Effect of CaCl₂ on the Structure of DODAB Vesicles Containing the Ganglioside GD3

Abstract

CaCl₂ has been successfully used in the formulation and stabilization of vaccines. Vesicles made of the disialoganglioside GD3 and the cationic lipid DODAB are promising immunotherapeutic tools, by combining the targeting properties of GD3 and the adjuvant properties of DODAB. In this work, the effects of CaCl₂ on the structure of DODAB and DODAB+GD3 vesicles have been compared by Differential Scanning Calorimetry (DSC) and Electron Paramagnetic Resonance (EPR) spectroscopy. EPR spectra show that CaCl₂ does not induce the formation of GD3 domains by lateral phase separation. DSC thermograms show that CaCl₂ hinders the formation of a subgel phase for vesicles cooled below 15 °C. CaCl₂ also decreases the hysteresis between gel–fluid and fluid–gel transitions in both pure and mixed vesicles. EPR data show that CaCl₂ increases the superficial rigidity of DODAB and DODAB+GD3 vesicles at temperatures below the gel–fluid transition. In this temperature range, CaCl₂ does not affect the packing of the DODAB vesicle core but increases the packing of the DODAB+GD3 vesicle core. EPR data also show that CaCl₂ decreases the rigidity of both types of vesicles at temperatures above the gel–fluid transition. These results might help overcome the challenges of vaccine and immunotherapy cold chain logistics.

1. Introduction

The disialoganglioside GD3 is the precursor of the “b” series gangliosides, being the most abundant ganglioside in human colostrum. In mice, it is important for early neurological development, being abundant in neural stem cells. GD3 is involved with apoptosis control and is overexpressed in several types of cancer, such as melanoma, glioblastoma, and neuroblastoma. On the other hand, it has a low and restricted expression in extra-neural adult tissues. Moreover, it has been shown to induce responses by iNKT cells. , This makes GD3 an attractive target for immunotherapy, and it has been tested in several clinical tests. The success of these tests might depend on the use of adequate adjuvants.

Dioctadecyldimethylammonium bromide (DODAB) is a synthetic cationic lipid that induces strong immune responses. DODAB vesicles have been successfully employed in drug delivery , and as vaccine adjuvants. ,, This makes DODAB a suitable carrier for GD3, and recently, DODAB vesicles containing GD3 had their structure characterized.

One of the challenges in the development of new immunotherapeutic tools, such as vaccines, is the dependence on cold chain logistics, since freezing and thawing might affect their structures and stabilities. , Indeed, there has been much effort in the development of protective excipients to stabilize vaccines.

CaCl₂ has been used in the formulation and stabilization of vaccines , and in the preparation of drug delivery systems. The ionic pair that forms this salt also has strong biological relevance. For instance, Ca²⁺ ions are important second messengers whose concentrations are strictly controlled in the cellular milieu. They are also involved in important processes such as membrane fusion and the formation of membrane domains. Indeed, Ca²⁺ ions can interact with gangliosides and induce the lateral phase separation necessary to form the domains. − Cl^– ions are also important to regulate cell cycle, proliferation, volume, and pH levels.

The effects of CaCl₂ on phospholipid membrane structures have been extensively studied and even employed in the preparation of drug and vaccine delivery systems, such as cochleates. However, the effects of CaCl₂ on cationic vesicles have not been addressed thus far.

In the present work, the effects of CaCl₂ on the structure of DODAB and DODAB+GD3 vesicles are characterized by differential scanning calorimetry (DSC) and electron paramagnetic resonance (EPR) spectroscopy. DSC gives information about the thermotropic behavior of lipid vesicles, and EPR spectroscopy gives information about lipid order and mobility at different depths of the assemblies. Indeed, DSC and EPR have been effectively combined to characterize mixtures of DODAB and other glycolipids. −

2. Experimental Methods

2.1. Materials

HEPES buffer (Lot# SLCB8158) and DODAB (Lot# BCBR1922 V) were purchased from Sigma Chemical Co. (St. Louis, MO, USA). CaCl₂ was purchased from Honeywell/Fluka (Lot No. H265A). GD3 from bovine milk (Lot no. 860060–01–010) and spin labels 1-palmitoyl-2-(n-doxylstearoyl)-sn-glycero-3-phosphocholine (n-PCSL, n = 5 or 16) were purchased from Avanti Polar Lipids (Birmingham, AL, USA). Figure shows the chemical structures of lipids and spin labels.

1.

Chemical structures of DODAB, 5-PCSL, 16-PCSL, and GD3.

2.2. Preparation of Vesicles

Vesicles were prepared as previously described. Briefly, chloroform/methanol solutions (2:1 v/v) containing the lipids were dried under a nitrogen stream and left under reduced pressure for 2 h. The resulting lipid films were then dispersed in HEPES buffer (10 mM, pH 7.4) containing the desired CaCl₂ concentration (0 mM, 1 mM, 10 mM, or 100 mM). The dispersion was achieved by heating the samples for 20 min at 87 °C in a water bath. Heating was interspersed with vortexing every 5 min to ensure a homogeneous dispersion.

