ACCELERATED COMMUNICATION

0090-9556/98/2610-0943 948$02.00/0 DRUG METABOLISM AND DISPOSITION Vol. 26, No. 10 Copyright 1998 by The American Society for Pharmacology and Experimental Therapeutics Printed in U.S.A. ACCELERATED COMMUNICATION COMPARISON OF IN VIVO AND IN VITRO DEPOSITION OF RHODAMINE AND FLUORESCEIN IN HAIR PETER R. STOUT AND JAMES A. RUTH Department of Molecular Toxicology and Environmental Health Sciences, University of Colorado Health Sciences Center (Received July 20, 1998; accepted August 17, 1998) This paper is available online at http://www.dmd.org ABSTRACT: A direct differentiation of the internal and external drug-deposition pattern into hair was made using two fluorescent dyes and fluorescence microscopy after systemic administration to mice or external exposure of untreated hair. Mice (23 days old, C57 and Balb/C) were administered either rhodamine or fluorescein intraperitoneally at varied doses on 3 consecutive days of 3 weeks, and hair was sampled 1 week later. Another group was given 10 mg/kg rhodamine or 100 mg/kg fluorescein and sampled at time points from 5 min to 168 hr. The time courses of external deposition of rhodamine and fluorescein into untreated hair were examined after hair was soaked in 0.1 mg/ml solutions at ph 3, ph 6, and ph 9 aqueous buffer or methanol. The hair was then extracted in ph 6 phosphate buffer or methanol for 24 hr. In vivo accumulation was distinguishable as fluorescent bands along the length of the hair for rhodamine and fluorescein. The pattern of in vivo deposition appears to arise from the rapid accumulation within the cortex and This work supported by NIH grant DA09545. 1 Abbreviations used are: IFs, intermediate filaments; IRS, internal root sheath; CMC, cell membrane complex; ERS, external root sheath. Send reprint requests to: James A. Ruth, Ph.D., University of Colorado Health Sciences Center, Department of Molecular Toxicology and Environmental Health Sciences, 4200 East. Ninth Avenue, Box C238, Denver, CO 80262. 943 medulla, with little deposition evident in the cuticle. Neither phosphate buffer nor methanol washes affected the intensity of fluorescence in the hair. External loading of rhodamine into the hair resulted in staining of the junctions of cuticle scales. This pattern persisted even after 12 hr of solution exposure. Extraction with ph 6 phosphate buffer or methanol did not remove rhodamine. Fluoroscein followed a similar pattern, with maximum fluorescence when hair was loaded in ph 6 100mM phosphate buffer and nominal staining when loaded in ph 9 100 mm Tris buffer or methanol. Soaking the hair in ph 6 buffer, but not methanol, removed some fluorescein. These results demonstrate that compounds in the circulation can rapidly diffuse into the forming cortex and medulla, where rapid associations occur with elongating intermediate filaments specific to the medulla and cortex. These compounds can become significantly occluded within the mature matrix and are resistant to removal in aqueous or methanolic solutions. The elucidation of drug/hair matrix interaction mechanisms after systemic drug exposure is an essential element in establishing the validity of hair as a substrate for detection of drugs of abuse. Differentiation of in vivo drug accumulation and external drug contamination is equally important. The hair follicle has a complex architecture of multiple cellular layers and a complex array of proteins produced in the process of hair formation (fig. 1A). These proteins may serve as the source of drughair association. Hair keratins comprise the majority of the proteins present in the follicle (Bertolino et al., 1988; Bertolino et al., 1990; Heid et al., 1986; Heid et al., 1988). A set of eight IFs 1 are unique to hair; four type I acidic keratins and four type II basic keratins have been described in living hair-forming cells (trichocytes) and appear to be unique from epidermal keratins (Bertolino et al., 1988; Bertolino et al., 1990; Heid et al., 1986; Heid et al., 1988). The keratins in mature hair are highly crosslinked by disulfide bonds and are tightly associated with other proteins. These unique keratins are found primarily in the cortex and medulla of the hair shaft without any notable epithelial keratins present. The formation of IFs and their subsequent crosslinking occurs as the hair matures, with few IFs present in the hair papilla or deep in the hair follicle (Heid et al., 1988). Though no epidermal cytokeratins have been described in the cortex and medulla of the hair shaft, the IRS, composed of Huxley s, Henle s, and cuticle layers, and ERS demonstrated the