Absorption and Evaporation of Pesticides from Human Skin in Vitro

UNIVERSITY OF CINCINNATI

Date:______

I, ______, hereby submit this work as part of the requirements for the degree of: in:

It is entitled:

This work and its defense approved by:

Chair: ______

Absorption and evaporation of pesticides from human

skin in vitro.

A dissertation submitted to the

Division of Research and Advanced Studies of the University of Cincinnati

in partial fulfillment of the requirements for the degree of

DOCTORATE OF PHILOSOPHY (Ph.D.)

In the Division of Pharmaceutical Sciences of the College of Pharmacy

2007

Varsha D. Bhatt

B.Pharm., S.N.D.T University, Mumbai, 1999

Committee Chair: Gerald B. Kasting, Ph.D.

ABSTRACT

Estimation of penetration rates of compounds through skin is an important for assessment of either the efficacy of a topical formulation, its irritation potential, or its potential

systemic exposure. Current risk assessment models often assume that 100 % of the applied dose is absorbed through the skin. This thesis provides experimental data to validate and calibrate a working diffusion/evaporation model for topically applied

chemicals based on their physicochemical properties, the known biological properties of

skin and principles of diffusion theory.

Four compounds with varying physicochemical properties were studied under different

conditions—benzyl alcohol (BA), diethyl-m-toluamide (DEET), tecnazene and

malathion. Absorption of 14C – BA, a perfume ingredient, was studied at different doses

(0.9 µg/cm2 – 10.6 mg/cm2) through human cadaver skin and silicone membrane. The

disposition of BA was satisfactorily described by the diffusion model using a variable

diffusivity coefficient.

The absorption and evaporation of the insect repellent 14C – DEET (127 µg/cm2) was

characterized at airflows ranging from 10 – 100 mL/min. The amount of DEET absorbed

through the skin systematically decreased as the airflow increased.

For tecnazene, a fungicide, and malathion, an insecticide, a Head space Solid Phase

Micro Extraction technique was developed to analyze the receptor fluid following skin

penetration experiments. The analysis was done using GC-MS.

Absorption rate of tecnazene, was studied following application of 103 µg/cm2 and 864

µg/cm2 under open and occluded conditions. The absorption was much higher under occlusion and the overall recovery was also better for the occluded treatment.

Additionally disposition of topically applied 14C – malathion (101, 0.5 and 0.1 µg/cm2) was studied under open and occluded conditions and compared to GC-MS results from

101 µg/cm2. A silicone membrane study was performed with all three doses of 14C –

malathion.

The model satisfactorily described the absorption of malathion and tecnazene through

human skin in vitro. Key parameters that needed modification were kevap (evaporation mass transfer coefficient), Ksc (partition coefficient of the stratum corneum, SC) and Psc

(permeability of SC). Future work entails conducting partitioning studies of lipophilic compounds to get better understanding of disposition of such compounds in the lower skin layers.

To my Grandparents:

Krishnalal Gopalji Bhatt and Sharda K. Bhatt

My parents:

Dilip Krishnalal Bhatt and Charu D. Bhatt

ACKNOWLEDGEMENT

Cincinnati has been home away from home since September 2000. At the end of my graduate life, I would be remiss if I did not express my sincere thanks and gratitude to everyone who helped mould my life at both professional as well as personal level.

Firstly, I would like to thank Dr.Kasting, or Dr.K as we fondly call him, for his mentorship, guidance, patience and support throughout my graduate life. I only hope to take on some of his virtues and do justice to his teachings wherever I go. I express my deepest regard for Dr. Wickett who admitted me into the program and from whose anecdotal tales; I’ve learnt not only science, but also ways of life. Collaboration with Dr.

Soman came about at a time when options for furthering my research were dim and thanks to his support and guidance, I could see my doctoral work to completion. Dr.

Talaska and Dr. Shenouda have been invaluable in both their input and suggestions for my research and I thank them both.

I would also like to take this opportunity to thank University of Cincinnati for granting me Graduate Scholarship through out. I would also like to express my gratitude to

College of Pharmacy, all professors and administrative staff especially, Marcie Silver,

Sue Ryan and Donna Taylor. Along side academics, my involvement with extra curricular activities enriched my living experience at Cincinnati. I express my special thanks to the Office of International Student Services for providing a platform for cross cultural interaction. I will cherish the memories from ISSO events and Cultural

Connections for life time! My involvement with Association for India’s Development

(AID) and Graduate Association for Pharmaceutical Sciences Students (GAPSS) helped me foster my people skills and will always have a special place in my heart.

During my stay at UC, I have come to know some wonderful people with whom I have developed a life long friendship. Among these are Andy, Arjun, Fair, Grettel, Hemali,

Leena, Penpan, Rachna, Rucha, Shagun and Tarun. I have been extremely fortunate to have had my closest friend, Namrata, with me from undergraduate to graduate days. I want to thank her and all my friends for the unwavering support and kindness! Others,

with whom I interacted with over the course of time at UC, I thank you for the friendship

and a sense of belonging, especially the entire group of Cosmetic Sciences.

Finally, my family, without their love and direction and faith I would not be here. My

Grandparents, parents, uncles, aunts, siblings and extended family have continuously

blessed me with good wishes and I cannot thank you enough! All I can say is that I love

you all. A special thanks to my sister, Kavita, who was patient with me during the hectic

and sometimes chaotic behavior towards the end of the graduate life.

Last but not at the very least, I thank the Almighty for showering his kindness and giving me the strength to face challenges in life.

TABLE OF CONTENTS

Page No.

Abstract

Acknowledgement

Table of Contents 1

List of Tables 2

List of Figures 4

List of Abbreviations 8

List of Symbols 12

Chapter 1 Introduction 13

Chapter 2 Objective 27

Chapter 3 Experimental methodology 29

Chapter 4 Working diffusion model 44

Chapter 5 Dose & airflow dependence of benzyl alcohol disposition on

skin 49

Chapter 6 Airflow dependence study of DEET: Results 81

Chapter 7 Absorption of tecnazene through human skin in vitro 97

HS-SPME/GC-MS study

Chapter 8 Dermal absorption of malathion in vitro: GC – MS results 127

Chapter 9 Dermal absorption of malathion through human cadaver skin

and silastic membrane: Radiochemical results 153

Chapter 10 Summary and future directions 180

Chapter 11 References 184

LIST OF TABLES

Table 1.1 Literature studies of dermal absorption of pesticides 17

Table 1.2 Physiochemical properties of compounds of interest 26

Table 3.1 Recent SPME pesticide studies 42

Table 5.1 Appearance of radioactivity in the receptor solution for dose-

dependence skin disposition study with 14C-benzyl alcohol 63

Table 5.2 Cumulative disposition of radioactivity at 24 h for the study

reported in Table 5.1 64

Table 5.3 Regression parameters for diffusion models of benzyl alcohol

disposition on skin. 68

Table 5.4 Physical properties and auxiliary skin permeability data for

benzyl alcohol 76

Table 5.5 Derived diffusion model parameters for benzyl alcohol. 77

Table 5.6 Percentage absorption of benzyl alcohol after 24 h estimated by

three different methods 79

Table 6.1 Physical Properties of DEET 87

Table 6.2 Mean cumulative evaporation of 14C- DEET from human skin

in vitro 88

Table 6.3 Mean absorption of 14C- DEET from human skin in vitro 89

Table 6.4 Appearance of radioactivity in the vapor trap for airflow-dependence

skin disposition study with 14C- DEET from human skin in vitro 90

Table 6.5 Appearance of radioactivity in the receptor solution for

airflow-dependence skin disposition study with 14C- DEET from

human skin in vitro 91

Table 6.6 Mass balance of 1% w/w solution of 14C-DEET from human skin

in vitro as a percent of dose applied 92

Table 7.1 Figures of merit for tecnazene method development 109

Table 7.2 Correction factors for predicted absorption flux of tecnazene 124

Table 7.3 Mass balance of topically applied tecnazene from human skin

in vitro as % applied dose 125

Table 8.1 Figures of merit for malathion method development 133

Table 8.2 Mass balance of topically applied malathion from human skin

in vitro as % applied dose 150

Table 9.1 Appearance of radioactivity in the receptor solution: Skin

disposition study with 14C-malathion through human skin in vitro 160

Table 9.2 Mass balance of disposition of radioactivity associated with 14C-

malathion from human skin in vitro, expressed as the percent of

dose applied 162

Table 9.3 Physical properties and input parameters of malathion 170

Table 9.4 Calculated model parameters for malathion absorption through

human skin in vitro 171

Table 9.5 Appearance of radioactivity in the receptor solution: Skin

disposition study with 14C-malathion through silicone membrane

in vitro 172

Table 9.6 Mass balance of disposition of 14C-malathion from silicone membrane

in vitro as a percent of dose applied 173

LIST OF FIGURES

Figure Page

1.1 Estimates of world pesticide consumption in 2000-2001 16

1.2 Structure of Benzyl alcohol 24

1.3 Structure of DEET 24

1.4 Structure of malathion 25

1.5 Structure of tecnazene 25

3.1 In vitro set up for volatile trapping 33

3.2 SPME fiber assembly 37

3.3 Schematic for steps involved in SPME 40

4.1 One dimensional diffusion model 45

5.1 Skin absorption rates of 14C benzyl alcohol from human skin in vitro from dose

dependant studies 60

5.2 Maximum flux and maximum time to flux for absorption of 14C benzyl

alcohol from human skin in vitro from dose dependant studies 61

5.3 Cumulative percent dose absorbed of 14C benzyl alcohol from human skin

in vitro 62

5.4 Skin evaporation and absorption rates of 14C benzyl alcohol from human skin

in vitro from air flow dependant studies 66

5.5 Flux of 14C benzyl alcohol from human skin in vitro from air flow

dependant studies 70

5.6 Model prediction Vs. Experimental absorption for absorptive flux of

14C benzyl alcohol from human skin in vitro 71

6.1 Structure of DEET 86

6.2 Disposition of 14C – DEET from human skin in vitro as a percent of

applied dose 93

6.3 Effect of airflow on evaporation rate of 14C – DEET from human skin

in vitro 94

6.4 Effect of airflow on absorption rate of 14C – DEET from human skin

in vitro 95

7.1 Structure and physical properties of tecnazene 101

7.2a Optimization of exposure conditions using 100µm PDMS fiber at 70ºC

for 15 minutes to a concentration of 20 ppb – Effect of salt addition 106

7.2b Optimization of exposure conditions using 100µm PDMS fiber with 0.6 g

NaCl to a concentration of 20 ppb – Effect of time 107

7.3 Recoveries of tecnazene from different matrices. 110

7.4a Representative chromatogram using Varian 3800 ion trap GC-MS 115

7.4b Representative chromatogram using HP 5890 Quadrupole GC-MS 116

7.5a Typical calibration curve using Varian 3800 GC-MS equipped with

8200 auto sampler 117

7.5b Calibration curve using 100 µm PDMS fiber for analysis of tecnazene

using hexachlorobenzene as an internal standard 118

7.6 Absorption rate of tecnazene through human skin in vitro under open and

occluded conditions 123

8.1 Structure and physical properties of malathion 129

8.2a Optimization of exposure conditions – Comparison of fibers 136

8.2b Optimization of exposure – Effect of salt addition 137

8.3 Recovery of malathion from different matrices 139

8.4a Representative chromatogram using Varian 3800 ion trap GC-MS 143

8.4b Representative chromatogram using HP 5890 Quadrupole GC-MS 144

8.5a Typical calibration curve using Varian 3800 GC-MS equipped with

8200 auto sampler 145

8.5b Calibration curve using 100 µm PDMS fiber for analysis of malathion

using fenitrothion as an internal standard 146

8.6 Absorption flux of tecnazene through human skin in vitro under open and

occluded conditions 148

8.7 Tissue levels of malathion in presence and absence of an occlusive barrier 149

9.1 Structure and physical properties of malathion 156

9.2 Cumulative absorption of 14C-malathion through human skin in vitro in

the presence and absence of an occlusive barrier 163

9.3 Absorptive flux of radioactivity associated with 14C-malathion in the

presence and absence of an occlusive barrier 164

9.4 a Cumulative absorption of 14C-malathion (77.4 μg/cm2) through human

cadaver skin in vitro 167

9.4 b Cumulative absorption of 14C-malathion (0.04 μg/cm2) through human

cadaver skin in vitro 168

9.4 c Cumulative absorption of 14C-malathion (0.002 μg/cm2) through human

cadaver skin in vitro 169

9.5 Cumulative absorption of 14C-malathion through silicone membrane

in vitro 174

9.6 Absorptive flux of radioactivity associated with 14C-malathion through

silicone membrane in vitro 175

LIST OF ABBREVIATIONS

2, 4-D = (2,4-dichlorophenoxy)acetic acid a.i. = Active ingredient

AUC = Area under the curve

BA = Benzyl alcohol

C = Concentration

CAS No. = Chemical Abstract Service Number

Csat = Saturation concentration

CV = Coefficient of variation

DDT = 1,1'-(2,2,2-trichloroethylidene)bis[4-chlorobenzene] de = Dermis

Dde = Diffusivity in dermis

Ded = Diffusivity in viable epidermis

DEET = N,N, diethyl-3-methylbenzamide

DEP = Diethyl phthalate

DMSA = Disodium methanearsoate

ds = donor solution

Dsc = Diffusivity in SC

DVB = Polydivinylbenzene

ed = Viable epidermis

EI = Electron impact

EPA = Environmental Protection Agency

EPTC = s-ethyl dipropylthiocarbamate

GC = Gas chromatograph

3 H2O = Tritiated water h = thickness of SC

HCB = Hexachlorobenzene hdep = deposition depth

HPLC = High performance liquid chromatography

HS = Headspace

HSSE = Headspace Sorptive Extraction

Jmax = maximum flux k1 = evaporation rate constant

Kde = Partition coefficient between dermis and water

Ked = Partition coefficient between viable epidermis and water kevap = evaporative mass transfer coefficient kg’ = proportionality constant kg = gas phase mass transfer coefficient ~ k p = permeability coefficient of partially hydrated skin kp = Steady state permeability coefficient

Koct = Octanol water partition coefficient ~ K SC / w = Estimated SC/water partition coefficient in partially hydrated skin

Ksc = Partition coefficient between SC and water

LOD = Limit of detection

LOQ = Limit of quantification

LSC = Liquid Scintillation Counting

M0 = Initial dose

MCPA-DMA = (4-chloro-2-methylphenoxy)acetic acid

Mr = reduced dose

MS = Mass Spectrometer

Msat = Saturation dose

MSD = Mass selective detection

MSMA = Monosodium methanearsoate

MW = Molecular weight

NIOSH = National Institute of Occupational Safety and Health

NIST = National Institute of Science and Technology

NTP = National Toxicology Program

PA = Polyacrylate

PBS = Phosphate buffered saline

PCP = 2, 4, 5, 2’, 4’, 5’-Hexachlorobiphenyl

PDMS = Polydimethylsiloxane

Psc = Permeability of SC

Pvp = Vapor pressure

R = gas constant r2 = squared correlation coefficient

RSD = relative standard deviation

R(t) = Retention time

SC = Stratum Corneum

SBSE = Stir Bar Sorptive Extraction

SDME = Single Drop Micro Extraction

SIM = Single ion monitoring

SPME = Solid Phase Micro Extraction

SSR = Residual Sum of Squares

SST = Total Sum of Squares

Sw = Water solubility

T = absolute temperature

TCMBT = 2-(Thiocyanomethylthio) benzothiazole

tmax = time to maximum flux

Vh = headspace volume

VX = S-2-(diisopropylamino)ethyl O-ethyl methylphosphonothioate

LIST OF SYMBOLS

ρ = Density

τ = reduced time f = fractional deposition depth

β = membrane capacity/dose ratio

χ = volatility parameter

v = airflow

CHAPTER ONE

INTRODUCTION

Introduction:

Understanding the barrier function of the skin is crucial for transdermal drug delivery as well as risk assessment dermal exposure. Such exposures result from accidental or intentional contact with products ranging from seemingly benign consumer care products to recognized health hazards like pesticides. The three primary routes by which systemic toxicity can arise are through inhalation, ingestion and dermal absorption. With respect to the latter, it is generally accepted that the outermost keratin rich layer of skin − the stratum corneum (SC) − provides primary resistance to entry of moderately lipophilic molecules 1.

Pesticides form an important group of hazardous materials. The U.S. Environmental

Protection Agency (EPA) defines a pesticide as “any substance or mixture of substances intended for preventing, destroying, repelling, or mitigating any pest.” Pests can be insects, mice and other animals, unwanted plants (weeds), fungi, or microorganisms like bacteria and viruses. The term pesticide also applies to herbicides, fungicides, and various other substances used to control pests. Thus, by their very nature, most pesticides create some risk of harm to humans, animals or the environment since they are designed to kill or otherwise adversely affect living organisms.

Pesticides constitute a heterogeneous group of structurally and physico-chemically variable compounds with variation in volatility, molecular weight, lipophilicity, water solubility and other physico-chemical properties. They are widely used in numerous

commercial, agricultural and homeowner applications. The most common way in which they are classified is based on the type of pest they affect, for example, a fungicide acts against fungi, a rodenticide acts against rodents and so on. Another way is differentiating them on basis of origin – i.e. chemical pesticides and biopesticides 2.

Chemical pesticides are mainly divided in to four divisions –

(a) Organophosphate, e.g., malathion, diazinon

(b) Organochlorine, e.g., aldrin, hexachlorobenzene

(d) Pyrethroids, e.g., cypermethrin, fenvalerate

According to 2000 and 2001 estimates, EPA reports that the world consumption of pesticides surpassed 5 billion pounds, while US use alone exceeded 1.2 billion pounds.

Figure 1 depicts the world and US pesticide consumption expressed as pounds of active ingredient (a.i).

Dermal absorption of pesticides remains a topic of continued interest among occupational hygienists and research scientists for the simple underlying reason of its potential impact on human health. There is a large body of literature on pesticide absorption in humans as well as animals. Table 1.1 presents a list of pesticide absorption studies that have been reported in the literature. This list, although extensive is not exhaustive. It should alert the reader as to the wide availability of dermal absorption data.

6000

5000

2000

1000 Millions of lbs a.i Millions

0 es s s ide c Other Total ticide gi erbicid un H Insec F

Figure 1.1 Estimates of world pesticide consumption in 2000-2001. Data re-plotted from

the EPA report. 3.

Table 1.1 Literature studies of dermal absorption of pesticides

Type of Type of

Ref. skin study Year Pesticides studied No animal (a) in vivo (c)

human (b) in vitro (d) 4 2006 Methyl Parathion, Parathion, Fenthion a d

5 2005 2,4-D a, b c, d

6 2005 Permethrin b c

7 2005 Glyphosate, Malathion, Parathion, Ethylenedioxide, 2-Butoxyethanol, VX a, b c, d

2, 4 –D, Alachlor, Azodrin, Baygon, Benz[o]pyrene, Chlordane, DDT, Ethion, 8 2005 a, b c, d Lindane, Malathion, Parathion

9 2004 Dimethoate, Methiocarb, Paclobutrazol, Pirimicarb, Prochloraz b d

10 2003 DDT a c

Atrazine, Azinphos-methyl, o-Benzyl-p-chlorphenyl, Cacodylic acid, 11 2003 a c Cyromazine, Desmedipham, Disulfoton, EPTC, Etofenprox, Maneb, Metam

Sodium, Metolachlor, MCPA-DMA, Molinate, Monocrotophos, Terbutryn,

Tetrachlorvinphos, Triphenyltin hydroxide, Vinclozolin

12 2002 Pentachlorophenol mixtures a d

13 2002 Atrazine, Alachlor, Trifluralin a d

14 2001 Malathion a c

Azinophos-methyl, Benomyl, Dinoconazole, Disulphoton, EPTC, Glufosinate-

15 2000 Ammonium, Hexachlorobenzene, Hydrogen Cyanamide, Iprodione, Lindane, a c, d

Mevinphos, Phosmet, TCMBT, Thiobencarb

16 2000 Propoxur a, b c, d

17 2000 Methiocarb, Paclobutrazol, Primicarb b d

18 1999 Chlorpyrifos b c

19-21 1994 Diazinon, 2,4-dichlorophenoxyacetic acid, DDT a, b b, d

22 1994 Pyrethrin, Piperonyl butoxide b c

23 1993 Diazinon b c

24 1993 Butachlor b d

25 1993 DEET, 2,4-D, Diazinon, DDT a, b d

26 1992 Fluazifop-butyl b c

27 1992 Isofenphos b c

28 1987 Cypermethrin a c, d

Atrazine, Captan, Carbayl, Carbofuran, Chlordecone, Chlorpyrifos, Dinozeb, 29 1987 a d DMSA, Folpet, MSMA, Nicotine, Parathion, PCB, Permethrin

30 1987 Fenitrothion, Aminocarb a c

31 1985 Fenvalerate, DDT, 2,4-D, Vamidothion, Permethrin a c

32 1980 Soman, VX a c

2,4-D, Aldrin, Azodrin, Baygon, Carbaryl, Dieldrin, Diquat, Ethion, Guthion, 33 1974 b c Lindane, Malathion, Parathion,

2, 4-D = (2,4-dichlorophenoxy)acetic acid DDT = 1,1'-(2,2,2-trichloroethylidene)bis[4-chlorobenzene] DEET = N, N-diethyl-methyltoluamide DMSA = Disodium methanearsoate EPTC = s-ethyl dipropylthiocarbamate MCPA-DMA = (4-chloro-2-methylphenoxy)acetic acid MSMA = Monosodium methanearsoate PCP = 2, 4, 5, 2’, 4’, 5’-Hexachlorobiphenyl TCMBT = 2-(Thiocyanomethylthio) benzothiazole VX = S-2-(diisopropylamino)ethyl O-ethyl methylphosphonothioate

Interest in percutaneous absorption as a measure for human exposure assessment has increased in the recent years. The National Toxicology Program (NTP) has undertaken efforts to evaluate the hazard and risk of environmental substances for Human Exposure

Assessment with an inter agency initiative that includes among others, the National

Institute of Occupational Safety and Health (NIOSH) and the EPA, to name a few 34.

Estimation of penetration rates and systemic absorption of compounds following accidental or intentional application to the skin is an important aspect of exposure assessment. However, the vast number of compounds of interest (an estimated 80,000 compounds are currently in use in the US, with up to 2000 new ones being introduced each year) makes quantification of all compounds difficult. Here, well conceived mathematical models play an important role in either screening newer chemicals or ranking existing compounds on the basis of their dermal penetration potential, thereby helping save time and resources 35. However, current models have limitations for compounds and exposure conditions that are commonly encountered in industrial environments and consumer products.

Existing risk assessment methods generally assume 100% absorption of the applied dose.

This is an obvious over estimate: evaporative loss of the volatile compounds can affect significantly the amount of absorbed through the skin 36. Furthermore, the evaporation

process is controlled by surface airflow and temperature. Evaporation and absorption are inversely related for obvious reasons. Several other factors including skin condition, site

of application, and degree of hydration also dictate the absorption of compounds through

the skin 37-39.

From a predictive stand point, the most commonly used model till date is that developed

by Potts and Guy 40. This model is based on the relationship between skin permeability

coefficient kp and two basic physical and chemical characteristics of a compound −

molecular weight (MW) and octanol-water partition coefficient (Koct). This relationship

was derived via linear regression using a large skin permeability database 41. The main

disadvantage of this is that it is based on steady-state, infinite dose permeation data from aqueous solutions 41. Although there are many alternative analyses in the literature, most

of them are modified versions of the Potts-Guy analysis and therefore suffer from the

same drawbacks 42,43.

