Modeling the percutaneous absorption of solvent-deposited solids over a wide dose range: II. Weak electrolytes.

Dermal absorption of weak electrolytes applied to skin from pharmaceutical and cosmetic compositions is an important consideration for both their efficacy and skin safety. We developed a mechanistic, physics-based framework that simulates this process for leave on applications following solvent deposition. We incorporated this framework into our finite dose computational skin permeation model previously tested with nonelectrolytes to generate quantitative predictions for weak electrolytes. To test the model, we analyzed experimental data from an in vitro human skin permeation study of a weak acid (benzoic acid) and a weak base (propranolol) and their sodium and hydrochloride salts from simple, ethanol/water vehicles as a function of dose and ionization state. Key factors controlling absorption are the pH and buffer capacity of the dose solution, the dissolution rate of precipitated solids into a lipid boundary layer and the rate of conversion of the deposited solid to its conjugate form as the nonionized component permeates and (sometimes) evaporates from the skin surface. The resulting framework not only describes the current test data but has the potential to predict the absorption of other weak electrolytes following topical application.

Solvent and Crystallization Effects on the Dermal Absorption of Hydrophilic and Lipophilic Compounds.

This study probes the mechanisms by which volatile solvents (water, ethanol) and a nonionic surfactant (Triton X-100) influence the skin permeation of dissolved solutes following deposition of small doses onto unoccluded human skin. A secondary objective was to sharpen guidelines for the use of these and other simple solvent systems for dermal safety testing of cosmetic ingredients at finite doses. Four solutes were studied - niacinamide, caffeine, testosterone and geraniol - at doses close to that estimated to saturate the upper layers of the stratum corneum. Methods included tensiometry, visualization of spreading on skin, polarized light microscopy and in vitro permeation testing using radiolabeled solutes. Ethanol, aqueous ethanol and dilute aqueous Triton solutions all yielded surface tensions below 36 mN/m, allowing them to spread easily on the skin, unlike water (72.4 mN/m) which did not spread. Deposition onto skin of niacinamide (32 µg·cm-2) or caffeine (3.2 µg·cm-2) from water and ethanol led to crystalline deposits on the skin surface, whereas the same amounts applied from aqueous ethanol and 2 % Triton did not. Skin permeation of these compounds was inversely correlated to the extent of crystallization. A separate study with caffeine showed the absence of a dose-related skin permeability increase with Triton. Permeation of testosterone (8.2 µg·cm-2) was modestly increased when dosed from aqueous ethanol versus ethanol. Permeation of geraniol (2.9 µg·cm-2) followed the order aqueous ethanol > water ∼ 2 % Triton >> ethanol and was inversely correlated with evaporative loss. We conclude that, under the conditions tested, aqueous ethanol and Triton serve primarily as deposition aids and do not substantially disrupt stratum corneum lipids. Implications for the design of in vitro skin permeability tests are discussed.

Transdermal and lateral effective diffusivities for drug transport in stratum corneum from a microscopic anisotropic diffusion model.

This paper presents a computational model of molecular diffusion through the interfollicular stratum corneum. Specifically, it extends an earlier two-dimensional microscopic model for the permeability in two ways: (1) a microporous leakage pathway through the intercellular lipid lamellae allows slow permeation of highly hydrophilic permeants through the tissue; and (2) the model yields explicit predictions of both lateral (D‾‖sc) and transdermal (D‾⊥sc) effective (average, homogenized) diffusivities of solutes within the tissue. We present here the mathematical framework for the analysis and a comparison of the predictions with experimental data on desorption of both hydrophilic and lipophilic solutes from human stratum corneum in vitro. Diffusion in the lipid lamellae is found to make the effective diffusivity highly anisotropic, with the predicted ratio D‾‖sc/D‾⊥sc ranging from 34 to 39 for fully hydrated skin and 150 to more than 1000 for partially hydrated skin. The diffusivities and their ratio are in accord with both experimental data and the results of mathematical analyses performed by others.

Opportunities and challenges in the diagnostic utility of dermal interstitial fluid.

