A novel liposome-based immunoassay platform utilizing the complement system was developed to enable a simple and rapid detection of antibodies in serum. While the concept of complement-induced lysis and serum stability (stealthiness) of liposomes was studied with fluorescent, chemiluminescent, and electrochemically active encapsulants, the majority of subsequent experiments were performed with fluorescent liposomes only, as they allowed an easier assay procedure and a more sensitive readout. Moreover, it was demonstrated that this platform can also be applied in a simple POC lateral flow assay (LFA) format beyond the intended microplate-based high-throughput screening (HTS) approach. As proof of principle, antibody-triggered complement lysis of liposomes was specifically induced through selected surface functionalizations (biotin, PEG, peptide). In addition, naturally occurring anti-PEG antibodies were detected. Storage stability of selected liposomes was monitored at 4 °C, RT, and 37 °C for up to 40 months to assess colloidal stability, liposome integrity, and serum stability.

Tuning of liposome serum stability by cholesterol and encapsulant content

Cholesterol is a key regulator of membrane fluidity; it influences the phase transition and rigidity of the lipid bilayer, thereby supporting liposome stability [36]. In previous works, liposomes containing 44 mol% cholesterol have been established as a versatile platform with different marker molecules for various applications, such as detection of SARS-CoV- 2 neutralizing antibodies [12], DNA from C. parvum [15] and nucleic acids from Influenza A, Influenza B, and SARS-CoV- 2 [14] employing fluorescent, electrochemiluminescent, or electrochemical readout strategies. However, the use of high-cholesterol liposomes in active complement serum led to complement-induced lysis of liposomes. Cholesterol is a known complement trigger when used in high concentrations in vesicles or as cholesterol crystals [37,38,39,40]. In the intended format, non-specific liposome lysis allows false-positive signals, which is why stealth (serum stable) liposomes are required. Therefore, a stepwise reduction of the cholesterol content (5, 10, 15, 20, 25, and 44 mol%) was thoroughly studied using 10 mM SRB-encapsulating liposomes (Fig. 1a). As expected, it was found that the less cholesterol that was used, the less lysis was obtained, and it can be concluded that the cholesterol content can be used to tune the liposome stealthiness in serum. Another crucial parameter for the serum stability of the liposomes was found to be the SRB content. While higher concentrations are desirable to provide a more sensitive readout, it was found that its encapsulation was affected by the cholesterol content (Fig. 1b). While 10 mM SRB-liposomes can easily be formed regardless of the cholesterol content, low-cholesterol concentrations do not support higher SRB concentrations. Specifically, at 5 mol% cholesterol and > 10 mM SRB, encapsulation efficiency goes toward zero (data not shown). Therefore, more than 5 mol% cholesterol is required to entrap 25 mM SRB or higher. Furthermore, SRB, albeit highly water soluble, probably associates with the lipid bilayer [41], which led to undesired interactions with serum, as stably formed liposomes with 50 mM SRB and 30 mol% cholesterol lysed even in inactive serum. Thus, the optimal composition for stealth liposomes, where the interaction with serum components is solely determined by the liposome surface chemistry, was found to be 5 mol% cholesterol and 10 mM SRB. This composition was used for all subsequent studies utilizing biotin-, PEG-, and carboxyl-functionalized lipids to facilitate liposome surface modification. Stealth 5 mol% cholesterol liposomes were also successfully synthesized using chemiluminescent (m-carboxy luminol) and electrochemical (ruthenium hexamine (RuHex)) encapsulants (Fig. 1c and Fig. S1a). As expected, 44 mol% cholesterol RuHex and m-carboxy luminol-liposomes also triggered complement-induced lysis (Fig. 1c and Fig. S1b) due to the high-cholesterol content. It should be noted that the ratio of liposomes to serum chosen for the RuHex-liposomes led to only 9% lysis (100 µM total lipids to 10 vol% complement serum) and will be discussed further below. In the case of the electrochemical readout, interference from serum components had to be considered when selecting the encapsulant. Among the electroactive components studied, only RuHex allowed selective and sensitive detection in serum (Figs. S2 and S3).

