In Vitro Studies of the Effects of Antithrombotic Zn-Dipicolylamine-Harboring Liposomes (DPALs) on Serum Albumin and Human Umbilical Vein Endothelial Cells

Abstract

Thrombosis remains a leading cause of cardiovascular morbidity and mortality. During thrombosis, activated platelets and endothelial cells expose phosphatidylserine (PS) on their outer membranes, creating a surface that accelerates clot formation. Current antithrombotic therapies, such as heparin and warfarin, carry significant bleeding risks, highlighting the need for safer alternatives. In response, we developed a PS-targeting liposomal formulation composed of Zn-dipicolylamine (DPA)-cyanine-3[22,22] and 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (molar ratio 3:97). This DPA-harboring liposome (DPAL) binds selectively to PS-rich surfaces such as activated platelets and has demonstrated efficacy in reducing thrombosis in mouse models, with minimal bleeding. In the present study, we examined the interaction of DPAL with albumin, the most abundant plasma protein and a key transporter in the bloodstream, to assess the potential for harmful protein aggregation or structural disruption. Using dynamic light scattering and intrinsic protein fluorescence, we found that, unlike warfarin and heparin, DPAL does not induce any large protein aggregates or cause significant conformational changes near the tryptophan residue when mixed with human serum albumin, suggesting a favorable interaction profile. In addition, we used transwell permeability assays and CyQUANT cell proliferation assays to assess the cytotoxicity of DPAL in cultured human umbilical vein endothelial cells (HUVECs). Our results showed that DPAL does not compromise endothelial barrier integrity in HUVEC monolayers nor the cells’ viability. Our current and previous findings together suggest that DPAL could offer a promising approach to modulate harmful coagulation pathways and provide a new targeted therapeutic strategy for managing thrombotic disorders.

1. Introduction

Anionic lipids such as phosphatidylserine (PS) in the plasma membrane of mammalian cells flip more from the inner monolayer to the outer monolayer when cells are dying, stimulated, or becoming cancerous. This membrane lipid behavior prompts the design of various chemical compounds or nanoparticle formulations targeted at anionic surfaces of cells for diagnostic or therapeutic purposes [1,2,3]. Zinc (II)-bis-dipicolylamine (Zn-DPA) moiety, due to its specific binding to anionic phospholipids, has been utilized to make such PS-specific agents [4,5].

Recently, our lab developed a novel Zn-DPA-containing liposome formulation called DPAL that can bind specifically with PS-rich surfaces such as the surfaces of MCF-7 breast cancer cells [6] and procoagulant human platelets [7]. This formulation uses a synthetic lipid analog of Zn-DPA (Scheme 1). This analog contains two 22-C hydrocarbon chains that can insert into lipid bilayers. Through a short linker, these two long hydrocarbon chains are linked to cyanine 3, which serves as an optical reporter. This new Zn-DPA lipid analog is abbreviated as Zn-DPA-Cy3[22,22]. Specifically, DPAL is composed of 3 mol% Zn-DPA-Cy3[22,22] and 97 mol% 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine (POPC) [6,7], with ~70% Zn-DPA moiety exposed to the extra-vesicular environment.

Intravenous administration of DPAL to mice exhibited strong antithrombotic activities with little bleeding [7]. In contrast to the current standard of care such as heparin and warfarin, in which bleeding is a major concern, DPAL appears to be a safer antithrombotic. We proposed that antithrombotic effects derive from blockade of PS exposed as a result of prothrombotic stimuli, preventing PS-dependent propagation of coagulation to reduce thrombotic risk [7].

Since DPAL is being considered for treating thrombosis within the vasculature (US patent number 11,090,309 [8]), it is of interest to thoroughly investigate how DPAL interacts with blood components. In the current study, we examined the effects of DPAL on albumin and endothelial cells in vitro. Albumin, synthesized in the liver by hepatocytes, is the most abundant protein in the blood. It plays a crucial role in modulating intravascular osmotic pressure and serves as a carrier for various hydrophobic compounds, including long-chain fatty acids and bilirubin. In the present study, we used dynamic light scattering to assess whether the interaction between the positively charged DPALs and the negatively charged albumin proteins [9] results in large aggregates, which could then block vascular capillaries causing detrimental effects to health. We also used intrinsic protein fluorescence to examine whether DPAL affects albumin structure. Endothelial cells, on the other hand, form a tightly packed layer that lines the inside of the blood vessels, regulating blood flow, vessel constriction/dilation, and the exchange of nutrients/wastes between tissues and the blood. The luminal surface of the endothelium is rich in negatively charged carbohydrate moieties such as proteoglycans [10]. In this study, we used the transwell and CyQUANT assays to assess if and to what extent DPAL affects the barrier properties of the endothelial layer and the proliferation of endothelial cells.