The volume of interstitial fluid (ISF) in the human body is three times that of blood. Yet, collecting diagnostically useful ISF is more challenging than collecting blood because the extraction of dermal ISF disrupts the delicate balance of pressure between ISF, blood and lymph, and because the triggered local inflammation further skews the concentrations of many analytes in the extracted fluid. In this Perspective, we overview the most meaningful differences in the make-up of ISF and blood, and discuss why ISF cannot be viewed generally as a diagnostically useful proxy for blood. We also argue that continuous sensing of small-molecule analytes in dermal ISF via rapid assays compatible with nanolitre sample volumes or via miniaturized sensors inserted into the dermis can offer clinically advantageous utility, particularly for the monitoring of therapeutic drugs and of the status of the immune system.

Impact of solvent dry down, vehicle pH and slowly reversible keratin binding on skin penetration of cosmetic relevant compounds: I. Liquids.

To measure progress and evaluate performance of the newest UB/UC/P&G skin penetration model we simulated an 18-compound subset of finite dose in vitro human skin permeation data taken from a solvent-deposition study of cosmetic-relevant compounds (Hewitt et al., J. Appl. Toxicol. 2019, 1-13). The recent model extension involved slowly reversible binding of solutes to stratum corneum keratins. The selected subset was compounds that are liquid at skin temperature. This set was chosen to distinguish between slow binding and slow dissolution effects that impact solid phase compounds. To adequately simulate the physical experiments there was a need to adjust the evaporation mass transfer coefficient to better represent the diffusion cell system employed in the study. After this adjustment the model successfully predicted both dermal delivery and skin surface distribution of 12 of the 18 compounds. Exceptions involved compounds that were cysteine-reactive, highly water-soluble or highly ionized in the dose solution. Slow binding to keratin, as presently parameterized, was shown to significantly modify the stratum corneum kinetics and diffusion lag times, but not the ultimate disposition, of the more lipophilic compounds in the dataset. Recommendations for further improvement of both modeling methods and experimental design are offered.

Absorption of solvent-deposited weak electrolytes and their salts through human skin in vitro.

Permeation of a weak acid (benzoic acid) and a weak base (propranolol) in various stages of ionization through human skin in vitro was measured from 0 to 72 h following solvent deposition of radiolabeled doses ranging from 11 to 11,000 nmol/cm2 and 1.93-1930 nmol/cm2, respectively. For the twenty combinations tested for each compound, mean permeation into the receptor fluid over 72 h ranged from 1.5 to 40.7 percent of dose for benzoic acid and 1.3-35.5 percent of dose for propranolol. For all but the lowest doses, permeation increased with increasing fraction of nonionized permeant in the dose solution. Generally, this trend became stronger as the dose increased. Recovery of radioactivity averaged 94.3 ± 5.5% for propranolol and was independent of ionization state and dose. Recovery of radioactivity for benzoic acid ranged from 40 to >100%, increasing with fraction nonionized and with dose. These effects can be qualitatively explained in terms of the low permeability of ionized species through stratum corneum, the volatility of free benzoic acid, and a buffer capacity of the stratum corneum deposition region on the order of 10-20 nmol/cm2.

A Framework for Incorporating Transient Solute-Keratin Binding Into Dermal Absorption Models.

Interpretation of experiments involving transient solute binding to isolated keratin substrates is analyzed and discussed in terms of their impact on transient permeation of topically-applied compounds through human stratum corneum. The analysis builds upon an earlier model (Nitsche and Frasch 2011 Chem Eng Sci 66:2019-41) by adding a second level of homogenization (ultrascopic-to-microscopic) prior to the microscopic-to-macroscopic conversion. Here "ultrascopic" refers to isolated keratin suspensions, "microscopic" to corneocyte interiors and "macroscopic" to tissue-averaged properties in the stratum corneum. Results are interpreted in the context of current parameterizations of the underlying ultrascopic binding parameters. The present analysis, which is limited to linear binding isotherms common in dilute solutions, reveals a maximum in the macroscopic forward binding rate constant as a function of solute lipophilicity, whereas the underlying equilibrium constant increases monotonically and the macroscopic reverse binding rate constant decreases monotonically. The size and location of the maximum depends upon the hydration state of the stratum corneum. Explicit equations expressing these findings allow both equilibrium and kinetic binding data in isolated keratins to be applied to the kinetics of transient absorption through the skin. They will enable more quantitative estimation of the long-recognized stratum corneum reservoir function.

Pirfenidone as a potential antifibrotic injectable for Dupuytren's disease.