Fig. 1

Optimization of the liposome composition to generate stealth liposomes. a Homogeneous complement assay of 10 mM SRB-encapsulating liposomes (10 µM total lipids; batches S1–6) showing the cholesterol dependency of liposome stealthiness in active serum (10 vol% PHS). Statistical analysis (two-sample independent t-test) revealed a significant difference between active and inactive serum samples at a cholesterol content of 10 mol% (p = 0.026) and above (p < 0.001). In contrast, no significant difference was observed at a cholesterol content of 5 mol% (p = 0.881). b Liposome stealthiness in active serum depending on SRB and cholesterol content (batches S7–10). 10 vol% human serum was used as a complement source (IRS41174). Statistical analysis (two-sample independent t-test) showed no significant difference between active and inactive serum samples of liposomes containing 10 mM SRB and 5 mol% cholesterol (p = 0.714), while in the other cases either a statistically significant amount of complement lysis (p < 0.001) or general serum stability was observed. Fluorescence measurements were carried out for 1 h at 37 °C in LCB, iaS, aS, and 30 mM OG + aS. Fluorescence intensities were normalized to the endpoint fluorescence of the positive control. λEx = 565(5 or 8) nm and λEm = 585(5 or 8) nm; gain 150. T = 37 °C. n = 3. c Heterogeneous complement assay of biotin-modified chemiluminescence liposomes (30 mM m-carboxy luminol; batches CL1–2). Statistical analysis (two-sample independent t-test) showed a significant difference between active and inactive serum samples of high-cholesterol liposomes. Low-cholesterol liposomes remained stealth and showed no statistical difference in 5 and 10 vol% serum, whereas 25 vol% led to minor lysis in active serum. Liposomes (50 μM total lipids) were immobilized overnight on a streptavidin-coated MTP and incubated for 1 h at 37 °C in either inactive serum or active serum (5, 10, or 25 vol% PHS in LCB). Chemiluminescence measurements were performed by adding 50 μL of 4 μM hemin and 50 μL 40 mM H2O2 in 0.01 M CBS, pH 10.5, 2 s integration time, gain 80, RH 1 mm, T = 25 °C, n = 3. d Flow chart of the homogeneous complement assay procedure for SRB-encapsulating liposomes

Proof-of-principle antibodies as external complement triggers and assay design setup

Antibodies are well-known complement triggers of the classical pathway and activate the complement system by binding the complement protein C1q to their Fc fragments [30]. Monoclonal and polyclonal antibodies directed against biotin, PEG, and a peptide were studied as model analytes. Biotinylation and PEGylation of liposomes were accomplished through the addition of functionalized phospholipids during liposome synthesis. The ARMS peptide originates from the age-related maculopathy susceptibility 2 (ARMS2) protein and is highly associated with age-related macular degeneration (AMD) [42]. It was conjugated to carboxyl-bearing liposomes using EDC/sulfo-NHS chemistry. Liposomes were characterized with respect to size, size distribution, ζ potential, and initial fluorescence signal as an indication for agglomeration and colloidal stability (Tables S2–4). Biotin- and PEG-liposomes revealed hydrodynamic diameters ranging from 104 to 135 nm and PDIs between 0.13 and 0.14, indicating colloidal stability. Furthermore, the successful surface modification with biotin and PEG was demonstrated using simple binding assays (Fig. S4), and in the case of ARMS by DLS. It was found that ARMS-modified liposomes tend to agglomerate, which is probably a result of the electrostatic interactions of the negatively charged liposome surface and the positively charged peptide under physiological conditions (pH 7.4).