2. Results and Discussion

2.1. Extent of Phospholipid Hydrolysis in DPAL Dispersions

To assess the interactions of albumin with DPAL, it is necessary to use the albumin and DPAL that contain little or no free fatty acids because free fatty acids can bind albumin to form complexes, which will complicate the data interpretation. In this study, we used albumins that are essentially fatty acid free according to the manufacturer. Yet to be tested is whether DPALs being studied contain any significant amount of free fatty acids due to spontaneous lipid hydrolysis. To address this issue, we used a modified ADIFAB assay to quantify the amount of free fatty acids present in DPAL dispersions. ADIFAB stands for 1-(6-(dimethylamino)naphthalen-2-yl)prop-2-en-1-one (Acrylodan) labeled intestinal fatty acid-binding protein. We took the fluorescence emission spectra of 9 μL of 60 μM ADIFAB in the presence of different amounts of oleic acid sodium salt (OAss) in 1.8 mL of alkaline water (pH 8.3 or pH 10.3) (Figure 1A,C). The fluorescence intensity at 425 nm (F₄₂₅) decreases with increasing OAss because the binding of OAss to the protein ADIFAB triggers a protein conformational change, moving the chromophore acrylodan to a more hydrophilic environment. All the measured F₄₂₅ values were normalized against F₄₂₅ without OAss. As shown in Figure 1B,D, the normalized F₄₂₅ changed linearly with the OAss content, generating a standard curve. The normalized F₄₂₅ values measured from the samples extracted from DPAL by organic solvents are marked as Xs in Figure 1B,D. It can be calculated from these F_425,X values and the standard curves that only 0.13% (pH 8.3, Figure 1D) or 0.16% (pH 10.3, Figure 1B) of fatty acyl chains in the POPC component of DPAL were hydrolyzed as free fatty acids after DPAL was stored in water containing 145 mM NaCl at room temperature (~22 °C) for 6 months. Such negligible amounts of free fatty acid detected also provide evidence that DPAL dispersions are stable for several months, consistent with the results from the particle size and polydispersity measurements [7].

2.2. Zeta Potential of DPAL

Zeta potential values of POPC liposomes with varying mole fractions of Zn-DPA-Cy3[22,22] at 25 and 37 °C are listed in

Table 1. Zeta potential of liposomes containing Zn-DPA-Cy3[22,22] and POPC.

Mole Fraction of Zn-DPA-Cy3[22,22]	Zeta Potential (mV)
	25 °C	37 °C
0.75%	−2.38 ± 0.25	0.65 ± 0.31
1.5%	1.19 ± 0.39	2.67 ± 0.36
3.0% (DPAL)	3.63 ± 0.43	6.64 ± 0.87
4.5%	5.22 ± 0.51	7.01 ± 0.75
6.0%	8.45 ± 0.79	9.22 ± 0.68
Buffer: 10 mM TES, pH 7.2, containing 145 mM NaCl

. For DPAL (i.e., 3 mol% DPA-Cy3[22,22] in POPC), its zeta potential in 10 mM TES buffer (pH 7.2) containing 145 mM NaCl is 3.63 ± 0.43 mV at 25 °C and 6.64 ± 0.87 mV at 37 °C. Since zeta potential reflects surface charges on the particles, the zeta potential values in Table 1 indicate that DPAL is moderately positively charged at physiologically relevant concentrations of NaCl (145 mM) and at neutral pH.

2.3. Particle Size/Size Distribution of DPAL in the Presence of Albumin

Since DPAL is moderately positively charged (Table 1) whereas the surface charge on serum albumin (pI = 4.9) is negative at physiological pH [11], it is necessary to test, before using DPAL as an antithrombotic agent [7,8], whether the charge attraction between DPAL and albumin leads to harmful large amyloid-like aggregates [12,13]. To address this issue, using dynamic light scattering, we studied particle size and size distribution of mixtures containing DPAL and albumin at their concentrations matching those used in the mice study [7]. Specifically, we examined the particle size and size distribution of mixtures containing 136 μL of DPAL ([phospholipid] = 2.647 mM) in 1.8 mL of 600 μM human serum albumin (HSA) in the presence of 145 mM NaCl. These concentrations were selected for use because they are within the physiological concentration range of albumin and NaCl in the circulation and because the amount of DPAL in the sample is concentration-wise equivalent to 200 μL of DPAL ([phospholipid] = 1.5 mM) injected into the bloodstream (~1.5 mL) of a ~21-g mouse used in previous in vivo studies [7].

Table 2. Hydrodynamic diameter (Z_ave) and polydispersity (PDI) of particles in DPAL-HSA mixtures as compared to DPAL and HSA alone. Aqueous solution: water containing 145 mM NaCl at pH 7.4.