Dupuytren's disease is a progressive fibrotic condition of the hand that causes contracture of fingers in later stages. Our previous in vitro studies suggest that the transformation of fibroblasts to myofibroblasts induced by transforming growth factor-beta can be inhibited by the addition of the antifibrotic drug, pirfenidone (PFD). We hypothesize that the local delivery of PFD directly to nodules can potentially prevent the progression to cords and, furthermore, that injection of PFD after the resection of cords can limit the recurrence of the disease. The purpose of this research was to develop a PFD injectable solution and to assess its safety in mice. Based on preformulation observations, a sterile solution containing up to 8 mg/0.4 mL of PFD was prepared in a phosphate buffer with and without 15%v/v N-methyl-2-pyrrolidone. Accelerated stability studies suggested that the product should be kept at refrigerated temperature (2-8 °C) for long-term storage. Safety studies involving subcutaneous administration to mice showed that 2-4 mg of PFD in 0.4 mL aqueous buffer did not elicit a significant inflammatory reaction. However, 4 mg PFD in 0.4 mL (FB) of buffer: NMP cosolvent system led to a significant increase in the influx of inflammatory cells and 8 mg PFD (FA) in the cosolvent system was lethal to the animals.

In Vitro Human Skin Absorption of Solvent-deposited Solids: Niacinamide and Methyl Nicotinate.

A quantitative understanding of the dose dependence of topical delivery is important to cosmetic and dermatological product development and to risk assessment for hazardous chemicals contacting the skin. Despite considerable research, predictive capability in this area remains limited. To this end we conducted an experimental skin absorption study of two closely related skin care agents, niacinamide (nicotinamide, NA) and methyl nicotinate (MN), and analyzed the results quantitatively using a transient diffusion model described separately (Yu et al. submitted for publication). Radiolabeled test compounds were solvent-deposited onto ex vivo human skin mounted in Franz diffusion cells over a dose range exceeding 4.5 orders of magnitude, and permeation was measured over a 1-4 day period. At low doses, the permeation rate of NA was approximately 60-fold lower than that of its lower melting, more lipophilic analog, MN; at high doses an even greater difference was observed. The difference can be qualitatively explained based on higher lipid solubility and lower crystallinity of MN relative to NA. Dissolution-limited mass transfer through a lipid layer at the SC surface is suggested. Relevance of the results to practical skin care formulations was confirmed by a parallel study of NA in an o/w emulsion.

Modeling the Percutaneous Absorption of Solvent-deposited Solids Over a Wide Dose Range.

The transient absorption of two skin care agents, niacinamide (nicotinamide, NA) and methyl nicotinate (MN), solvent-deposited on ex vivo human skin mounted in Franz diffusion cells has been analyzed according to a new variation on a recently published mechanistic skin permeability model (Yu et al. 2020. J Pharm Sci 110:2149-56). The model follows the absorption and evaporation of two components, solute and solvent, and it includes both a follicular transport component and a dissolution rate limitation for high melting, hydrophilic solids deposited on the skin. Explicit algorithms for improving the simulation of transient diffusion of solvent-deposited solids are introduced. The simulations can account for the ex vivo skin permeation time course of both NA and MN over a dose range exceeding 4.5 orders of magnitude. The model allows one to describe on a mechanistic basis why the percutaneous absorption rate of NA is approximately 60-fold lower than that of its lower melting, more lipophilic analog, MN. It furthermore suggests that MN perturbs stratum corneum barrier lipids and increases their permeability while NA does not, presenting a challenge to molecular modelers engaged in simulating biological lipid barriers.

Surfactant Penetration into Human Skin from Sodium Dodecyl Sulfate and Lauramidopropyl Betaine Mixtures.

