid phase markers following incubation of NWs and correlation with C3 deposition.

Fluid phase markers showed a significant increase compared to baseline serum and a strong correlation (Pearson r) with surface-deposited C3 for C3a (A), C5a (B), and sC5b-9 (C). Data are means of 3 technical replicates. Paired, two-sided t-test *p-value<0.05.

CONCLUSIONS

DBI is a robust assay for measuring complement C3 deposition on nanoparticles. While several methods exist to measure nanoparticle complement activation²², DBI offers several advantages: 1) Unlike other assays, anti-C3 antibodies are available for mice, rats, dogs, pigs, primates, and humans; 2) it is more economical than commercial ELISA (up to $1000/plate for some markers), allowing for large-scale screens. The ability to compare many samples simultaneously enhances the efficiency of the assay, providing a broader scope for analysis. Establishing a standard curve allows quantification of the amount of deposited protein and comparison between assays^{10, 12}; 3) C3 is the key opsonin, and its deposition can be directly linked to the immune uptake; 4) DBI is not limited to C3 deposition as it can be used to measure other proteins in the nanoparticle corona, including immunoglobulins and clotting factors, providing a comprehensive view of the immune response. In that regard, the analysis of protein corona for rare serum biomarkers associated with nanoparticles after in vivo injection is an interesting avenue of research.²³

There are certain limitations and challenges associated with the DBI: 1) Possible underestimation due to the crowding of nanoparticles on the membrane, impacting the accuracy of measurements. This problem can be overcome by appropriately diluting the particles before dotting to ensure linearity; 2) The method requires technical skill to ensure reproducible results; 3) while effective for certain nanoparticles, ultracentrifuge washing may not universally work for low-density materials like lipid nanoparticles; 4) there are concerns regarding the potential loss of “soft corona” proteins during washing steps¹⁶. This could be overcome by consistently treating all samples and minimizing the number of washes. The correlation between C3 deposition and fluid phase markers observed for SPIO NWs might not necessarily hold for other nanoparticles and coatings, especially in cases where nanosurface binding and activity of serum regulators Factor H and Factor I are enhanced.

In conclusion, DBI has advantages in terms of cost-effectiveness and versatility, but researchers should be mindful of its limitations. Current work is focused on DBIs that do not require washing steps. For example, anti-C3 antibodies that do not react with native C3 but rather with the cleaved isoforms (e.g., C3d) could be used. Efforts to refine the DBI will contribute to a comprehensive understanding of complement activation dynamics by nanoparticles.

Data Availability Statement:

All the raw data are available upon request.