Samples	n	Zave (nm)	PDI	Peak by Intensity (nm)	Intensity Area (%)
25 °C
DPAL (with [POPC] = 200 μM)	5	159.30 ±2.50	0.19 ±0.02	Monomodal size distribution Peak: 198.0 ± 10.6	100
600 μM HSA	6	11.00 ± 0.20	0.40 ± 0.07	Peak 1: 8.40 ± 0.09 Peak 2: 71.62 ± 4.25 Peak 3: 2934 ± 238	Peak 1: 70.33 ± 0.79 Peak 2: 18.92 ± 1.08 Peak 3: 10.75 ± 0.52
600 μM HSA + DPAL (with [POPC] = 200 μM)	4 5	39.04 ± 0.10 41.85 ± 2.77	1.0 1.0	Peak 1: 199.88 ± 9.58 Peak 2: 7.88 ± 0.19 Peak 3: not detected Peak 1: 188.40 ± 14.18 Peak 2: 7.97 ± 0.29 Peak 3: 3893 ± 2172	Peak 1: 79.50 ± 0.57 Peak 2: 20.50 ± 0.57 Peak 3: 0 Peak 1: 77.36 ± 0.93 Peak 2: 20.58 ± 0.74 Peak 3: 2.12 ± 0.74
37 °C
DPAL (200 μM POPC)	7	145.29 ±1.01	0.12 ± 0.03	Monomodal size distribution Peak: 165.8 ± 3.1	100
600 μM HSA	6	11.64 ± 0.20	0.37 ± 0.01	Peak 1: 8.34 ± 0.05 Peak 2: 67.38 ± 3.65 Peak 3: 2256 ± 680	Peak 1: 69.27 ± 0.91 Peak 2: 18.52 ± 0.76 Peak 3: 11.18 ± 2.45
600 μM HSA + DPAL (200 μM POPC)	6 3	41.72 ± 0.55 42.32 ± 1.28	1.0 1.0	Peak 1: 199.60 ± 8.81 Peak 2: 8.10 ± 0.27 Peak 3: not detected Peak 1: 196.47 ± 10.88 Peak 2: 8.14 ± 0.71 Peak 3: 1416 ± 2400	Peak 1: 78.15 ± 0.55 Peak 2: 21.77 ± 0.63 Peak 3: 0 Peak 1: 76.2 ± 0.99 Peak 2: 21.3 ± 0.71 Peak 3: 2.45 ± 1.63

shows that, in the present study, the hydrodynamic diameter (Z_ave) of DPAL at 200 μM phospholipid concentration is 159 nm at 25 °C and 145 nm at 37 °C with a PDI value less than 0.2, showing a monomodal size distribution. These results are like those obtained from DPAL, which was measured at much lower phospholipid concentrations (~10 μM) [6,7]. The consistency suggests that, even at high concentrations such as 200 μM phospholipids, DPAL does not aggregate or fuse to larger vesicles.

For the sample of 600 μM HSA, three intensity peaks were observed at both 25 and 37 °C. The major component (~70%, designated as Peak 1 in Table 2) has a size around 8 nm. This component likely contains mainly monomeric serum albumin, which is known to have a hydrodynamic diameter of about 7–8 nm [14,15], or it can be a mixture of monomeric and dimeric serum albumin [16]. The second-largest component (~18–19%) has a size of 67–71 nm, which may arise from a mixture of oligomers of HSA [12,16,17]. The third peak (10–11%) showed a size of about 2–4 μm, which suggests that there probably exist some (10–11%) large protein aggregates in 600 μM HSA. Micrometer-sized serum albumin, either in the form of amyloid-like or amorphous aggregates, can occur at high albumin concentrations as previously reported [18,19].

It is interesting to note from Table 2 that these large micrometer-sized aggregates of HSA either are not detected or are diminished in the presence of 200 μM DPAL. As presented in Table 2, four sets of data (n = 4) indicate that adding DPAL (with 200 μM POPC) to the 600 μM HSA solution at 25 °C made larger HSA aggregates not detectable, with the intensity area of Peak 3 reduced to 0%, leaving only two intensity-weighted peaks; one is around 8 nm due to the presence of HSA, and the other is about 200 nm, which most likely originates from DPAL (Table 2). Five other sets of data measured at 25 °C (Table 2) show that, in the presence of DPAL, large HSA aggregates could still be detected as the third peak. However, the intensity area of Peak 3 was significantly reduced from 10–11% without DPAL to 2.12% with DPAL. Similar results were seen from the 37 °C data (Table 2), where six sets of data showed Peak 3 not detected, while three sets indicated a decrease in the intensity area of Peak 3 from 11.18% without DPAL to 2.45% after adding DPAL.