Surfactant mixtures are used in a variety of personal care and cosmetic applications but are known to be harsh on the skin. The purpose of this study was to examine anionic surfactant penetration into human skin from nonideal surfactant mixtures under short-time exposure conditions that are relevant to realistic exposure scenarios. This was done by measuring the penetration of a radiolabeled probe (14C-SDS) into human cadaver skin in Franz diffusion cells in vitro from the mixtures of sodium dodecyl sulfate (SDS) and lauramidopropyl betaine (LAPB). Monomer and micelle concentrations in the SDS/LAPB/14C-SDS mixtures were predicted using a regular solution theory approximation. We confirmed that the mixtures of SDS and LAPB exhibit nonideal behavior with a net attraction between the two surfactants. Penetration of 14C-SDS into excised human skin from the mixtures of SDS and LAPB was found to decrease in a log-linear manner with increasing mole fraction of LAPB in the bulk solution (R2 = 0.97, p < 0.001). Additionally, the penetration of 14C-SDS into excised human skin from the mixtures of SDS and LAPB was found to correlate linearly and strongly with the predicted values of 14C-SDS monomer concentration in SDS/LAPB/14C-SDS mixtures (R2 = 0.95, p < 0.01). 14C-SDS penetration from the mixed surfactant composition could be quantitatively reconciled with that from an SDS-only composition by postulating a secondary, positive contribution from LAPB related to its own penetration and binding to skin components that increased SDS penetration at low concentrations. This research therefore supports a monomer penetration theory of surfactant penetration into the skin, combined with a measurable impact of favorable surfactant interactions within the tissue.

Rheology Control Using Nonionic Cosurfactants and pH Titration in an Amino Acid-Derived Surfactant Composition.

Sulfate-based formulations can be easily thickened by adding salt or amphoteric cosurfactants. However, sulfate-free and amino acid-based surfactants cannot. We explored an alternative thickening mechanism by studying the thickening effect of adding nonionic cosurfactants to a mixture of an amino acid-based surfactant, sodium lauroyl sarcosinate (SLSar), and a zwitterionic cosurfactant, cocamidopropyl hydroxysultaine (CAHS) at a 6:9 weight ratio. To characterize the formulations, we combined traditional rheometry with a state-of-the-art mesoscopic analysis of micelle dynamics obtained via diffusing wave spectroscopy. In addition, the formulations were characterized by cross-polarized light microscopy and dynamic light scattering. The cosurfactants studied included fatty alcohols, alkanediols, a fatty acid, and fatty alcohol ethoxylates (CnE3 and CnE6). Adding the nonionic cosurfactants increased the zero-shear viscosity up to 350 times the viscosity of the no-additive system at neutral pH. When pH titration was incorporated as a second thickening mechanism, the viscosity maximum was lower than the no-additive mixture. Furthermore, the pH of the viscosity maximum was shifted to higher pH for all systems except for CnE6, which shifted the maximum to lower pH. The nonionic amphiphiles also broadened the viscosity maximum, particularly in the C10OH system. Consequently, the C10OH system had a more favorable profile for development as a practical thickening system for an amino acid-based cleanser. Analysis according to the Zou and Larson micelle dynamics model revealed that the broadening effect was associated with substantially longer breakage times for the C10OH system (4-208 ms) compared to the no-additive system (4-38 ms).

The effect of prolonged exposure on sodium dodecyl sulfate penetration into human skin.

The purpose of this study was to examine the effect of prolonged surfactant exposure on mechanisms of anionic surfactant penetration into human skin. A radiolabeled probe (14-carbon sodium dodecyl sulfate (14C-SDS)) was used to trace the penetration of a model anionic surfactant, sodium dodecyl sulfate (SDS), into excised human skin and into an inert membrane composite in vitro. SDS dose varied from 0.03 to 15 mg/cm2, mimicking the exposure of a rinse-off cleanser on skin. Two surfactant exposure lengths were tested, 2 min and 5 h. SDS penetration into excised human skin was constant from 50 to 600 mM for skin samples exposed to SDS for 2 min. For skin samples exposed to SDS for 5 h, SDS penetration into skin increased log-linearly with increasing SDS concentration. SDS penetration into the inert membrane composite was constant from 50 to 600 mM SDS regardless of length of surfactant exposure. Penetration of the radiolabeled probe into skin and into the inert membrane correlated well with the monomeric concentration of the radiolabeled probe in the applied surfactant solution. These results support that monomer concentration is the driving force for initial SDS penetration into upper layers of the stratum corneum over a wide range of concentrations. With prolonged exposure, SDS penetrates the skin in a dose-dependent manner due to surfactant-induced damage to the skin.

Effect of pH on the Structure and Dynamics of Wormlike Micelles in an Amino Acid-Derived Surfactant Composition.