The ability of the liposomes to bind to their respective antibodies and the subsequent triggering of the complement system was studied using a time-resolved fluorescence readout. To separate binding and complement activation steps in these preliminary studies, liposomes and antibodies were initially incubated for 1 h prior to serum addition (Fig. 1d). For all modifications, namely ARMS modification (Fig. 2d), biotinylation (Fig. 2e), and PEGylation (Fig. 2f), antibody-triggered complement activation was successfully demonstrated. Liposome lysis started after 15–30 min in the presence of a complement trigger before reaching saturation after 50–60 min, while liposomes remained stealth in the absence of triggering antibodies (Fig. 2a–c). As a positive control, providing 100% liposome lysis, a detergent was added to the liposome, analyte, and serum mixture. As negative controls, either liposomes in complement assay buffer (LCB) or liposomes in inactivated serum were used. Both negative controls serve as indicators of possible liposome instability per se or that caused by serum components. An observed offset of the inactive serum signal to the buffer control is likely due to agglomeration of liposomes with serum proteins and thus enhanced scattering. A similar effect was observed in active serum samples of stealth liposomes (Fig. 2a and b). Most nanoparticles and nanovesicles form a protein corona when in contact with biological fluids. This interaction is influenced by surface charges and overall chemical composition and leads to some minor degree of agglomeration [43, 44]. Not surprisingly, PEGylated liposomes are less likely to agglomerate due to PEG’s shielding effect, which reduces non-specific interactions and thus serum protein adsorption (Fig. 2c) [45]. To minimize agglomeration and avoid further scattering effects, several additives (trehalose, sucrose, D-galactose, lactose, NaCl, and BSA) were tested for their effect on the background signal. Sucrose (200 mM) was the most promising candidate, reducing the background signal by 77% (Fig. S5). In the case of the positive control, general effects on the fluorophores can be observed over the incubation period as fluorescence is highly dependent on the environment, such as the solvent, pH, and temperature. The initial dramatic rise of the signal is attributed to liposome lysis as temperature rises from 4 to 37 °C (Fig. 2). The subsequent minor decrease is likely due to quenching effects of the detergent, which may cause the active serum sample to exceed the positive control. In addition, serum proteins create a more hydrophobic environment, leading to increased fluorescence signals [46]. Consequently, free SRB or lysed SRB-liposomes exhibited 65–141% higher fluorescence intensities in serum than in buffer solutions (Fig. S6). However, the fluorescence-enhancing effect of serum proteins appears to be less pronounced in OG-containing samples, which leads to an even greater discrepancy between the endpoint signal of the positive control and the active serum sample. For the simplified endpoint evaluation, fluorescence intensities were normalized to the endpoint of the positive control, allowing for lysis values exceeding 100% (Figs. 2d and 3). While the assay platform is envisioned to be an endpoint assay, the time-resolved measurement enables precise monitoring and thus a better opportunity for data interpretation throughout the development of the technology.

Fig. 2

Homogeneous complement assay of 10 mM SRB-liposomes using antibodies as complement trigger. Time-resolved normalized fluorescence intensities of 0.5 mol% ARMS-modified liposomes (10 µM total lipids; batch S11) a without or d with 0.5 mol% complement-triggering anti-ARMS antibody (A626). ARMS-liposomes were incubated with the A626 antibody for 1 h at 300 rpm. 10 vol% human serum was used as complement source (IRS35577). Time-resolved fluorescence intensities of biotinylated liposomes (1 µM total lipids; batch S12) b without or e with 0.5 mol% complement-triggering anti-biotin antibody. Biotin-liposomes were incubated with the anti-biotin antibody for 1 h at 300 rpm. 5 vol% human serum was used as complement source (IRS45270). Time-resolved fluorescence intensities of PEGylated liposomes (1 µM total lipids; batch S1) c without or f with 0.05 mol% complement-triggering anti-PEG antibody (clone RM105). PEG-liposomes were incubated with the anti-PEG antibody for 1 h at 300 rpm in 5 µL HSS. 5 vol% human serum was used as complement source (IRS45270). Fluorescence measurements were carried out for 1 h at 37 °C in LCB, iaS, aS and 30 mM OG + aS and normalized to the endpoint fluorescence of the positive control. λEx = 565(8) nm and λEm = 585(8) nm; gain 150. T = 37 °C. n = 3

Fig. 3

Dose–response curves of liposome lysis dependent on the antibody concentration. Corrected lysis values of a Biotin- (batch S12) or b, c PEG-biotin-liposomes (batch S1) (10 mM SRB, 1 µM total lipids) in a homogeneous complement assay performing an anti-biotin or anti-PEG (clone RM105 or clone 6.3) antibody titration. d Different combinations of these antibodies (0.03 mol% clone RM105, 0.1 mol% clone 6.3, 0.3 mol% anti-biotin antibody) with PEG biotin-liposomes (batch S1). A one-way ANOVA, including a post hoc Tukey test, was performed for statistical analysis (p < 0.05) of the antibody combinations. Liposomes were incubated with the respective amount of antibodies for 1 h at 300 rpm. 5 vol% human serum was used as a complement source (IRS45270). Fluorescence intensities were adjusted by the iaS signal and normalized to the endpoint fluorescence of the positive control. λEx = 565(8) nm and λEm = 585(8) nm; gain 150. T = 37 °C. n = 3

