Many drugs interact with membranes in order to pass into cells to exert their intracellular effects or to penetrate to hydrophobic sites on membrane-associated channels and enzymes. Local anesthetics (LAs), class I antiarrhythmic agents, and other amphipathic agents exert their primary actions on ionic Na⁺ channels from the cytosol after permeating the plasma membrane or from within the membrane phase itself [1] [2] [3] [4] [5]. Equally important may be the proposed role of the membrane in controlling the rate-limiting step for dissociation of the drug from its target protein, e.g. via the so-called “hydrophobic pathway” [6] [7].

Even more important for clinical applications, LAs must pass through epidural and perineural sheaths in order to reach the nerve fibers and block conduction [8] [9]. Delivery from the extraneural space, where injectate is deposited, to the intraneural compartment, is very inefficient, with only about 5% of the LA molecules in an injected 1% lidocaine solution (37 mM) actually entering the nerve [10]. But the high concentration of injectate is essential for the effective block; in vivo studies in rats show that, for bolus injection, at least 1% - 2% (37 - 75 mM) lidocaine is required for the transient conduction block of all fibers whereas at equilibrium (30 - 60 min exposure) only 1 - 2 mM is fully effective [11] [12]. Therefore, improving the efficiency of LA membrane permeation would allow the use of lower concentrations and, possibly, reduce adverse side effects.

The equilibria of drug:membrane interactions have been extensively investigated previously using a variety of physical methods, including X-ray diffraction [13] [14], NMR [15] [16] [17] [18], ESR [19] [20] [21] and FTIR [22] [23] [24]. These methods, while providing good signal resolution, are far too slow to allow measures of the rates of membrane:drug interactions. Moreover, NMR usually requires a high concentration of the test substance, whereas high concentrations of tetracaine may significantly change a membrane’s structure [16] [20] [25] [26].

To avoid these limitations we have previously used the intrinsic fluorescence of the LA tetracaine, at low concentrations, to study its equilibrium interactions with detergent micelles [27] [28], and lipid bilayer membranes [29]. In the present work, we have extended these equilibrium studies by examining the adsorption and permeation rates of the local anesthetics tetracaine and procaine with model membranes. Membranes formed as unilamellar vesicles were made of dipalmitoylphosphatidylcholine (DPPC), without or with cholesterol (27.5 mol%), a normal component of plasma membranes, were used to answer the following questions: 1) How does LA hydrophobicity and ionization state affect membrane adsorption and permeation? 2) How does the molecular composition or physical state of the membrane (solid gel vs disordered) affect rates of LA adsorption into and permeation through a membrane?

Briefly, the adsorption rate was obtained from the increase in the LA’s intrinsic fluorescence (at 350 nm/302nm excitation) with membrane binding [29]. The permeation rate of LA across the membrane was obtained from the decrease in fluorescence intensity (509 nm) of intravesicular pyranine, resulting from both direct interaction of intravesicular pyranine with LAs as they translocated into the intra-vesicular aqueous compartment, and from the pH rise caused by the basic LA’s accumulation inside the vesicle [30] [31].

The role of drug hydrophobicity is explored here by comparing moderately hydrophobic procaine with its more highly hydrophobic homologue, tetracaine. Drug structures are shown in Figure 1, and Partition Coefficients between octanol and buffer and between membranes and buffer are listed in Table 1.

Procaine’s membrane location. The fluorescence of LA’s results from light emission from the excited N-substituted aromatic region of the molecules (Figure 1). Fluorescence of neutral PRO (measured at pH 11.0) was greatly enhanced and the emission fluorescence maximum (302 nm excitation) was blue-shifted, from 355 nm for aqueous PRO to 344 nm when bound to DPPC

*Rate of mole fractions for LA-Ocranol and LA-DPPC systems. ⁺From Strichartz et al. [30]. **From Zhang et al. [29].

membranes (data not shown). The equivalent dielectric constant of membrane-bound PRO, determined by the maximum blue-shift calibrated by the same method of dioxane-water system used for TTC [29], was 10 - 15, far less than an aqueous solvent’s (80) and slightly higher than that of membrane-bound neutral TTC, 6 - 10, implying that neutral PRO binds less deeply in the membrane than neutral TTC. However, for PROH⁺ no fluorescence maxima shift and only a 20% increase in fluorescence intensity was observed upon membrane binding, suggesting that PROH⁺ is located at an even more superficial location than PRO.