The PDI value of the DPAL/HSA mixture is high (~1) (Table 2), which is not surprising because of the presence of DPAL and various serum albumin oligomers and aggregates [16]. When PDI is near one, particle sizing determined by dynamic light scattering (DLS) may not be relied upon for precise quantification. However, the current DLS technique with the detection range up to 10 μm can still provide a qualitative estimate of the particle size range. With this understanding in mind, we drew our conclusion from the DLS data (Table 2). First, if adding DPAL to HSA induced the formation of additional large amyloid-like albumin aggregates, Z_ave would have jumped to micrometer sizes because larger particles would be overwhelming in the DLS measurements as light scattering is proportional to the 6th power of the diameter. However, from our DLS data, we observed the Z_ave values of DPAL/HSA mixture to be only ~39–42 nm, which are far below (~two orders of magnitude lower) the micrometer sizes. In this case, the absolute value of Z_ave may not be precisely determined due to high PDI, but the range of the observed Z_ave provides useful information, which is indicative of a lack of a significant number of micrometer-sized particles induced by mixing DPAL with albumin. Second, we also drew our conclusion based on the intensity peak distributions of our DLS data. When comparing the intensity distributions in DPAL alone, HSA alone, and the DPAL/HSA mixture together (Table 2), we found the data were consistent and informative. If DPAL and albumin remained as their original particles, then the mixture would have four different size populations (one population for DPAL and three populations for HSA), according to Table 2. HSA alone exhibited three populations: Peak 1 is likely due to monomeric albumin, Peak 2 oligomers, and Peak 3 large aggregates. The albumin large aggregate contributes about 11% of the total intensity in the sample of HSA alone. In the DPAL/HSA mixture, monomeric albumin was detected, and a liposome population was also detected. But the albumin large aggregate was either not detected or detected with a much smaller percentage (2.10–2.45%). From Table 2, we can see that about 78% of the total intensity of the mixture came from DPAL and about 22% of the intensity originated from albumin. Since 11% of the albumin intensity came from large aggregates, large aggregates contribute 2.4% (=11% × 22%) of the total intensity in the DPAL/HSA mixture. This means if large albumin aggregates remain unchanged in the presence of DPAL, then the micrometer-sized peak should contribute about 2.4% total intensity from the DPAL/HSA mixture. If additional large aggregates are induced by the presence of DPAL, micrometer-sized population should significantly exceed 2.4%. What we observed is 0–2.45% (Table 2) of the intensity from the micrometer-sized particles, suggesting that the percentage of large albumin aggregates is either reduced slightly by DPAL or remains virtually unchanged. In either case, there is no evidence from the DLS intensity distribution data to indicate that DPAL can induce additional amyloid-like large aggregates (μm size) of HSA under physiologically relevant conditions used in our study (i.e., 600 μM albumin, pH 7.4, 145 mM NaCl, and 37 °C). Instead, the data in Table 2 lead to the speculation that DPAL might be beneficial in terms of reducing large albumin aggregates that already exist. Since under certain conditions HSA aggregates could be turned to conduct self-seeding [12,20] or cross-seeding [21], leading to irreversible formation of amyloid-like structures, DPAL’s ability to reduce large serum albumin aggregates might have other therapeutic benefits in addition to antithrombotic actions. According to Vetri et al. [19], at pH 7.4, large aggregates of serum albumin may arise from relatively small yet ordered aggregates (oligomers) through many intermolecular β-sheet interactions. It is possible that DPAL disrupts or attenuates those β-sheet interactions, thus reducing large albumin aggregates (Peak 3).

2.4. Effects of DPAL on Protein Conformation of Serum Albumin

In the study of the effect of DPAL on the protein structure of serum albumin, both HSA and BSA were utilized. HSA and BSA have 76% sequence homology and similar protein structures [22]. HSA has only one tryptophan (Trp214) while BSA has two (Trp134 and Trp213).

Figure 2 shows the emission and uncorrected excitation spectra of intrinsic tryptophan fluorescence obtained from 600 μM HSA (top panel) and 600 μM of BSA (bottom panel) in 1.8 mL water (pH 7.4) containing 145 mM NaCl with and without DPAL ([phospholipid] = 200 μM). As discussed earlier, these concentrations were selected for use because they are within the physiological concentration range and mimic the amount of DPAL used in our previous mice study [7].

The emission spectra were obtained using excitation at 297 nm, by which only tryptophan, not tyrosine, fluorescence was detected. Trp214 is in a hydrophobic area in Domain II of HSA [23]. As shown in Figure 2 (top panel), DPAL did not cause any significant change (no more than 1–2 nm) in emission maxima or excitation maximum of HSA, indicating that the polarity of the local environment of Trp214 remained virtually unchanged in the presence of DPAL.

Similar spectral observations were made on BSA/DPAL mixtures (Figure 2, lower panel). Trp213 in BSA is embedded in a hydrophobic pocket of Domain II, just like Trp214 in HSA, whereas Trp134 in BSA is in a hydrophilic region near the surface of Domain I [24]. According to Figure 2 (lower panel), DPAL did not lead to any significant change in emission maxima or excitation maximum of BSA. This result suggests that DPAL does not induce any major structural changes to both the hydrophobic core and the hydrophilic surface where the BSA tryptophan residues reside.

To further study the effect of DPAL on the local environment of the tryptophan-containing peptide segment in serum albumin, we measured the steady-state polarization of tryptophan fluorescence in HSA and BSA in the absence and presence of DPAL. As shown in

Table 3. Effects of DPAL (200 μM phospholipids) on polarization of tryptophan fluorescence obtained from 600 μM serum albumin in water (pH 7.4) containing 145 mM NaCl measured at 37 °C using λ_ex = 297 nm. The polarization values are expressed as average ± standard deviations of 30 measurements.

λem (nm)	− DPAL	+ DPAL	p Value
HSA and DPAL (Batch 1)
350	0.288 ± 0.008	0.292 ± 0.008	0.202
BSA and DPAL (Batch 1)
350	0.287 ± 0.005	0.288 ± 0.006	0.493
BSA and DPAL (Batch 2)
350	0.283 ± 0.005	0.282 ± 0.005	0.653

, the polarization values of intrinsic fluorescence of serum albumin with and without DPAL were virtually the same (p > 0.2). Assuming that the fluorescence lifetime of tryptophan in serum albumin changes little by DPAL, the polarization data (Table 3) suggests that DPAL does not induce any significant structural changes in the immediate environment surrounding the tryptophan residues.