Models that incorporate biophysical properties of skin, realistic skin exposure conditions and parameters obtained from microscopic characteristics of skin transport offer a potential advantage to Potts and Guy model. Research in the Kasting laboratory is focused towards this effort. Starting from simple kinetic models for perfume ingredients, and expanding to a diffusion model that predicts the absorption and evaporation of compounds from skin, a considerable progress has been made in the group 44-50.

This dissertation describes the absorption and evaporation of three compounds − two

pesticides (malathion and tecnazene) and one insect repellent (DEET) − from human skin

in vitro. An additional study to characterize the dose dependency of benzyl alcohol (BA)

disposition on human skin and silastic membrane in vitro is also presented. The

compounds were selected in order to test a range of physicochemical properties including

lipophilicity and vapor pressure. Experimental methods and skin absorption data are

presented for each of the compounds. Figures 1.2 − 1.5 show the structure of the compounds studied. Table 1.2 lists their physical and chemical properties.

O H

Figure 1.2 Structure of benzyl alcohol (BA)

Me C NEt 2

Figure 1.3 Structure of N, N, diethyl - 3- methylbenzamide (DEET)

Figure 1.4 Structure of malathion

Cl NO 2

Cl Cl

Figure 1.5 Structure of tecnazene

Table 1.2 Physicochemical properties of compounds of interest

a b c Compound CAS No. MW ρ, log Koct Sw Pvp

g/cm3 g/L torr

BA 100-51-6 108.1 1.04 1.10 42.9d 0.094

DEET 134-62-3 191 0.98 2.02 9.9d 0.0017

Malathion 121-75-5 330 1.23 2.36 0.145e 0.0000234

Tecnazene 117-18-0 261 2.02 4.38 0.0021f 0.0036

a Density; b Water solubility; c Vapor pressure; d 25°C; e 20°C; f 22 °C

CHAPTER TWO

OBJECTIVE AND SPECIFIC AIMS

2.1 Objective:

The overall aim of the research group is to develop a predictive mechanistic mathematical model based on biophysical properties of SC and physicochemical properties of compounds to satisfactorily describe the absorption and evaporation of topically applied compounds.

2.2 Specific Aims:

The aims of this project were to obtain experimental data to characterize the disposition of pesticides and fragrance compounds applied to skin surface under different conditions

– at different airflows and in presence and absence of occlusive barrier. The specific aims of the project were as follows:

I. To study the absorption and evaporation rates of small doses of volatile

compounds, pesticides (DEET, tecnazene, malathion) and a fragrance compound

(benzyl alcohol) applied to human skin in vitro.

II. To develop analytical methods to detect and quantify the disposition of tecnazene

and malathion from human skin in vitro using GC-MS.

III. To evaluate a working computational model by comparing the experimental

values with the predicted values.

CHAPTER THREE

EXPERIMENTAL METHODOLGY

3.1 Skin membrane:

Dermatomed split thickness (300µm) human cadaver skin was obtained from U.S Tissue

and Cell (Cincinnati, OH). Skin from back, thigh and abdomen was used for the in vitro

studies. The skin was treated with 10% glycerol solution and kept at −80 ºC prior to use.

The skin was gently thawed by immersing it in Dulbecco’s phosphate buffered saline

(PBS) (Sigma-Aldrich, St. Louis, MO), pH 7.4 containing 0.02% (w/v) sodium azide as a preservative. The tissue was cut in to smaller pieces (roughly 1.5 cm2 × 1.5 cm2) and mounted on glass receptor chamber of modified Franz® diffusion cells with SC facing up.

The bottom chambers of the cells were filled with PBS with a small magnetic stir bar to

aide uniform mixing. The top and the bottom chambers were clamped and the cells were placed in a thermostatted aluminum heating and a stirring module. The cells were typically equilibrated until they reached 37 ± 2 ºC to yield a skin surface temperature of

30-32 ºC 51. Care was also taken to keep the receptor chamber free from bubbles.

3.1.1 Tissue integrity procedure:

Post equilibration, the tissue was screened by exposing each 0.79 cm2 skin piece to 150

µL of tritiated water (0.4 µCi/mL) for 5 minutes after which the dose was removed with

cotton tip applicators 52. The receptor fluid was collected 60 minutes post-dose and

analyzed by liquid scintillation counting. Samples with water permeation greater than 1.2

µL/cm2 were discarded as per the recommendation from previously reported study 53.

Following the test, the receptor solution was exchanged twice to ensure complete

removal of residual radioactivity. The top and bottom chambers were then sealed with

Parafilm®. The cells were randomized over treatments using previously established

procedures 52. They were equilibrated overnight and dosed the following morning with the dose solution. Prior to dosing, care was taken to ensure complete removal of air

bubbles.

3.2 Evaporation/Vapor trapping equipment:

The in vitro set up using modified Franz® diffusion cells for trapping volatile fractions of compounds applied to the skin is shown in Figure 3.1. The cells were custom made by

Dana Enterprises (West Chester, OH). A modification was made in the donor chamber by introducing an additional inlet to facilitate the circulation of air in the headspace over the skin. The top of the donor chamber was connected to a large Omnifit® universal

connector (Alltech Associates Inc., Deerfield, IL) by means of a glass connector (Dana

Enterprises). This in turn was connected to a Tenax® cartridge (Supelco, St. Louis, MO).

The cartridge was oriented to enable proper directionality of airflow in the set up. The

free end of a Tenax® cartridge was attached to a 210 series SKC pocket pump (SKC Inc.,

Eighty Four, PA) by means of silicone tubing. A constant air pressure controller was

used in conjunction with a dual low flow holder (SKC Inc., PA) to avoid airflow

fluctuations.

A similar set up was used for additional cells using a Universal Sample Pump (PCXR 4

series, SKC Inc., PA). In this way 4 cells could be studied. Prior to finalizing the set up

a calibration of airflows was done by attaching the low flow holder to a bubble flow meter- Accuflow® digital calibrator (SKC Inc., PA). The airflows were varied from 10 to

100 mL/min. After finalizing the airflow, the cells were allowed to equilibrate for a

minimum 30 minutes prior to application of dose solution. The top and the bottom chambers of the modified Franz® cells were sealed with Parafilm® to minimize loss of

analyte.

Figure 3.1 In vitro apparatus for volatiles trapping 45

3.3 Dosing and Sample Collection:

3.3.1 Radiochemical samples:

The dose solutions made in either ethanol or acetone were dispensed on the surface of the mounted skin using an Eppendorf® Micro Pipette (Eppendorf North America Inc., NY).

The receptor fluid was collected into glass scintillation vials at pre determined times post dose.

For evaporation experiments, the Tenax® cartridges were also replaced simultaneously at pre determined collection times. The cartridges were thermally desorbed into a scintillation vial containing 12 mL of scintillation fluid. This was done at by encasing the cartridge in a heating sleeve up to 220 ºC and bubbling Ultra Pure Nitrogen gas (Wright

Brothers Inc., Cincinnati OH) at 40 ml/min for 10 to 15 minutes.

3.3.2 GC samples:

For absorption studies employing GC analysis, the receptor fluid was collected at pre determined times. At the end of the experiment the skin was dissected to obtain epidermis and dermis separately. The Franz® cell top and the receptor chamber were rinsed with acetone and the wash was collected in a separate vial. Parafilm® and aluminum foil were also collected in individual vials. All samples were stored at −4 ºC until the day of analysis.

3.4 Sample Analysis:

3.4.1 Radiochemical Method:

For radioactive chemicals, the receptor fluid vials were filled with 12 mL of Ultima®

Gold Scintillation Fluid (Perkin Elmer Inc., Wellesley, MA). The radioactivity was measured by Beckman LS 6500 TM scintillation counter (Beckman Coulter Inc., Fullerton,

CA).

3.4.2 Analytical Method:

The receptor fluid was analyzed by Solid Phase Microextraction in the head space mode

(HS-SPME). Section 3.5 describes the methodology in brief with some literature review for pesticide detection.

3.4.2.1 Sample preparation:

The samples were thawed at room temperature prior to analysis. The receptor fluid

(5mL) was transferred to a Kimble® glass conical vial with a screw cap top.

Polytetrafluoroethylene lined septa were used for sealing the vials and were regularly

replaced to avoid loss of analyte resulting from repeated needle piercing of the fiber

assembly. The vial was placed in a thermostatted base maintained at 70 ºC and the fiber

was exposed for a fixed period of time. After determining the optimum conditions such

as time of exposure and salt content, a needle was pierced through the septum and fiber was exposed to the headspace above the vial for a pre determined time.

3.5 Solid Phase Microexctraction (SPME)

SPME is a solventless sample preparation technique that enables pre-concentration of analytes prior to analysis. It does so by allowing partitioning of compounds between the bulk phase and a polymeric coating fused on a metal rod. Traditional sample preparation methodologies often involve multiple steps, are cumbersome and utilize an appreciable volume of solvents. By design this technique integrates several steps like sampling, extraction, pre concentration and ease of sample introduction into a single step procedure.

The technique was first described for environmental water analysis by Berlardi and

Pawliszyn in 1989 54. Since then it has emerged as a powerful analytical tool with applications in numerous scientific disciplines. These include environmental science, forensic science, food and flavor analyses, toxicology, forensic science, clinical science, physical chemistry, microbiology and fundamental SPME development.

The basic fiber assembly is shown in Figure 3.2 55. It consists of a hollow cylinder that encases a fiber and needle assembly which in turn is attached to a plunger. The fiber consists of a thin rod coated with a polymeric sheet of inert material like the ones used for coating gas chromatograph (GC) columns.

Figure 3.2 SPME fiber assembly 55

The selectivity of the technique comes in part by choosing the right fiber with desired

polarity and thickness. There are several types of fibers that are commercially available.

The most commonly used are polydimethylsiloxane (PDMS), polydivinylbenzene(DVB),

and polyacrylate(PA). PDMS fibers are suitable for a wide range of moderately polar

compounds and are often a good starting point. This makes them most commonly used

fibers. PA fibers, on the other hand, work well for more polar compounds. There are

several combination fibers like Carbowax-PDMS and Carbowax-DBV that are designed

for mixture analyses. Custom coating of fibers is also a possibility for highly specific

applications. There are several studies that compare different types of fibers with respect

to their efficiencies and sensitivities for detecting several types of compounds 56-58.

There are three modes of fiber exposure –

1. Direct Immersion – The fiber is exposed directly to the bulk of the sample.

2. Headspace – The fiber is introduced into the headspace above the sample.

3. Membrane-protected – The fiber is enclosed in a protective sheath and then

exposed to the sample, usually by an immersion technique.

More recently, modification of techniques like Single Drop Micro Extraction (SDME),

Headspace Sorptive Extraction (HSSE) and Stir bar Sorptive Extraction (SBSE) have been developed as well 59,60. Headspace (HS) mode is preferred when the analyte of

interest is volatile or semi volatile and also when interference from sample matrix is

anticipated for example from humic substances and from biological fluids. Since the

fiber is not in direct contact with the sample matrix, headspace sampling results in

longevity of the fiber, i.e., more determinations per fiber. However, HS sampling is very sensitive to the volume of HS in addition to factors such as temperature and time of exposure. Thus care must be taken to ensure accuracy to obtain reproducible results.

Figure 3.3 represents a schematic of steps involved in SPME. These are listed below.

1. Depress the plunger to pierce the septum and expose the fiber through the needle

2. Keep the fiber exposed for a specific period of time

3. Retract the fiber

4. Pierce the chromatography septum

5. Expose the fiber into the heated injection port to enable desorption of analytes

from the fiber

6. Retract the fiber

Figure 3.3 Schematic descriptions of steps in SPME

By controlling the exposure conditions like time, temperature, salt content, pH and

stirring, quantitative results can be achieved. The assembly is readily mountable and

amenable to automation as well.

Since its first commercial use by Supelco in 1993, there has been an exponential growth

in the literature with different applications. A large collection of publications can be

found at http://sciborg.uwaterloo.ca/chemistry/pawliszyn

To impress upon the reader the vastness of the literature, following is a breakdown of the

articles currently listed on the website. This was last updated in November 2006.

o General articles – 104

o Fundamental development – 224

o Food analysis and Botanical applications – 600

o Clinical and Forensic applications – 366

o Environmental applications – 955

Table 3.1 lists a select few articles that describe pesticide determination and

quantification of pesticides based on SPME and HS-SPME methodology. These are

mostly recent publications and offer the reader an insight on the variety of applications

available

Table 3.1 Recent SPME pesticide studies

Reference: Year Type of Sample Interface

61 2006 honey GC

62 2006 whole milk GC

63 2006 water GC

64 2006 ground water HPLC-UV

65 2005 pesticides HPLC

food, water, soil, 66 2000 Several biological

Despite the growing acceptance of SPME as an analytical technique, there is a surprising

scarcity of its application in the area of skin absorption. Curran et al have reported a

study to characterize odor from human skin which employs HS-SPME 67. A recent article from 2005 demonstrates the use of SPME to study in vitro permeation of diethyl

phthalate (DEP), a stabilizer in cosmetic products, through hairless guinea pig skin 68.

Using immersion SPME and 85 µm PA fibers, permeation of DEP was characterized by determining kp values. In this regard, use of HS-SPME to characterize absorption of

topically applied compounds is a novel application that will be discussed for two

compounds- tecnazene and malathion in chapters 6 and 7 respectively.

CHAPTER FOUR

WORKING DIFFUSION MODEL

FOUR LAYER MODEL

kevap

Vehicle Permeant Layer (Case 2) z = 0 C(z,0) = g(z)

Stratum D , K h corneum sc sc sc

D , K ed ed h Epidermis ed z

D , K Dermis de de hde

Clearance kde

Figure 4.1 One dimensional diffusion model

Figure 4.1 is a schematic of a working one-dimensional diffusion model 69. It is comprised of four layers − (1) Donor solution layer (ds), (2) Stratum corneum (sc), (3)

Viable epidermis (ed) and (4) Dermis (de). The model uses physical and chemical properties of the compound such as molecular weight (MW), density (ρ), water solubility

(S w), vapor pressure (P vp), octanol water partition coefficient (K oct/w) as input

parameters to calculate the flux and amount of compound being absorbed in to the skin.

This model uses the effective medium, macrotransport approach described by Brenner

and Edwards 70 to describe the transport properties of complex and heterogenous systems,

such as skin membrane system, in a simpler way. Thus, for SC, the model considers the

uniform effective medium of thickness h; whose properties are a composite of lipid and

protein phases of the SC 47,49,69,71.

The model addresses the transport of solvents or solvent deposited compounds and the

upper two layers donor solution and SC are described in detail elsewhere 47. Briefly, for

a permeant with initial concentration C, Csat is the maximum concentration in the upper

layers of the SC. This also represents the solubility limit of the permeant within the SC.

It is assumed to be uniformly deposited in the upper SC to a depth hdep up to its solubility

Csat. Msat is the saturation dose ≅ Csat × hdep. There are two scenarios possible depending

upon the dose M0, wherein Case 1 corresponds to M0 ≤ Msat and Case 2 corresponds to

M0 > Msat. In Case 1, the permeant evaporates from the skin at a rate proportional to its

concentration in the skin while in Case 2, along with the evaporation, the permeant

saturates the upper layers of SC and maintains this condition until the permeant gets

depleted, after which it reverts to Case 1.

The diffusivity (Dsc) is calculated from permeability of SC (Psc), partition coefficient

between the SC and water (Ksc) and thickness h, using equations described in detail

48,49,72 elsewhere . Psc calculation takes into account the hydration state of the SC,

including lipid disruptions, geometric swelling and changes in the corneocyte

permeability. The Ksc is calculated from an equation which arises from a two phase partition model based on a comprehensive compilation of literature SC/w partitioning

data 48. Best fit analyses were carried out to give rise to correlations that correspond to

average lipid phase and corneocyte phase occupancy.

For the dermis, Kde is estimated using a calculation incorporating ionization of weak

acids and bases, non covalent binding to extravascular serum proteins (mainly albumin) and partitioning into lipids associated with the dermis 73. This is a marked improvement

over previous treatment of dermis as an aqueous layer, especially for highly lipophilic

compounds like tecnazene as demonstrated later (Chapter 7). For the purposes of comparing in vitro results to model comparisons, capillary clearance within the tissue is

taken to be zero and the thickness of the dermis is taken to be 300μm. Dde is calculated

from a formula that arises from regression of literature data of measured dermis

diffusivities of compounds against their molecular weight 73.

The viable epidermis is treated as unperfused dermis. Thus, Ded and Ked have the same

value as Dde and Kde. The thickness of the viable epidermis layer is taken to be 100μm.

The evaporative loss component of the model calculates the mass transfer coefficient

using chemical spills literature 74 that factors in the effect of wind velocity 47.

Finally, this calculation is incorporated into Visual Basic code incorporated in a

Microsoft Excel add in. Comparison of the model calculations and experimental data for tecnazene and malathion are presented in Chapters 7 and 9 respectively.

CHAPTER FIVE

DOSE AND AIRFLOW DEPENDENCE OF BENZYL ALCOHOL DISPOSITION ON SKIN

Matthew A. Miller, Varsha Bhatt and Gerald B. Kasting* College of Pharmacy The University of Cincinnati Medical Center

Journal of Pharmaceutical Sciences 2006, Volume 95, p 281-291

5.1 Introduction

Quantitative dermal risk assessment for volatile chemicals is routinely carried out for

cosmetic and personal care products. Fine fragrances employ concentrated solutions of

volatile ingredients from natural and synthetic sources in an even more volatile solvent, usually ethanol or phenoxyethanol. The composition of these products is complex 75.

Fragrances are used at lower levels in a wide variety of consumer products. Risk assessment for these products requires a careful study of each component, with top priority often given to fragrances and preservatives 76,77. Estimation of the absorbed dose is key to identifying systemic risks, whereas absorption rate and skin concentrations are important for understanding contact sensitization thresholds 78. The objective of the

present study is to characterize the skin disposition of a representative fragrance

ingredient, benzyl alcohol (BA), in terms of the diffusion model described in the

accompanying paper 79. To do this we conducted an in vitro study to determine the dose

dependence of BA absorption following topical application in ethanol and analyzed these

data in conjunction with a previous study of BA skin disposition which evaluated the

airflow dependence 80.

5.2 Methods

5.2.1 Chemicals

Carbonyl-[14C] benzyl alcohol (55 mCi/mmol; 0.1 mCi/mL in ethanol) was purchased

from Moravek Biochemicals (Brea, CA). The radiochemical purity was stated by the manufacturer to be 98.3%. Unlabeled benzyl alcohol and calcium-free Dulbecco’s phosphate-buffered saline were purchased from Sigma-Aldrich (St. Louis, MO).

Soluene-350® was from Perkin-Elmer Biosciences. Ethanol (95%) was from Aaper

(Shelbyville, KY).

5.2.2 Dose-Dependence Study

The method has been previously described 81. Split-thickness human cadaver skin (~300

μm) was mounted in modified Franz diffusion cells (0.79 cm2) and placed in

thermostatted aluminum blocks maintained at 37°C in a fume hood with a partially drawn

sash. Skin was obtained from three donors, with 4-5 replicates per donor at each dose.

The receptor solution (magnetically stirred) was Dulbecco’s phosphate-buffered saline, pH 7.4, containing 0.02% sodium azide to inhibit microbial growth. The integrity of the

3 82 2 2 tissue was verified using H2O . BA doses ranging from 0.9 μg/cm to 10.6 mg/cm dissolved in ethanol were applied to each skin sample. The dose volume was 5 μL for the smaller doses, 10 μL for the 3.24 mg/cm2 dose and 20 μL for the 10.6 mg/cm2 dose. The

applied dose appeared to spread rapidly and evenly across the skin surface, which was

visually dry within 10-30 seconds depending on the dose volume. The receptor solution

was removed for analysis at 1, 2, 4, 8, 12 and 24 hours post-dose and replaced with fresh

buffer. Following the last receptor collection, the skin samples were removed from the

diffusion cells and dissolved in 2 mL of Soluene®. All samples were analyzed by liquid

scintillation counting (LSC). All data were first averaged by dose for each donor and

were then averaged across donors to obtain the reported results.

5.2.3 Airflow-Dependence Study

This study has been previously reported 80. The methodology was similar to the dose-

dependence study, except that a single dose of 127 μg/cm2 of BA dissolved in 10 μL of ethanol was applied to the skin samples and a modified top was fitted to the Franz cells.

The setup allowed room air to be drawn over the samples at a controlled flow rate.

Volatiles were collected in absorbent cartridges at regular intervals and analyzed (along with the receptor solution samples) by LSC following thermal desorption.

5.2.4 Data Analysis

The cumulative absorption and evaporation data from both experiments were analyzed according to the diffusion model in Ref. 79. Nonlinear regression analysis was performed

in Microsoft Excel® as described elsewhere 81,83. The quantity minimized was the

2 residual sum of squares (SSR) normalized by the degrees of freedom, or χν as defined by Bevington,84

1 n 2 χ 2 = w []y (obs ) − y (fit ) (1) v − ∑ i i i n p i =1

where n is the number of observations, p is the number of adjustable parameters, wi is the weight and the yi’s are the cumulative amounts absorbed or evaporated for each observation. In order to assess the stability of the model parameters and the importance of the penetration enhancement effect in the dose-dependence study, several variations on the fitting procedure were conducted. These included constant diffusivity (Model 1) vs. variable diffusivity (Model 2), separate analyses of the dose- and airflow-dependence

2 data, and a combined analysis. Both equal weighting and 1/ yi weighting were investigated. Each of these variations encompassed periods of time in which the residual

BA in the system fell into the large dose (Case 2) and small dose (Case 1) limits as described in Ref. 79. A small headspace correction was applied to the evaporation data from the airflow-dependence experiment 85. This adjustment allowed the inclusion of two early time evaporation data at low airflows that had been omitted in a previous analysis 80. As this correction was of minor consequence for BA, the details will be reserved for a subsequent report. Values of the squared correlation coefficient r2 were calculated by treating each type of observation (absorption from the dose study, absorption from the airflow study, and evaporation from the airflow study) as a separate data type having its own mean value. Squared deviations of the observed values from these means were then calculated and summed to produce a total sum of squares (SST).

2 2 The sum of squared residuals (SSR) was calculated from eq. 1 as SSR = (n − p)χν and r followed from r2 = 1 − SSR/SST. This procedure resulted in lower r2 values than those reported in a previous analysis 80 and gave more meaningful estimates of the fraction of explained variance.

Absorptive and evaporative fluxes were estimated from the cumulative data by a two- point difference method. Values so calculated were plotted at the midpoint of the collection time interval for graphical presentation. This procedure yields approximations to the true flux vs. time profiles due to the discrete nature of the data. It did not affect the fitted model parameters since these were determined from the cumulative data rather than from flux.

For the case of constant diffusivity D, the mathematical model 79 involves four independent, dimensionless parameters (τ, f, β and χ) and a scaling parameter (h or h2/D).