Antibody-triggered complement lysis was also investigated with m-carboxy luminol-liposomes; however, in the case of CL quantification, only endpoint detection is possible. Initial experiments for CL quantification of the liposomes indicated strong interference of the radical-dependent CL reaction due to the radical scavenging nature of serum, leading to an about 64-fold higher limit of detection (data not shown). Moreover, m-carboxy luminol undergoes continuous degradation through reactions with naturally occurring H₂O₂ and hemin in serum, depending on the release rate of the encapsulant, which complicates reproducible detection in a homogeneous assay format. Therefore, a heterogeneous format was developed (Fig. S7). Here, biotinylated m-carboxy luminol-liposomes were immobilized in an MTP via the biotin-streptavidin interaction. After incubation with anti-biotin antibodies and serum or serum solely, the wells were washed, and the remaining intact liposomes were lysed and quantified. In the presence of complement-triggering anti-biotin antibodies, a lower CL signal was observed in 10 and 25 vol% active serum due to complement-induced liposome lysis (Fig. S8). However, the higher concentration of liposomes or greater steric hindrance of a heterogeneous approach rendered the assay less sensitive than the fluorescence approach.

In the case of the EC liposomes (assay procedure Fig. S9), no antibody-triggered liposome lysis was detected (data not shown). Since complement-induced lysis depends on the liposome to serum ratio, the absence of lysis in the case of EC liposomes is likely due to the higher liposome concentration (100 µM total lipids) required for EC detection. This finding was later confirmed using SRB-liposomes by investigating the ratio between total lipids and serum concentration needed to avoid oversaturation of the complement system with liposome surfaces. This high liposome concentration was required because the EC approach proved to be three orders of magnitude less sensitive than fluorescent detection, with an LOD of 13 µM total lipids for RuHex-liposomes compared to 49 nM for SRB-liposomes. In the future, the EC liposome approach will only be used to study complement activity via high-cholesterol liposomes, rather than antibody quantification via ligand binding.

Quantification of antibodies and analyzing complement activation potential for therapeutic applications

For the quantification of antibodies, a typical dose–response curve with respect to antibody concentration was recorded to determine the limit of detection and the corresponding EC50 values. Two monoclonal antibodies against PEG (clone RM105 and clone 6.3) and a polyclonal antibody against biotin were investigated (Fig. 3a–c). Typical binding curves were recorded to provide reliable and quantifiable data for comparative studies. As mentioned above, specific binding of the antibodies to the fluorescent liposomes was confirmed in a normal heterogeneous binding assay where the antibodies were immobilized via adsorption to the MTP surface and liposomes bound to them via their surface modifications (PEGylation or biotinylation) (Fig. S4). In the case of clone RM105, adsorption worked poorly, and immobilization via a secondary anti-rabbit antibody was chosen instead. Interestingly, the clone RM105 antibody, which is directed against the terminal methoxy group of the PEG chain, also revealed binding to PEGylated liposomes with a terminal biotin moiety, suggesting binding through cross-reactivity toward other PEG structures, in this case, the PEG backbone [47]. As expected, the clone 6.3 antibody, which is directed against the PEG backbone, showed binding to both PEGylated liposomes, and the anti-biotin antibody bound strongly to the biotinylated liposomes. It should be noted that differences in the overall fluorescence signals (Fig. S4) were due to the use of differently sized liposomes with different marker encapsulation efficiencies.

Comparing the findings of the heterogeneous binding assay with the homogeneous assay, it is obvious that more information can be gained from the latter. While the heterogeneous assay relies on an immobilized antibody to function well (e.g., clone RM105 adsorbed vs. secondary antibody), the homogeneous assay allows for natural binding events and simplifies the assay procedure. In addition, it provides information on the complement-triggering ability of the antibodies studied. For example, clone RM105 was found to trigger the complement system very well [48]. This aligns well with our results, as clone RM105 had an EC50 value more than four times lower than that of clone 6.3 or the anti-biotin antibody (Fig. 3a–c), whereas it performed poorly in the heterogeneous binding assay compared to the other antibodies (Fig. S4). Thus, we postulate that our assay can provide such functional information in a simple assay format. The complement-triggering capability of antibodies generally depends on clonality, isotype, subclass, hinge flexibility, and affinity toward the antigen [49, 50]. In particular, glycosylation of the antibody’s Fc region is essential for C1 complex binding and thus further complement activation [51]. This theoretically allows tuning of complement-triggering capabilities by glyco-engineering of the antibodies [52] and hence adaptation for specific therapeutic applications.