For EPR experiments, 0.8 mol % of 5-PCSL or 0.3 mol % of 16-PCSL was added to the chloroform/methanol solutions before lipid film formation, and the samples were dispersed as described above. The final lipid concentrations were 2 mM DODAB and 0.1 mM GD3. At these concentrations, the preparations resulted in unilamellar vesicles.

2.3. Differential Scanning Calorimetry (DSC)

DSC thermograms were obtained in a VP-DSC Microcalorimeter (Microcal Inc., Northampton, MA). Heating and cooling rates were set at 20 °C/h, and a heating (annealing) scan at 90 °C/h was performed before all experiments. Vesicles were equilibrated at 5 °C for 15 min before the heating scans. Scans were performed with at least two samples prepared on different days. Thermograms were analyzed using the Microcal Origin software, and the following thermodynamic parameters were determined: the phase transition temperature (T_m), phase transition enthalpy (ΔH), and width at half-maximum (ΔT1/2). T_m is the highest value of heat capacity variation (ΔC_p). ΔH is the enthalpy change per mol of lipid, and its values were obtained by integrating the area under the thermograms. ΔT_1/2 values were obtained from the most intense peaks and reflect the phase transition cooperativity: the less cooperative a transition is, the larger the ΔT _1/2 needed to complete it. All experiments were performed at least in duplicate with samples prepared on different occasions.

2.4. Electron Paramagnetic Resonance (EPR) Spectroscopy

In this work, the spin labels 5-PCSL and 16-PCSL were used to probe the regions near the surface and the core of the vesicles’ bilayers (Figure ). These spin labels contain a nitroxide radical whose hyperfine structure consists of 3 lines. These lines have equal heights if the nitroxide group is moving rapidly in the EPR time scale. An increase in the immobilization of the spin label leads to differential line broadening in its EPR spectrum, and this anisotropy can be used to examine the environment of the probe. A detailed discussion of the EPR technique regarding its use to investigate the structure of cationic vesicles is given in ref .

A Bruker EMX spectrometer equipped with the ER4119HS high-sensitivity cavity was used to obtain CW-EPR spectra in the X-band. Sample temperatures were kept within 0.1 °C using a Bruker BVT-2000 variable temperature device. The values of maximum (A_max) and minimum (A_min) hyperfine splittings, and of low (h₊₁), central (h₀) and high (h_‑1) field amplitudes, were measured directly from the spectra (see Figures and ).

3.

EPR spectra of 5-PCSL and 16-PCSL embedded in 2 mM DODAB vesicles or in 2 mM DODAB + 0.1 mM GD3 vesicles, with or without 100 mM CaCl₂, at temperatures below the gel–fluid transition. The maximum hyperfine splitting (A_max) and the amplitudes of low (h₊₁) and central (h₀) field lines are indicated. The total spectra width is 100 G.

5.

EPR spectra of 5-PCSL and 16-PCSL embedded in 2 mM DODAB vesicles or in 2 mM DODAB + 0.1 mM GD3 vesicles, with or without 100 mM CaCl₂, at temperatures above the gel–fluid transition. The maximum and minimum hyperfine splittings (A_max and A_min), and the amplitudes of central (h₀) and high (h_‑1) field lines are indicated. The total spectral width is 100 G.

The effective order parameter, S_eff, was calculated from the expression

$$S_{eff}=\frac{A_{\parallel }-A_{\perp }}{A_{zz}-(1/2)(A_{xx}+A_{yy})}\frac{a_o^{\prime }}{a_o}$$

where theoretical and experimental isotropic hyperfine splittings, ${\mathrm{a}}_{\mathrm{o}}^{\prime }$ and a _o respectively, can be calculated from the expressions: ${\mathrm{a}}_{\mathrm{o}}^{\prime }=(1/3)(A_{xx}+A_{yy}+A_{zz})$ , a_o = (1/3)­(A _//+2A _⊥), A_// (= A_max) is the maximum hyperfine splitting directly measured in the spectrum (see Figure ), ${\mathrm{A}}_{\perp }={\mathrm{A}}_{\min }+1.4[1-\frac{{\mathrm{A}}_{//}-{\mathrm{A}}_{\min }}{{\mathrm{A}}_{zz}-(1/2)({\mathrm{A}}_{xx}+{\mathrm{A}}_{yy})}]$ , A_min is the measured inner hyperfine splitting (see Figure ) and A _xx , A _yy, and A _zz are the principal values of the hyperfine tensor for doxylpropane.

Each experiment was performed at least twice with fresh samples, and the standard deviations are shown as error bars.

3. Results

3.1. Effect of CaCl₂ on the DODAB and DODAB+GD3 Vesicles Phase Behavior

The heating and cooling thermograms of pure DODAB vesicles and mixed DODAB + GD3 vesicles in the presence or absence of CaCl₂ are shown in Figure .

2.

Effect of CaCl₂ on the DSC thermograms of DODAB (A) and DODAB+GD3 (B) vesicles. Scan rates for heating and cooling were 20 °C/hour.

When heated, pure DODAB vesicles in the absence of CaCl₂ display a small and broad endothermic peak at 29.9 °C, corresponding to the subgel-gel pretransition, and a sharp endothermic peak at 46.9 °C, corresponding to the main gel–fluid phase transition of unilamellar vesicles (Figure A). These peaks have been extensively described. , In the presence of 1 mM CaCl₂, a less intense subgel-gel peak is still observed, but in the presence of 10 mM and 100 mM CaCl₂, this pretransition peak disappears (Figure A). The main gel–fluid peak remains nearly unchanged at 46.9 °C in the presence of all CaCl₂ concentrations tested (Figure A, Table ).