The former are related to the dimensional model parameters as follows:

τ = Dt h 2 reduced time (2)

= f hdep h fractional deposition depth (3)

β = hCsat M 0 membrane capacity/dose ratio (4)

χ = ρ hkevap DCsat volatility parameter (5)

In eqs. 2-5, D is the diffusivity of the permeant in the membrane, t is time, h is the membrane thickness, hdep is the thickness of the deposition region in the upper part of the membrane, Csat is the solubility of the permeant in the membrane, M0 is the dose, kevap is

the evaporative mass transfer coefficient and ρ is the density of the pure liquid (or solid) form of the permeant. Two useful, dose-related parameters can be defined from these variables: a saturation dose, Msat,

= = β ⋅ M sat hdepCsat ( f ) M 0 (6)

and a “reduced dose”, Mr,

= β −1 = M r ( f ) M 0 / M sat . (7)

The former is the amount of permeant required to saturate the deposition region and the latter is the (inverse) ratio of this quantity to the applied dose. For values of Mr ≤ 1 (Case

1) the permeant fully dissolves into the membrane upon dosing, whereas for Mr > 1 (Case

2) a liquid (or solid) film of permeant remains on the skin surface. These two cases lead to different solutions to the diffusion equation in the membrane and to correspondingly different absorption and evaporation profiles. Another important relationship defines the gas phase mass transfer coefficient for evaporation, kg:

P MW k ρ = k vp (8) evap g 0.76RT

where Pvp is the vapor pressure of the permeant, MW is its molecular weight, R is the gas constant and T is absolute temperature. This equation allows the evaporation rate to be expressed in terms of a number (kg) that is fundamentally easier to estimate than kevap.

The units of all parameters are given in Ref. 79 and their values are discussed later.

Consideration of this problem shows that, for the case of constant D, there are only four parameters that can be independently estimated. (This follows since the scaling parameter is incorporated into eqs. 2-5). We found it most straightforward to fit the

2 model to experimental data in terms of the dimensional parameters h /D, kg and Msat and the dimensionless ratio f. The remaining model parameters can then be calculated from these values and known physical properties of the permeant using eqs. 2-8. The parameter kg is the diffusion model analog of the evaporation rate constant k1 in a previous analysis,80 which was shown to vary linearly with airflow velocity v to within the precision of the experiment. Hence we chose the airflow dependence for kg to be

= ′ ⋅⎛ v ⎞ k g k g ⎜ ⎟ (9) ⎝ 42 mL/min ⎠

′ where v is airflow in mL/min (airflow dependence study) and k g is a constant of proportionality. An equivalent value of v = 42 mL/min was determined for the dose- dependence study (which was performed in a fume hood) by matching the absorption values to those in the airflow dependence study at a comparable dose. This value was consistent with results from a previous study 81. Equation 9 thus leads to the equality

= ′ k g k g for the dose-dependence study.

The fractional deposition depth was varied over the range f = 0.05 – 0.3. A broad optimum was found near the value f = 0.1, which also resulted in the least systematic variation in the other model parameters when dose data or airflow data were fitted separately. This value was used in all subsequent analyses.

Initial values of the time constant h2/D were estimated from the dose-dependence data by separately fitting the model at each individual dose. The value of h2/D was found to decrease systematically with increasing dose, as had been observed in a similar analysis of 14C-DEET disposition on skin 81. Rather than reporting this variation, we developed a sigmoidal expression for a concentration-dependent diffusion coefficient which varied

(approximately) from D0 at low concentrations to Dsat at the saturation concentration Csat.

The inflection point of the sigmoid curve was designated Ctrans and the slope factor as m.

Thus,

D − D D = D + sat 0 (10) 0 + []()− 1 exp m 1 C / Ctrans where C = C(x) is the local concentration of permeant in the membrane. Equation 10 is one of many possible forms that such a dependence D = D(C) could take. It was not possible to determine a unique, underlying form of D(C) from the analysis; however, it was possible to estimate the range over which D must vary and an approximate concentration at which the variation must occur to provide a quantitative explanation of the dose-dependence results. Equation 10 gave representative results.

5.3 Results

5.3.1 Dose-Dependence Study

Table 5.1 shows the mean absorption for radioactivity associated with 14C-BA as a function of time and applied dose. Flux profiles associated with these data are shown in

Figure 5.1. Maximum absorption rates were reached within the first hour post-dose for doses ≤ 300 μg/cm2 and at regularly increasing times for larger doses. This may be clearly seen in the plot of Jmax and tmax shown in Figure 5.2. Maximum flux (Jmax) increased with dose at a slightly greater than dose-proportional rate for doses ≤ 300

μg/cm2 and more gradually at larger doses. Consequently, for the smaller doses, the percentages of radioactivity absorbed at any given time increased systematically with dose (Figure 5.3). At longer times this trend was continued through the highest dose in the study (Table 5.2). Cumulative absorption ranged from 20% of the applied dose at 0.9

μg/cm2 to 29% at 10600 μg/cm2. Thus, BA increased its own permeation rate through skin under the conditions of the test, i.e., it was a mild penetration enhancer.

The amount of radioactivity remaining in the tissue 24 h after application was less than

2% for all doses except the highest, for which 3.8% was retained (Table 5.2). At the lower doses, skin retention was essentially independent of dose, suggesting that the dissipation process was complete by 24 h. Thus, the small amounts of retained 14C appear to be bound to the tissue. Similar results were obtained in the airflow-dependence

study 80. Further analysis of these data was conducted under the assumption that radioactivity missing from the system at 24 h had evaporated. Evaporation estimated by this procedure ranged from 67-79% of the dose (Table 5.2).

14 (a) 12 10 8 6 97 4 29 2 9.4 0.9

h 2.8 2 0 g/cm

, μ 0.5 1.0 1.5 2.0 2.5 3.0

abs 500

J Hours Post-Dose (b) 400

10600 300

200 3240 100 990 290 0

0 5 10 15 20 Hours Post-Dose

Figure 5.1. Skin absorption rates for radioactivity associated with 14C-benzyl alcohol in the dose-dependence study (mean ± SE of 3 donors, n = 4-5/donor). The numbers on the graph are the applied dose of BA in μg/cm2 and the lines are a guide to the eye. (a) Low doses; (b) High doses.

1000 5

100 4 h 2

g/cm 10 3 μ

1 2 (max), hours (max), abs t abs J 0.1 1

0.01 0 1 10 100 1000 10000 Dose, μg/cm2

Figure 5.2. Maximum flux (•) and time-to-maximum flux (○) for absorption of radioactivity in dose-dependence skin disposition study with 14C-benzyl alcohol. The error bars on the closed symbols reflect sample-to-sample variation, whereas those on the open symbols reflect the discrete sampling interval. The curves were calculated from the diffusion model using the bolded parameters in Table 5.3. Solid line – Calculated maximum flux; Light dashed line – Apparent maximum flux if measured over experimental sampling interval; Heavy dashed line – time-to-maximum flux.

4 h

2 h 10

1 h Percent of Dose Absorbed 1 1 10 100 1000 10000 μ 2 Dose, g/cm

Figure 5.3. Cumulative percentage of radioactivity in the receptor solutions (mean ± SE of 3 donors, n = 4-5/donor) for dose dependence skin disposition study with 14C-benzyl alcohol. •1 h post-dose; ○ 2 h post-dose; ▼4 h post-dose. The curves were calculated from the diffusion model using the bolded parameters in Table 5.3.

Table 5.1: Appearance of Radioactivity in the Receptor Solution for Dose-dependence

Skin Disposition Study with 14C-Benzyl Alcohol (mean of 3 donors, n = 4-5/donor).

Dose Percentage of Dose

µg/cm2 0-1 h 1-2 h 2-4 h 4-8 h 8-12 h 12-24 h

0.9 10.2 5.2 2.6 1.1 0.4 0.3

2.8 11.1 5.8 2.5 1.2 0.3 0.2

9.4 11.6 6.5 2.4 0.8 0.3 0.2

29 13.1 5.0 2.0 0.9 0.4 0.4

97 12.4 5.0 1.8 0.5 0.1 0.1

290 13.7 7.7 3.2 1.1 0.3 0.2

990 8.5 9.9 5.0 1.2 0.2 0.2

3240 3.3 7.9 9.4 2.7 0.5 0.4

10600 1.3 3.3 7.2 11.3 5.2 0.9

Table 5.2. Cumulative Disposition of Radioactivity at 24 h for the Study Reported in

Table 5.1. Results are expressed as percentage of the applied dose (mean ± SE of 3 donors).

Dose, Evaporated

µg/cm2 Absorbed Skin (estimated)

0.9 19.8 ± 2.9 1.6 ± 0.1 78.6 ± 3.0

2.8 21.1 ± 1.7 1.0 ± 0.0 77.9 ± 1.7

9.4 21.9 ± 1.1 1.7 ± 0.7 76.4 ± 0.5

29 21.8 ± 0.9 0.6 ± 0.2 77.6 ± 1.1

97 20.0 ± 2.6 0.6 ± 0.0 79.4 ± 2.5

290 26.3 ± 3.2 1.1 ± 0.3 72.6 ± 3.5

990 25.0 ± 4.7 0.6 ± 0.1 74.4 ± 4.8

3240 24.2 ± 3.0 0.7 ± 0.1 75.1 ± 2.9

10600 29.2 ± 3.0 3.8 ± 2.5 67.0 ± 0.6

5.3.2 Airflow-Dependence Study

These data have been previously reported 80. Representative evaporation and absorption rates are shown in Figure 5.4. The evaporation rate of 14C-BA increased systematically with increasing airflow over the skin surface (Fig. 5.4a), with a corresponding decrease in the absorption rate (Fig. 5.4b). The maximum absorption rate occurred within 30 min post-dose for airflows of 50-100 mL/min and between 30 and 60 min for airflows of 10-

40 mL/min. Maximum evaporation rate occurred within 5 min for the higher airflows and between 5 and 15 min for the lower airflows. The delay for the lower airflows is related to the headspace between the skin surface and the vapor trap 80,86. These results can be described in terms of compartmental models in which the rate constant for evaporation is proportional to airflow 80. No information on skin permeability enhancement was obtained in this study as only a single dose of BA was tested.

1200 400 (a) 30 1000 h 2 300 20 h 2 800 g/cm

μ 200 , 10 g/cm

600 evap

μ 100 J ,

evap 400 0 J 0.0 0.2 0.4 0.6 0.8 1.0 200 50 Hours Post-Dose 100 0 0.0 0.2 0.4 0.6 0.8 1.0 60 (b) 50 10

h 40 2 30 20 g/cm μ , 20

abs 30 J 50 10 100 0

012345

Hours Post-Dose

Figure 5.4. Skin evaporation rates (a) and absorption rates (b) for radioactivity associated with 14C-benzyl alcohol in the airflow-dependence study (mean of 2 trials).80

The numbers on the graph are the airflow over the skin surface in mL/min. The lines are a guide to the eye.

5.3.3 Test of Diffusion Model

The diffusion model 79 was fit to the dose- and airflow-dependence data as described under Methods. The parameters resulting from this process are shown in Table 5.2.

Model calculations are shown in Figures 5.2, 5.3, 5.5 and 5.6.

For the dose-dependence study, the time constant h2/D was found to decrease with increasing dose. The average value of 5.2 h obtained by assuming a constant diffusivity

(Model 1) gave a poor representation of the absorption at high doses. This is consistent with the penetration enhancement effect noted earlier (Table 5.2 and Figure 5.3) and has also been seen with DEET 81. Better fits were obtained using Model 2, in which diffusivity was considered to be a function of the local concentration of BA in the skin.

The optimum fit was obtained using a model in which diffusivity increased by a factor of

2.5 as the local BA concentration increased from 0 to its saturation value, Csat. The value

2 −4 of Csat for this fit (calculated from eq. 6 with hdep = 1.5 μm) was 127 μg/cm /(1.5 × 10 cm) × (mg/1000 μg) = 850 mg/cm3. The degrees of freedom-adjusted residual sum of

2 squares ( χυ ) was almost 4-fold lower with Model 2 than Model 1, an improvement significant at p < 0.001 84. A corresponding increase in the r2 value of the fit from 0.56 to

0.89 was also obtained.

Table 5 3. Regression Parameters for Diffusion Models of Benzyl Alcohol Disposition on Skin. A fractional deposition depth f = 0.1 was used in these analyses.

Model 1: Constant Diffusivity Parameters Units Dose Airflow Combined 2 h /D h 5.2 2.4 2.2 ′ k g cm/h 1130 930 860 a 3 b a Vh cm [0] 45 42 2 Msat µg/cm 142 41 42 v mL/min [42] −c −c,d n − 63 104 167 s % of dose 4.1 4.1 5.0 r2 − 0.56 0.95 0.92 2 2 χυ (% of dose) 17.1 17.1 24.9

Model 2: Variable Diffusivity 2 h /D0 h 5.7 2.4 2.6 ′ k g cm/h 1190 1090 880 a 3 a Vh cm [0] 44 35 2 Msat µg/cm 127 39 34 v mL/min 42 −c −c,d Dsat / D0 − 2.5 2.4 3.0 Ctrans / Csat − 0.61 [0.61] [0.61] m 5.0 [5.0] [5.0] n − 63 104 167 s % of dose 2.1 4.3 4.2 r2 − 0.89 0.94 0.94 2 2 χυ (% of dose) 4.5 18.9 17.7 a Vh is a headspace volume parameter used to account for the evaporation time lag. The

value applies only to data obtained in the airflow-dependence study. bBrackets denote parameters that were not optimized for this dataset. cValue of v varied systematically from 10-100 mL/min for airflow-dependence study (cf.

Ref. 80). dValue of v was fixed at 42 mL/min for dose-dependence study.

For the airflow-dependence study, comparable fits were obtained using Models 1 and 2.

2 The optimum values of h /D and Msat were lower than those in the dose study by factors of 2 and 3, respectively. When both studies were analyzed together, the airflow study tended to drive the results. This was due in part to the larger number of observations and in part to the fact that the dose study provided no direct data on evaporation. The fit of

Model 2 to the combined dataset (bolded values in Table 5.3) was taken to be the best overall representation of the BA disposition data, explaining 94% of the variance in the dataset and yielding an rms deviation of observed and fitted values of 4.2%.

Using the parameters in Table 5.3, the diffusion model failed to accurately describe the extended “tails” of the evaporation and absorption profiles. Representative comparisons are shown in Figures 5.5 and 5.6. Diffusion through a homogeneous membrane with a constant diffusivity leads to terminal exponential behavior of the flux, with the decay constant determined by the leading eigenvalue of the sum of exponentials solution 79,87.

Thus, such models predict a linear decay of the logarithm of evaporative or absorptive flux with time. Both the dose and airflow studies with BA showed a curvilinear decay

(Figures 5.5 and 5.6), indicative of prolonged retention of the radiolabel. Similar behavior has been observed in other studies of fragrance disposition on skin 88. This phenomenon is a poorly understood aspect of BA disposition in these in vitro experiments that is further addressed in the Discussion.

Figure 5.5. Flux of radioactivity associated with 14C-BA from the surface of the skin in the airflow study. ∇ 20 mL/min; ▲ 100 mL/min (a) evaporation; (b) absorption. The curves are calculations using the bolded parameters in Table 5.3 (Model 2).

Figure 5.6. Absorptive flux of radioactivity associated with 14C-BA for the dose- dependence study. ○ 2.8 µg/cm2; ● 3240 µg/cm2. The heavy lines are calculations using the bolded parameters in Table 5.3 (Model 2). The thinner lines are calculations using the parameters associated with the fit of Model 2 to the dose data only (Column 3, Table

5.3).

5.4 Discussion

The general features of BA disposition on skin following solvent deposition from ethanol include (1) rapid absorption with peak flux within one hour except at very large doses

(Figs. 5.1 and 5.2); (2) evaporation of 45-85% of the applied dose from the skin surface at rates proportional to airflow over the skin 80,86 (cf. Table 5.2 and Fig. 5.4); (3) a 47% increase in percentage absorption at 24 h with dose over the dose range 1-10000 μg/cm2

(Table 5.2); and (4) prolonged “tails” to the absorption and evaporation profiles indicative of non exponential processes (Figs. 5.5 and 5.6). The first three of these phenomena were well described by the diffusion model 79 using a variable diffusivity coefficient that varied by a factor of 2.4-3.0 over the range of estimated concentrations in the skin (Table 5.3, Model 2). The non exponential behavior was not predicted. In vivo studies of other fragrance ingredients 88 and in vitro work with DEET 81,85 suggest that the prolonged release characteristics of BA from skin under similar test conditions may be shared by other volatile organics. While of little consequence for the cumulative absorption/evaporation ratios, a slow release profile could contribute to prolonged efficacy of either fragrances or insect repellent products.

In order to test whether the prolonged absorption rates were associated with the tissue or with the experimental conditions, we conducted a dose-dependence study for 14C-BA using a silicone membrane (0.020” or 500 μm) in place of the skin. A constant diffusivity model (Model 1) was then fit to the absorption data using the same procedure as in the skin penetration studies. Representative results are shown in Figure 5.7. BA penetration through silicone membrane showed even more prolonged absorption rates

than had skin, demonstrating that this phenomenon was not associated with the heterogeneous nature of skin or some mysterious “stratum corneum reservoir”.

Anissimov and Roberts’ analysis 87 shows that accumulation of solute in the receptor solution can lead to prolonged flux profiles similar to those in Figs. 5.5b, 5.6 and 5.7.

However, the receptor solutions were exchanged completely at each sampling point in our studies and the concentration of BA never exceeded 0.6% of its solubility. Therefore, it is unlikely that this feature was associated with a failure to maintain sink conditions. It may be associated with lateral diffusion of radiolabel into the area of the membrane clamped by the ground glass joint, and a corresponding slow release from this region; however, this is still a matter of conjecture. A similar phenomenon could also have occurred in the in vivo studies in which a volatiles trap was strapped continuously to the skin 86,89. A definitive answer to this question may require an in vivo test in which a volatiles trap is applied intermittently rather than continuously. Such experiments require high analytical sensitivity due to the low evaporation rates later in the study.

Other, relatively minor differences between theory and experiment may be understood in terms of the experimental design. The theory predicts somewhat higher values of the maximum absorptive flux than calculated from the data (Fig. 5.2); however, the sampling frequency in the experiment lead to averaging of the flux over times comparable to the width of the peak, broadening and lowering the peak shape. When the theoretical flux was averaged over an interval corresponding to the sampling interval (light dashed line in

Fig. 5.2), the agreement between theory and experiment was substantially improved. In any case, this sampling artifact was not a factor in the data analysis because model

parameters were fit to cumulative data rather than flux. There was also a small delay in the evaporation rate profiles for low airflows in the airflow-dependence study (Fig. 5.4a) as has been previously discussed 80,86. We accounted for this delay in the model by adding a correction related to the headspace between the skin surface and the vapor trap.

As this feature was of minor consequence for the BA analysis, discussion is deferred to a study in which it plays a larger role 85.

At the highest dose (10.6 mg/cm2) the theory predicts a shorter elapsed time for attainment of the maximum absorption rate tabs(max) than was observed experimentally

(Fig. 5.2). In fact, the theoretical flux at this dose is nearly constant from 1.5 h to 8 h

(data not shown), whereas the experimental flux peaked between 2 and 4 h and fell only slightly during the 4-8 h time period (Table 5.1 and Fig. 5.1). Thus, the difference in tabs(max) is less substantial than might be inferred by merely studying Fig. 5.2. The key finding was that a prolonged, nearly constant absorption rate was obtained from large doses of BA applied to skin, consistent with the Case 2 scenario in the diffusion model.

5.4.1 Independent estimation of diffusion parameters

As noted in the accompanying paper 79 (Evaluation of Parameters section) the parameters associated with the diffusion/evaporation model for an arbitrary permeant can be estimated from existing data and correlations without the need of conducting an experiment. A detailed procedure for this calculation was proposed 79. It is instructive to compare the parameters resulting from such a calculation for BA with those arising from

the detailed model fitting process described herein. By so doing, one can gain a feel for whether the model in its present form is capable of making useful predictions.

Admittedly, this test is for a single compound; however, additional comparisons are possible 81,83.

Following the procedure outlined in Ref. 79, we first obtained relevant physical properties for BA and selected appropriate environmental conditions. These values are given in

Table 5.4. In order to compare with diffusion cell data obtained in a laboratory fume hood (dose-dependence study) we chose a skin temperature of 30°C and a wind velocity of 1.5 m/s, corresponding to a typical outdoor environment 90. The four required diffusion model inputs were then estimated as follows: The values h = 13.4 μm and f =

0.1 were selected as discussed elsewhere 79. The steady-state permeability coefficient for

91 BA in aqueous solution, kp = 0.0169 cm/h, was obtained from the literature and an ~ estimate of the corresponding value for partially hydrated skin, k p = 0.00563 cm/h, was

79 made by dividing kp by 3 . An estimate of the SC/water partition coefficient for BA in

~ 79 partially hydrated skin, K SC / w = 8.71, was made using eq. 43 in Ref. . The remaining parameters were then calculated from eqs. 2-8 or the additional relationships in Ref. 79

(eqs. 36-45). The results of the calculation (Tables 5.5 and 5.6) are presented in comparison to values obtained using the directly fitted model parameters from Table 5.3.

Table 5.4. Physical Properties and Auxiliary Skin Permeability Data for Benzyl Alcohol

Parameter Units Value Ref.

Physical and environmental properties

MW g/mol 108.1 −

ρa g/cm3 1.04 92

b 93 log Koct 1.10

c d 94 Pvp torr 0.149

e 3 f 95 Sw g/cm 0.0429

RT (30°C) Latm/mol 24.89 −

ug m/s 1.5 90

Measured or estimated parameters

h μm 13.4 79

f − 0.1 79

91 kph cm/h 0.0169

~ i 79 K SC / w 8.71 aDensity; bOctanol/water partition coefficient; cVapor pressure; dExtrapolated from 25°C value of 0.0940 torr by Antoine method. 94; eWater solubility; f25°C value; gWind velocity; hSteady-state permeability coefficient from aqueous solution; iStratum corneum/water partition coefficient for partially hydrated skin (eq. 43 in Ref. 79)

Table 5.5 Derived Diffusion Model Parameters for Benzyl Alcohol. Values were calculated from eqs. 2-9 in the text unless otherwise specified.

Parameter Units Source

Table 5.3 (bolded values) Table 5.4

a hdep μm 1.50 1.34

~ b c d k p cm/h 0.00305 0.00563

2 D0Csat/h μg/cm h 131 242

2 h /D0 2.6 2.07

2 −1 −10 −10 D0 cm s 2.40 × 10 2.41 × 10

3 Csat mg/cm 227 374

2 Msat μg/cm 34 50

e kg cm/h 880 1820

−4 −3 kevap cm/h 7.17 × 10 1.49 × 10

χ − 5.70 6.41

aDeposition region; bSteady-state permeability coefficient for partially hydrated skin79; c d e 79 Calculated as (D0Csat/h) / Sw; Calculated as kp/3; Calculated from eq. 37 in Ref. .

Table 5.3 shows a general agreement between the model parameters obtained via the independent parameter estimation method and the experimentally-derived parameters.

Although the skin capacity measures Msat and Csat derived by the independent method were 50-60% larger than those derived experimentally, these differences were offset by higher evaporation rate parameters kg and kevap, leading to comparable estimates for the dimensionless ratio χ and, hence, of total absorption at low-to-moderate doses (Table

5.6). As dose increased, the independent method underpredicted absorption by a wider margin, since it did not account for increased skin permeability to BA. This effect could be anticipated in a model calculation (e.g., for risk assessment) by building in an appropriate variable diffusivity factor using eq. 10. Thus, we view the comparisons in

Tables 5.5 and 5.6 as supportive of a predictive capability for the diffusion model. Such a method, according to which transient disposition profiles of either volatile and non volatile chemicals on skin could be estimated from physical and environmental properties combined with steady-state permeability data, would have substantial value.