Furthermore, the lack of a significant change in emission maximum (Figure 2) and the lack of a significant increase in steady-state polarization of tryptophan fluorescence (Table 3) as well as the retaining of a high population (~20%) of particles with a monomeric Z_ave value of 7.7–8.4 nm (Table 2) suggest that serum albumin and DPAL do not readily form complexes under our current in vitro and prior in vivo study [7] conditions, despite albumin and DPAL having opposite charges on their surfaces. This result is not surprising, as Dubin and other colleagues previously demonstrated [25] that complexation between BSA and positively charged polyelectrolytes occurs only when localized “charge patches” on the BSA surface, rather than the global protein charge density, could reach a critical effective charge density. It is possible that such a critical “patch” charge density for BSA to form a complex with DPAL is not reached in our experimental conditions.

Figure 3A,B show that concentration dependent self-quenching of tryptophan fluorescence occurred in HSA solution in the absence of DPAL. Fluorescence intensity increased when HSA concentration increased from 10 to 200 μM (Figure 3A,C (dark squares)). However, fluorescence intensity decreased with increasing protein concentration, when [HSA] exceeded 200 μM (Figure 3B,C, dark squares). It is well known in the literature that such fluorescence self-quenching can occur when protein molecules oligomerize or aggregate, resulting in fluorophores climbing together in more proximity and consequently reducing fluorescence intensity (e.g., [26]).

Figure 3D,E show that the presence of DPAL makes subtle changes to the concentration dependence of the emission spectrum of HSA intrinsic fluorescence. In the presence of DPAL, self-quenching of HSA intrinsic fluorescence emerged but it occurred when [HSA] exceeded 100 μM (Figure 3C, open triangles), rather than 200 μM in the case without DPAL (Figure 3C, dark squares). This subtle difference may be explained by the proposition that DPAL reduces larger aggregates of albumin (Table 2; discussed earlier) but has little effect on oligomers (small aggregates) and that intrinsic fluorescence quenching associated with large aggregation and oligomerization have different protein concentration dependences.

In addition to self-quenching, the shape of the emission spectrum of HSA is informative. Figure 3A,B show that, in the absence of DPAL, the emission spectrum of HSA at 600 μM (Figure 3B) exhibited two emission peaks: one centered at 331 nm and the other at 367 nm. The ratio of these two intensities (F331/F367) increased when HSA was diluted, yielding F331/F367 = 1.08 at 600 μM HSA, 1.29 at 60 μM, and 1.60 at 10 μM (Figure 3F, dark squares). Dilution should reverse oligomers back to monomers. Thus, the peak at 367 nm, which decreased upon dilution, likely originates from HSA oligomers/aggregates. In the presence of DPAL, the ratio of F331/F367 also decreased with increasing [HSA] (Figure 3F, open triangles). However, the values of F331/F367 in the presence of DPAL (Figure 3F, open triangles) were slightly higher than those in the absence of DPAL (Figure 3F, dark squares). A higher F331/F367 for the cases with DPAL than the cases without DPAL supports the idea that DPAL reduces certain large albumin aggregates that already exist.

The present dynamic light scattering and fluorescence data suggests that DPAL does not have strong interactions with serum albumins, especially with HSA. This result is in sharp contrast to the interactions of warfarin and heparin (two commonly used anticoagulants) with albumin. Warfarin is 99% bound to serum albumin under therapeutic conditions, with a high binding constant, e.g., 2.5–9.0 × 10⁵ M⁻¹ [27]. The binding of warfarin alters the structure of HSA and leads to a red shift of the emission maximum and a significant decrease in intensity of HSA tryptophan fluorescence due to fluorescence energy transfer between albumin tryptophan and warfarin [28,29,30]. Heparin also binds strongly to serum albumin, forming a heparin–albumin complex through hydrogen bonding and van der Waals interactions, with a binding constant ranging from 3.9 × 10³ to 7.2 × 10⁴ M⁻¹ [31]. It has also been reported that heparin accelerates albumin aggregation [32]. In conjunction with heparin binding, the intrinsic fluorescence of BSA is quenched through the mechanism of static quenching and a blue shift (5 nm) in emission maximum occurs, indicating that heparin induces BSA structural changes near tryptophan to a more hydrophobic local environment [31]. Strong binding and the resulting structural changes to albumin have a negative impact on the bioavailability and anticoagulation activities of warfarin and heparin. DPAL is most likely not to have this problem as its interaction with serum albumin is rather weak, as discussed earlier (Table 2 and Table 3 and Figure 2 and Figure 3).