1. Heating Cycle Thermodynamic Parameters (± Standard Deviations) of 2 mM DODAB and 2 mM DODAB + 0.1 mM GD3 Dispersions in the Absence and Presence of CaCl₂ .

		Temperature range			Temperature range			Temperature range
		5–20 °C			20–35 °C			35–80 °C
Dispersion	[CaCl2] (mM)	Tm (°C)	ΔH (kcal/mol)	ΔT1/2 (°C)	Tm (°C)	ΔH (kcal/mol)	ΔT1/2 (°C)	Tm (°C)	ΔH (kcal/mol)	ΔT1/2 (°C)
		Heating cycle
DODAB	0	-	-	-	29.9 ± 0.3	3.4 ± 0.9	4 ± 1	46.9 ± 0.1	14 ± 2	0.7 ± 0.1
DODAB	1	-	-	-	29.3 ± 0.3	1.3 ± 0.8	2.9 ± 0.4	46.7 ± 0.1	14 ± 2	0.6 ± 0.1
DODAB	10	-	-	-	-	-	-	47.0 ± 0.1	15.0 ± 0.6	0.6 ± 0.1
DODAB	100	-	-	-	-	-	-	47.0 ± 0.1	14.3 ± 0.8	0.6 ± 0.1
+ GD3	0	12 ± 1	-2.4 ± 0.2	7 ± 1	29.3 ± 0.3	5.2 ± 0.6	4.1 ± 0.1	45.8 ± 0.1	12.5 ± 0.3	1.7 ± 0.1
+ GD3	1	12.9 ± 0.1	-2 ± 1	9 ± 1	28.5 ± 0.1	3.7 ± 0.9	3.3 ± 0.2	45.5 ± 0.1	15 ± 2	1.7 ± 0.1
+ GD3	10	-	-	-	-	-	-	46.2 ± 0.1	13.8 ± 0.8	1.6 ± 0.2
+ GD3	100	-	-	-	-	-	-	45.7 ± 0.1	15 ± 1	2.2 ± 0.1

When cooled, the pure DODAB vesicles display a single sharp exothermic peak at 41.5 °C, corresponding to the fluid-gel transition (Figure A), as previously described. In the presence of 1 mM CaCl₂ the peak remains at 41.5 °C, but in the presence of 10 mM and 100 mM CaCl₂ it is shifted to somewhat higher temperatures (Figure A, Table ).

2. Cooling Cycle Thermodynamic Parameters (± Standard Deviations) of 2 mM DODAB and 2 mM DODAB + 0.1 mM GD3 Dispersions in the Absence and Presence of CaCl₂ .

		Temperature range			Temperature range			Temperature range
		5–20 °C			20–35 °C			35–80 °C
Dispersion	[CaCl2] (mM)	T m (°C)	ΔH (kcal/mol)	ΔT 1/2 (°C)	T m (°C)	ΔH (kcal/mol)	ΔT 1/2 (°C)	T m (°C)	ΔH (kcal/mol)	ΔT 1/2 (°C)
		Heating cycle
DODAB	0	-	-	-	-	-	-	41.5 ± 0.1	-11.7 ± 0.7	0.8 ± 0.1
DODAB	1	-	-	-	-	-	-	41.5 ± 0.1	-13 ± 2	0.7 ± 0.1
DODAB	10	-	-	-	-	-	-	41.8 ± 0.1	-14 ± 1	0.9 ± 0.1
DODAB	100	-	-	-	-	-	-	42.2 ± 0.1	-14 ± 2	1.0 ± 0.1
+ GD3	0	-	-	-	-	-	-	40.6 ± 0.1	-11.3 ± 0.4	1.3 ± 0.6
+ GD3	1	-	-	-	-	-	-	40.4 ± 0.1	-13.3 ± 0.1	1.0 ± 0.1
+ GD3	10	-	-	-	-	-	-	41.2 ± 0.2	-13.5 ± 0.7	1.1 ± 0.1
+ GD3	100	-	-	-	-	-	-	42.3 ± 0.1	-12 ± 2	1.0 ± 0.1

When heated, the mixed DODAB + GD3 vesicles display a broad and shallow exothermic peak around 12 °C, followed by a more intense endothermic subgel-gel pretransition peak at 29.3 °C (Figure B). An endothermic peak at 45.8 °C corresponds to the main gel–fluid transition (Figure B); it is broader and less intense than the one observed for pure DODAB vesicles (Figure A). This complex thermotropic behavior upon heating has already been described and discussed for mixed DODAB+GD3 vesicles.

The DODAB + GD3 heating thermogram profile is not significantly changed in the presence of 1 mM CaCl₂, except that the endothermic pretransition peak is shifted to 28.5 °C (Figure B). In the presence of 10 mM and 100 mM CaCl₂, the exothermic peak around 12 °C and the endothermic pretransition peak disappear (Figure B). With these higher CaCl₂ concentrations, the endothermic peak corresponding to the main gel–fluid transition remains around 45.8 °C (Figure B).