The membrane:buffer partition coefficient, expressed as: [molesPRO/molesDPPC]/[molesPRO/moles H₂O], is 2.5 greater for PRO than PROH⁺, the same ratio occurs for unionized to ionized tetracaine, although tetracaine partitions to DPPC membranes 2.8 times more than procaine (Table 1). In other words, LA charge affects equilibrium adsorption to liquid crystal bilayers to the same extent, regardless of the difference in a molecule’s hydrophobi city or location on the membrane’s surface. Importantly, partition coefficients of LAs to membranes are 50 - 10,000 greater than for octanol, demonstrating how poor a system that solvent is for modelling membranes.

Kinetics of membrane adsorption of TTC.

Effect of Ionization on adsorption rates. Typical curves showing the increased fluorescence of tetracaine when rapidly mixed with DPPC membranes are shown in Figure 2. At time 0 the mixing of an equal volume of 2 mM DPPC

LUVETs with 40 μM TTC was complete and the fluorescence changes reflect the binding of tetracaine to the membrane, at pH values ranging from 10.5 (upper curve) to 6.6. The curves were fit by a single exponential function to extract time constant (τ) and amplitude changes (DF; see Methods). Both rates and steady-state changes (after 0.3 sec) increase with higher pH, showing the faster and greater adsorption of the uncharged drug species. The plot of final DF against pH (inset, Figure 2) is fit with a pK_a of 8.12 ± 0.1, in good agreement with the pK_a of 8.17 for DDPC adsorption reported previously [29], and lower than the pK_a of tetracaine in buffer, 8.59 ± 0.04 [30]. Time constants for a simple first-order process, such as adsorption: desorption, are related to the rate constants for those processes by Equation (1):

τ − 1 = k on [ TTC ] + k off (1)

where k_on is the first-order rate constant for adsorption and k_off the desorption rate constant. The plot of inverse time constant (τ⁻¹) versus uncharged TTC concentration is shown in Figure 3, based on τ values determined from fits to the data like those in Figure 1. Linear regress ion fit allows calculation of k_on, from the slope, and k_off, from the Y-axis intercept. The neutral TTC concentrations were calculated using a pK_a value of aqueous tetracaine [30] and the free [TTC] = 20 mM. (The correlation coefficients of the fit, 0.87 and 0.89, are the same (p > 0.10, t-test) whether the pK_a of aqueous or of membrane-bound tetracaine is used to determine charged and neutral concentrations.) A similar analysis for ionized TTCH⁺ yields rate constants for the charged species. These rate constants are displayed in Table 2, along with the initial rates of adsorption determined by the initial slope of the traces in Figure 2. (The negative values of k_on is real and probably results from the electrostatic repulsion of free TTCH⁺ by the

charged TTCH⁺ bound to the membrane surface; see Discussion). For neutral TTC adsorption the calculated k_on from fits to Figure 3 are the same as those from the initial slope (p > 0.05), about 2 - 5 × 10⁴ M⁻¹∙s⁻¹. Therefore, at 20 µM TTC the product of concentration and adsorption rate constant equals an adsorption rate of 0.4 - 1.0 s⁻¹ compared to ~16 s⁻¹ for the desorption rate. Using the initial binding rate to estimate k_on for TTCH⁺ shows, by the same analysis, an adsorption rate of 0.012 s⁻¹ and a desorption rate of 20 s⁻¹. Thus, desorption of the neutral and charged species of tetracaine from DPPC membranes occur at about the same rate whereas adsorption of the charged species is about 8-fold slower than that of the neutral form. But regardless of the ionization state of tetracaine, the desorption reaction is the dominant term controlling equilibrium partitioning.