2.5. Effects of DPAL on Permeability Across Endothelial Layers Made of HUVECs

Cytotoxicity is a critical issue that needs to be addressed when developing new therapeutics. Wurtzel et al. previously reported that no apparent spontaneous bleeding and no other adverse effects in mice were observed following intravenous infusion of DPAL [7]. Using lactate dehydrogenase (LDH), we also demonstrated that DPAL has little cytotoxicity against primary HUVEC cell cultures [7]. In the present study, we further tested if DPAL had any adverse effects on HUVECs by using the transwell assay and the CyQUANT assay. Using multiple assays to assess cytotoxicity against cultured cells is important because different assays report different aspects of cell viability. LDH assay reflects the leakage of the cytosolic enzyme LDH due to damage to the cells’ plasma membrane. CyQUANT assay measures cell proliferation. The transwell assay tests if DPAL damages the endothelial layer made of HUVECs.

Endothelial barrier damage by DPAL was assessed by the amount of FITC-dextran (originally 0.1 mg/mL) leakage into the basal chamber of the transwell one hour after incubation (F530). The results are shown in Figure 4. We chose to monitor the transwell permeability for only one hour because our previous study showed that clearance of DPAL in mouse plasma is very fast, reducing to 25% in 3 h, and that DPAL is readily detectable in circulating mouse blood within one hour after intravenous injection of DPAL [7]. Max fluorescence shown in Figure 4 was generated by allowing for complete dye passage through the insert without a HUVEC monolayer. Each DPAL-treated group received a liposome concentration of 0.176 mM, consistent with prior experiments [7]. No statistically significant differences in FITC-dextran permeability (inferred by F530) were observed between DPAL-treated (“Dye + DPAL”; Figure 4) and the control groups (“Dye” alone; Figure 4) (p > 0.05).

Effects of DPAL on cell proliferation of HUVECs. Cytotoxicity was also assessed by quantifying cell proliferation, monitored by the fluorescence-based CyQUANT assay. The results are presented in Figure 5, which shows the effect of DPAL doses on the proliferation of HUVECs. Three doses were tested: 0.2 mM [1×], 0.4 mM [2×], and 0.6 mM [3×]. The 1× DPAL group came close to the liposome concentrations used in prior in vitro and in vivo experiments. No significant differences in cell proliferation were observed between DPAL-treated groups and the media control (p > 0.05), indicating that DPAL within the doses examined imposes little risk to HUVEC cell viability.

We also used the CyQUANT assay to evaluate whether the size of DPAL makes a difference in HUVEC cell proliferation. DPAL with Z_ave values of 78, 153, and 224 nm were examined. Cell proliferation was expressed relative to the media control (set at 100%). As shown in Figure 6, there were no significant differences observed among the DPAL-treated groups regarding liposome size and cell proliferation.

3. Material and Methods

3.1. Preparation of DPAL

Zn-DPA-Cy3[22,22] was obtained from Molecular Targeting Technologies (West Chester, PA, USA). The concentration of the Zn-DPA-Cy3[22,22] stock solution in ethanol was determined spectroscopically using the extinction coefficient of 127,100 M⁻¹cm⁻¹ in ethanol at 556 nm. POPC was purchased from Avanti Polar Lipids (Alabaster, AL, USA). Liposomes composed of Zn-DPA-Cy3[22,22] and POPC were prepared by the dry film and extrusion method as previously described [6]. Specifically, we prepared a liposome formulation composed of 3 mol% Zn-DPA-Cy3[22,22] and 97 mol% POPC, designated as DPAL, which is the same liposome formulation used in our recent antithrombotic studies on mice [7]. In brief, multilamellar vesicles were formed via vortexing the lipid mixture, in either water containing 145 mM NaCl (Batch 1) or 10 mM 2-[Tris(hydroxymethyl)-methylamino]-ethanesulfonic acid (TES) buffer (Batch 2) at pH 7.4 containing 145 mM NaCl, for ~3 min at ~40 °C. The multilamellar vesicles were extruded (Lipex Biomembranes, Vancouver, BC, Canada) 15 times through two stacked Nucleopore polycarbonate membranes (membrane pore size: 200 nm for both Batches 1 and 2) at ~45 °C under N₂ gas pressure to make unilamellar vesicles (SUVs).

Particle size and size distribution of DPAL in water were determined on a Malvern Zetasizer Nano ZS instrument (Worcestershire, UK). The hydrodynamic diameter (Z_ave) was determined using light scattering and the Stokes-Einstein equation. Particle size distribution was reflected by the polydispersity index (PDI). Z_ave and PDI of Batch 1 were 159 nm and 0.19, respectively, whereas Z_ave and PDI of Batch 2 were 153 nm and 0.10. In addition, DPAL with two other sizes (Z_ave = 78 and 224 nm) was prepared by extrusion with 50 nm and 400 nm Nucleopore polycarbonate membranes, respectively. After being flushed with nitrogen, DPAL dispersions were stored in the dark in a screw-capped Pyrex glass tube wrapped with Teflon tape and parafilm prior to use.

3.2. Serum Albumin

Bovine serum albumin (BSA, catalog # A0281 or A6003, essentially free from fatty acids) and human serum albumin (HSA, catalog # A1887, essentially free from fatty acids) were purchased from Sigma (St. Louis, MO, USA). Albumins were dissolved in deionized water containing 145 mM NaCl (pH 7.4). The BSA and HSA solutions were made at a 600 μM concentration mimicking physiological conditions.