When cooled, the mixed DODAB + GD3 vesicles have a single sharp exothermic peak at 40.6 °C, corresponding to the fluid-gel transition (Figure B). This peak shifts to higher temperatures in the presence of CaCl₂ concentrations above 1 mM (Figure , Table ).

As observed in Figure , the heating and cooling cycles result in very different thermogram profiles for both pure and mixed vesicles. Hence, the thermodynamic parameters obtained from these thermograms are summarized in Table for the heating cycles and in Table for the cooling cycles.

The thermodynamic analyses were performed in separate temperature ranges to facilitate the discussion of the thermal events observed in the thermograms. A similar approach was previously used for the analyses of mixed DODAB + GD3 vesicles. For instance, the temperature range of 5 to 20 °C is suitable for the analysis of the exothermic peak at 12 °C that appears in the heating scans of mixed DODAB + GD3 vesicles (Figure B). The T_m, ΔH, and ΔT_1/2 values for this peak are not significantly affected by the presence of 1 mM CaCl₂ (Table ). However, in the presence of higher CaCl₂ concentrations, this peak is absent (Figure B).

In the temperature range of 20 to 35 °C, it is possible to observe that 1 mM CaCl₂ causes a decrease in the pretransition ΔH for both pure DODAB and mixed DODAB + GD3 vesicles (Table ). This concentration also causes a slight increase in pretransition cooperativity for both kinds of vesicles, since ΔT_1/2 values are decreased in the presence of 1 mM CaCl₂ (Table ). In mixed vesicles, 1 mM causes a shift in the position of the pretransition peak from 29.3 to 28.5 °C (Table ), as observed in Figure B. In the presence of higher CaCl₂ concentrations, there is no pretransition peak (Figure A and B).

In the temperature range of 35 to 90 °C, increasing CaCl₂ concentrations does not significantly affect the T_m, ΔH, and ΔT_1/2 values of the gel–fluid transition for pure DODAB vesicles (Table ). For mixed DODAB + GD3 vesicles, the increase in CaCl₂ concentrations does not significantly affect the T_m values but causes a slight increase in ΔH and ΔT_1/2 values (Table ). This effect of CaCl₂ concentration on ΔH and ΔT_1/2 is more evident in the presence of 100 mM CaCl₂ (Table ).

For the cooling cycles (Table ), it is possible to observe that CaCl₂ does not affect the ΔH and ΔT_1/2 values for pure DODAB vesicles, but it increases the T_m values for CaCl₂ concentrations above 1 mM. A similar behavior is observed for mixed DODAB+GD3 vesicles is the presence of CaCl₂ (Table ).

For both pure DODAB and mixed DODAB + GD3 vesicles, the gel–fluid and fluid–gel transitions do not occur at the same temperature (Figure , Tables and ). This temperature hysteresis can be quantified as the difference between T_m values for heating and cooling cycles, as shown in Table .

3. Temperature Hysteresis (± Standard Deviations) between Heating and Cooling Cycles for 2 mM DODAB and 2 mM DODAB + 0.1 mM GD3 Dispersions in the Absence and Presence of CaCl₂ .

Dispersion	[CaCl2] (mM)	Tm heating (°C)	Tm cooling (°C)	Hysteresis (Tm heating – Tm cooling) (°C)
DODAB	0	46.9 ± 0.1	41.5 ± 0.1	5.4 ± 0.2
DODAB	1	46.7 ± 0.1	41.5 ± 0.1	5.2 ± 0.2
DODAB	10	47.0 ± 0.1	41.8 ± 0.1	5.2 ± 0.2
DODAB	100	47.0 ± 0.1	42.2 ± 0.1	4.8 ± 0.2
+ GD3	0	45.8 ± 0.1	40.6 ± 0.1	5.2 ± 0.2
+ GD3	1	45.5 ± 0.1	40.4 ± 0.1	5.1 ± 0.2
+ GD3	10	46.2 ± 0.1	41.2 ± 0.2	5.0 ± 0.3
+ GD3	100	45.7 ± 0.1	42.3 ± 0.1	3.4 ± 0.2

Pure DODAB vesicles, in the absence of CaCl₂ have a thermal hysteresis of 5.4 °C, as previously described. , For these vesicles, a slight decrease in hysteresis is observed as CaCl₂ concentrations are increased (Table ). Mixed DODAB + GD3 vesicles have a hysteresis of 5.2 °C, very similar to that of pure DODAB vesicles, as previously described. For mixed vesicles, the hysteresis also decreases as CaCl₂ concentration increases, and this effect is more evident in the presence of 100 mM CaCl₂, since hysteresis is reduced by almost 2 °C (Table ).

Considering that the thermodynamic differences observed in Figure and Tables , , and are more evident in the presence of 100 mM CaCl₂, this saturating concentration was chosen to investigate the effect of CaCl₂ on the structure of the vesicles by EPR spectroscopy.

3.2. Effect of CaCl₂ on the Structure of DODAB and DODAB + GD3 Vesicles at Temperatures below the Gel–Fluid Transition

The structures of DODAB or DODAB + GD3 vesicles, in the presence and absence of 100 mM CaCl₂ were compared using the paramagnetic probes 5-PCSL and 16-PCSL (Figure ). These probes give information about the microviscosity and packing near the interface and the core of lipid bilayers, respectively. The very small concentrations of probes used in this work do not affect the structure or thermotropic behavior of the vesicles. Indeed, the presence of these probes does not affect the thermograms of the vesicles (see Supporting Information Figure 1).