Table 5.6 Percentage absorption of BA after 24 h estimated by three different methods

Dose Experimental Diffusion Model Calculations

µg/cm2 (Table 5.2) (Table 5.5, Column 3)a (Table 5.5, Column 4)b

0.9 19.8 ± 2.9 19.4 17.5

29 21.8 ± 0.9 20.3 17.5

990 25.0 ± 4.7 23.0 13.2

10600 29.2 ± 3.0 23.8 13.1

aModel estimates obtained by fitting parameters to experimental data (Franz diffusion cells in a fume hood); bModel estimates obtained via independent parameter estimation method as described in Discussion and Ref. 79.

5.5 Conclusions

Many features of the dose and airflow dependence of BA disposition on human skin in vitro following solvent deposition have been found to be consistent with a finite dose diffusion model involving evaporative loss from the skin surface and a concentration- dependent diffusivity of BA in the skin. The extent of agreement of these results with independent, model-based estimates of BA skin disposition is supportive of an eventual predictive use of the model.

5.5.1 Acknowledgements

Financial support for this study was provided by NIOSH/CDC grant R01 OHO7529.

CHAPTER SIX

AIRFLOW DEPENDENCE STUDY OF DEET: RESULTS

6.1 Introduction:

DEET is the most widely used insect repellant. It can be found in a variety of products and in various formulations like lotions, solutions, pressurized and non pressurized aerosols and more recently in the form of towelettes and wrist bands. It was developed by the US army in 1946 and introduced for public use in 1957. The consumer products containing DEET, which are registered with the EPA, range in concentrations from 4 to

100%. It is estimated that DEET is used in approximately 30% of the US population annually 96. The compound has been studied extensively and is generally considered safe and effective for consumers 97,98. Schoenig and Osimitz have compiled the toxicity studies conducted in animals and humans using DEET 99.

Most formulations of DEET are ethanol based and recently products containing DEET and sunscreen agents have also been introduced. Studies have shown DEET to be a permeation enhancer when co-administered with other compounds like sunscreens 100 and other pesticides like permethrin 101. DEET has also been shown to increase its own absorption at higher doses 46. Effect of formulation additives 102,103, airflow 36 , dose 46 and temperature 104 and solvents 104,105 on absorption of DEET has been previously reported. The objective of the study was to characterize the evaporation and absorption of DEET from human skin in vitro. A related study with a dose-dependency of DEET has been separately reported 46. The results and data of the airflow study are presented in the following sections. Comparison of the model calculations and experimental data will soon follow.

6.2 EXPERIMENTAL

6.2.1 Materials

Carbonyl-[14C] DEET (52 mCi/mmol) was custom synthesized by Vitrax (Placentia, CA).

The radiochemical purity was stated by the manufacturer to be 99%. Unlabeled DEET

(CAS No.134-62-3), and calcium-free Dulbecco’s phosphate-buffered saline were purchased from Sigma-Aldrich (St. Louis, MO). Soluene-350® was purchased from

Perkin-Elmer Biosciences. Ethanol (95%) was purchased from Aaper (Shelbyville, KY).

Split thickness (300 µm) human cadaver skin (back, thigh and abdomen) was procured from U.S. Tissue and Cell (Cincinnati, OH). The skin was preserved in 10% glycerol and kept at −80 ºC until use.

6.2.2 Skin penetration study

The excised skin tissue was mounted on modified Franz® diffusion cells 51. The receptor solution was Dulbecco’s phosphate − buffered saline (pH 7.4) containing 0.02% (w/v) sodium azide w/v to inhibit microbial growth. Receptor solutions were maintained at 37

± 2ºC on thermostatted heating and stirring modules, yielding a skin surface temperature

3 of 30-32ºC. The tissue was screened using H2O permeability test described in Chapter

3, section 3.3.1. Samples with water permeation greater than 1.2 µl/cm2 were discarded

53. The receptor solution was exchanged twice to ensure complete removal of residual radioactivity. The top and bottom chambers were then sealed with Parafilm®. The cells

were randomized over treatments using previously established procedures 52. They were equilibrated overnight and dosed the following morning with 10 µL of 1%w/w DEET solution in ethanol resulting in an average dose of 127 µg/cm2.

6.2.3 Airflow-Dependent Study

The method has been previously reported to describe the disposition of benzyl alcohol 80.

The modifications used in the DEET airflow study have been described in detail in

Section 3.2. A single dose of 127 μg/cm2 of DEET dissolved in 10 μL of ethanol was applied to the skin samples and a modified top was fitted to the Franz cells, to which a

Tenax® cartridge was attached. The setup allowed room air to be drawn over the samples at a controlled flow rate. The airflows were varied from 10 to 100 mL/min. After finalizing the airflow, the cells were allowed to equilibrate for a minimum 30 minutes prior to application of dose solution. The top and the bottom chambers of the modified

Franz® cells were sealed with Parafilm® to minimize loss of analyte. Receptor fluid was collected at 0.25, 0.75, 2, 4, 8, 12 and 24 hours post dosing. Tenax® cartridges were also replaced at these times. The cartridges were thermally desorbed into 10 mL of scintillation fluid at 220 °C for 15 minutes and measured using liquid scintillation counting (LSC). At the conclusion of each experiment, the upper components of the trapping unit were rinsed with ethanol and stored separate vial. The pieces of Parafilm® and skin pieces of each cell were also collected into vials and measured along with receptor fluid and ethanol washings by LSC.

6.3 Results:

The structure of DEET is shown in Figure 6.1. The physicochemical properties of DEET are shown in Table 6.1. The mean cumulative evaporation and absorption of 14C − DEET recovered from the trap and receptor fluid is shown in Tables 6.2 and 6.3 respectively.

The appearance of radioactivity associated with 14C DEET in the volatiles trap and receptor fluid over the time course is presented in Tables 6.4 and 6.5 respectively. The over all mass balance is shown in Table 6.6. The total recovery was ranged from 78 to 95

% of applied radioactive dose. It is likely that the missing fraction may have evaporated.

Evaporation of DEET is likely to have occurred between the dosing and enclosing the trap over the modified Franz® cell. It is also likely that some radioactivity was lost while replacing of the Tenax® cartridges through out the experiment.

Figure 6.2 shows the overall disposition of DEET as a function of airflow (v = 10 – 100 mL/min). As seen from the graph, the amount of DEET absorbed through the skin decreases as the airflow increases. This is an expected outcome; as evaporation increases with increase in the airflow, resulting in decreased absorption from the surface of the skin. The evaporation and absorption rates of DEET at different airflows are shown

Figure 6.3 and 6.4 respectively. The evaporation rate of 14C-DEET increased systematically with increasing airflow over the skin surface, with a corresponding decrease in the absorption rate. The time required to reach maximum absorption rate and maximum evaporation rate decreased with increasing airflow velocity, excepting the airflow of 10 mL/min for absorption and 30 mL/min in the evaporation case. This delay

for the lower airflows could be related to the headspace between the surface of the skin and trap, and subsequent condensation of DEET on the insides on the headspace. As seen from the graphs, the evaporative fluxes peaked between 0 and 3 hours for most airflows studied. The lag times generally increased with decreasing air flow. The absorption fluxes peaked between 0.5 and 3 hours with no specific correlation to airflow.

The overall absorption and evaporation of DEET appears to be complete at the end of 24 hours.

Me C NEt 2

Figure 6.1 Structure of DEET

Table 6.1 Physical Properties of DEET

Parameter Units Value Reference

MW g/mol 191.3 −

ρa (30°C) g/cm3 0.984b 106

c 107 log Koct 2.02

d e 97 Pvp (30°C) torr 0.00267

f 3 97 Sw g/cm 9.9

aDensity; bCalculated for 30°C from reported value of 0.996 at 20°C using an estimated thermal expansivity of 0.0012 K-1; cOctanol/water partition coefficient; dVapor pressure; eExtrapolated from 25°C value of 0.00167 torr by modified Grange method. 94; fWater solubility

Table 6.2 Mean cumulative evaporation of 14C- DEET from human skin in vitro (n = 2-6/airflow)

Airflow Percent Applied Dose mL/min 0.25h 0.75h 2h 4h 8h 12h 24h

10 0.2 ± 0.1 0.6 ± 0.3 2.5 ± 3.0 7.6 ± 5.6 13.0 ± 7.1 14.3 ± 7.5 15.8 ± 7.7

20 0.1 ± 0.1 0.7 ± 0.3 4.1 ± 0.8 9.9 ± 2.3 19.5 ± 6.6 24.2 ± 8.2 29.1 ± 11.0*

30 0.2 ± 0.1 2.86 ± 1.8 8.6 ± 7.5 21.4 ± 10.6 33.8 ± 15.4 39.6 ± 19.5 45.2 ± 25.1

40 0.7 ± 0.6 4.3 ± 2.8 13.6 ± 5.2 28.5 ± 5.9 38.3 ± 7.1 41.0 ± 8.3 42.4 ± 8.8

50 1.2 ± 1.4 6.3 ± 5.5 19.1 ± 17.2 33.3 ± 25.6 46.1 ± 29.1 48.4 ± 28.6 51.1 ± 29.5

60 2.2 ± 0.3 10.3 ± 1.9 22.2 ± 0.4 36.0 ± 5.0 40.2 ± 2.8 42.2 ± 4.4 35.2 ± 4.5*

70 5.0 ± 1.6 16.5 ± 3.5 34.9 ± 3.7 47.8 ± 8.7 56.0 ± 12.7 57.7 ± 13.1 58.8 ± 13.3

80 9.0 ± 5.5 24.0 ± 8.5 40.3 ± 16.1 47.8 ± 19.1 49.7 ± 19.6 51.6 ± 19.1 52.3 ± 19.3

100 4.1 ± 3.6 12.4 ± 9.9 26.6 ± 16.8 31.4 ± 19.6 34.7 ± 19.9 36.5 ± 20.8 37.6 ± 21.1

* Experiment terminated at 25h

Table 6.3 Mean absorption of 14C- DEET from human skin in vitro (n = 2-6/airflow)

Percent Applied Dose Airflow mL/min 0.25h 0.75h 2h 4h 8h 12h 24h

10 0.5 ± 0.4 9.5 ± 5.4 32.9 ± 11.0 50.6 ± 14.5 63.8 ± 14.2 66.2 ± 13.9 68.6 ± 13.3

20 0.0 ± 0.0 0.4 ± 0.4 4.9 ± 3.6 16.0 ± 9.7 35.1 ±15.0 43.6 ± 15.1 51.5 ± 12.3*

30 0.1 ± 0.1 1.7 ± 1.8 7.5 ± 6.3 16.0 ± 11.4 26.1 ± 16.7 30.8 ± 19.3 36.4 ± 21.9

40 0.1 ± 0.1 1.4 ± 1.2 8.0 ± 4.6 18.8 ± 5.4 27.7 ± 5.5 30.3 ± 4.1 31.9 ± 3.5

50 0.0 ± 0.1 0.6 ± 0.7 4.5 ± 3.7 12.1 ± 9.3 20.2 ± 13.0 22.1 ± 13.3 24.6 ± 12.3

60 0.0 ± 0.0 0.7 ± 0.2 8.2 ± 1.4 18.4 ± 0.6 24.9 ± 3.9 25.6 ± 4.3 26.5 ± 4.5*

70 0.0 ± 0.2 0.5 ± 0.3 4.8 ± 1.7 11.6 ± 3.5 17.8 ± 5.0 18.9 ± 5.4 18.3 ± 5.3

80 0.3 ± 0.3 3.2 ± 2.1 9.1 ± 3.5 15.1 ± 1.9 17.9 ± 3.3 18.7 ± 3.6 19.4 ± 3.6

100 2.5 ± 4.8 9.3 ± 14.5 20.4 ± 19.1 29.2 ± 17.5 33.1 ± 16.1 34.2 ± 16.3 35.5 ± 16.1

* Experiment terminated at 25h

Table 6.4 Appearance of radioactivity in the vapor trap for airflow-dependence skin disposition study with 14C- DEET from human

skin in vitro (n = 2-6/airflow)

Airflow Percent Applied Dose mL/min 0-0.25h 0.25-0.75h 0.75-2h 2-4h 4-8h 8-12h 12-24h

10 0.21 0.35 1.93 5.13 5.42 1.30 1.50 20 0.11 0.57 3.42 5.75 9.69 4.69 4.90* 30 0.23 2.63 5.71 12.86 12.4 5.77 5.63 40 0.72 3.53 9.33 14.93 9.80 2.64 1.47 50 1.23 5.06 12.76 14.28 12.75 2.32 2.68 60 2.16 8.12 11.87 13.83 4.21 1.97 0.99* 70 5.02 11.45 18.47 12.84 8.24 1.72 1.07 80 8.99 15.05 16.30 7.44 1.94 1.84 0.76 100 4.09 8.35 14.13 4.80 3.32 1.85 1.04

* 25 hour sample

Table 6.5 Appearance of radioactivity in the receptor solution for airflow-dependence skin disposition study with 14C- DEET from

human skin in vitro (n = 2-6/airflow)

Airflow Percent Applied Dose mL/min 0-0.25h 0.25-0.75h 0.75-2h 2-4h 4-8h 8-12h 12-24h

10 0.52 9.00 23.39 17.69 13.19 2.36 2.44

20 0.00 0.39 4.55 11.10 19.01 8.53 7.89*

30 0.08 1.62 5.75 8.57 10.05 4.72 5.60

40 0.05 1.34 6.61 10.78 8.87 2.69 1.60

50 0.04 0.59 3.88 7.62 8.07 1.89 2.49

60 0.02 0.64 7.52 10.24 6.43 0.74 0.95*

70 0.01 0.49 4.30 6.82 5.55 1.73 0.98

80 0.32 2.87 5.88 6.04 2.83 0.79 0.71

100 2.49 6.83 11.08 8.76 3.96 1.11 1.22

* 25 hour sample

Table 6.6 Mass balance of 1% w/w solution of 14C-DEET from human skin in vitro as a percent of dose applied

A B

Airflow aEt-OH Receptor Total mL/min Cartridge Rinsing Parafilm bTissue Fluid Recovery 10 15.8 ± 3.8 1.2 ± 0.3 0.7 ± 0.5 8.0 ± 2.9 68.8 ± 6.6 94.5 ± 1.1 20 29.1 ± 5.5 1.4 ± 0.3 0.0 ± 0.0 7.4 ± 0.6 51.8 ± 6.0 89.7 ± 5.1 30 41.3 ± 10.5 1.9 ± 0.4 0.4 ± 0.2 11.9 ± 4.0 36.7 ± 9.7 92.2 ± 4.0 40 42.4 ± 3.6 0.6 ± 0.1 0.0 ± 0.0 5.9 ± 1.0 32.0 ± 1.4 81.0 ± 3.9 50 51.1 ± 14.8 1.4 ± 0.3 0.1 ± 0.0 13.5 ± 8.3 24.7 ± 6.1 90.8 ± 7.0 60 44.9 ± 3.0 0.9 ± 0.2 0.0 ± 0.0 5.8 ± 1.1 26.5 ± 3.1 78.1 ± 1.2 70 58.8 ± 6.7 0.8 ± 0.3 0.4 ± 0.3 11.3 ± 8.2 20.0 ± 2.6 92.0 ± 1.0 80 52.4 ± 7.3 0.7 ± 0.1 0.8 ± 0.5 13.4 ± 6.1 19.7 ± 1.3 87.0 ± 3.6 100 37.6 ± 10.6 0.5 ± 0.2 1.9 ± 1.8 13.6 ± 9.2 21.6 ± 6.6 89.2 ± 3.7

Average 88.3 ± 1.8

A = % Dose Evaporated, B = % Dose penetrated, a Ethanol rinsing of Evaporation trap, modified Franz® cell top and connecting tubes

(where applicable), b % dose recovered from skin and tape strip (where applicable) at end of experiment i.e. 24/25 hrs.

90 80 70 60 50 40 % Evaporation 30 20 10 0 20406080100120 90 80 70 60 50 40 %Absorption 30 20 10 0 20406080100120

Airflow v (mL/min)

Figure 6.2 Disposition of 14C – DEET from human skin in vitro as a percent of applied dose (Mean ± SE, n = 2-6 / airflow)

80 80

60 40 .hr

2 20

40 0 g/cm

μ 0 5 10 15 20 25 30 , trap J 20

02468

Tmid, hours

Figure 6.3 Effect of airflow on evaporation rate of 14C – DEET from human skin in vitro

(●) = 10 mL/min, (○) = 20 mL/min, (▼) = 30 mL/min, (∇) = 40 mL/min, (■) = 50 mL/min, (□) = 60 mL/min, (♦) = 70 mL/min, (◊) = 80 mL/min, (▲) = 100 mL/min

30 30 25 25 20 15 20 10 5 .hr 2 0 15

g/cm 0 5 10 15 20 25 30 μ , 10 abs J 5

02468 T , hours mid

Figure 6.4 Effect of airflow on absorption rate of 14C – DEET from human skin in vitro

(●) = 10 mL/min, (○) = 20 mL/min, (▼) = 30 mL/min, (∇) = 40 mL/min, (■) = 50 mL/min, (□) = 60 mL/min, (♦) = 70 mL/min, (◊) = 80 mL/min, (▲) = 100 mL/min

6.4 Conclusion

The general features of disposition of 127 μg/cm2 DEET on skin following solvent deposition from ethanol include (1) a 0.5 to 1 hour time lag for absorption with peak flux within three hours (Figure 6.4); (2) a delay in the evaporation rate profiles, with peak flux within four hours (Figure. 6.3) (3) evaporation of 53-97% of the applied dose from the skin surface at rates proportional to airflow over the skin (cf. Tables 6.2, 6.4); (4) a 2-fold increase in percentage absorption at 24 h with dose over the airflow range 10-100 mL/min (Table 6.6); (5) high amounts absorbed through the skin in both the lowest (10 mL/min) and highest (100 mL/min) relative to the other airflows; and (6) prolonged

“tails” to the absorption and evaporation profiles indicative of non exponential processes

(Figures 6.2 and 6.3).

CHAPTER SEVEN

ABSORPTION OF TECNAZENE THROUGH HUMAN SKIN IN

VITRO:

HS-SPME/GC-MS RESULTS

7.1 Introduction:

The three primary routes by which systemic exposure to a hazardous chemical can arise are through inhalation, ingestion and dermal absorption. Interest in the dermal absorption component has increased in the recent years. Most commonly, skin permeation has been characterized by studying the penetration of chemicals in vitro using human or animal skin 1. Cultured skin substitutes and model membranes with varying components

(isopropyl myristate to egg shells to synthetic silastic membranes) that potentially mimic the stratum corneum barrier have also been examined 108-111. Most in vitro studies employ radioactive compounds to characterize skin disposition. While radiochemical assays provide an attractive and simple methodology, their biggest shortcoming is the lack of specificity in the detection of the chemical entity, i.e., the parent compound cannot usually be distinguished from a metabolite or impurity. Therefore, in order to accurately ascertain the chemical form of the analyte, a more selective approach must be employed. High performance liquid chromatography (HPLC) techniques have been the most commonly used analytical tool especially from in vivo studies. However, it requires use of organic solvents and is not a method of choice to study volatile and semi volatile compounds. This report describes method development for the determination of tecnazene absorbed through human skin in vitro using headspace solid phase micro extraction (HS-SPME) followed by GC-MS analysis. Penetration results for two doses of tecnazene applied to skin in acetone solution are also presented.

Tecnazene is a dual acting pesticide. It is used as a plant growth regulator to suppress sprouts on potatoes during storage and as a fungicide to prevent dry rot on the tubers. Its structure and physicochemical properties are shown in Figure 7.1. Tecnazene is commercially available as granules and as dispersible dust formulations. Although it has been studied extensively as a fungicide and its toxicology and metabolism in animal models including rats, dogs and mice has been documented 112,113 its disposition following application to human skin has not been reported to date. A case report published in 1981 addresses the occurrence of contact dermatitis following topical exposure to dichloronitrobenzene which is found to be present in small quantities in tecnazene 114. Clearly, if the compound is used as a pesticide, the study of its disposition following topical exposure is warranted. This paper describes the fate of topically administered tecnazene on human skin in vitro.

SPME is a simple and convenient sample preparation technique that uses a thin fused silica rod to pre-concentrate analytes on to a polymeric fiber 115. It does so by enabling favorable partitioning of the compound onto an absorbing or adsorbing polymer coating.

The fiber is exposed to potential analytes by using one of the three possible modes of

SPME − immersion, headspace or membrane − protected mode 115. The pre - concentrated analytes are then thermally desorbed into a heated injection port attached to a chromatographic device for separation and quantification. Since its commercial introduction by Supelco in 1993, there have been numerous research papers exploring various applications of SPME. These include, but are not limited to, environmental studies to screen and detect compounds from soil, water, fruits, vegetables, wines; to

trace element determinations, to analyzing flavors and aromas, to forensic applications to more recent application to biological monitoring 116-120.

Several excellent review articles address SPME and its role in sample preparation and analysis 121,122. A recent article describes the use of SPME to study in vitro permeation of diethyl phthalate (DEP), a stabilizer in cosmetic products through hairless guinea pig skin

68. This study employed immersion SPME and 85 µm polyacrylate fibers. The present report demonstrates the applicability of HS-SPME combined with GC-MS to determine in vitro human skin permeation of tecnazene.

100

CAS No. = 117-18-0

MW = 261 Da

mp = 99 ºCa

a log K oct = 4.38

-3 a Sw (20 ºC) = 2.1 × 10 g/ L

ρ = 2.019 g/cmb

-4 c Pvp (25 ºC) = 3.95 × 10 torr

-4 a Pvp (25 ºC) = 2.97 × 10 torr

Figure 7.1 Structure and physical properties of tecnazene a EPIWIN Suite® (Modified Grain Method) b calculated using Imirzi and Perini’s method for density of solids, c Report on tecnazene presented at WHO in 1994

101

7.2 Methods

7.2.1 Materials

Tecnazene (2, 3, 5, 6-tetrachloronitropbenzene) in the form of PESTANAL®, CAS No

[117-18-0], was purchased from Reidel-de-Haen (Germany). The purity was estimated at

99.8% by GC analysis. Hexachlorobenzene, CAS No. [118-74-1], 99.99 % purity was purchased from Supelco (Bellefonte, PA). Polyoxyethylene − 20 oleyl ether, reagent grade sodium chloride and calcium-free Dulbecco’s phosphate saline were obtained from

Sigma-Aldrich (St. Louis, MO). Pesticide grade acetone and GC-Resolve methylene chloride were purchased from Fisher Scientific (Pittsburgh, PA). Split thickness (300

µm) human cadaver skin (back, thigh and abdomen) was procured from U.S. Tissue and

Cell (Cincinnati, OH). The skin was preserved in 10% glycerol and stored at −80 ºC until use.

7.2.2 Skin penetration study

The excised skin tissue was mounted on modified Franz® diffusion cells 51. The receptor solution was Dulbecco’s phosphate − buffered saline (pH 7.4) containing 1% (w/v)

Oleth-20 as a solubilizer and 0.02% (w/v) sodium azide w/v to inhibit microbial growth.