Effect of temperature on adsorption rates. In order to study the effect of membrane phase, experiments were also conducted at 45˚C, where the membranes are in the liquid-crystal state rather than the solid-gel phase that exists at 23˚C (the phase transition temperature for DPPC membranes is 41˚C). However, we found that the membrane binding of TTC at this higher temperature was too fast to be measured. Figure 4 shows the fluorescence intensity for TTCH⁺ (40 µM, pH = 5.97) after mixing with 2 mM DPPC, at 45˚C (open squares). Fluorescence traces for dilution of 40 µM TTC in buffer (stars) and for 2 mM DPPC mixed with buffer (open circles) are also shown, along with the equilibrium spectra of those mixtures after 10 mins (inset). The enhanced intensity and blue shift of TTC plus membranes (inset) shows that adsorption did occur, but too rapidly to be detected (mixing time = 0.008 sec). Therefore, the adsorption of tetracaine to liquid-crystal membranes is much faster than adsorption to solid-gel membranes (e.g., Figure 2), paralleling the larger partition coefficient in the liquid crystal [29].

Inclusion of cholesterol in DPPC membrane accelerates membrane binding. Cholesterol’s ability to reduce the mobility of phospholipids in liquid crystal membranes both by increasing hydrocarbon chain ordering and increasing membrane thickness of bilayers is well documented [35] [36] [37] [38] [39]. Inclusion of cholesterol in pure DPPC membranes in the solid-gel state at 23˚C also changes the dynamics of the lipids, disordering the otherwise tightly packed acyl chains and causing steric crowding near the polar headgroups [38] [39]. We found that the adsorption of tetracaine into DPPC-Chol membranes was too fast to be detectable both at 23˚C and 45˚C (data not shown). Apparently, the

loosening of tight packing of the phospholipid acyl chains, by the addition of cholesterol or by the change in phase at high temperature, both increase the adsorption to undetectably fast rates. Since, as noted above, the time constant for the membrane binding reaction is dominated by the desorption rate, the now undetectably fast rate means that the desorption rate is much greater than in the Chol stiffened membranes, an observation consistent with the reduction of TTC’s equilibrium partitioning by Chol inclusion [29]. Tetracaine does not bind as tightly to membranes when Chol is present because it dissociates from the membrane more rapidly.

Local anesthetic permeation through membrane.

Interaction of pyranine with TTC. As shown in Figure 5, the fluorescence of pyranine in solution (20 µM) was progressively reduced at increasing concentrations of TTC. The maximum inhibition was reached at a TTC concentration of about 40 mM (inset, Figure 5), allowingdynamic response to TTC at the concentrations studied here. Similar inhibitions of pyranine fluorescence were observed at other experimental conditions, e.g., pH 11 and 45˚C, and pH 7.4 and 23˚C. A control kinetic experiment with TTC added to pyranine showed that the inter action was complete within 8 ms, the dead time of the mixing accessory, which is much faster than the permeation time constants (shown below) and will not limit their measurement.

Effect of ionization of TTC on permeation through solid-gel membrane.

Permeation of tetracaine was monitored by changes in intravesicular pyranine, which does not pass across the membrane [30] [31]. Typical curves of

pyranine fluorescence after addition of TTC to DPPC LUVETS are shown in Figure 6, for a range of pHs, at 23˚C. In every case, the intra-vesicular and external solution were at the same initial pH. The initial difference in pyranine fluorescence (at t = 0) is due to its inherent pH-dependence, but the subsequent drop in fluorescence reflects the rate at which TTC enters the intra-vesicular space, faster at higher pHs than at lower pHs, suggesting that the neutral species permeates these solid-gel DPPC membranes more rapidly than the charged species. The curves were fitted with single exponential functions to extract a permeation time constant (τ). A graph of τ⁻¹ against the concentration of neutral tetracaine, calculated for the different pH values used in Figure 6 at 20 µM TTC, is shown in Figure 7. The slowest rate, with τ⁻¹ = 0.028 ± 0.01 s⁻¹, occurs for ionized TTCH⁺, at pH values less than 6.5; rates increase with increasing fraction of TTC in the neutral form. Note that the rates for permeation, indicated by these inverse time constants, are more than 100 times smaller than those for membrane adsorption (Figure 3).