To demonstrate the capabilities of this assay, combinations of antibodies (anti-biotin, anti-PEG clone 6.3 and clone RM105) were investigated using liposomes with both a biotin and methoxy-PEG moiety. Antibody concentrations in the dynamic range were specifically selected for this purpose (Fig. 3d and Fig. S11). It was found that not only did the complement-triggering capabilities of the individual antibodies add up, but that their joint use also led to an enhancing effect in complement-mediated lysis. This is well known in the complement field, where the use of two or more antibodies targeting either different antigens on the membrane surface or different epitopes of an antigen facilitates a synergistic effect and promotes complement activation [53]. Here, maximum liposome lysis (166%) was obtained when all three antibodies were applied. This observation also suggests that two or more antibodies, that do not trigger complement lysis on their own or are present at too low concentrations, might be able to trigger complement lysis when used in combination.

To obtain the traditional antigen-binding ability of an antibody in the homogeneous assay format, the readout must be decoupled from its complement-triggering activity. This was achieved by using a secondary antibody that binds to the antigen-specific antibody, thereby taking over the major role in complement activation. This was demonstrated for the goat anti-biotin antibody (Fig. S12) using a secondary donkey anti-goat antibody. Furthermore, by using saturating concentrations of the secondary antibody, complement activation and hence liposome lysis and signal readout can be maximized. Finally, this also demonstrates that the liposome platform technology is applicable to analytes other than antibodies and can be used for competitive and sandwich assay strategies alike. This will be further explored in future studies.

Detection of naturally occurring anti-PEG antibodies in human sera

PEG is widely used in various industrial products, particularly in cosmetics and pharmaceutical drugs [54]. In cosmetics, for instance, PEG is applied as a surfactant, cleansing agent, emulsifier, or skin conditioner [55], while PEGylation is a common and popular conjugation strategy in drug delivery, since it decreases renal, proteolytic, and phagocytotic clearance, resulting in higher circulation times and a reduction of adverse effects [54, 56, 57]. Therefore, numerous FDA-approved PEGylated drugs are currently on the market. Not surprisingly, frequent exposure to PEGylated materials triggers the human immune system to produce anti-PEG antibodies. While free PEG has no or a weak immunogenic effect, it can trigger an immune response when conjugated to macromolecules or nanoparticles [54, 58, 59]. The induced antibodies are typically directed against specific motifs of PEG, such as the backbone or terminal groups [47].

The natural presence of these complement-active antibodies in human serum samples makes them an ideal model analyte for further optimizing the homogenous liposome-based complement assay and for demonstrating its application potential. Initially, we focused on the overall detection of such antibodies to show the capability of this novel diagnostic platform. Here, the stealthiness of liposomes in dependency on the human serum concentration was investigated using biotinylated liposomes, as no natural anti-biotin antibodies were present in human serum samples. Liposomes remained stealth over a broad range of serum to liposome ratios, whereas increasing serum concentrations—and thus the amount of available complement proteins—led to enhanced antibody-triggered lysis (Fig. S13). This shows that the liposome stealthiness or triggerability does not only depend on the amount of trigger but can also be tuned by the liposomes to serum ratio. Similar observations were made for PEGylated liposomes, which remained stealth when 1 µM total lipids were used in 1 vol% human serum but were lysed by naturally occurring anti-PEG antibodies in 10 vol% serum (Fig. 4). Unlike the approach using biotinylated liposomes, where the trigger amount was kept constant, the quantity of trigger antibodies plays a major role in this case. Here, we also demonstrated that the liposomes remained stealth when both the amount of serum (10 vol%) and liposomes (10 µM total lipids) were increased (Fig. S14) equally, maintaining a constant liposome to serum ratio. This further proves that the liposomes to serum ratio is key for successful complement activation. Eventually, the presence of naturally occurring anti-PEG antibodies in different human sera could be detected using PEGylated liposomes (Fig. 4). EC50 values were obtained for two of the pooled commercial sera (IRS35577, IRS31758), one had none present (IRS41174) and one (IRS45270) would require more in-depth analysis due to a slight increase in lysis at high serum concentrations. The EC50 values provide insight into the amount of complement-active anti-PEG antibodies in the sera, with lower values indicating a higher presence of these antibodies. As expected, non-PEGylated liposomes were shown to be stealth in active serum under the same conditions, demonstrating that the observed complement lysis originated from naturally occurring anti-PEG antibodies, which only bind to PEGylated liposomes (Fig. S15). Considering that in recent years the scientific community has intensively discussed the occurrence of anti-PEG antibodies in the population as a result of the SARS-CoV- 2 mRNA-based vaccinations, functional, simple assays such as the liposome platform technology are timely developments. Scientists have found that vaccinations with PEGylated lipid nanoparticles boosted the overall anti-PEG antibody levels in individuals. Ju et al. studied plasma from 130 adults vaccinated with either BNT162b2 (Pfizer-BioNTech) or mRNA- 1273 (Moderna) and found that anti-PEG IgG levels increased by a mean of 13.1-fold or 1.78-fold, respectively [60]. Anti-PEG IgM levels were boosted 68.5-fold or 2.64-fold following mRNA- 1273 and BNT162b2 vaccination, respectively [60]. However, the clinical relevance of anti-PEG antibodies has not yet been fully researched. We therefore suggest that the functionality of the liposome platform technology, which not only detects the presence but also the complement-triggering capability of the antibodies, will be very useful.