Receptor solutions were maintained at 37 ± 2 ºC on thermostatted heating and stirring modules, yielding a skin surface temperature of 30-32 ºC. The tissue was screened using

3 H2O permeability test described in Chapter 3, section 3.1.1. Samples with water

102

permeation greater than 1.2 µL/cm2 were discarded 53. The receptor solution was exchanged twice to ensure complete removal of residual radioactivity. The top and bottom chambers were then sealed with Parafilm®. The cells were randomized over treatments using previously established procedures 52. They were equilibrated overnight and dosed the following morning with 10 µL solution of either 1% (w/w) or 9% (w/w) tecnazene in acetone resulting in an average dose of 103 and 864 µg/cm2 respectively.

Each dose was studied under open and occluded conditions, yielding a total of four treatments. The occluded cells were covered with aluminum foil, which was wrapped securely around the top immediately after dosing. The cells and the thermostatted blocks were placed in a fume hood with the sash raised exactly at 18”. This condition simulates a wind velocity of 1.5 m/s based on a previous study 46. The entire receptor content (4.5 mL) was collected after 4, 8, 12, 24 and 48 hours post-dose. At this point the skin was dissected to obtain epidermis and dermis samples. The diffusion cells were rinsed with acetone and the wash was collected in a separate vial. Parafilm® and aluminum foil were also collected in individual vials. All samples were stored at −4ºC until analysis. The skin was obtained from 4 donors with 1-4 replicates for each treatment. The data reported here were averaged for each treatment for each donor and then averaged across donors.

7.2.3 Sample preparation and analysis

The samples were thawed at room temperature prior to analysis. The receptor fluid was transferred to a conical glass derivatization vial with a screw cap top.

103

Polytetrafluoroethylene-lined septa were used for sealing the vials. They were regularly replaced to avoid loss of analyte resulting from repeated needle piercing of the fiber assembly. The vial was placed in a thermostatted base maintained at 70 ºC and the fiber was exposed for a period of 15 minutes. Samples were analyzed using HS-SPME coupled with HP 5980 gas chromatograph and 5972 series mass selective detector. The response was noted and the amount of tecnazene in the 5mL was calculated using a standard curve that was constructed separately for each fiber on each day of analysis.

Due to the time consuming nature of the experiment, only single measurements were used to construct the calibration curve. The high precision of the analysis justified this decision, as demonstrated later.

Epidermis, dermis, Parafilm® and foil samples were extracted with methylene chloride and analyzed using a Varian 3800 ion trap gas chromatograph coupled with Saturn 2000

GC/MS/MS detector and 8200 Auto sampler. The wash samples were directly analyzed without any modification. All samples were analyzed in triplicate and the mean value was used for quantification of tecnazene. The injection volume for all samples was 2 µL.

7.2.4 SPME fibers

The SPME device was set up as per the manufacturer’s instructions (Supelco, Bellefonte,

PA, USA). A 100 µm polydimethylsiloxane (PDMS) fiber was used for receptor phase analysis using HS-SPME. It has been suggested that 100 µm PDMS fibers work well for semi volatile and volatile analytes 123,124. The fibers were pre-conditioned as per the

104

manufacture’s recommendations. A blank analysis was performed at the beginning of each day to ensure absence of carry over effect for all fibers used during the course of the analyses.

7.2.5 Optimization Study

The extraction conditions for the 100 µm PDMS fibers used for analyzing tecnazene in headspace mode were optimized by modifying the salt content, time of exposure and temperature of the receptor fluid. The salt concentration was varied from 0.2 to 0.8 g per

5 mL of the receptor fluid. The time of exposure was varied from 5 to 30 minutes. Once the conditions were set, they were used to analyze all subsequent samples and standards.

The results are shown in Figures 7.2a and 7.2b. Each determination was made once for each fiber. This was considered to be acceptable considering the low CV obtained during the precision study.

105

4.0 3.5 Fiber 1 3.0 Fiber 2 5 2.5 2.0 AUC x 10 1.5 1.0 0.5 0.10.20.30.40.50.60.70.80.9 Amount of NaCl, g

Figure 7.2a Optimization of exposure conditions using 100µm PDMS fiber at 70ºC for

15 minutes to a concentration of 20 ppb - Effect of salt addition

106

2.5 6

2.0 AUC x 10 0.5

0.0 0 5 10 15 20 25 30 35 40

Time of exposure, minutes

Figure 7.2b Optimization of exposure conditions using 100µm PDMS fiber with 0.6 g

NaCl to a concentration of 20 ppb - Effect of time

107

7.2.6 Precision Study

Each fiber was exposed to an identical concentration of analyte and internal standard.

The amount of salt, time of exposure and temperature of the apparatus were kept constant. The response from each fiber was measured for 5 consecutive exposures. For

Varian ion trap MSD, a standard solution containing a known amount of analyte and internal standard was placed into an auto sampler vial. Ten repeated injections of identical volume were made onto the Varian chromatograph and the response of the detector was measured. The results were calculated as relative standard deviation (RSD) and coefficient of variation (CV = 100 × RSD) have been reported in Table 7.1.

7.2.7 Recovery Study

A recovery study was performed using epidermis, dermis, Parafilm® and aluminum foil.

A known amount of tecnazene was spiked onto the samples and was immediately extracted with methylene chloride and analyzed. This was done in the same manner as a sample would be treated. % Recovery was calculated from a calibration curve. This was repeated three times for each matrix (n = 3), and each sample was analyzed in triplicate to obtain the % recovery values as shown in Figure 7.3.

108

Table 7.1 Figures of merit for tecnazene analysis. The data are reported in terms of coefficient of variation – CV. The limits of detection and quantification (LOD and LOQ) were determined by standard calculations using 3σ and 10LOD value.

HS-SPME 1 Liquid Samples 2

Precision, CV (n = 5) Precision, CV (n = 10)

R(t) T R(t) HCB Ratio AUC R(t) T R(t) HCB Ratio AUC

0.032 0.037 1.5 0.1 0.1 5.5

LOD LOQ Linearity LOD LOQ Linearity µg/mL µg/mL µg/mL µg/mL µg/mL µg/mL

0.08 - 0.10 0.8 - 1.0 0 – 5.2 0.03 - 1.2 1.6 - 11 0 - 100

1 HP 5890 GC-MS

2 Varian 3800 GC-MS

109

140 120

100

80 60 % Recovery 40 20

0 Epidermis Dermis Foil Parafilm

Figure 7.3 Recoveries of tecnazene from different matrices. (mean ± SE, n = 3 each matrix, each reading in triplicate)

110

7.2.8 Chromatographic conditions

7.2.8.1 Quadrupole GC-MS

A Hewlett Packard (HP) 5890 Series II Gas Chromatograph (GC) coupled to a HP 5972

Mass Selective Detector (MSD) was used for analyzing the HS-SPME standards and samples. The fiber was introduced into a specialized glass inlet for SPME and small sample volumes (Supelco, Part No. 26375-01). This was done by piercing the needle through Thermogreen® LB-2 septa − lowest bleed and predrilled septa designed for

SPME. The analytes were separated on a ZB-5 column (5%/95% methyl methyl- polysiloxane stationary phase, 30 m × 0.32 mm o.d. × 0.25 mm i.d) purchased from

Zebron, Bellefonte, PA. The column temperature was held at 70ºC for 2 minutes and then programmed at 70ºC min−1 to 140ºC (held for 2 min), then 20º min−1 to 230ºC (held for 3 min). The injection port was maintained at 250º C. Helium was used as a carrier gas, at a flow of 1.5 mL min−1. The MSD was operated in positive electron impact (EI) ionization mode. The ion source temperature was maintained at 175 ºC, and the energy for ionization was 70eV.

To further improve the sensitivity of the detection, the mass spectrometer was operated in a single ion-monitoring mode (SIM). The data obtained was processed using the standard

HP Chemstation Software. A maximum sensitivity auto tune procedure was performed daily. Hexachlorobenzene (HCB) was used as an internal standard for the quantitative analysis of tecnazene. A single ion characteristic to tecnazene (m/z 203) and HCB (m/z

111

284) was used for quantification. This was done based on standard reference database and published reports 125.

7.2.8.2 Ion trap GC-MS

Chromatographic analysis was carried out using a Varian 3800 capillary gas chromatograph equipped with 8200 Autosampler. Separation of analytes was carried out on a ZB-5 column coated with 5% phenyl and 95% dimethylpolysiloxane, 30 m × 0.32 mm o.d. × 0.25 mm i.d (Zebron, Bellefonte, PA). The column temperature was held at

100ºC and then programmed at 25º min–1 to 150ºC, then 40º min–1 to 230ºC, (held for 2 min). The injection port was maintained at 250ºC. Helium was used as a carrier gas, at a flow of 1.5 mL min–1. The mass selective detection was carried out using a Saturn 2000 series detector with an ion trap mass spectrometer operating in positive electron ionization mode. The mass spectra were obtained at 70eV. To further improve the sensitivity of the detection, SIM was employed. Three ions characteristic to the analyte

(m/z 201, 203, 261), and two ions for hexachlorobenzene (m/z 282, 284) were used for quantification based on the NIST reference spectrum for the most relative abundant ion and the published literature 125.

112

7.3 Results

Table 7.1 lists the figures of merit of the developed method. A very good coefficient of variation was obtained for samples analyzed by HS SPME and liquid sample analyses.

Figure 7.2a demonstrates optimization of exposure conditions by varying amount of salt in the receptor fluid. The response of each of two fibers was determined at different salt concentrations while keeping constant exposure time, temperature and concentration of analyte. The response was the highest at 0.6 g sodium chloride, which was used for all subsequent samples. Figure 7.2b shows the effect of exposure time on the signal. The signal was highest at 15 minutes exposure time, and it nicely complemented the chromatographic run time. Figure 7.2c depicts the effect of temperature on AUC.

The recovery of tecnazene from a variety of matrices used in the experiment is shown in

Figure 7.3. The percentage recoveries were 111.7 ± 9.3, 89.3 ± 8.8, 100.0 ± 1.5 and 51.6

± 19.7 from epidermis, dermis, foil and Para film® respectively (mean ± SE, n = 3).

Figure 7.4a is a representative reconstructed chromatogram from Varian Ion trap GC-MS.

A clean separation of the target analyte and internal standard was achieved. This resolution was seen from all the samples analyzed including those obtained from epidermis, dermis, wash and foil. Figure 7.4b is a typical reconstructed chromatogram obtained from the HS-SPME analysis procured using the SIM mode.

Figures 7.5a and 7.5b are typical calibration curves obtained from ion trap and quadrupole GC-MS respectively. The correlation coefficient, r2, for the daily standard

113

curves varied from 0.9881 to 0.9993 for liquid samples and from 0.8930 to 0.9997 for

HS-SPME samples over. The linearity of range for the calibration curves were 0.1 – 5.2 mg/L for HS-SPME and 0.5 – 100 mg/L.

Figure 7.6 shows the absorption rate of two doses of tecnazene through human cadaver skin under occluded and unoccluded conditions. With the exception of the low dose application under occlusion (which peaked at 8 hours), the absorption rate of tecnazene peaked at 12 hours post–dose. The absorption of tecnazene was incomplete especially for the higher dose. The low dose in occluded application appeared to reach a steady state flux.

114

10 Tecnazene 8

6 Hexachlorobenzene

4 3.0 2.5 2.0 m/z 201, 203, 261 1.5 1.0 0.5 Abundance x 10 0.0 2.0

1.5

1.0 m/z 282,284

0.5

0.0

3.5 4.0 4.5 5.0 5.5 6.0 6.5

Time, min

Figure 7.4a Representative chromatogram using Varian 3800 ion trap GC-MS in Single Ion Monitoring (SIM) mode.

115

4 HCB (m/z 284) 4 Tecnazene 3 (m/z 203)

1 Abundance x 10 0

7.5 8.0 8.5 9.0 9.5 10.0 10.5

R(t), minutes

Figure 7.4b Representative chromatogram obtained by desorption of tecnazene and hexachlorobenzene from 100 µm PDMS fiber on to HP 5980 GC-MS

116

6 y = 0.1495 x + 0.0604 R2 = 0.996

2 AUC Ratio

0 102030405060

Concentration, ppm

Figure 7.5a Typical calibration curve using Varian 3800 GC-MS equipped with 8200 auto sampler. (mean ± SE, each reading in triplicate)

117

50 y = 104.27 x – 1.6727 r2 = 0.9931 40

20 AUC Ratio 10

-10 0.0 0.1 0.2 0.3 0.4 0.5 0.6

Concentration, ppm

Figure 7.5b Calibration curve using 100 µm PDMS fiber for analysis of tecnazene using hexachlorobenzene as an internal standard.

118

7.4 Discussion

Understanding the absorption of compounds across skin is a crucial step in obtaining a complete picture of percutaneous absorption. This is true for both scenarios- either when the desired effect is an intentional increase in penetration for transdermal drugs and cosmetic applications or the exact opposite case where minimal absorption of hazardous compounds like pesticides, solvents, and allergens is the ultimate goal. Our group has been involved in development of a mathematical model to predict absorption and evaporation of compounds from human skin 46,47,49,50,126. The working model is a one dimensional model with four layers comprising of a vehicle layer, stratum corneum (SC), viable epidermis and dermis. It is based on physicochemical properties of the compound and transport parameters are derived from microscopically based components of stratum corneum and dermis. The model addresses the transport of solvents or solvent deposited compounds and the upper two layers donor solution and SC are described in detail

47 elsewhere . Briefly, for a permeant with initial concentration C, Csat is the maximum concentration in the upper layers of the SC. This also represents the solubility limit of the permeant within the SC. The permeant is assumed to be uniformly deposited in the upper SC to a depth hdep up to its solubility Csat. Msat is the saturation dose ≅ Csat × hdep.

In case of tecnazene absorption, the two doses are similar in their behavior despite difference in the doses. This can be easily explained by taking into consideration the Msat values, which according to the model is 0.07 -0.08 µg/cm2. Thus, with regards to the

119

saturation of the compound in the upper layers of SC, both doses are large doses and differ only in the duration of absorption.

As is evident from Figure 7.6, the model under predicts absorption of tecnazene for all four cases. The correction factors used to match the experimental curves are shown in

Table 7.2. The first correction was based on altering the permeability coefficient of SC -

Psc, and partition coefficient of SC – Ksc for tecnazene. The average correction factor of

3 – 4 was needed for both the parameters with the exception of the low dose occluded condition. It is noteworthy that, except for the low dose occluded treatment, absorption of tecnazene from all other treatments could be explained without modifying the diffusivity explicitly. For the low dose occluded treatment, the diffusivity had to be decreased in order to match the experimental results. The second correction was introduced by adjusting the evaporation mass transfer coefficient kevap. Typically, the values of kevap were reduced by one to three orders of magnitude to match up with the observed experimental data. This was done after observing that the model predicted the disposition of tecnazene to approach completion at a much faster rate. This modification can be justified as follows. The model incorporates the Peress correlation to calculate the mass transfer coefficient 74. However, this calculation tends to overestimate the evaporative loss. Also, in case of occlusive treatments, the presence of the foil is retarding the evaporation by providing a physical barrier and thus it is reasonable to expect that the evaporation rate of tecnazene for these cells would be slower.

120

These findings reinstate the fact that in order to have a clearer picture of absorption of lipophilic molecules, transport properties in lower skin tissues have to be addressed. This bears a significant impact on risk assessment calculations since under predicting absorption may lead to erroneous predictions with little contribution to existing knowledge. It must also be noted that addition of 1% surfactant in the receptor fluid is likely to increase the absorption of tecnazene across the in vitro set up. Modification of receptor fluid to facilitate absorption of very lipophilic molecules by addition of surfactants such as Volpo 20 and up to 50% ethanol has been reported in the past 28,127.

This is an added factor that the model calculation does not account for.

The high Ko/w of tecnazene indicates that the compound has high affinity for the lipids and subsequently may be bound to the stratum corneum lipids causing a slow release of the compound. The appreciable amount of tecnazene recovered from epidermis and dermis at the end of 48 hours supports this conjecture. The amount of tecnazene that penetrated into the skin was higher under occluded conditions. This is an expected outcome as has been previously reported for other non polar and lipophilic compounds such as steroids and parabens for topical delivery 128-130. The noticeable difference in the levels of tecnazene in both epidermis and dermis for low and high dose supports this outcome. The effect varied from a slight 1.5–fold increase in the epidermal concentration for the low dose occluded treatment to a more pronounced 34–fold increase in the dermal concentration for the high dose occluded treatment (Table 7.3).

121

The percentage of applied dose recovered from aluminum foil was 11% and 3% for low and high dose respectively. This is indicative of loss due to volatilization. The difference in the amount of tecnazene absorbed in presence and absence of an occlusive barrier is also apparent from the overall mass balance, which was substantially higher for the occluded treatments than for the unoccluded treatments (Table 7.3). A part of the reason may be that the foil prevents the volatilization of tecnazene from the surface of the skin as well as from the surface of the Franz® cell top and the junction of the top and the bottom chambers that are sealed with Parafilm®. The increased amounts of tecnazene collected from washes and Parafilm® for occluded conditions support this argument. The missing fraction may be a result of loss due to decomposition of the analyte at the HS-

SPME analysis temperature of 70°C. Also, there may be evaporative losses in this multi step sample collection and analysis technique. An alternative experimental design for trapping volatile fractions 45 or an innovative cell design which facilitates direct SPME monitoring of the headspace above the skin surface, followed by immediate thermal desorption, may be logical next steps.

122

7 30 6 25 5 20 4 15 3 10 2 1 5

2 0 0 -1 g/cm -5 μ 0 1020304050 0 1020304050

16 35

Amount, 14 30 12 25 10 20 8 15 6 10 4 5 2 0 0 -5 0 1020304050 0 1020304050

Hours, post dose

Figure 7.6 Absorption rate of tecnazene through human cadaver skin in vitro. (mean ± SE, 4 donors, n = 2-4 cells per donor). (○, ●) 103.4 μg/cm2 (U,▲) 873.8μg/cm2 The open and close symbols represent open and occluded conditions respectively. The solid line represents model prediction while the broken line represents corrected estimates.

123

Table 7.2 Correction factors used to match experimental absorption curves of tecnazene. The wind velocity u = 1.5 m/s for open treatments and 0.05 m/s for occluded treatments.

Correction factor a Treatment kevap Psc Ksc Dsc cm/h cm/h cm2/s 1% open 2E-06 2.5 2.5 1 1% occluded 2E-07 3 6 0.5 9 % open 1E-05 4.5 4.5 1 9 % occluded 1E-08 3 3 1 a calculated kevap 1% ,9% open = 1.51E-04, 1% occluded = 1 E-05, 9% occluded 1.06E-

124

Table 7.3 Distribution of topically applied tecnazene at the end of 48 hours (mean ± SE, 4 donors, n = 2-4 cells per donors )

Mean Dose Mean Dose

102.8 µg/cm2 864.3 µg/cm2

% Dose Open Occluded Open Occluded

Evaporated Foil N/A 10.6 ± 3.4 N/A 3.1 ± 0.6

In/on Skin Wash 0.4 ± 0.2 9.0 ± 3.2 9.2 ± 4.7 41.2 ± 7.9

Para 2.5 ± 0.8 17.6 ± 5.8 6.0 ± 2.2 7.2 ± 3.5

Epidermis 2.0 ± 0.5 3.2 ± 0.4 5.2 ± 3.2 12.8 ± 4.1

Dermis 1.0 ± 0.4 4.2 ± 0.7 0.6 ± 0.5 20.5 ± 3.4

Absorbed Rec. Fluid 5.3 ± 1.7 29.8 ± 4.7 1.8 ± 0.3 3.9 ± 0.6

Total 10.5 ± 1.3 71.1 ± 5.5 17.1 ± 5.2 82.5 ± 9.9

Missing 89.5 ± 1.3 28.9 ± 5.5 82.9 ± 5.2 17.5 ± 9.9

125

7.5 Conclusion

A GC-MS method to describe the absorption of tecnazene through human skin in has been described using HS-SPME. The method has good precision and accuracy and provides an alternative to routine radiochemical assays and has potential for applications in simple and complex mixture analysis. Absorption of tecnazene through human skin in vitro could be satisfactorily described using a working diffusion model. The model calculations were with in a factor of 3-4 of the observed experimental data. Possible explanations for the differences and modifications made to correct the model calculations are described. This study also addresses the need to study more lipophilic compounds in order to get a clearer picture of dermal absorption from a predictive model stand point.

126

CHAPTER EIGHT

DERMAL ABSORPTION OF MALATHION IN VITRO – GC-MS

RESULTS

127

8.1 Introduction:

The dermal route represents a significant pathway for penetration of compounds into the human body after intentional or accidental contact. Many pesticides have a high likelihood of penetrating the skin due to their physical and chemical properties and thereby pose a substantial health risk. Malathion is a broad spectrum organophosphorous pesticide that is commonly used in agricultural, industrial and household setting.

Percutaneous absorption of malathion has been studied extensively in vitro and in vivo in various animal models and human volunteers 33,104,131,132. Many of these studies involve tracking radioactive material from its application site.

The objective of this study was to study the disposition of topically applied malathion from human skin in vitro using GC-MS analysis. Solid phase micro extraction (SPME) is becoming a widely used screening technique for analyzing compounds of interest. The head space mode is a preferred mode for analyzing volatile and semi volatile compounds from different matrices 56. Several studies have reported employing SPME and HS-

SMPE to detect and quantify malathion from soil, natural waters, biological fluids, and from wine and fruit juices 56,133-135.

128

CAS No. 121-75-5

MW = 330 Da

mp = 2.8 ºCa

a log K oct = 2.36

-1 a Sw (22 ºC) = 1.45 × 10 g/ L

ρ = 1.23 g/cma

-6 a Pvp (25 ºC) = 3.38 × 10 torr

Figure 8.1 Structure and physical properties of malathion a EPIWIN Suite®

129

8.2 Experimental section

8.2.1 Materials:

Malathion [O, O-dimethyl-S-(1,2-dicarbethoxyethyl) phosphorodithioate] CAS No [125-

75-5] in the form of PESTANAL® was purchased from Reidel-de-Haen (Germany). The purity was estimated at 100 % by GC analysis. Fenitrothion, CAS No. [118-74-1] was purchased from Accustandards Inc., (New Haven, CT) with 100 % percent purity.

Reagent grade sodium chloride and calcium-free Dulbecco’s phosphate saline were obtained from Sigma-Aldrich (St. Louis, MO). Pesticide grade acetone and GC-Resolve methylene chloride were purchased from Fisher Scientific (Pittsburgh, PA). Split thickness (300 µm) human cadaver skin (back, thigh and abdomen) was procured from

U.S. Tissue and Cell (Cincinnati, OH). The skin was preserved in 10% glycerol and kept at −80 ºC until use.

8.2.2 Skin penetration study

The excised skin tissue was mounted on modified Franz® diffusion cells (0.79 cm2) 51.

The receptor solution was Dulbecco’s phosphate − buffered saline (pH 7.4) containing

0.02% (w/v) sodium azide w/v to inhibit microbial growth. Receptor solutions were maintained at 37 ± 2ºC on thermostatted heating and stirring modules, yielding a skin

3 surface temperature of 30-32ºC. The tissue was screened using H2O permeability test described in Chapter 3, Section 3.1.1. Samples with water permeation greater than 1.2

130

µL/cm2 were discarded 53. The receptor solution was exchanged twice to ensure complete removal of residual radioactivity. The top and bottom chambers were then sealed with Parafilm®. The cells were randomized over treatments using previously established procedures 52. They were equilibrated overnight and dosed the following morning with 10 µL of 1% (w/w) malathion solution in acetone resulting in an average dose of 100 µg/cm2. The dose was studied under open and occluded conditions, yielding two treatments. The occluded cells were covered with aluminum foil, which was wrapped securely around the top immediately after dosing. The cells and the thermostatted blocks were placed in a fume hood with the sash raised exactly at 18”.