3.3. ADIFAB Assay

1-(6-(dimethylamino)naphthalen-2-yl)prop-2-en-1-one (Acrylodan) labeled intestinal fatty acid-binding protein (ADIFAB) (FFA Sciences, San Diego, CA, USA) was used to determine the amount of free fatty acids present in DPAL dispersions. ADIFAB assay is one of the most sensitive methods to detect free fatty acids (sensitivity: 1 nM) [33]. However, due to the presence of cyanine 3 (Cy3) in DPAL, fluorescence energy transfer occurs from acrylodan to cyanine 3, which renders the traditional ADIFAB assay [33,34] not useful for this study. Instead of using the fluorescence intensity ratio at 505 nm over 432 nm [33,34], we used the ADIFAB fluorescence intensity at 425 nm (F₄₂₅) and oleic acid sodium salt (OAss, from Sigma, Catalog # O-7501) in alkaline water (pH 8.3 or 10.3) to generate a standard curve relating F₄₂₅ to the amount of OAss in the sample. F₄₂₅ was selected because at this wavelength there is no overlap of acrylodan emission with the Cy3 excitation spectrum; thus, the energy transfer problem can be circumvented. Specifically, 0, 50, 100, 150, and 200 μL of 50 nM OAss freshly made in alkaline water were pipetted to 1.8 mL of alkaline water (pH 8.3 or 10.3) containing 9 μL of 60 μM ADIFAB in HEPES measuring buffer (FFA, San Diego, CA, USA) and then the fluorescence emission spectra were recorded upon excitation at 370 nm. A linear standard curve of normalized F₄₂₅ (against F₄₂₅ without OAss) versus OAss content was established (Figure 1). To determine the unknowns, 2 μL of DPAL dispersions (where the concentration of the POPC component ([POPC]_original) equals 2.657 mM) was mixed with 5 mL of water and 5 mL of chloroform:methanol (2:1, v/v). The organic layer was collected and dried first with a stream of nitrogen and then by high vacuum (~15 mbar) for one hour. The dried materials were redispersed in 1.8 mL alkaline water (pH 8.3 or 10.3) plus 9 μL of 60 μM ADIFAB in HEPES measuring buffer and the normalized F₄₂₅ values for the unknowns were determined. The amount of free fatty acid present in DPAL (the unknowns) was calculated by comparing the normalized F₄₂₅ values of the unknown with those in the standard curve.

3.4. Measurements of Intrinsic Albumin Fluorescence

Fluorescence excitation and emission spectra as well as steady-state fluorescence polarization of serum albumin with and without DPAL were measured on an ISS K2 fluorometer (Champaign, IL, USA). Polarization measurements were made using an L-format optical arrangement with a Glan-Thompson polarizer placed after the excitation monochromator and before the emission monochromator. Polarization values were determined after background fluorescence values from DPAL alone were subtracted from all fluorescence sample readings. Slit widths of the excitation and emission monochromators were set at 8 nm. The emission spectra and the steady-state polarization measurements of albumin tryptophan fluorescence with and without DPAL were obtained upon excitation at 297 nm.

3.5. Cell Culture

Primary human umbilical vein endothelial cells (HUVECs) were purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured according to the supplier’s recommendations. Cells were maintained in vascular cell basal medium (ATCC) supplemented with the endothelial cell growth kit-VEGF (ATCC) and 1% penicillin–streptomycin-Neomycin (Thermo-Fisher, Philadelphia, PA, USA), for the complete growth medium.

Cells were cultured in 100 mm tissue-culture treated dishes with a grip ring (Celltreat Scientific Products, Ayer, MA, USA) and maintained at 37 °C with 5% CO₂. The culture medium was refreshed every other day. Cells were detached at 80–90% confluence using 0.05% trypsin-EDTA (Corning, Manassas, VA, USA) and washed with 1× Dulbecco’s phosphate-buffered saline (DPBS) (Gibco, Palsley, UK). Cells were centrifuged (500 rpm, 5 min) and counted manually using trypan blue and a hemocytometer. Only cells under passage 6 were utilized for experiments to ensure consistency.

3.6. Cell Proliferation Assay

HUVECs were seeded in 96-well flat-bottom clear-bottom black plates at a density of 10,000 cells per well in 100 μL of complete culture medium. Cells were left to adhere and grow overnight at 37 °C in a 5% CO₂ incubator.

Prior to treatment incubation, medium containing unattached cells was removed from all wells and fresh medium was added. DPAL was diluted with fresh medium to a final concentration of 0.2 mM (DPAL 1×), 0.4 mM (DPAL 2×), and 0.6 mM (DPAL 3×) phospholipids for treatment wells and 1% TX-100 (Sigma-Aldritch, St. Luis, MO, USA) diluted with deionized water was added to positive control wells. Each treatment condition had 6 wells. Cells were incubated under test conditions at 37 °C in a 5% CO₂ incubator for 1–2 h. Following the incubation period, the medium was removed from all wells. Cells were washed once with 1× DPBS, and the plate was kept at −20 °C overnight to allow for cell lysis until further analysis. DPAL of varying particle sizes (Z_ave: 78 nm, 153 nm, and 224 nm) was assessed for cytotoxicity following the above-described protocol.