We studied the impact of pH as a thickening mechanism on the structure and dynamics of wormlike micelles in a mixture of sodium lauroyl sarcosinate (SLSar) and cocamidopropyl hydroxysultaine (CAHS). The viscoelastic properties were obtained using mechanical rheometry and diffusing wave spectroscopy, which provided access to a wide range of frequencies. By using a mesoscopic simulation method [Zou; Larson. J. Rheol. 2014, 58 (3), 681-721], characteristic micelle lengths and times were extracted including contour length, persistence length, entanglement length, reptation time, breakage time, breakage rate, and breakage rate constant. The interplay of pH-dependent reptation times (10-1000 ms) and breakage times (4-38 ms) leads to a minimum in the ratio of reptation time to breakage time of about 0.02 at pH 4.8. This minimum was closely associated with the sharp increase and decrease of the observed viscosity maximum at pH 4.8 in this system. These values may be contrasted with much longer breakage times (20-300 ms) that have been measured in more easily thickened sulfate-based systems. The low breakage times of the SLSar/CAHS system were attributed to the high and pH-sensitive breakage rate constants (0.01-0.17 ms-1 µm-1).

Computer Simulation of Skin Permeability of Hydrophobic and Hydrophilic Chemicals - Influence of Follicular Pathway.

A recently published mechanistic skin permeability model (Kasting et al., 2019. J Pharm Sci 108:337-349) that included a follicular diffusion pathway has been extended to describe transient diffusion and finite dose applications. The model follows the disposition of two components, solute and solvent, so that solvent deposition processes can be explicitly represented. Experimentally-calibrated permeability characteristics of the follicular pathway leading to the permeation of highly hydrophilic permeants are further refined. Details of the refinements and a comparison with the earlier model using two large experimental datasets are presented. An example calculation shows the marked difference between the time scales for achievement of near steady-state diffusion for large hydrophilic and lipophilic compounds, with the former being more than 100-fold faster than the latter. However, the true steady state for the hydrophilic compound is not reached until much later due to the very slow filling of the corneocyte phase.

Evaluation of Heat Effects on Fentanyl Transdermal Delivery Systems Using In Vitro Permeation and In Vitro Release Methods.

Experimental conditions that could impact the evaluation of heat effects on transdermal delivery systems (TDS) using an in vitro permeation test (IVPT) and in vitro release testing (IVRT) were examined. Fentanyl was the model TDS. IVPT was performed using Franz diffusion cell, heating lamp, and human skin with seven heat application regimens. IVRT setup was similar to IVPT, without using skin. Dissolution study was conducted in a modified dissolution chamber. The activation energy of skin permeation for fentanyl was determined using aqueous solution of fentanyl. In IVPT, the increase of temperature from 32 °C to 42 °C resulted in a 2-fold increase in flux for fentanyl TDS, consistent with the activation energy determined. The magnitude of flux increase was affected by the heat exposure onset time and duration: higher flux was observed when heat was applied earlier or following sustained heat application. Heat induced flux increases could not be observed when inadequate sampling time points were used, suggesting the importance of optimizing sampling time points. Drug release from TDS evaluated using IVRT was fast and the skin was the rate-limiting barrier for TDS fentanyl delivery under elevated temperature.

Evaluation of Heat Effects on Transdermal Nicotine Delivery In Vitro and In Silico Using Heat-Enhanced Transport Model Analysis.

A combined experimental and computational model approach was developed to assess heat effects on drug delivery from transdermal delivery systems (TDSs) in vitro and nicotine was the model drug. A Franz diffusion cell system was modified to allow close control of skin temperature when heat was applied from an infrared lamp in vitro. The effects of different heat application regimens on nicotine fluxes from two commercial TDSs across human cadaver skin were determined. Results were interpreted in terms of transport parameters estimated using a computational heat and mass transport model. Steady-state skin surface temperature was obtained rapidly after heat application. Increasing skin surface temperature from 32 to 42°C resulted in an approximately 2-fold increase in average nicotine flux for both TDSs, with maximum flux observed during early heat application. ANOVA statistical analyses of the in vitro permeation data identified TDS differences, further evidenced by the need for a two-layer model to describe one of the TDSs. Activation energies associated with these data suggest similar temperature effects on nicotine transport across the skin despite TDS design differences. Model simulations based on data obtained from continuous heat application were able to predict system response to intermittent heat application, as shown by the agreement between the simulation results and experimental data of nicotine fluxes under four different heat application regimens. The combination of in vitro permeation testing and a computational model provided a parameter-based heat and mass transport approach to evaluate heat effects on nicotine TDS delivery.