Figure shows the spectra of these probes embedded in vesicles in the presence or absence of CaCl₂ at temperatures below the gel–fluid transition. For clarity, spectra at other temperatures in this range are shown in Supporting Information Figure 2.

The spectra of 5-PCSL and 16-PCSL are very different at all temperatures and for all samples tested (Figure ). While the spectra of 5-PCSL are characteristic of an anisotropic environment, in which the movement of the probe is more restricted due to increased molecular packing, the spectra of 16-PCSL are typical of a more isotropic environment, where the probe has higher mobility and lower order due to decreased molecular packing (Figure ).

Hence, the spectra of 5-PCSL and 16-PCSL show the presence of a gradient of increased flexibility from the surface to the core of the bilayers, which is typical of noninterdigitated lamellar structures. This flexibility gradient has been described for DODAB vesicles in the gel phase, , and from Figure it is possible to observe that it is not affected by the presence of GD3, CaCl₂, or both.

The spectra display features of a single EPR signal at all conditions tested (Figure , Supplementary Figure 2). Considering that both probes are miscible with DODAB and GD3, , the presence of a single EPR signal shows that these lipids are miscible and that no lateral phase separation is induced by CaCl₂ at these temperatures. The presence of a single EPR signal allows for the determination of empirical parameters from the spectra. These parameters can be used to compare the structures of vesicles under different conditions.

The maximum hyperfine splitting (A_max), for instance, can be directly obtained from the 5-PCSL spectra (Figure ). A_max is sensitive to the packing and viscosity of the probe microenvironment: a decrease in packing or viscosity results in smaller A_max values. In contrast, A_max values cannot be accurately determined from 16-PCSL spectra due to increased isotropy (Figure ). For this probe, structural information can be obtained from the ratio of the low- and central-field-line amplitudes (h₊₁/h₀) (Figure ). The values of the h₊₁/h0 ratio increase as the bilayers become less packed.

Figure shows the A _max and h₊₁/h₀ values for DODAB and DODAB+GD3 vesicles in the presence and absence of CaCl2 at temperatures below the gel–fluid transition.

4.

Effect of 100 mM CaCl₂ on the maximum hyperfine splitting (A_max) of 5-PCSL (A, C) and on the ratio of the low and central field line amplitudes (h₊₁/h₀) of 16-PCSL (B, D) embedded in 2 mM DODAB or 2 mM DODAB + 0.1 mM GD3 vesicles at temperatures below the gel–fluid transition. Error bars indicate standard deviations.

As expected, A_max values decrease as the temperature increases because membrane packing decreases as temperature increases (Figure A,C). In the presence of CaCl₂, the A_max values for pure DODAB vesicles increase, indicating that CaCl₂ increases the superficial packing of these vesicles (Figure A). A similar but less intense effect is observed for DODAB+GD3 vesicles (Figure C).

The decrease in packing with temperature increase is also observed in the membrane core, since h₊₁/h₀ values increase (Figure B,D). For pure DODAB vesicles, CaCl₂ does not have a clear effect on membrane core packing (Figure B). For DODAB+GD3 vesicles, h₊₁/h₀ values decrease in the presence of CaCl₂, indicating that CaCl₂ induces an increase in DODAB+GD3 membrane core packing. Hence, CaCl₂ has a rigidifying effect on vesicles formed by a mixture of DODAB and GD3 at temperatures below the gel–fluid transition.

3.3. Effect of CaCl₂ on the Structure of DODAB and DODAB + GD3 Vesicles at Temperatures above the Gel–Fluid Transition

The spectra of 5-PCSL and 16-PCSL embedded in vesicles, in the presence or absence of CaCl₂ at temperatures above the gel–fluid transition, are shown in Figure .

When compared to the spectra at temperatures below the gel–fluid transition (Figure ), the spectra of 5-PCSL shown in Figure have thinner features, which indicate that the movement down the long axis of this probe is faster. This is typical of a fluid yet microscopically ordered structure close to the membrane interface. Similarly, the spectra of 16-PCSL at higher temperatures display sharp peaks (Figure ), indicating a fast and nearly isotropic movement of the nitroxide label, which is typical of the fast motional regime observed in fluid bilayers.

The spectra of both probes are very similar under all conditions tested (Figure ), so the empirical parameters can be more informative about the structural organization of DODAB and DODAB+GD3 vesicles in the presence and absence of CaCl₂ above the gel–fluid transition.

The maximum and minimum hyperfine splittings (A_max and A_min) can be accurately measured in the 5-PCSL spectra (Figure ). With their values, it is possible to calculate the effective order parameter (S_eff, see Section ), which gives information about the hydrocarbon chain order. ,

The 16-PCSL spectra sense a less ordered environment, so the hyperfine splittings are not clearly defined (Figure ). However, the amplitudes of the high (h_‑1) and central field lines (h₀) can be accurately measured (Figure ). The ratio of these amplitudes (h_‑1/h₀) can be used to assess the viscosity and packing in the same manner as described for h ₊₁ /h ₀ in the previous section: the values of h_‑1/h₀ increase as the membrane viscosity decreases.