CyQUANT reagent (Invitrogen, Carlsbad, CA, USA) was prepared according to the manufacturer’s instruction, and the plate was allowed to thaw to room temperature. Each well received 200 μL of CyQUANT reagent and the plate was incubated for 5 min at room temperature in the dark. The fluorescence intensity was measured using a microplate reader, SpectraMax i3x (Molecular Devices, San Jose, CA, USA), with an excitation wavelength of 480 nm and an emission wavelength of 520 nm.

Background fluorescence values were subtracted from all fluorescence readings. Cell numbers were then determined using a standard curve generated in parallel during the same experiments. The control condition (no DPAL added) was defined as 100% cell proliferation, and proliferation in experimental conditions was expressed as a percentage relative to the control. Statistical significance was determined using a two-tailed t-test, with six replicates per condition. Differences were considered significant at p < 0.05 between control and test conditions.

3.7. Transwell Assay

HUVECs were seeded at a density of 80,000–100,000 cells per insert (8.0 μm pore size, 12 mm diameter; Millipore, Burlington, MA, USA) in 400 μL of complete culture medium in the apical chamber with 600 μL of complete culture medium in the basal chamber for each transwell system. Cells were allowed to attach and form a monolayer on the insert over 14 days and were maintained at 37 °C with 5% CO₂. Both apical and basal chambers received fresh medium every other day.

Medium was refreshed in both chambers on the day of the experiment to ensure consistent volume across all wells. For the treatment group, the apical chamber received 0.176 mM of DPAL and 0.1 mg/mL of fluorescein isothiocyanate-dextran (FITC-dextran, FD40S, average molecular weight 40,000, Sigma-Aldrich, St. Louis, MO, USA). To generate maximum fluorescence control data, only 0.1 mg/mL of FITC-dextran was added to the apical chamber with no HUVEC monolayer, allowing for complete dye diffusion. Wells were kept at 37 °C with 5% CO₂ during the 1-h incubation period.

After incubation, 200 μL of the medium from the basal chambers of each well was collected and diluted with 1.6 mL of 145 mM NaCl. The fluorescence readings were scanned for emission at 500–630 nm, with an excitation wavelength of 485 nm and the peak fluorescence intensity was recorded at 530 nm. Readings were obtained through an ISS K2 fluorometer (Champaign, IL, USA).

Background fluorescence generated through the liposome itself is subtracted from all readings as DPAL contains an intrinsic fluorophore (i.e., cyanine 3) with an overlapping emission range of FITC-dextran. Statistical significance between the dye group and the DPAL + dye group was determined using a two-tailed t-test, with four replicates per condition. Differences were considered significant at p < 0.05 between control and test conditions.

4. Concluding Remarks

The development of antithrombotic agents that minimize the risk of hemorrhage remains a critical unmet need, particularly in the context of acute thrombotic events such as ischemic stroke and venous thromboembolism, where timely intervention and inhibiting recurrence are essential but current therapies are limited by bleeding complications. As the incidence of thromboembolic disorders continues to rise with an aging population, there is increased clinical demand for treatment options that provide effective antithrombotic protection while still preserving hemostatic function. In our previous study [7], we characterized DPAL as a novel antithrombotic agent with selective PS-blocking activity and a favorable safety profile. DPAL demonstrated targeted binding to activated platelets and inhibition of the tissue factor (TF)-driven extrinsic coagulation pathway, while sparing the thrombin-mediated intrinsic pathway. Specifically in murine models, DPAL exhibited rapid onset of action, attenuated injury-induced arterial occlusion in a dose-dependent manner but did not compromise hemostatic plug stability. Furthermore, treated animals showed no significant alterations in blood cell counts or prolongation of bleeding time in tail-bleed assays, indicating a broad therapeutic window. These findings support the continued investigation of DPAL as a potential hemostasis-sparing antithrombotic agent.

In the present study, our results demonstrated that mixing DPALs with serum albumin does not induce any significant protein conformational changes near tryptophan residues nor any larger aggregates, indicating stability and biocompatibility. Instead, our data suggests that DPAL might have the ability to reduce the population of large, micrometer-sized albumins. Furthermore, our data showed that DPAL does not cause increased dye permeability across HUVEC layers and that HUVEC cell proliferation remains normal when treated with 1–3× doses of DPAL (1× = 0.2 mM phospholipids) and with DPAL of varying diameters (78–224 nm). To this end, our in vitro studies showed that DPAL has little cytotoxicity against HUVECs based on three different assays, i.e., the transwell (Figure 4), CyQUANT (Figure 5 and Figure 6), and lactate dehydrogenase assay [7]. In an earlier study, we also showed that DPAL (3 mol% DPA-Cy3[22,22] in POPC) has no significant cytotoxicity against MCF-7 breast cancer cells and MCF-12A breast non-cancer cells [6].

These findings together suggest that DPAL could offer a promising approach to modulate harmful coagulation pathways, providing a new targeted therapeutic strategy for managing thrombotic disorders and improving patient outcomes. While DPAL is a promising new antithrombotic agent (US patent number 11,090,309 [8]), the underlying molecular and mechanistic understanding of its therapeutic effect remains elusive.