This condition simulates a wind velocity of 1.5 m/s based on a previous study 46. The entire receptor content (4.5 mL) was collected at 4, 8, 12, 24, 48, 72, 96 and 120 hours post-dose. At this point the skin was dissected to obtain epidermis and dermis samples.

The diffusion cells were rinsed with acetone and the wash was collected in a separate vial. Parafilm® and aluminum foil were also collected in individual vials. All samples were stored at − 4ºC until analysis. The skin was obtained from 3 donors with 3 replicates for each treatment. The data reported here were averaged for each treatment for each donor and then averaged across donors.

8.2.3 Sample preparation and analysis

The samples were thawed at room temperature prior to analysis. The receptor fluid was transferred to a conical glass derivatization vial with a screw cap top.

Polytetrafluoroethylene-lined septa were used for sealing the vials. They were regularly

131

replaced to avoid loss of analyte resulting from repeated needle piercing of the fiber assembly. The vial was placed in a thermostatted base maintained at 70 ºC and the fiber was exposed for a period of 15 minutes. Samples were analyzed using HS-SPME coupled with an HP 5980 gas chromatograph and 5972 series mass selective detector.

The response was noted and the amount of malathion in the 5mL sample was calculated using a standard curve that was constructed separately for each fiber on each day of analysis. Due to the time-consuming nature of the experiment, only single measurements were used to construct the calibration curve. The high precision of the analysis justified this decision, as demonstrated later.

Epidermis, dermis, Parafilm® and foil samples were extracted with methylene chloride and analyzed using a Varian 3800 ion trap gas chromatograph coupled to a Saturn 2000

GC/MS/MS detector and 8200 Autosampler. The wash samples were directly analyzed without any modification. All samples were analyzed in triplicate and the mean value was used for quantification of tecnazene. The injection volume for all samples was 2 µL.

132

Table 8. Figures of merit for GC-MS analysis. The data are reported in terms of coefficient of variation – CV. The limits of detection and quantification (LOD and LOQ) were determined by standard calculations using 3σ and 10LOD values.

HS-SPME 1 Liquid Samples 2 Precision, CV (n = 5) Precision, CV (n = 7)

R(t) T R(t) HCB Ratio AUC R(t) M R(t) F Ratio AUC 0.032 0.037 1.5 0.04 0.03 3.2

LOD LOQ Linearity LOD LOQ Linearity µg/mL µg/mL µg/mL µg/mL µg/mL µg/mL 0.08 – 0.10 0.8 – 1.0 0 – 5.2 0.03 – 1.2 1.6 – 11 0 – 100

1 HP 5890 GC-MS

2 Varian 3800 GC-MS

133

8.2.4 SPME fibers

The SPME device was set up according to the manufacturer’s instructions (Supelco,

Bellefonte, PA, USA). A 100 µm PDMS fiber was chosen for analysis of malathion.

Numerous studies have been reported for detection of pesticides and a wide variety of fibers have been used. Malathion specifically has been detected and quantified using 100

µm PDMS, 85 µm PA fibers and 65 µm PDMS-DVB fibers 57,58,134. A 100 µm PDMS fiber was used for receptor phase analysis using HS-SPME. It has been suggested from literature that 100 µm PDMS fibers work well for semi volatile and volatile analytes

123,124. The fibers were pre-conditioned according to the manufacture’s recommendations. A blank analysis was performed at the beginning of each day to ensure lack of carry over effect for all fibers used during the course of the analyses.

8.2.5 Optimization Study:

The extraction conditions for the 100 µm PDMS fibers for analyzing malathion in headspace mode were optimized by modifying the salt content and time of exposure. The salt concentration was varied from 0.2 to 0.8 g per 5 mL of the receptor fluid. The exposure time was 20 minutes at a temperature of 70ºC. Although the range of reported analysis temperatures has varied from room temperature to 80ºC, high temperatures have been found to decrease the response for malathion that due to increased breakdown

56,133,134. Thus a moderate temperature of 70ºC was chosen for the analysis.

134

Optimization results are shown in Figures 8.2a, and 8.2b. Each determination was made once for each fiber. This was considered to be acceptable following the low CV obtained during precision study.

135

25 PA PDMS 20

AUC Ratio 10

0 0 5 10 15 20 25 Malathion concentration, ppm

Figure 8.2a Comparison of sensitivity of signal obtained from 85 µm PA and 100 µm

PDMS fiber at 70ºC for 20 minutes with 0.8 g NaCl.

136

3.0

2.5 Fiber 1 Fiber 2 2.0

1.5

AUC Ratio 1.0

0.5

0.0 0.0 0.2 0.4 0.6 0.8 1.0

Amount of NaCl, g

Figure 8.2b Effect of salt addition on GC-MS response to malathion: Optimization of exposure conditions using 100 µm PDMS fiber at 70ºC for 20 minutes at a malathion concentration of 0.8 ppm

137

8.2.6 Precision Study:

Each fiber was exposed to same concentration of analyte and the internal standard. The amount of salt, time of exposure and temperature were kept constant. The response from each fiber was measured for 5 consecutive exposures. For ion trap studies a standard solution containing a known concentration of analyte and internal standard was placed into an auto sampler vial. Ten repeated injections of same volume were made on to the chromatographic column and the response of the detector was measured. The results, calculated as relative standard deviation (RSD) and coefficient of variation (CV = 100 ×

RSD), have been reported in Table 8.1.

8.2.7 Recovery Study:

A recovery study was performed using epidermis, dermis, Parafilm® and aluminum foil.

A known amount of malathion was spiked onto the samples and was immediately extracted with methylene chloride and analyzed. This was done in the same manner as a sample would be treated. The results were calculated from a calibration curve. This was repeated three times for each matrix (n = 3) and each sample was analyzed in triplicate to obtain % recovery as shown in Figure 8.3.

138

120

100

% Recovery 40

0 Epidermis Dermis Foil Parafilm

Figure 8.3 Mean percent recoveries of malathion from different matrices,

(mean ± SE, n = 3)

139

8.2.8 Chromatographic conditions:

8.2.8.1 Quadrupole GC-MS

Chromatographic analysis was carried out using an HP 5890 capillary gas chromatograph. The analytes were separated on a ZB-5 column (5% phenyl and 95% dimethylpolysiloxane, 30 m × 0.32 mm o.d. × 0.25 mm i.d; Zebron, Bellefonte, PA). The column temperature was held at 70ºC for 2 minutes and then programmed at 70ºC min−1 to 140ºC (held for 2 min), then 10ºC min−1 to 230ºC, (held for 3 min). The injection port was maintained at 250ºC. The carrier gas was ultra high purity helium at a flow of 1.5 mL min−1. The mass selective detection was carried out using an HP 5972 series detector with a quadrupole mass spectrometer in a positive electron ionization mode. The mass spectra were obtained at 70eV. Fenitrothion was added as an internal standard for malathion. To further improve the sensitivity of the detection, single ion-monitoring mode (SIM) was employed. A single ion characteristic to malathion (m/z 173) and the internal standard (m/z 277) was used for quantification. This was done by referring to a standard reference library for the most relative abundant ion and published works

125,133,136,137.

8.2.8.2 Ion trap GC-MS

Chromatographic analysis was carried out using Varian 3800 capillary gas chromatograph, equipped with 8200 Auto sampler. Separation of analytes was carried on

140

a ZB-5 column coated with 5% phenyl and 95% dimethylpolysiloxane, 30 m × 0.32 mm o.d. × 0.25 mm i.d (Zebron, Bellefonte, PA). The column temperature was held at 150ºC and then programmed at 40ºC min−1 to 210ºC, then 10ºC min−1 to 230ºC, (held for 2 min).

The injection port was maintained at 250ºC. The carrier gas was helium, at a flow of 1.5 mL min−1. The mass selective detection was carried out using a Saturn 2000 series detector with an ion trap mass spectrometer in a positive electron ionization mode. The mass spectra were obtained at 70eV. To further improve the sensitivity of the detection,

SIM was employed. Three ions characteristic to malathion (m/z 93,127, 173) and three ions for fenitrothion (m/z 109, 260, 277) were used for quantification based on literature reports and standard reference spectra125.

141

8.3 Results:

Table 8.1 lists the figures of merit of the developed method. Figure 8.2a demonstrates better response obtained from the 100µm PDMS fiber as compared to the 85 µm PA fiber. Both have been used previously to quantify malathion. Salt addition has been a common technique used to improve the sensitivity of semi volatile and volatile analytes especially in the HS mode 138. Figure 8.2b shows the effect of exposure time on the signal. The response of malathion increased with addition of the salt to the buffered receptor solution. However, since some salt remained undissolved in the buffer at 1.2 g

(24% w/v in buffered saline), 0.8 g of NaCl was chosen. An exposure time of 20 minutes was chosen to complement the chromatographic run time.

The recovery of malathion from the matrices used in the experiment is shown in Figure

8.3. The percentage recoveries were 105.2 ± 2.0, 77.7 ± 4.9, 90.0 ± 4.1 and 64.6 ± 5.0 from epidermis, dermis, foil and Para film® respectively (mean ± SE, n = 3). Figure 8.4a is a representative reconstructed chromatogram from Varian Ion trap GC-MS. Clean separation of the target analyte and internal standard was achieved. This resolution was observed for all the samples analyzed including those obtained from epidermis, dermis, wash and foil. Figure 8.4b is a typical reconstructed chromatogram obtained from the

HS-SPME analysis procured using the SIM mode. Representative calibration curves from Varian ion trap GC-MS and HP GC-MS are shown in Figures 8.5a and 8.5b respectively. The correlation coefficients for HS-SPME analysis ranged from 0.9343 to

0.9999 and 0.9593 to 0.9997 for liquid samples.

142

M 11 F (m/z 93, 127,173)

4 (m/z 109, 260, 277)

9 Abundance x 10 8

7 11.3 11.4 11.5 11.6 11.7 11.8 11.9 12.0 12.1

R(t), minutes

Figure 8.4a Representative chromatogram obtained from Varian 3800 ion trap GC-MS

143

14 M 12 (m/z 173) F

3 10 (m/z 277) 8 6 4

Abundance x 10 2 0

15.8 16.0 16.2 16.4 16.6 16.8

R(t), minutes

Figure 8.4b Representative chromatogram obtained by desorption of malathion(M) and

fenitrothion (F) from 100μm PDMS fiber on to HP 5980 GC-MS

144

50 y = 0.9955 + 0.5963 2 40 R = 0.9972

20 Ratio AUC 10

0 102030405060

Concentration of malathion, ppm

Figure 8.5a Typical Calibration curve using Varian 3800 GC-MS equipped with 8200

auto sampler. (Mean ± SE, each reading in triplicate)

145

1.8 1.6 y = 0.0017 x - 0.0356 1.4 R2 = 0.9963 1.2 1.0 0.8 0.6 AUC Ratio AUC Ratio 0.4 0.2 0.0 -0.2 0 200 400 600 800 1000 1200

Concentration of malathion, ppm

Figure 8.5b Typical Calibration curve using 100μm PDMS fiber for analysis of

malathion using fenitrothion as an internal standard

146

A parallel study was performed using 14C-malathion and is reported separately (Chapter

9). Absorption rates of malathion through human skin from both GC analysis and radiochemical analysis are shown in Figure 8.6. The mass balances are shown in Table

8.2. Figure 8.6 presents the absorptive flux of malathion through human skin. For the

GC study, in presence and absence of an occlusive barrier, the maximum absorption rate of malathion occurred 72 hours post dose. This suggests that the absorption of the compound is a slow process, stretching over days. These observations are mirrored in the radiolabel study for the unoccluded treatment. The results show that percutaneous absorption of malathion increased under occlusion. This is an expected result; similar findings have been reported with moderately lipophilic compounds such as hydrocortisone 37,39. Correspondingly, Figure 8.7 shows a marked difference (7 fold increase) in dermal levels of malathion the presence of an occlusive barrier at the end of

120 hours.

147

1.4 1.2 1.0 0.8 0.6 0.4

1 0.2 − .hr

2 0.0 −

0 20406080100120 10

pplied dose.cm 8 %Α 6

0 20406080100120

T , hours mid

Figure 8.6 Absorption of malathion through human skin in vitro. The open and closed symbols represent open and occluded treatments. (∇) (▼) GC study (○) (●) 14C- malathion study

148

30 Open Occluded 25

10 % Applied Dose

0 Epidermis Dermis

Figure 8.7 Tissue levels of malathion 120 hours post-dose in the presence and absence of an occlusive barrier. The dose was 100 μg/cm2.

149

Table 8.2 Distribution of topically applied 1% malathion solution (10μL/ 0.79 cm2) at the end of 120 hours expressed as percent of dose applied (mean ± SE, 3 donors, n = 3-4 cells per donor)

GC Analysis RC Analysis

% Applied Dose Open Occluded Open Occluded

Evaporated Foil 0.6 ± 0.2 0.4 ± 0.1

In/on Skin Wash 8.1 ± 2.3 12.5 ± 2.7 10.2 ± 1.9 6.6 ± 1.5

Potentially Epidermis 24.3 ± 8.5 8.5 ± 3.5 20.6 ± 4.1 14.5 ± 4.9 absorbed Dermis 0.1 ± 0.1 7.7 ± 3.1 5.3 ± 1.2 22.1 ± 2.8

Absorbed Rec. Fluid 9.9 ± 1.9 13.4 ± 2.4 22.3 ± 4.1 32.1 ± 3.8

Total 42.4 ± 7.9 42.8 ± 7.9 53.0 ± 2.6 74.0 ± 2.8

Missing 57.6 ± 7.9 57.2 ± 7.9 47.0 ± 2.6 26.0 ± 2.8

150

8.4 Discussion:

The overall recovery of malathion was much higher for with the radiolabel study as compared to the GC study. The amount of 14C-malathion absorbed through the skin and into the receptor fluid was about 2.4 fold higher for both open and occluded treatments than that obtained from the GC analysis. This can be explained on the basis of the analysis techniques. The receptor fluid samples for the GC study were analyzed at 70°C using HS-SPME. Hence, loss of analyte can occur due to thermal degradation. Also, higher amounts of malathion were recovered from the dermis (5 – fold) and epidermis (2

– 3 fold) in the radiolabel study. This could be a result of loss of analyte due to evaporation during the extraction of the tissue using methylene chloride in the GC analysis. The general features of malathion absorption in terms of time to peak flux and shape of the flux plots matched with those obtained from radiochemical assay.

The objective of the study was to explore an alternative analysis for determining percutaneous absorption of malathion through human skin. This was achieved successfully. A method was developed using HS-SPME. The method has good precision and accuracy with potential for application for future analyses for other compounds and possibly mixtures of compounds. However, from a risk assessment point of view, this type of analysis would yield lower values for absorption and thereby underestimate a candidate compound’s potential to cross the human skin. Thus, factors such as thermal degradation and evaporative loss associated with sample preparation must be taken in to account while determining percutaneous absorption by the HS-SPME/ GC-MS method.

151

8.5 Conclusion

Absorption of malathion thorough human skin in vitro was studied using standard radiochemical assay and an analytical method. An alternative method to determine the absorption of malathion through human skin in vitro was developed using GC-MS. A novel application of HS-SPME was established. The differences in the total recovery of malathion from two techniques was explained in terms of losses due to volatilization and thermal degradation. The method has good precision and accuracy with potential for application for routine analysis of other compounds and possibly mixtures of compounds.

152

CHAPTER NINE

ABSORPTION OF MALATHION FROM HUMAN SKIN AND

SILICONE MEMBRANES IN VITRO – RADIOCHEMICAL

RESULTS

153

9.1 Introduction:

Absorption through skin plays a crucial role in risk assessment of chemicals encountered in the workplace. Pesticides constitute a group of hazardous compounds that has a high likelihood of dermal exposure and thereby a risk of toxicity associated with the dermal route. Malathion is a broad spectrum organophosphate insecticide. Organophosphates mainly exert their action by inhibiting cholinesterase, thus interfering with cholinergic synapses. Since it was first introduced in 1956, malathion has become the most widely used pesticide for commercial, agricultural, industrial and homeowner applications. The current estimate for annual malathion usage in the United States is 15 million pounds from over 200 registered products 139. This is due in large part to its relative low toxicity to mammals. Malathion is also commercially available as a 0.5% solution Ovide® to treat head lice in children (Medics-Phoenix, AR). The physical and chemical properties of malathion are listed in Figure 9.1. The moderate lipophilic character, Ko/w 2.36, combined with its moderate molecular weight and low melting point, makes it an ideal candidate for penetrating the skin.

Percutaneous absorption of malathion has been studied extensively in vitro and in vivo using various animal models and human volunteers.33,104,131,132. Many of these studies involve tracking radioactive material from its application site. The objective of this study was three-fold. The first aim was to characterize the absorption of topically applied malathion through human skin in vitro at different doses that were above, near and below the saturation dose (Msat), i.e., the dose required to saturate the upper layers of the stratum

154

corneum (SC) 47. It is postulated that the upper 10-20% of the SC is the fractional deposition depth for solvent deposited compounds 47 140. The data thus obtained were used to test a working diffusion model 47. The second objective of the study was to compare the effects of occlusion for each dose. Finally, a parallel study was performed with silicone membrane with identical doses to compare the disposition of malathion through skin with that through an inert membrane.

155

CAS No. = 121-75-5

MW = 330 Da

mp = 2.8ºCa

a log Koct = 2.36

-1 a Sw (20ºC) = 1.45 × 10 g/ L

ρ = 1.23 g/cm3

-6 a Pvp (25ºC) = 3.38 × 10 torr

Figure 9.1 Structure and physical properties of malathion a EPIWIN Suite®

156

9.2 Experimental section

9.2.1 Materials:

Malathion 2, 3-14C (5.9 mCi/mmol; 0.05 mCi) was purchased from Sigma Chemicals

(St.Louis, MO). The radiochemical purity was stated by the manufacturer to be 97.1%.

All experiments were done within a month of receiving the radiochemical. Unlabelled malathion [O,O-dimethyl-S-(1,2-dicarbethoxyethyl)phosphorodithioate], CAS No [125-

75-5], in the form of PESTANAL® was purchased from Reidel-de-Haen (Germany). The purity was estimated at 100% by GC. Calcium-free Dulbecco’s phosphate saline was obtained from Sigma-Aldrich (St. Louis, MO). Pesticide grade acetone was purchased from Fisher Scientific (Pittsburgh, PA). Split thickness (300 µm) human cadaver skin

(back, thigh and abdomen) was procured from U.S. Tissue and Cell (Cincinnati, OH).

The skin was preserved in 10% glycerol and kept at −80 ºC until use.

9.2.2 Skin penetration study:

Excised split-thickness human cadaver skin (300 µm) was mounted on modified Franz® diffusion cells. The receptor solution was Dulbecco’s phosphate buffered saline (pH 7.4) with 0.02% sodium azide w/v added as an antimicrobial. Cells were maintained at 37 ±

2ºC on thermostatted heating and stirring modules. The tissue was screened using permeability test described in Chapter 3, Section 3.3.1. Tissue samples with permeability greater than 1.2 µL/cm2 were discarded 53. The receptor solution was changed twice to

157

ensure complete removal of residual radioactivity. The top and the bottom chambers were sealed with Parafilm®. The cells were randomized using previously established procedures 52. The cells were equilibrated overnight and dosed the subsequent morning with 10 µL of either 1% (w/w), 0.005% or 0.001% solution of malathion in acetone. The lower concentration solutions were made by serial dilution of the 1% w/w solution. This resulted in an average dose of 101 µg/cm2, 0.5 µg/cm2 and 0.1 µg/cm2 respectively. Each solution was separately spiked with equal amount of radioactive malathion. Each dose was studied under occluded and non occluded conditions, yielding a total of six treatments. Occlusion was achieved by covering the top of the donor chamber with aluminum foil, which was wrapped securely around the top immediately after dosing.

The skin was obtained from 3 donors with 3-4 replicates for each treatment. The data reported here were averaged for each treatment for each donor and then averaged across donors to obtain the reported results.

A parallel absorption study was performed using silicone membrane, yielding a total of nine treatments (the occluded condition was not studied with the silicone membrane).

The cells and the thermostatted blocks were placed in a fume hood with the sash raised to

18”. This condition simulates a wind velocity of 1.5 m/s according to a previous study 46.

The entire receptor content (4.5 mL) was collected at 4, 8, 12, 24, 48, 72, 96 and 120 hours post-dose. At this point the skin was dissected to obtain epidermis and dermis samples, which were dissolved separately in 2 mL of Soluene®. The silicone membrane was rinsed with acetone and collected in a separate vial. The diffusion cells were rinsed with acetone and the wash was collected in a separate vial. Parafilm® and aluminum foil

158

were also collected in individual vials. All samples were analyzed by liquid scintillation counting (LSC). The silicone membrane study was repeated thrice with three samples in each run.

159

9.3 Results

9.3.1 Skin disposition study

The mean absorption of radioactivity associated with malathion as a function of time and dose is listed in Table 9.1. The cumulative absorption of malathion and instantaneous flux are shown in Figure 9.2 and 9.3 respectively. The maximum absorption rate was achieved between 6-8 hours post dose for all treatments except for 101 µg/cm2 unoccluded cells, for which the maximum flux was reached at 24 hours post dose. The rate of absorption for the latter case remained fairly constant for the 5 day duration of the experiment. At a given dose, the cumulative absorption of malathion from occluded treatments was higher than that for the corresponding non occluded treatments. This was an expected outcome, since it has been shown that occlusion enhances the absorption of moderately lipophilic compounds due to increased skin permeability, possible alteration in the lipid phase organization 130,141,142 and by reducing evaporation.

The mass balance for the human skin study is presented in Table 9.2. The cumulative absorption of malathion ranged from 20 to 25% in non occluded cells and from 32 to

46% in occluded cells. An appreciable amount of radioactivity was recovered from both epidermis and dermis at the end of 120 hours, suggesting a long residence time of malathion in the skin layers. A marked difference in the dermal levels of malathion was observed in the occluded cells for all doses − 4 fold to 16 fold increase from the highest to the lowest dose.

160

Table 9.1 Appearance of radioactivity in the receptor solution: Skin disposition study with 14C-malathion through human skin in vitro

(Mean of 3 donors, n = 3-4/donor).