The rate constants of TTC permeation in and out the DPPC bilayer at 23˚C can be obtained from the slopes of plots of τ⁻¹ vs TTC concentration (cf. Equation (1), above) and are collected in Table 3. Negative values of k_in represent the downward curve of permeation rate with concentration in Figure 7, suggesting that the permeation process reaches a rate-limiting step at relatively low TTC

concentrations. The reverse rate constant, k_out, at pH 11 is twice the value at pH 5.9 showing that the neutral species of TTC has a net permeation twice as fast as the charged species.

Inclusion of cholesterol slows TTC’s permeation through solid-gel membranes. The effect of cholesterol on tetracaine’s permeation through DPPC membranes at pH 9 is shown in Figure 8(a). Permeation through the DPPC-Chol membrane is more than 10-times slower than that through the DPPC membrane. Figure 8(b) shows pyranine fluorescence time course for TTC permeation through DPPC-Chol membrane at 23˚C at different pHs. (Note the difference in the Y axis scale with Figure 8(b)). No significant fluorescence signal decays were observed at pH 5.98 to 10.7, showing that cholesterol inclusion greatly slows permeation for both charged and neutral tetracaine species.

Comparison of TTC and PRO permeation through solid-gel membranes.

The permeation time-courses for tetracaine and procaine through DPPC solid-gel membranes (23˚C) at physiological pH (7.4) are shown in Figure 9(a). Permeation of the two LAs occur at virtually identical rates; single exponential fitting to data gave time constants of 58 and 60 s, at 40 µM LA. Interestingly, cholesterol inclusion slowed PRO’s permeation about 16-fold, 4-times more of an effect than that experienced by TTC, which was slowed 4.4-fold (Figure 9(b)). The crowding of DPPC phospholipids by cholesterol appears to have a greater effect on the transmembrane transport of more superficially bound procaine than on more deeply bound tetracaine.

Local anesthetic molecules interact with membranes in order to cross them and reach intracellular sites and to reach hydrophobic loci on membrane-intrinsic proteins. Local anesthetics are not freely dissolved in membranes but are predominantly adsorbed near the membrane’s surface, with the ionizable amine group facing towards the aqueous (extracellular or intracellular) compartment and the aromatic moiety directed towards the membranes apolar hydrocarbon core (Figure 1). Ionic interactions between charged amino groups and polar head groups of phospholipids, dipole-dipole interactions between the LA ester or amide bind and the glycerol-lipid ester bonds of the membrane lipids, and hydrophobic interactions between the LA’s aromatic regions and the acyl side chains of the phospholipids all contribute to the energetics of binding to and transport across the membrane.

We have found that the homologous LAs tetracaine and procaine bind rapidly and reversibly to simple bilayer membranes and are transported across them at a much slower rate. The unionized form of each LA partitions 2.5-times that of the ionized form, and more hydrophobic tetracaine partitions 2.8-times more than procaine, comparing both ionized and ionized species. Shifts in the fluorescence emission spectra upon adsorption to the membranes indicate that neutral and protonated tetracaine are in the same region of low dielectric constant, close to that of the phospholipid acyl chains (16 carbons for DPPC), despite the effect of ionization on partitioning. In contrast, procaine, whose neutral form appears to reside more superficially than TTC’s, perhaps close to the glycerol ester bonds of DPPC, is located even further onto the surface of the membrane when ionized.

Rates for tetracaine’s partitioning to solid-gel membranes are about 8 times faster for the neutral than the ionized species, reflected in a slower adsorption rate for the charged species. Rates for tetracaine’s adsorption to liquid-crystal membranes are too fast to be measured, for either ionization state, suggesting that the relative immobility of the lipids in the solid gel slows the desorption process and tends to stabilize the LA on the membrane. Inclusion of cholesterol into the membrane also speeds the adsorption process to undetectable rates, perhaps due to the crowding of phospholipids in the head group region and the resultant weaker adsorption energy, consistent with the weaker equilibrium partitioning caused by cholesterol [29].