The values of S_eff and h_‑1/h₀ for DODAB and DODAB+GD3 vesicles, in the presence and absence of CaCl₂ at temperatures above the gel–fluid transition, are shown in Figure .

6.

Effect of 100 mM CaCl₂ on the effective order parameter (S_eff) of 5-PCSL (A, C) and on the ratio of the central and high field line amplitudes (h_‑1 /h₀) of 16-PCSL (B, D) embedded in 2 mM DODAB or 2 mM DODAB + 0.1 mM GD3 vesicles at temperatures above the gel–fluid transition. Error bars indicate standard deviations.

S_eff values are smaller in the presence of CaCl₂ for both DODAB and DODAB+GD3 vesicles (Figure A,C), showing that CaCl₂ decreases the superficial order of these vesicles. Similarly, CaCl₂ decreases the molecular packing of the membrane core of both DODAB and DODAB+GD3 vesicles, as seen from the higher h_‑1/h₀ values in the presence of this salt (Figure B,D). In summary, CaCl₂ decreases the rigidity of DODAB and DODAB+GD3 vesicles in the fluid phase.

4. Discussion

Pure DODAB vesicles have a rich phase behavior, which depends on temperature, concentration, and the presence of additives. In the diluted regime tested here, DODAB vesicles can form subgel, gel, and fluid phases. ,

The subgel phase is formed when the vesicles are cooled below 15 °C. , This cooling enables a reorganization of DODAB molecules from a close-to-hexagonal to a close-to-triclinic 2D lattice. , This new molecular arrangement results in more tightly packed acyl chains, with a consequent freezing of the chains’ motion and their further stretching. ,

When subgel phase vesicles are heated, a pretransition peak appears on the thermograms as a reflex of the energy needed to complete the order–disorder transition to the gel phase (Figure A). For pure DODAB vesicles prepared in water, the pretransition peak is centered at 36 °C and is almost as intense as the main gel–fluid transition peak. , For vesicles prepared in HEPES buffer, the pretransition peak is shifted to 29.9 °C, and the peak is significantly decreased when compared to the gel–fluid transition peak, as observed in Figure A and Table , and as previously described. These differences suggest that the increase in ionic strength might affect the formation of the subgel phase due to electrostatic shielding. Indeed, the pretransition peak decreases and disappears as CaCl₂ concentration increases in the buffered samples (Figure A, Table ).

Although the chloride ion (Cl^–) has a smaller polarizability and binding affinity to the quaternary ammonium headgroup when compared to bromide ions (Br^–), − at saturating concentrations, it might effectively shield the electrostatic repulsion between cationic DODAB headgroups and could inhibit the transformation of a gel-phase hexagonal arrangement to a subgel triclinic arrangement. In agreement with that, CaCl₂ does not significantly affect the acyl chain packing at the membrane core (Figure B), but it does increase the superficial packing of the vesicles (Figure A). Similarly, the cationic lipid didodecyldimethylammonium bromide (DDAB) displays a phase-separated region with two coexisting lamellar phases. , One of these phases disappears when Br^– ions are replaced by Cl^– ions. ,

The mixed DODAB+GD3 vesicles also display a pretransition peaking at 29.3 °C when heated, which is preceded by a broad and shallow exothermic peak around 12 °C (Figure B, Table ). This suggests that GD3 does not interfere with the formation of the subgel phase, but that this subgel phase is possibly metastable, with a superficial condensation effect caused by the bulky and charged headgroup and a decrease in packing of the core due to steric effects of the hydrophobic chains of GD3. Stepwise changes that reflect head/tail differences in molecular interactions have already been described for DODAB during the fluid-to-coagel phase transition.

Both exothermic and pretransition peaks disappear as CaCl₂ concentrations increase (Figure B, Table ). The disappearance of these peaks is probably related to the electrostatic shielding effects of CaCl₂. As discussed above, Cl^– shielding of cationic headgroups could inhibit the rearrangement of DODAB molecules into a subgel phase. Moreover, Ca²⁺ could interact with the sialic acid residues of GD3 (Figure ), changing their conformation and transmitting this conformational change to the GD3 hydrocarbon chains. It was shown that CaCl₂ induces a micellar-to-lamellar change in dispersions of pure GD3, with an increase in superficial and core rigidity. Although the vesicles are already arranged as lamellae, a similar effect is observed here (Figure ): in the presence of CaCl₂ there is a superficial packing increase (Figure C) and an increase in packing of the membrane core (Figure D).

After the pretransition is completed, pure DODAB vesicles are at gel phase, and a single peak at 46.9 °C corresponding to the gel–fluid transition is observed (Figure A and Table ). This transition temperature is higher than the one at 44 °C described for DODAB vesicles prepared in water, − probably because the HEPES buffer has an electrostatic shielding effect that increases the molecular packing of the gel phase.

CaCl₂ does not significantly affect the T_m, ΔH, and ΔT_1/2 values of the gel–fluid transition for pure DODAB vesicles at any of the concentrations tested (Table ), but a substantial increase in superficial packing is observed in the presence of 100 mM CaCl₂ (Figure A). On the other hand, CaCl₂ does not affect the molecular packing of the membrane core (Figure B). This suggests that although Cl^– ions can shield the electrostatic repulsion between DODAB headgroups and increase their packing, the Cl^– ions do not affect the overall molecular packing and thermotropic behavior of a gel phase that is already more packed due to HEPES buffer when compared to pure water, as discussed above.