Dose Percent of dose

µg/cm2 0-1 hr 1-2 hr 2-4 hr 4-6 hr 6-8 hr 8-12 hr 12-24 hr 24-48 hr 48-72 hr 72-96 hr 96-120 hr

a 101 0.01 0.03 0.15 0.24 0.29 0.64 2.19 4.60 5.25 4.29 3.33

b 101 0.06 0.15 0.56 0.70 0.77 1.63 4.68 7.26 6.35 4.24 3.45

a 0.5 0.01 0.11 0.58 0.96 1.16 2.36 6.33 6.81 3.41 1.76 0.71

b 0.5 0.15 0.77 2.90 3.43 3.25 5.26 10.27 10.40 5.54 2.85 1.82

a 0.1 0.01 0.09 0.53 0.93 1.07 2.09 5.30 5.62 2.52 1.00 0.51

b 0.1 0.08 0.49 1.86 2.27 2.28 3.73 8.05 8.70 4.57 2.76 1.86

a = Open cells, b = Occluded cells

161

Table 9.2 Mass balance of disposition of radioactivity associated with 14C-malathion from human skin in vitro, expressed as the

percent of dose applied (Mean ± SE, 3 donors, n = 3-4/donor)

Mean Dose Mean Dose Mean Dose

101 µg/cm2 0.5 µg/cm2 0.1 µg/cm2

% Dose Open Occluded Open Occluded Open Occluded

Evaporated Foil N/A 0.4 ± 0.1 N/A 0.7 ± 0.2 N/A 0.5 ± 0.2

Potentially Wash 10.2 ± 1.9 6.6 ± 1.5 1.9 ± 1.0 2.6 ± 0.8 0.8 ± 0.4 2.4 ± 0.8

Absorbed Para 0.4 ± 0.2 7.7 ± 3.4 − 8.6 ± 4.1 0.1 ± 0.0 4.6 ± 0.7

Epidermis 20.6 ± 4.1 14.5 ± 4.9 5.5 ± 1.2 7.0 ± 1.4 4.6 ± 0.7 7.6 ± 1.9

Absorbed Dermis 5.3 ± 1.2 22.1 ± 2.8 3.9 ± 1.4 21.4 ± 2.0 1.3 ± 0.5 20.7 ± 3.0

Rec. Fluid 22.3 ± 4.1 32.1 ± 3.8 25.0 ± 2.4 46.4 ± 4.6 20.1 ± 2.7 34.5 ± 3.5

Total Recovery 53.0 ± 2.6 74.0 ± 2.8 33.6 ± 2.6 80.6 ± 1.0 25.2 ± 2.8 76.1 ± 1.4

Missing 47.0 ± 2.6 26.0 ± 2.8 66.4 ± 2.6 19.4 ± 1.0 74.8 ± 2.8 23.9 ± 1.4

162

40 50 (a) (b) 40 (c) 30 40 30 20 30 20 20 10 10 10

% Applied dose 0 0 0

0 20406080100120 0 20406080100120 0 20 40 60 80 100 120

Hours post-dose

Figure 9.2: Cumulative absorption of 14C-malathion through human skin in vitro in the presence and absence of an occlusive barrier.

(Mean ± SE of 3 donors, n = 3-4/donor) Open and closed circles represent unoccluded and occluded cells respectively. (a) 101

μg/cm2, (b) 0.5 μg/cm2, (c) 0.1 μg/cm2.

163

0.5 2e-4 6e-6 (a) (b) (c) 0.4 5e-6 1e-4 .hr 2 0.3 4e-6 1e-4

g/cm 3e-6

μ 0.2 5e-5 2e-6 0.1 Flux, 1e-6 0.0 0 0

0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20406080100120 T ,hours mid

Figure 9.3 Absorptive flux of radioactivity associated with 14C-malathion in the presence and absence of an occlusive barrier (Mean ±

SE of 3 donors, n =3-4/donor). The open and filled circles represent non occluded and occluded cells respectively. The lines are a guide to the eye. (a) 101 μg/cm2, (b) 0.5 μg/cm2, (c) 0.1 μg/cm2.

164

A modest increase of epidermal levels (1.3 to 1.7 fold) was also observed for the near saturation and below saturation doses while the above saturation dose presented itself with a decrease in epidermal levels.

Figures 9.4 a, b and c show the comparison of the experimental data with the calculations from the working skin diffusion model 143. This model calculates the absorption and evaporation of a compound from human skin based on its physico-chemical properties.

As can be seen from the graphs, the model under predicts absorption for the highest dose and over predicts for the lower doses. In case of the former, by increasing the partitioning of the compound between the SC and water by a factor or 3.5-4, the model calculations match the experimental observations. We defer possible explanations to the discussion.

9.3.2 Silicone membrane study:

Table 9.3 shows the appearance of radioactivity associated with 14C malathion through silicone membrane for all three doses. The associated mass balance is shown in Table

9.4. Recovery of radioactivity (86-91%) was higher than that in the skin disposition study. The missing radiolabel is likely to have evaporated over the period of the study.

Since the missing fraction is much smaller than those missing from the skin disposition study it seems possible that the added losses in the skin study may be due to metabolic processes. The cumulative absorption and absorptive flux of 14C malathion are shown in

165

Figures 9.5.and 9.6 respectively. Disposition of malathion from all three doses is nearly complete by 120 hours post dose as seen from the figures.

166

25 30 (a) (b) 20 25 2 2 20 15 g/cm g/cm 15 μ μ 10 10 5 5 Amount, Amount, 0 0

-5 -5 0 20406080100120140 0 20 40 60 80 100 120 140

Hours, post-dose Hours, post-dose

Figure 9.4a Cumulative absorption of 14C-malathion (101 μg/cm2) through human cadaver skin in vitro. (Mean ± SE of 3 donors, n =

3-4 cells per donor). The open circles (a) and close circles (b) represent open and occluded conditions respectively. The solid line represents model calculation while the dashed line represents corrected estimates.

167

0.14 0.5 (a) (b) 0.12 0.4 2 2 0.10 0.08 0.3 g/cm g/cm μ μ 0.06 0.2 0.04 0.1 Amount, 0.02 Amount, 0.00 0.0 -0.02 -0.1 0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140

Hours, post-dose Hours, post-dose

Figure 9.4b Cumulative absorption of 14C-malathion (0.5 μg/cm2) through human cadaver skin in vitro. (Mean ± SE of 3 donors, n =

3-4 cells per donor). The open (a) and close (b) symbols represent open and occluded cells respectively. The solid line represents model calculation while the dashed line represents corrected estimates.

168

30 100 (a) (b) 25 2

2 80 20 g/cm g/cm 60 μ μ , , -3

15 -3 40 10 20 5 Amount x 10 0 Amount x 10 0

-5 -20 0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140

Hours, post-dose

Figure 9.4c Cumulative absorption of 14C-malathion (0.1 μg/cm2) through human cadaver skin in vitro. (Mean ± SE of 3 donors, n =

169

Table 9.3 Physical properties and input parameters of malathion

Parameter Units Value MW a g/mol 330 ρ b g/cm3 1.23 c log K oct 2.36 d −6e P vp torr 7.14 × 10 f 3 −4 S w g/cm 1.45 × 10 RT (32°C) Latm/mol 25.03 u g m/s 1.5 h, 0.05 I h j μm 13.4 h, 43.4 I ƒ k − 0.1 l h I K sc/w − 7.91 , 21.22 a Molecular weight b Density c Octanol water partition coefficient d vapor pressure e Extrapolated using the EPIWIN ® program f water solubility g wind velocity h unoccluded condition I occluded condition j Stratum corneum thickness k fractional deposition depth l Stratum corneum/ water partition coefficient calculated from 47.

170

Table 9.4 Calculated model parameters for malathion absorption through human skin in vitro

Amount M sat C sat D sc k evap ,cm/h

μg/cm2 μg/cm2 g/cm3 cm2/s Unmodified Modified

Unoccluded

− − − − 101 0.411 3.08 × 10 3 1.90 × 10 11 a 1.26 × 10 7 1.26 × 10 7

− − − − 0.5 0.411 3.08 × 10 3 4.76 × 10 12 1.26 × 10 7 1.26 × 10 9

− − − − 0.1 0.411 3.08 × 10 3 4.76 × 10 12 1.26 × 10 7 1.26 × 10 9

Occluded

− − − − 101 0.498 1.15 × 10 3 2.32 × 10 10 b 8.9 × 10 9 8.9 × 10 11

− − − − 0.5 0.498 1.15 × 10 3 1.74 × 10 10 c 8.9 × 10 9 1.5 × 10 7

− − − − 0.1 0.498 1.15 × 10 3 1.16 × 10 10 8.9 × 10 9 8.9 × 10 7

M sat = Saturation dose, C sat = Saturation concentration, D sc0, D ed, D de are diffusivities in SC, viable epidermis and dermis respectively, K ed/w = partition coefficient between viable epidermis and water, K de/w = partition coefficient between dermis and water. a 4 – fold higher, b 2 – fold higher, c 1.5 – fold higher than initial model calculations.

171

Table 9.5 Appearance of radioactivity in the receptor solution: Skin disposition study with 14C-malathion through silicone membrane

in vitro (Mean of 3 runs, n = 3/run).

Dose Percent of dose

µg/cm2 0-1 hr 1-2 hr 2-4 hr 4-6 hr 6-8 hr 8-12 hr 12-24 hr 24-48 hr 48-72 hr 72-96 hr 96-120 hr

101 7.29 8.71 12.07 11.60 9.64 10.77 8.68 5.05 3.00 2.10 1.6

0.5 16.98 17.32 13.23 9.67 6.19 4.88 3.93 2.69 1.47 0.97 0.70

0.1 9.76 10.68 12.93 9.66 6.53 6.62 5.85 4.15 3.18 2.19 1.8

172

Table 9.6 Mass balance of disposition of 14C-malathion from silicone membrane in vitro

as a percent of dose applied (Mean ± SE, 3 runs, n = 3/run)

14C malathion through silicone membrane

%Applied Dose 101 µg/cm2 0.5 µg/cm2 0.1 µg/cm2

Evaporated Foil N/A N/A N/A

Wash 0.2 ± 0.1 0.2 ± 0.1 0.2 ± 0.1 In/on membrane Parafilm 0.1 ± 0.0 0.6 ± 0.3 0.1 ± 0.0

Membrane 11.9 ± 2.2 10.5 ± 2.5 14.6 ± 2.0

Absorbed Receptor Fluid 74.9 ± 4.4 76.8 ± 5.4 67.3 ± 2.5

Total 90.5 ± 2.2 88.1 ± 3.6 85.7 ± 2.5

Missing 9.5 ± 2.2 11.9 ± 3.6 14.3 ± 2.5

173

100 100 100

80 80 80

60 60 60

40 40 40

% Applied Dose 20 20 2 μ 2 20 101 μg/cm 0.5 g/cm 0.1 μg/cm2 0 0 0 0 20 40 60 80 100 120 0 20 40 60 80 100 120 0 20406080100120

Hours, post dose

Figure 9.5: Cumulative absorption of 14C-malathion through silicone membrane in vitro (Mean ± SE, 3 runs, n = 3/run)

174

8 0.010 2.5e-4 0.008 2.0e-4 .hr 6 2

2 2 101 μg/cm 0.5 μg/cm μ 2 0.006 1.5e-4 0.1 g/cm 4 g/cm

μ 0.004 1.0e-4 2 0.002 5.0e-5 Flux 0 0.000 0.0

0 20406080100120 0 20406080100120 0 20406080100120

010203040 0 10203040 0 10203040

Time, hours post dose

Figure 9.6 Absorptive flux of radioactivity associated with 14C-malathion through silicone membrane in vitro.

(Mean ± SE, 3 runs, n = 3/run)

175

9.4 Discussion

Malathion absorption through skin appears to be slow process, stretching over days. This can be deduced from continuous appearance of radioactivity following topical application of 14C-malathion for at least 120 hours and from the appreciable tissue levels from all three doses (Tables 9.1, 9.2). These results also support the theory that SC acts like a

“sink” rather than an impermeable barrier for pesticides 144. The recovery of radioactivity is higher for the occluded cases, as has been reported previously 130. However, about 20-

26% of the dose remains unaccounted for in the occluded case. It is possible that some evaporative loss occurred after applying the dose, but before covering the cell top with the foil. It is noteworthy that the mass balance is the best from the silicone membrane study (86-90% recovery). This suggests that in addition to evaporative losses, either metabolic breakdown or degradation of the malathion is occurring in the presence of skin.

This has been seen with chemically related compounds such as chemical warfare agents.

Fredriksson et al demonstrated hydrolysis of soman, sarin and tabun in guinea pig homogenates by phosphorylphosphate enzymes145. Blank et al reported hydrolysis of sarin in excised human skin 146. Thus, it is plausible that malathion breaks down during the process of absorption through human skin in vitro, resulting in incomplete recovery at the end of the experiment.

The effect of occlusion on the permeation of moderately lipophilic compounds like hydrocortisone esters, parabens and pesticides has been well established 128-130,144. Of the nine pesticides studied, all were reported to have an increased absorption under

176

occlusion. Malathion incurred the greatest increase – 9.2 fold 144. A modest but a distinct increase in the amount of malathion in the receptor fluid was seen in the occluded treatments in the present study (Table 9.2).

The value of predicting absorption rates of chemicals traversing through skin in dermal risk assessment is clear. Despite considerable efforts in this area, accurate prediction of transport prior to experimental evaluation remains elusive, especially for small or transient skin exposures. Towards this objective, a skin diffusion model for volatile compounds has been developed in our research lab. Transport parameters in the model are calculated from physico-chemical properties of the permeant (ρ, MW, Koct/w, Sw, Pvp)

47. The model framework derives from the structure and microscopic transport properties of the SC, epidermal and dermal layers 48,69,72,147. Briefly, the model is described in terms of four layers – a solvent deposited layer on top of SC, SC, viable epidermis and dermis.

Each layer has certain a partition coefficient, a diffusivity and a thickness. For the present discussion we will describe the parameters used for absorption of malathion.

Consider a permeant deposited on the surface of the skin with a finite dose M0. Csat is the maximum concentration in the upper layers of the SC which also represents the solubility limit of the permeant within the SC. The compound is assumed to be uniformly deposited in the upper SC to a depth hdep up to its solubility Csat. Msat is the saturation dose ≅ Csat × hdep. Other important parameters are kevap, (the mass transfer coefficient);

Dsc (diffusivity of the permeant with respect to SC) and Ksc (partition coefficient of the permeant with respect to SC). The input parameters and model calculations are shown in

Table 9.4.

177

In the present study, the highest dose was much above the saturation dose and the other two doses were much below the saturation dose. As can be seen from the Figures 9.4a,

9.4b and 9.4c, the model under-predicts for the highest dose and over predicts for the lower two doses. The only exception was the calculation for 0.5 μg/cm2 dose in open conditions which matched the experimental data.

In order to match the model calculations to the experimental data, the diffusivity of malathion was altered by changing the permeability and the partitioning of malathion by a factor of 1.5 – 4. A second correction was made by modifying the mass transfer coefficient of malathion from the surface of the skin. In all occluded treatments, the kevap was reduced by two orders of magnitude to match the data (Table 9.4). It is plausible that the rate of evaporation was slower for occluded cells, since the presence of the foil physically retarded the evaporation.

It has been reported that approximately 50% of malathion was bound to all layers of human skin when partitioned against an aqueous buffer 131. This result was observed over a 100-fold range of doses. The investigators also reported the preference of malathion partitioning in to plasma proteins over aqueous buffer. The current model incorporates reversible protein binding in the lower skin layers, but does not include irreversible binding to tissue. This could be one possible explanation for over-predicting the absorption of the lower doses. For the highest dose treatments, the impact of binding was not large enough to affect the amount of malathion being absorbed, unlike the pronounced effect seen on the lower doses.

178

9.5 Conclusion

This study describes the absorption of malathion through human skin and silastic membrane. 20-25 % applied dose was recovered in the receptor fluid for unoccluded treatments. This was increased to 32 – 46 % for occluded cells. If the amounts recovered from the washing, epidermis and dermis were included, up to 66% of applied dose was found to be potentially absorbed from unoccluded cells which increased to up to 80% of applied dose under occlusion. The absorption of malathion was described in terms of a working diffusion model and plausible explanations were offered for discrepancies between experimental observations and model calculations. The current model calculations were within the acceptable ranges of prediction and in the worse case scenario, within a factor of four from the observed experimental data.

179

CHAPTER TEN

SUMMARY AND FUTURE WORK

180

10.1 Summary

The objective of this project was to obtain experimental data to characterize the disposition of pesticides and fragrance compounds applied to skin surface under different conditions – at different airflows and in presence and absence of occlusive barrier. The main findings from this work are listed below:

o Absorption and evaporation rates of small doses of volatile compounds, pesticides

(DEET, tecnazene, malathion) and a fragrance compound (benzyl alcohol) from

human skin in vitro were obtained. Absorption of DEET decreased systematically

with increase in the airflow. Benzyl alcohol demonstrated a dose dependent

behavior for absorption through skin. Absorption of tecnazene and malathion

increased in presence of occlusion.

o Analytical methods using a novel application of HS-SPME to detect and quantify

the disposition of tecnazene and malathion from human skin in vitro using GC-

MS were developed and validated. Although a method is in place for detection

and quantification, results from malathion absorption study show that for the

purpose of risk assessment, radiochemical assays provide better estimates. This is

mainly due to the added loss of analyte incurred due to thermal degradation and

loss due to volatilization. Also, radiochemical assays are simpler and less time

consuming.

181

o All data were described in terms of a working computational model by comparing

the experimental values with the predicted values. The model described the

absorption characteristics adequately.

o A key correction to improve the viable tissue model to better explain the behavior

of highly lipophilic compounds in lower layers of skin was incorporated. The

results obtained from tecnazene absorption study provided the impetus to

incorporate such a correction. This correction was made by incorporating

ionization of weak acids and bases, reversible binding to extravascular proteins

like albumin and partitioning of the compounds into lipids associated with the

dermis. As a result of this, the model estimates were significantly improved.

Currently, the model estimates are within a factor of 3 – 4 of the experimental

data (malathion and tecnazene).

o Evaporation estimates were revised.

182

10.2 Suggestions for Future work

• To conduct partitioning studies of highly lipophilic compounds to get a better

understanding of the distribution of such compounds in the lower layers of the

skin i.e., viable epidermis and dermis.

• To apply the developed methodology to analyze more compounds and to extend

the method to the quantification of mixtures.

• To create a modified cell design that enables continuous monitoring of the

headspace above the Franz® cell using sample preparation tools like SPME

coupled with GC-MS. This mode of direct sampling will help in minimizing the

loss of analyte due to evaporation and give a complete picture of the disposition

of the compound.

183

CHAPTER ELEVEN

REFERENCES

184

1. Scheuplein RJ, Blank IH 1971. Permeability of the skin. Physiological Reviews

51:702-747.

2. . http://www.epa.gov/pesticides/about/types.htm. ed.

3. . http://www.epa.gov/oppbead1/pestsales/. ed.

4. Van der Merwe D, Brooks JD, Gehring R, Baynes RE, Monteiro-Riviere NA,

Riviere JE 2006. A physiologically based pharmacokinetic model of prganphosphate dermal absorption. Toxicological Sciences 89(1):188-204.

5. Ross JH, Driver JH, Shelley A, Maibach HJ 2005. Dermal absorption of 2,4-D: a review of species differences. Regulatory Toxicology and Pharmacology 41(1):82-91.

6. Tomlik-Scharte D, Lazar A, Meins J, Bastian B, Ihrig M, Wachall B, Jetter A,

Tantcheva-Poor I, Mahrle G, Fuhr U 2005. Dermal absorption of permethrin following topical administration. European Journal of Clinical Pharmacology 61(5-6):399-404.

7. Wester RC, Quan D, Maibach HI, Wester RM. 2005. Percutaneous absorption of hazardous chemicals from fabric into and through human skin. In Bronaugh RL, Maibach

HJ, editors. Percutaneous Absorption: Drugs--Cosmetics--Mechanisms--Methodology,

Fourth ed.: Marcel and Dekker. p 303-310.

8. Wester RC, Maibach HI. 2005. Skin contamination and absorption of chemicals from water and soil. In Bronaugh RL, Maibach HJ, editors. Percutaneous Absorption:

Drugs--Cosmetics--Mechanisms--Methodology, Fourth ed.: Marcel and Dekker. p 107-

121.

9. Nielsen JB, Nielsen F, Sorensen JA 2004. In vitro percutaneous penetration of five pesticides- effects of molecular weight and solubility characteristics. Annals of

Occupational Hygiene 48(8):697-705.

185

10. Tos-Luty S, Tokarska-Rodak M, Latuszynska J, Przebirowska D 2002.

Distribution of dermally absorbed 14C-DDT in the organs of Wistar rats. Annals of

Agricultural and Environmental Medicine 9(2):215-223.

11. Zendzian RP 2003. Pesticide Residue on/in the washed Skin and its Potential

Contribution to Dermal Toxicity. Journal of Applied Toxicology 23:121-136.

12. Baynes RE, Brooks JD, Mumtaz M, Riviere JE 2002. Effect of chemical interactions in pentachlorophenol mixtures on skin and membrane transport.

Toxicological Sciences 69(2):295-305.

13. Brand RM, Mueller C 2002. Transdermal Penetration of Atrazine, Alachlor,

Trifluralin: Effect of Formulation. Toxicological Sciences 68:18-23.

14. Dary CC, Blancato JN, Saleh MA 2001. Chemomorphic Analysis of Malathion in

Skin Layers of the Rat: Implications for the Use of DermatopharmacokineticTape

Stripping in Exposure Assessment to Pesticides. Regulatory Toxicology and

Pharmacology 34:234-248.

15. Zendzian RP 2000. Dermal Absorption of Pesticides in the Rat. AIHAJ 61:473-

483.

16. Van der Sandt JJM, Meuling WJM, Elliot GR, Cnubben NHP, Hakkert BC 2000.

Comparitive in vitro-in vivo percutanrous absorption of the pesticide propoxur.

Toxicological Sciences 58(1):15-22.

17. Nielsen JB, Nielsen F 2000. Dermal in vitro penetration of methiocarb, paclobutrazol and primicarb. Occupational and Environmental Medicine 57(11):734-

737.

186

18. Griffin P, Mason H, Heywood K, Cocker J 1999. Oral and dermal absorption of chlorpyrifos: a human volunteer study. Occupational and Environmental Medicine

56(1):10-13.

19. Moody RP, Nadeau B 1994. In vitro dermal absorption of pesticides : IV. In vivo and in vitro comparison with the organophosphate insecticide diazinon in rat, guinea pig, pig, human and tissue-cultured skin. Toxicology in vitro 8(6):1213-1218.

20. Moody RP, Nadeau B, Chu I 1994. In vitro dermal absorption of pesticides : V. In vivo and in vitro comparison of the herbicide 2,4-dichlorophenoxyacetic acid in rat, guinea pig, pig, human and tissue-cultured skin. Toxicology in vitro 8(6):1219-1224.

21. Moody RP, Nadeau B, Chu I 1994. In vitro dermal absorption of pesticides : VI.

In vivo and in vitro comparison of the organochlorine insecticide DDT in rat, guinea pig, pig, human and tissue-cultured skin. Toxicology in vitro 8(6):1225-1232.

22. Wester RC, Bucks D, Maibach HJ 1994. Human in vivo percutaneous absorption of pyrethrin and piperonyl butoxide. Food and Chemical Toxicology 32(1):51-53.

23. Wester RC, Sedik L, Melendres J, Logan F, Maibach HI, Russel I 1993.

Percutaneous Absorption of Diazinon in Humans. Food and Chemical Toxicology

31(8):5659-5572.

24. Ademola JI, Wester RC, Maibach HJ 1993. Absorption and metabolism of 2- chloro-2,6-diethyl-N-(butoxymethyl)acetanilide (butachlor) in human skin in vitro.

Toxicology and applied pharmacology 121(1):78-86.

25. Moody RP 1993. In vitro dermal absorption of pesticides: a cross species comparison including testskin. Journal of Toxicology -Cutaneous and Ocular Toxicology

12(2).