Modeling Temperature-Dependent Dermal Absorption and Clearance for Transdermal and Topical Drug Applications.

A computational model was developed to better understand the impact of elevated skin temperatures on transdermal drug delivery and dermal clearance. A simultaneous heat and mass transport model with emphasis on transdermal delivery system (TDS) applications was developed to address transient and steady-state temperature effects on dermal absorption. The model was tested using representative data from nicotine TDS applied to human skin either in vitro or in vivo. The approximately 2-fold increase of nicotine absorption with a 10°C increase in skin surface temperature was consistent with a 50-65 kJ/mol activation energy for diffusion in the stratum corneum, with this layer serving as the primary barrier for nicotine absorption. Incorporation of a dermal clearance component into the model revealed efficient removal of nicotine via the dermal capillaries at both normal and elevated temperatures. Two-compartment pharmacokinetic simulations yielded systemic drug concentrations consistent with the human pharmacokinetic data. Both in vitro skin permeation and in vivo pharmacokinetics of nicotine delivered from a marketed TDS under normal and elevated temperatures can be satisfactorily described by a simultaneous heat and mass transfer computational model incorporating realistic skin barrier properties and dermal clearance components.

Anionic Surfactant-Induced Changes in Skin Permeability.

Anionic surfactants compromise skin's barrier function by damaging stratum corneum lipids and proteins. The objective of this study was to examine anionic surfactant-induced changes in the skin's polar and transcellular pathways and the resulting impact on surfactant penetration into the skin. Three anionic surfactant formulations and one control formulation were each applied to split-thickness human cadaver skin in vitro for 24 h. Electrical conductivity of the skin, determined using a four-terminal resistance method, and water permeation across the skin, determined using a radiolabeled water tracer, were simultaneously measured at several points over the experimental period. Surfactant permeation across the skin was similarly measured using a radiolabeled sodium dodecyl sulfate tracer. Anionic surfactants rapidly enhanced skin electrical conductivity and water permeability in the excised human skin, resulting in nonlinear enhancements in surfactant permeation across the skin over time. Surfactant penetration into the skin was found to increase linearly with increasing surfactant monomer concentration. Surfactant zeta potential was found to correlate well with skin conductivity, water permeation across the skin, and surfactant permeation across the skin, particularly with long surfactant exposures. Micelle charge is a significant predictor of anionic surfactant-induced damage to the human skin, with more highly charged surfactants inducing the most damage.

Formulation and Artificial Sebum Effects on the Percutaneous Absorption of Zinc Pyrithione through Excised Human Skin.

BACKGROUND: Zinc pyrithione (ZnPT) is deposited on the skin as a fine particulate and must reach microorganisms localized in the stratum corneum and hair follicles in molecular form to exert its broad-spectrum antimicrobial/antifungal activity. Dissolution of ZnPT particles followed by molecular speciation results in the organic portion, i.e. pyrithione, being more susceptible to skin penetration than the inorganic component, i.e. zinc, or the chelate itself, i.e. ZnPT. OBJECTIVES: To further test the hypothesis that ZnPT skin penetration is rate-limited by dissolution and molecular speciation, the effect of different formulations and artificial sebum on the in vitro percutaneous absorption of radiolabel associated with Zn[14C]PT was investigated. METHOD: In vitro penetration of [14C]PT into and through excised human skin was measured following application of Zn[14C]PT prepared as suspensions in distinct vehicles including water-based carboxymethylcellulose (CMC), diluted body wash comprised of surfactants, and castor oil, in the presence and absence of artificial sebum. RESULTS: The steady-state flux and cumulative absorption of Zn[14C]PT increased 4- to 5-fold when deposited from a body wash or castor oil compared to a water-based CMC suspension. Tritiated water flux measured before and after treatment showed that neither the surfactant vehicle nor castor oil significantly altered barrier function versus water alone. An artificial sebum layer on the skin potentiated Zn[14C]PT and 3H2O absorption when dosed from both aqueous formulations, but not from castor oil. CONCLUSION: These data are consistent with the hypothesis that ZnPT percutaneous absorption, as measured by [14C]PT kinetics, is controlled by particle dissolution and molecular speciation.