In the mixed DODAB+GD3 vesicles, the gel–fluid phase transition temperature and cooperativity are decreased when compared to pure vesicles (Figure B,A, Table ). This could be explained by a condensation effect of the bulky and charged GD3 headgroup, that increases the superficial rigidity of DODAB vesicles, combined with a decrease in membrane core rigidity caused by the V-shaped configuration of GD3 ceramide that occupies a large area and disrupts hydrocarbon chain interactions of DODAB molecules, as previously described.

CaCl₂ does not affect the T_m of the mixed vesicles, and a slight increase in ΔH and ΔT_1/2 values is only observed in the presence of 100 mM CaCl₂ (Figure B, Table ). The increase in ΔH values could be explained by the rigidifying effect of CaCl₂ on the surface and core of the mixed vesicles (Figure C and D). This rigidifying effect could be due to conformational changes in the headgroups and hydrocarbon chains induced by Ca²⁺ interactions with the sialic acid residues of GD3 (Figure ), as discussed before.

The increase in ΔT_1/2 values indicates that 100 mM CaCl₂ decreases the gel–fluid transition cooperativity in mixed vesicles (Table ). This result is quite unexpected, since the increase in ΔH (Table ) and the rigidity of both the membrane surface and core (Figure C and D) would point instead to an increase in cooperativity. It is known that cooperativity does not correlate well with other thermodynamic parameters because it is sensitive to the specific character of intermolecular chain packing, such that the ability to communicate chain disordering can be very inefficient in either very tight or very weak packing lattices. Indeed, cooperativity is dependent on the morphology of bilayers, and it was shown that DODAB bicelles prepared by sonication present a sharp decrease in cooperativity when compared to DODAB vesicles. , These bicelles exhibit EPR spectra with two signals corresponding to the coexistence of a gel planar domain and a fluid rim domain. Hence, this decrease in cooperativity suggests that 100 mM CaCl₂ could be changing the morphology of mixed DODAB+GD3 vesicles, but this possible change does not seem to be so drastic as to induce the formation of bicelles, because the EPR spectra shown in Figure and Supplementary Figure 2 display a single signal. The formation of curled or faceted vesicles is more likely. ,

In the cooling cycles, concentrations of CaCl₂ above 1 mM induce an increase in the fluid-gel transition temperature for pure DODAB vesicles (Table ), indicating that it is not necessary to cool down the samples as much as in the absence of CaCl₂ to transition to the gel phase. This increase in the T_m for cooling results in a decrease in hysteresis from 5.4 °C in the absence of CaCl₂ to 4.8 °C in the presence of 100 mM CaCl₂ (Table ). Temperature hysteresis is common in charged lipids because the paths for gel–fluid and fluid-gel transitions are different: near the transition temperature, domains of one phase are present in the bulk matrix of the other phase, resulting in strain energy differences. If all these domains were collected into a single large domain, hysteresis would be reduced because the strain energy would also be reduced. Hence, the decrease in hysteresis caused by CaCl₂ suggests that this salt can reduce strain energy, possibly by facilitating the formation of larger domains near the transition temperature.

The effect of different counterions on the structure of DODAB vesicles has been explored for the gel phase. , It was shown that the smaller and less hydrated Br^– ion has a larger binding affinity for the quaternary ammonium headgroup than Cl^–, as expected from the Hofmeister series effects. , This means that Br^– as a counterion yields a “softer”, less tightly packed vesicle when compared to Cl^–. The EPR data in Figure A and B, however, show that in the presence of CaCl₂, the DODAB vesicles are less tightly packed than in absence of CaCl₂. A possible explanation for this difference is that specific ion (Hofmeister series) effects are dependent on temperature.

Similar effects of increasing fluid-gel transition temperature (Table ), decreasing hysteresis (Table ), and decreasing the fluid phase packing (Figure C and D), are observed for mixed DODAB+GD3 vesicles in the presence of CaCl₂. In this case, the effects are stronger than for the pure DODAB vesicles (Tables and , Figure C and D), suggesting that the interaction of Ca²⁺ ions with the sialic acid of GD3 headgroups might be playing a synergistic role with the Cl^– interacting with the DODAB headgroups.

GD3 is an attractive target for cancer immunotherapy, but clinical trials exploring this ganglioside did not achieve major impacts on patient outcomes so far. This could be due to the use of inadequate adjuvants, since cancer immunotherapy strongly depends on strong cellular responses. Since DODAB is able to induce the formation of both humoral and cellular responses, , vesicles of DODAB containing GD3 had their structures recently characterized to enable the development of new immunotherapeutic tools.

DODAB vesicles would be interesting as adjuvants for GD3-based therapies because cationic vesicles are known to be more potent adjuvants than neutral and anionic ones. Moreover, vesicles with higher gel–fluid transition temperatures usually elicit stronger immune responses, and Table shows that all vesicles tested here have gel–fluid phase transitions well above the physiological temperature of 37 °C.