These findings are consistent with published studies of equilibrium binding of these LAs to membranes of various lipid composition. Tetracaine has been shown by NMR and X-ray diffraction to reside at a deeper position within the membranes [40] [41] [42] [43], and to be able to shift rapidly between multiple binding loci [42], and procaine to be located at a more superficial site [40] [42].

The rates of permeation of tetracaine and procaine through membranes are also responsive to drug ionization and membrane state. Permeation occurs far more slowly than the equilibration reaction of partitioning, so the rate-limiting step cannot be membrane binding. Neutral species of these LAs permeate more rapidly than charged species because of the high energy required for a charged form to pass into and through the core hydrocarbon with its very low dielectric constant. Despite their different hydrophobicities and equilibrium partitioning, tetracaine and procaine permeate pure DPPC (solid-gel) membranes at the same rate. Lower partitioning of procaine corresponds to a weaker binding energy in the membrane and, perhaps a lower activation energy to cross the membrane’s hydrocarbon core, whereas the higher partitioning of tetracaine will concentrate more drug on the membrane, and at a deeper location from which it may experience a lower energy in crossing the core. Inclusion of cholesterol, which is a common component of all mammalian plasma membranes, lowers partitioning and slows permeation of both LAs, but of procaine more than tetracaine. Cholesterol’s crowding of the polar head groups in phospholipid membranes reduces lateral mobility and thus may slow LA penetration between DPPC molecules, at a locus which is more populated by procaine than by tetracaine molecules. But a reduction in acyl chain rotational mobility also occurs with cholesterol inclusion and may also contribute to the reduced transport.

It is essential for an understanding of LA interactions with membranes to realize that the membrane is itself directly affected by the bound drug. At the very low concentrations (20 µM) used in this study, tetracaine equilibrium partitioning results in a ratio of 1 TTC molecule per 250 DPPC molecules, and 1 TTCH⁺ molecule per 580 DPPC molecules (see Equation (1), Results). At the much higher concentrations needed for conduction block in exposed nerves, about 1 mM [11], each TTC molecule is present for every 5 - 10 lipid molecules. Membrane lipid fluidity, the rotational mobility and lateral diffusion of lipid molecules, are altered by the presence of LAs [23] [43] [44]. The dynamic exchange of lipids between the liquid crystal and solid gel domains that occurs in membranes, particularly real biological membranes composed of many different phospholipids and cholesterol, is shifted towards the fluid phase by the adsorbed LAs [44] [45] [46] [47], changing the structure and activity of membrane-intrinsic proteins such as ion transporters and pumps. In addition, the positive charge on ionized LAs is attracted to negative surface potential of membranes due to anionic phospholipids, and LA adsorption at the surface lowers that potential and thus alters the local concentration of charged molecules in solution [48] [49] [50]. Through these different mechanisms bound LAs will reduce their own membrane affinity; with increasing concentrations greater energy is required for subsequent binding, leading to a down-turning curve in the concentration dependence of LA binding. Such behavior likely accounts for the negative association rate constants observed in the tetracaine binding experiments reported here.

In summary, local anesthetics bind rapidly and reversibly to membranes at their outer surface and are transported more slowly across the hydrocarbon core. Drug ionization lowers binding to membranes of neutral phospholipids and drug hydrophobicity raises that binding; ionization slows drug permeation but greater hydrophobicity does not affect it. The fluidity of membranes, characterized by a dynamic fluctuation between solid-gel and liquid-crystal domains, favors LA binding and permeation and itself is shifted towards the more fluid behavior by the adsorbed LA molecules.

All experiments reported here and much of the analysis was conducted by Dr. Zhingzhong Zhang, The author was not able to contact Dr. Zhang, who therefore could not approve of its content but who unquestionably would have been included as an author.

Research supported by USPHS grant R01-GM064792. The secretarial support of Ms. Ellen Jacobson is greatly acknowledged.

The author declares no conflicts of interest regarding the publication of this paper.

Strichartz, G. (2022) Kinetics of Local Anesthetic Interactions with Model Membranes: Adsorption and Permeation. Open Journal of Anesthesiology, 12, 55-75. https://doi.org/10.4236/ojanes.2022.121006