Fig. 4

Screening of anti-PEG antibodies in human sera. Corrected lysis values of PEGylated 10 mM SRB-liposomes (1 µM total lipids; batch S1) in a homogeneous complement assay performing a serum titration of different commercial sera (IRS batches 31,758, 35,577, 41,174, and 45,270). 0.1–10 vol% of human sera were used. Fluorescence intensities were adjusted by the iaS signal and normalized to the endpoint fluorescence of the positive control. λEx = 565(8) nm and λEm = 585(8) nm; gain 150. T = 37 °C. n = 3

Translation of the high-throughput assay to the POC

Considering that the readout result of the liposome platform technology is solely based on the separation of intact from lysed liposomes, a simple detection system can easily be envisioned through a standard LFA. Here, we demonstrate with a proof-of-principle assay that such separation can be accomplished with a visual readout of the SRB-entrapping liposomes. Liposomes containing 44 mol% cholesterol were biotinylated. The high concentration of cholesterol ensured complement activation, whereas biotin enabled capture on a streptavidin test line. Using the same negative and positive controls as in the MTP assay, the expected results were obtained. Intact liposomes provided a strong positive test line signal, whereas lysed liposomes were not captured and hence showed no test line signal (Fig. S16) Such a simple readout strategy may be a valuable tool in the future for companion diagnostics in therapies, e.g., for quick and simple monitoring of the occurrence of anti-PEG antibodies.

Long-term storage stability

In a long-term storage stability study, the integrity and triggerability or stealthiness of SRB-encapsulating liposomes was monitored at 4 °C, RT, and 37 °C for up to 40 months. The three most relevant types of liposomes were investigated, including low-cholesterol (5 mol%) PEGylated liposomes at a liposome to serum ratio ensuring stealthiness, i.e., 10 µM total lipid with 10 vol% serum, carboxyl-functionalized liposomes, and high-cholesterol (44 mol%) liposomes. Low-cholesterol liposomes were expected to be stealthy, while high-cholesterol liposomes were expected to show lysis throughout the entire study.

It could be shown that storage in buffer at 4 °C ensured colloidal stability and lipid bilayer integrity over the entire study period of 40 months (Fig. 5 and Fig. S17). Interestingly, the last data point (40 months) stands out due to the use of a different pooled serum that probably contained higher concentrations of complement proteins. This may have led to non-stealth behavior of low-cholesterol carboxyl-liposomes and enhanced complement lysis in the case of high-cholesterol liposomes. The latter also showed an increased background signal in inactive serum, which was most likely caused by agglomeration of the liposomes with serum proteins and hence scattering. This suggests that the ratio between total lipids and serum concentration has to be adjusted to avoid non-specific complement lysis, which can also be seen at the highest serum concentration used for biotinylated liposomes (Fig. S13) and newly synthesized carboxyl-liposomes (Fig. S18), ruling out the possibility that it is due to an aging effect of the investigated liposomes, but rather an effect of the serum source.

Fig. 5

Long-term storage stability study of 10 mM SRB-liposomes. Normalized fluorescence intensities of SRB-liposomes (10 µM total lipids) with different lipid compositions: a low-cholesterol, carboxylated (batch S13); b low-cholesterol, PEGylated (batch S1); and c high-cholesterol, biotinylated (batch S6) in LCB, aS, or iaS throughout the long-term storage stability study up to 40 months at 4 °C. 10 vol% human serum was used as a complement source (PHS for 0–10 months and IRS45270 for the 40 months datapoint). Fluorescence intensities were normalized to the endpoint fluorescence of the positive control. λEx = 565(5) nm and λEm = 585(5) nm; gain 150. T = 37 °C. n = 3. d Summary of the liposome storage stability at 4 °C, RT, and 37 °C

Liposomes stored at RT were stable for at least 10 months (Figs.