Temperature is also critical for effective vaccination programs because vaccines have an inherent thermal instability that requires a cold chain of storage and distribution. , While mRNA-based vaccines are highly unstable when not frozen, most other types of vaccines rely on a cold chain with temperatures in the range of 2 to 8 °C. This is the case for nonreplicating vaccines, such as inactivated viruses/bacteria, purified subunit protein and carbohydrate antigens, virus-like particles, and vesicles that can be formulated as liquid solutions and stored in glass vials or prefilled syringes. These kinds of vaccines usually need stabilizers for the cold chain. ,,

The present work expands the characterization of DODAB+GD3 vesicles by exploring the effects of CaCl₂, which has been used to stabilize vaccines. ,, For the initial thermodynamic characterization using DSC, 1 mM, 10 mM, and 100 mM CaCl₂ were tested (Figure and Tables , and ). These concentrations were chosen because CaCl₂ is a divalent salt: for instance, 1 mM CaCl₂ would dissociate into 2 mM negative charges from Cl^– anions and 2 mM positive charges from Ca²⁺ cations. This would allow for the negative Cl^– charges to interact in an equimolar proportion to the monovalent cationic headgroup of 2 mM DODAB (Figures and ), and for the positive cationic Ca²⁺ charges to interact in a 10:1 proportion to the two anionic sialic residues of 0.1 mM GD3 (Figures and ).

It was shown that the stabilizing effects of CaCl₂ are not linear with its concentration: for instance, foot-and-mouth disease vaccines have increased thermal stability with the addition of up to 10 mM CaCl₂, while a decreasing trend was observed for concentrations above 30 mM CaCl₂. Hence, it was important for this work to test higher concentrations. The initial concentration of 1 mM CaCl₂ was increased by 10- and 100-fold, and the thermodynamic differences observed in Figure and Tables , , and are more evident in the presence of 100 mM CaCl₂. Consequently, this saturating concentration was chosen to investigate the effect of CaCl₂ on the structure of the vesicles by EPR spectroscopy, as previously mentioned.

As with other liquid vaccine formulations, the DODAB and DODAB+GD3 vesicles would most likely be kept in a 2 to 8 °C cold chain, which coincides with the temperature range of the subgel phase (Figure ). , The temperature of 5 °C to begin the DSC and EPR experiments, was chosen for this reason, and the disappearance of the subgel phase with increasing CaCl₂ concentrations (Figure ) could have impacts on vaccine distribution and/or potency.

The formation of the subgel phase creates a line tension with the gel phase, possibly because these phases have different thicknesses, which result in enhanced hydrophobic defects at the nucleating boundary. Since gel-phase DODAB vesicles are known to have many hydrophobic defects, , which can be observed in the faceted aspect of gel-phase giant unilamellar vesicles, it is possible that these defects are further enhanced in the subgel phase. Since hydrophobic defects destabilize membranes and facilitate membrane fusion, it is possible that the subgel phase is more prone to fusion events than the gel phase. Hence, the disappearance of the subgel phase with the addition of CaCl₂ might stabilize the DODAB vesicles at cold-chain temperatures.

Understanding the mechanisms of vaccine inactivation is very important to preserve their potency, , just as understanding the structure, stability, and fluidity of vesicles is key for their use in vaccines. Hence, characterizing the effects of CaCl₂ on the structure and the thermotropic phase behavior of DODAB and DODAB+GD3 vesicles is very important for their application in immunotherapy.

5. Conclusions

The effect of CaCl₂ on the structure of DODAB and DODAB+GD3 vesicles was characterized by DSC and EPR spectroscopy. EPR spectra show that CaCl₂ does not induce the formation of GD3 domains by lateral phase separation. DSC thermograms show that CaCl₂ abolishes the DODAB subgel pretransition and the heating scan’s exothermic peak in vesicles containing GD3. It is possible that Cl^– ions hinder the transformation of a gel to a subgel phase by shielding the cationic headgroups’ electrostatic repulsion in vesicles cooled below 15 °C.

The thermograms also show that CaCl₂ decreases the hysteresis between gel–fluid and fluid–gel transitions in both pure and mixed vesicles.

Empirical analysis of EPR spectra shows that CaCl₂ increases the superficial rigidity of DODAB and DODAB+GD3 vesicles at temperatures below the gel–fluid transition. In this temperature range, CaCl₂ does not affect the packing of the DODAB vesicle core but increases the packing of the core of DODAB+GD3 vesicles.

The empirical analysis also shows that CaCl₂ decreases the superficial order and the molecular packing of the core of DODAB and DODAB+GD3 vesicles at temperatures above the gel–fluid transition.

Although the effects of CaCl₂ on phospholipid membrane structures have been extensively studied, this is the first work to investigate its effects on cationic membranes containing gangliosides. The results shown here might help develop stable dispersions that can withstand the challenges of vaccine and immunotherapy cold-chain logistics.

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.6c03945.

• Thermograms of vesicles containing spin labels and additional EPR spectra at temperatures below the gel–fluid transition (PDF)

The Article Processing Charge for the publication of this research was funded by the Coordenacao de Aperfeicoamento de Pessoal de Nivel Superior (CAPES), Brazil (ROR identifier: 00x0ma614).

The authors declare no competing financial interest.