187

26. Ramsey JD, Woollen BH, Auton TR, Batter PL, Leeser JF 1992.

Pharmacokinetics of Fluazifop-butyl in Human Volunteers II: Dermal Dosing. Human and Experimental Toxicology 11:247-254.

27. Wester R, Maibach HJ, Melendres J, Sedik L, Knaak JB, Wang R 1992. In vivo and in vitro percutaneous absorption and skin evaporation of isofenphos in man.

Fundamental and Applied Toxicology 19:521-526.

28. Scott RC, Ramsey JD 1987. Comparison of the in vivo and in vitro percutaneous absorption of lipophilic molecule (Cypermethrin, a Pyrethroid Insecticide). J Invest

Dermatol 89(2):142-146.

29. Shah PV, Fisher HL, Sumler MR, Monroe RJ, Chernoff N, Hall LL 1987.

Comparison of the penetration of 14 pesticides through the skin of young and adult rats.

Journal of Toxicology and Environmental Health 21(3):353-366.

30. Moody RP, Franklin C, A. 1987. Percutaneous absorption of the insecticides fenitrothion and aminocarb in rats and monkeys. Journal of Toxicology and

Environmental Health 20(1-2):209-218.

31. Grissom REJ, Brownie C, Guthrie FE 1985. Dermal absorption of pesticides in mice. Pesticide Biochemistry and Physiology 24(1):119-123.

32. van Hooidonk C, Ceulen BI, Kienhuis H, Bock J. International Congress on

Toxicology, Brussels, Belgium July 6-11, 1980 1980, pp 646-646.

33. Feldmann RJ, Maibach HJ 1974. Percutaneous Penetration of Some Pesticides and Herbicides in Man. Toxicology and applied pharmacology 28:126-132.

34. . http://iccvam.niehs.nih.gov/docs/guidelines/validate.pdf. ed.

188

35. Robinson PJ. 1998. Prediction: Simple risk models and overview of dermal risk assessment. In Roberts MS, Walters KA, editors. Dermal Absorption and Toxicity

Assessment, 91 ed., New York: Marcel Dekker. p 203-244.

36. Reifenrath WJ, Spencer TS. 1985. Evaporation and penetration from skin. In

Percutaneous Absorption, ed., New York: Marcel Dekker. p 313-334.

37. Feldmann RJ, Maibach HI 1967. Regional variations in percutaneous penetration of 14C-cortisol in man. Journal of Investigative Dermatology 48:181-183.

38. Scott RC, Corrigan MA, Smith F, Mason H 1991. The Influence of Skin Structure on Permeability: An Intersite and Interspecies Comparison with Hydrophilic Penetrants.

Journal of Investigative Dermatology 96(6):921-925.

39. Wester RC, Maibach HI 1976. Relationship of topical dose and percutaneous absorption in rhesus monkey and man. Journal of Investigative Dermatology 67:518-

520.

40. Potts RO, Guy RH 1992. Predicting skin permeability. Pharmaceutical Research

9(5):663-669.

41. Flynn GL. 1985. Mechanism of percutaneous absorption from physicochemical evidence. In Bronaugh RL, Maibach HI, editors. Percutaneous Absorption, ed., New

York: Marcel Dekker. p 27-51.

42. Mitragotri S 2003. Modeling skin permeability to hydrophilic and hydrophobic solutes based on four permeation pathways. Journal of Controlled Release 86(1):69-92.

43. Geinoz S, Guy RH, Testa B, Carrupt P-A 2004. Quantitative Structure-

Permeation Relationships (QSPeRs) to Predict Skin Permeation: A Critical Evaluation.

Pharmaceutical Research 21(1):83-92.

189

44. Saiyasombati P, Kasting GB 2001. A physicochemical properties-based model for estimating evaporation and absorption rates of perfumes from skin. International Journal of Cosmetic Science 23(1):49-58.

45. Saiyasombati P, Kasting GB 2003. Two-stage kinetic analysis of fragrance evaporation and absorption from skin. International Journal of Cosmetic Science

25(5):235-243.

46. Santhanam A, Miller MA, Kasting GB 2005. Absorption and evaporation of N,N- diethyl-m-toluamide from human skin in vitro. Toxicology and applied pharmacology

204(1):81-90.

47. Kasting GB, Miller M 2006. Kinetics of finite dose absorption through skin 2.

Volatile compounds. J Pharm Sci 95:268-280.

48. Nitsche JM, Wang T-F, Kasting GB 2006. A two-phase analysis of solute partitioning into the stratum corneum. J Pharm Sci 95(3):649-666.

49. Wang T-F, Kasting GB, Nitsche JM 2006. A multiphase microscopic model for stratum corneum permeability. I. Formulation, solution and illustrative results for representative compounds. J Pharm Sci 95:620-648.

50. Wang T-F, Kasting GB, Nitsche JM 2006 (submitted). A multiphase microscopic model for stratum corneum permeability. II. Estimation of physicochemical parameters and application to a large permeability database. J Pharm Sci.

51. Merritt EW, Cooper ER 1984. Diffusion apparatus for skin penetration. Journal of

Controlled Release 1:161-162.

190

52. Kasting GB, Filloon TG, Francis WR, Meredith MP 1994. Improving the sensitivity of in vitro skin penetration experiments. Pharmaceutical Research

11(12):1747-1754.

53. Franz TJ, Lehman PA 1990. The use of water permeability as a means of validation of skin integritity in in vitro percutaneous absorption studies. Journal of

Investigative Dermatology 94:525.

54. Belardi RP, Pawliszyn JB 1989. The application of chemically modified fused silica fibers in the extraction of organics from water matrix samples and their rapid transfer to capillary columns. Water Pollution Research Journal of Canada 24(1):179-

191.

55. Wercinski EbSAS. 1999. Solid Phase Microextraction: A Practical Guide. ed.,

New York: Marcel Dekker, Inc.

56. Lambropoulou DA, Sakkas VA, Albanis TA 2002. Validation of an SPME method, using PDMS, PA, PDMS-DVB, and CW-DVB SPME fiber coatings, for analysis of organophosphorous insecticides in natural waters. Anal Bioanal Chem

374:932-941.

57. Tomkins BA, Ilgner RH 2002. Determination of atrazine and four organophosphorus pesticides in ground water using solid phase microextraction (SPME) followed by gas chromatography with selected-ion monitoring. 972(2):183-194.

58. López FJ, Pitarch E, Egea S, Beltran J, Hernández F 2001. Gas chromatographic determination of organochlorine and organophosphorus pesticides in human fluids using solid phase microextraction. Analytica Chimica Acta 433(2):217-226.

191

59. Lambropoulou DA, Psillakis E, Albanis TA, Kalogerakis N 2004. Single-drop microextraction for the analysis of organophosphorous insecticides in water. Analytica

Chimica Acta 516(1-2):205-211.

60. Baltussen E, Sandra P, David F, Cramers C 1999. Stir bar sorptive extraction

(SBSE), a novel extraction technique for aqueous samples: Theory and principles.

Journal of Microcolumn Separations 11(10):737-747.

61. Campillo N, Penalver R, Aguinaga N, Hernandez-Cordoba M 2006. Solid-phase microextraction and gas chromatography with atomic emission detection for multiresidue determination of pesticides in honey. Analytica Chimica Acta 562(1):9-15.

62. Cardeal Z, Paes C 2006. Analysis of organophosphorus pesticides in whole milk by solid phase microextraction gas chromatography method. Journal of Environmental

Science and Health, Part B: Pesticides, Food Contaminants, and Agricultural Wastes

41(4):369-375.

63. You J, Lydy MJ 2006. Determination of pyrethroid, organophosphate and organochlorine pesticides in water by headspace solid-phase microextraction.

International Journal of Environmental Analytical Chemistry 86(6):381-389.

64. Torres Padron ME, Sosa Ferrera Z, Santana Rodriguez JJ 2006. Optimisation of solid-phase microextraction coupled to HPLC-UV for the determination of organochlorine pesticides and their metabolites in environmental liquid samples.

Analytical and Bioanalytical Chemistry 386(2):332-340.

65. Aulakh JS, Malik AK, Kaur V, Schmitt-Kopplin P 2005. A review on solid phase micro extraction-high performance liquid chromatography (SPME-HPLC) analysis of pesticides. Critical Reviews in Analytical Chemistry 35(1):71-85.

192

66. Beltran J, Lopez FJ, Hernandez F 2000. Solid-phase microextraction in pesticide residue analysis. Journal of Chromatography, A 885(1-2):389-404.

67. Curran AM, Rabin SI, Prada PA, Furton KG 2005. Comparison of the Volatile

Organic Compounds Present in Human Odor Using Spme-GC/MS. Journal of Chemical

Ecology 31(7):1607-1619.

68. Frasch FH, Barbero AM 2005. Application of solid phase micro extraction to in vitro skin permeation experiments: example using diethyl phthalate. Toxicology in vitro

19(2):253-259.

69. Basketter DA, Pease C, Kasting GB, Kimber I, Casati S, Cronin MTD, Diembeck

W, Gerberick GF, Hadgraft J, Hartung T, Marty JP, Nikolaidis E, Patlewicz GY, Roberts

D, Roggen E, Rovida C, van de Sandt H. in press. Skin sensitisation and epidermal disposition. The relevance of epidermal disposition for sensitisation hazard identification/risk assessment. . ATLA, ed.

70. Brenner H, Edwards DA. 1993. Macrotransport Processes. ed., Boston:

Butterworth-Heinemann.

71. Wang T-F KGaNJ 2007. A multiphase microscopic model for stratum corneum permeability. II. Estimation of physicochemical parameters and application to a large permeability database. J Pharm Sci In Press.

72. Nitsche JM, Wang T-F, Kasting GB 2006. A two-phase analysis of solute partitioning into the stratum corneum. J Pharm Sci 95:649-666.

73. Kretsos K, Zamora-Estrada G, Miller M, Kasting GB 2006 (submitted).

Diffusivity and clearance of skin permeants in the dermis. Int J Pharm.

74. Peress J 2003. Estimate evaporative losses from spills. Chem Eng Prog:32-34.

193

75. Mookherjee BD, Patel SM, Trenkle RW, Wilson RA 1998. A novel technology to study the emission of fragrance from the skin. Perfumer & Flavorist 23:1-11.

76. Gerberick GF, Robinson MK 2000. A skin sensitization risk assessment approach for evaluation of new ingredients and products. American Journal of Contact Dermatitis

11(2):65-73.

77. Basketter DA 1998. Skin sensitization: risk assessment. International Journal of

Cosmetic Science 20:141-150.

78. Basketter DA, Cookman G, Gerberick GF, Hamaide N, Potokar M 1997. Skin sensitization thresholds: determination in predictive models. Food and Chemical

Toxicology 35:417-425.

79. Kasting GB, Miller MA 2005. Kinetics of finite dose absorption through skin 2.

Volatile compounds. Journal of Pharmaceutical Sciences submitted.

80. Saiyasombati P, Kasting GB 2003. Disposition of benzyl alcohol following topical application to human skin in vitro. Journal of Pharmaceutical Sciences

92(10):2128-2139.

81. Santhanam A, Miller MA, Kasting GB 2005. Absorption and evaporation of

N,N-diethyl-m-toluamide (DEET) from human skin in vitro. Toxicology and Applied

Pharmacology 204:81-90.

82. Kasting GB, Filloon TG, Francis WR, Meredith MP 1994. Improving the sensitivity of in vitro skin penetration experiments. Pharmaceutical Research

11(12):1747-1754.

194

83. Santhanam A. 2004. Mathematical model to predict the skin disposition of DEET and other volatile compounds. College of Pharmacy, ed., Cincinnati, OH: University of

Cincinnati.

84. Bevington PR. 1969. Data Reduction and Error Analysis for the Physical

Sciences. ed., New York: McGraw-Hill.

85. Bhatt V, Santhanum A, Miller MA, Kasting GB. American Association of

Pharmaceutical Scientists National Meeting, Baltimore, MD, 2004.

86. Saiyasombati P, Kasting GB 2004. Evaporation of benzyl alcohol from human skin in vivo. Journal of Pharmaceutical Sciences 93(2):515-520.

87. Anissimov YG, Roberts MS 2001. Diffusion modeling of percutaneous absorption kinetics: 2. Finite vehicle volume and solvent deposited solids. Journal of

Pharmaceutical Sciences 90:504-520.

88. Saiyasombati P, Kasting GB 2003. Two-stage kinetic analysis of fragrance evaporation and absorption from skin. International Journal of Cosmetic Science 25:235-

243.

89. Vuilleumier C, Flament I, Sauvegrain P 1995. Headspace analysis study of evaporation rate of perfume ingredients applied onto skin. International Journal of

Cosmetic Science 17:61-76.

90. Peress J 2003. Estimate evaporative losses from spills. Chemical Engineering

Progress (April):32-34.

91. Johnson ME, Blankschtein D, Langer R 1997. Evaluation of solute permeation through the stratum corneum: lateral bilayer diffusion as the primary transport mechanism. Journal of Pharmaceutical Sciences 86(10):1162-1172.

195

92. Budavari S editor 1996. The Merck Index. 12 ed., Whitehouse Station: Merck &

Co.

93. Hansch C, Leo A. 1999. MEDCHEM database and CLOGP. 2.0 ed.: BioByte,

Inc.

94. Meylan W. 1999. MPBPWIN. 1.40 ed., North Syracuse: Syracuse Research

Corporation.

95. Yalkowski SH. 1993. AQUASOL database. 6 ed., Tucson: University of Arizona.

96. . http://www.epa.gov/pesticides/factsheets/chemicals/deet.htm. ed.

97. Qui H, Jun HW, McCall JW 1998. Pharmacokinetics, formulation and safety of insect repellent N,N-diethyl-3-methylbenzamide (deet): A review. . Journal of American

Mosquito Control Association 14:12-27.

98. . http://www.epa.gov/oppsrrd1/REDs/0002red.pdf. ed.

99. Schoenig; GP, Thomas. OG. 2001. DEET. In Krieger R, editor Handbook of

Pesticide Toxicology, 2nd ed.: Academic Press. p 1439-1459.

100. Hayden C, G.J.;, Roberts MS, Benson H, A.E. 1997. Systemic absorption of sunscreen after topical application. Lancet 350(19):863 – 864.

101. Baynes RE, Halling KB, Riviere JE 1997. The influence of diethyl-m-toluamide

(DEET) on percutaneous absorption of permethrin and carbaryl. Toxicology and Applied

Pharmacology 144(2):332-339.

102. Reifenrath WG, Hawkins GS, Kurtz MS 1989. Evaporation and skin penetration characteristics of mosquito repellant formulations. Journal of the American Mosquito

Control Association 5(1):45-51.

196

103. Proniuk S, Liederer BM, Dixon SE, Rein JA, Kallen MA, Blanchard J 2002.

Topical formulation studies with DEET (N,N-diethyl-3-methylbenzamide) and cyclodextrins. . Journal of Pharmaceutical Sciences 91(1):101-110.

104. Hawkins GH, Jr., Reifenrath WG 1984. Development of an in Vitro Model for

Determining the Fate of Chemicals Applied to Skin. Fundamental and Applied

Toxicology 4:S133-S144.

105. Stinecipher J, Shah J 1997. Percutaneous permeation of N,N-diethyl-m-toluamide

(DEET) from commercial mosquito repellants and the effect of solvent. Journal of

Toxicology and Environmental Health 52(2):119-135.

106. Santhanam A. 2004. Mathematical Model to Predict the Skin Disposition of Deet and Other Volatile Compounds Pharmaceutical Sciences, ed., Cincinnati: University of

Cincinnati. p 80.

107. Hansch C, Leo, A., . 1999. MEDCHEM database and CLOGP program, Vers.

2.0.0. . ed.: BioByte, Inc., Claremont, CA.

108. Brain KR, Walters KA, Watkinson AC. 1998. Investigation of skin permeation in vitro. In Roberts MS, Walters KA, editors. Dermal Absorption and Toxicity Assessment, ed., New York: Marcel Dekker, Inc. p 161-189.

109. De Jager M, Groenink W, Van der Spek J, Janmaat C, Gooris G, Ponec M,

Bouwstra J 2006. Preparation and characterization of a stratum corneum substitute for in vitro percutaneous penetration studies. Biochimica et Biophysica Acta, Biomembranes

1758(5):636-644.

197

110. Washitake M, Takashima Y, Tanaka S, Anmo T, Tanaka I 1980. Studies on drug release from ointment. Part I. Drug permeation through egg shell membranes. Chemical

& Pharmaceutical Bulletin 28(10):2855-2861.

111. Houk J, Guy R 1988. Membrane models for skin permeation studies. Chemical

Reviews 88(3):455-471.

112. 1994. Pesticide residues in food : 1994 Evaluations Part II Toxicology. ed.:

FAO/WHO.

113. Lappin GJ, Pritchard D, Moore RB, Laird WJD 1996. Metabolism of 2,3,5,6- tetrachloronitrobenzene (tecnazene) in rat. Xenobiotica 26(1):65-77.

114. Cotterill JA 1981. Contact dermatitis following exposure to tetrachloronitrobenzene. Contact Dermatitis 7(6):353.

115. Pawliszyn J. 1997. Solid Phase Microextraction: Theory and Practice. ed.: Wiley-

VCH, Inc. p 247.

116. Sergi D, Bayona JM 2006. Trace Element Determination of Combining Solid-

Phase Microextraction Hyphenated to Elemental and Molecular Detection Techniques.

Journal of Chromatographic Science 44(7):458-471.

117. Flavio Belliardo CB, Chiara Cordero, Erica Liberto, Patrizia Rubiolo and Barbara

Sgorbini 2006. Headspace-Solid Phase Microextraction in the Ananlysis of the Volatile

Fraction and Medicinal Plants. Journal of Chromatographic Science 44(416-429).

118. Halasz A, Hawari J 2006. SPME in Environmental Analysis: Biotransformation

Pathways. Journal of Chromatographic Science 44(7):379-386.

119. Jelen HH 2006. Solid-Phase Microextraction in the Analysis of Food Taints and

Off-Flavors. Journal of Chromatographic Science 44(7):399-415.

198

120. Fucci N, De Giovanni N, Chiarotti M 2003. Simultaneous detection of some drugs of abuse in saliva samples by SPME technique. Forensic Science International

134(1):40-45.

121. Graham A.Mills, Valerie W 2000. Headspace solid-phase microextraction procedures for gas chromatographic analysis of biological fluids and materials. Journal of

Chromatography A 902:267-287.

122. Heather Lord JP 2000. Evolution of solid-phase microextraction technology.

Journal of Chromatography A 885:153-193.

123. Lambropoulou D.A. , Sakkas V.A., T.A. A 2002. Validation of an SPME method, using PDMS, PA, PDMS-DVB, and CW-DVB SPME fiber coatings, for analysis of organophosphorous insecticides in natural waters. Anal Bioanal Chem 374:932-941.

124. Pawliszyn J. 1999. Applications of Solid Phase Microextraction. ed.: The Royal

Society of Chemistry.

125. . NIST Standard Reference Database Number 69, June 2005 Release. ed.

126. Miller M, Bhatt V, Kasting GB 2006. Absorption and evaporation of benzyl alcohol from skin. J Pharm Sci 95:281-291.

127. Bronaugh RL, Stewart RF 1984. Methods for in vitro persutaneous absorption studies. III. Hydrophobic compounds. J Pharm Sci 73:1255-1258.

128. Cross S, Roberts M 2000. The effect of occlusion on epidermal penetration of parabens from a commercial allergy test ointment, acetone and ethanol vehicles. J Invest

Dermatol 115(5):914-918.

199

129. Barry BW, Southwell D, Woodford R 1984. Optimization of Bioavailability of

Topical Steroids: Penetration Enhancers Under Occlusion. Journal of Investigative

Dermatology 82(1):49-52.

130. Bucks D, McMaster J, Maibach H, Guy R 1988. Bioavailibility of topically administered steroids: a 'mass balance' technique. J Invest Dermatol 91(1):999-1005.

131. Menczel E, Bucks D, Maibach HJ, Wester R 1983. Malathion Binding to Sections of Human Skin: Skin Capacity and Isotherm Determinations. Arch Dermatol Res

275:403-406.

132. Wester RC, Maibach HJ, Bucks DAW, Guy RH 1983. Malathion percutaneous absorption after repeated administration to man. Toxicology and applied pharmacology

68:116-119.

133. Ng WF, Teo MJK, Lakso H-Å 1999. Determination of organophosphorus pesticides in soil by headspace solid-phase microextraction. Fresenius' Journal of

Analytical Chemistry 363(7):673 - 679.

134. Tsoukali H, Theodoridis G, Raikos N, Grigoratou I 2005. Solid phase microextraction gas chromatographic analysis of organophosphorus pesticides in biological samples. Journal of chromatography B, Analytical technologies in the biomedical and life sciences 822(1-2):194-200.

135. Zambonin CG, Quinto M, De Vietro N, Palmisano F 2004. Solid-phase microextraction - gas chromatography mass spectrometry: A fast and simple screening method for the assessment of organophosphorus pesticides residues in wine and fruit juices. Food Chemistry 86(2):269-274.

200

136. Namera A, Yashiki M, Nagasawa N, Iwasaki Y, Kojima T 1997. Rapid analysis of malathion in blood using head space-solid phase microextraction and selected ion monitoring. Forensic Science International 88(125-131).

137. Beltran J, Pitarch E, Egea S, Lopez FJ, Hernandez F 2001. Gas chromatographic determination of selected pesticides in human serum by head-space solid-phase microextraction. Chromatographia 54(11-12):757-763.

138. Dron J, Garcia R, E. M 2002. Optimization of headspace solid-phase microextraction by means of an experimental design for the determination of methyl tert.- butyl ether in water by gas chromatography-flame ionization detection. J Chromatogr A

963(1-2):259-264.

139. 2005. http://www.epa.gov/oppsrrd1/REDs/malathion_red.pdf. ed.

140. RayChaudhuri S, Gajjar R, Kasting GB 2007. Percutaneous Absorption of

Volatile Solvents Following Transient Liquid Phase Exposures II. Experimental Studies

& Model Validation. Chemical Engineering Science Submitted.

141. Haftek M, Teillon MH, Schmitt D 1998. Stratum corneum, corneodesmosomes and ex-vivo percutaneous absortion. Microscopy Research and Technique 43:242-249.

142. Zhai H, Maibach MD 2003. Effects of Occlusion: Percutaneous Absorption.

Cosmetics and Toileteries 118(11):22-29.

143. Kasting DGB, Miller MA, Nitsche JM. submitted. Absorption and Evaporation of

Volatile Compounds Applied to Skin. In Walters;, Roberts., editors. Dermatological and

Cosmecuetical Development, ed.

201

144. Wester RC, Maibach HJ 1985. In Vivo Percutaneous Absorption and

Decontamination of Pesticides in Humans. Journal of Toxicology and Environmental

Health 16:25-37.

145. Fredriksson T 1958. Percutaneous absorption of Sarin and two allied organophosphorus cholinesterase inhibitors. . Acta Dermato-Venereologica 38((Suppl.

41)).

146. Blank IH, Griesemer RD, Gould E 1957. Penetrartion of an anticholinesterase agent (Sarin) into skin. I. Rate of penetration into excised human skin. The Journal of

Investigative Dermatology 29:299-309.

147. Kretsos K, Kasting GB 2007. A geometrical model of dermal capillary clearance.

Math Biosci In Press.

202