presence of some epithelial cytokeratins (Bertolino et al., 1988; Bertolino et al. 1990). Numerous theories of how drugs become incorporated into the hair have been postulated. These theories involve incorporation into the hair from the circulation by passive diffusion from the vascular hair papilla (fig. 1A). Diffusion barriers, drug lipophilicities, and hair pigmentation may further complicate mechanisms of drug deposition into hair through this pathway. Melanin, produced by melanocytes at the crest of the hair papilla (fig. 1A) under the control of hormones and local tissue growth factors (Hirobe, 1995), has been postulated as a major binding site for drugs incorporated into hair (Potsch et al., 1997; Cone and Joseph, 1996; Joseph et al., 1997). Alternate routes of drug deposition into the hair involve incorporation of drugs carried to the hair by sweat or sebum. Potsch et al. (1997) have suggested that this pathway may account in part for the high parent-compound concentrations often seen in hair. It is unclear if drugs deposited by these pathways are distinguishable from drug

944 STOUT AND RUTH FIG. 1.Diagrams of mouse hair structure. (A) Cutaway section of a hair follicle showing the general structure of a mouse hair follicle, including layers of the internal root sheath and external root sheath as well as forming hair shaft and hair papilla (Hp). The structure is similar to human follicles, which have a smaller intermittent medulla. (B) Cutaway section of a mature mouse hair. Mouse hair has all of the same structures as human hair except that mouse hair is continuously medullated and the medulla is honey-combed, with air spaces left as medulla cells pull apart during maturation (Hojiro, 1972). deposited in the forming hair. Kidwell and Blank (1996) found that a significant amount of externally applied cocaine could not be recovered from the hair. Stout et al. (1998) also found that the majority of externally applied fentanyl could not be recovered. Such results suggest that externally applied drug can be tightly held in the hair. Potsch and Moeller (1996) examined the external deposition of rhodamine and reported that the deposition pattern was along the CMC. The CMC comprises residual cell membranes, lipid bilayers, intercellular proteinaceous material, and polysaccharides. Thus a multi-compartmental system within the hair is possible. They suggest that a preferential deposition pathway may exist between cells along the CMC rather than through keratinized cells. The mature hair in the upper isthmus of the follicle and above is different from the forming hair bulbar region. It is more compressed, nuclear material has been lost, and more fibers are evident (Hojiro, 1972). Thus the deposition of the same compound could present a very different pattern when endogenously deposited from the systemic circulation rather than from external deposition into mature hair. The objective of this study was to use fluorescence microscopy to investigate the deposition of rhodamine and fluorescein from the systemic circulation into the cytoarchitecture of the hair and compare this to the deposition of these compounds from an external solution. Materials and Methods In Vivo Study. Twelve Balb/C and twelve C57 male mice were obtained from Jackson Labs (Bar Harbor, ME) and maintained under approved protocols in the University of Colorado Health Sciences Center, Animal Care Facility. All animals injections were initiated at 23 days of age. Rhodamine B and fluorescein were obtained from Sigma (St. Louis, MO). One group of two Balb/C and two C57 mice were administered rhodamine intraperitoneally (0.9% saline solutions) on Wednesdays, Thursdays, and Fridays for 3 weeks (week one, 50 mg/kg; week two, 1 mg/kg; week three, 10 mg/kg). Fluoroscein was administered in the same fashion to two separate Balb/C and two C57 mice (week one, 100 mg/kg; week two, 10 mg/kg; week three, 50 mg/kg). One week after the final injection, animals were sacrificed and hairs plucked and skin sampled from the dorsal region. Two separate groups of six Balb/C and six C57 mice were administered one 10 mg/kg dose of rhodamine or one 100 mg/kg dose of fluorescein. At 5 min, 20 min, 2 hr, 20 hr, 44 hr, and 168 hr, animals were sacrificed and hairs plucked and skin sampled from the dorsal region. Plucked hair samples were mounted whole on glass slides in Permount (Fisher, Pittsburgh, PA) for fluorescence microscopy. Skin samples were fixed in formalin for 24 hr and embedded in paraffin. Multiple sections were cut and mounted unstained in Permount for fluorescence microscopy. Additional sections were cut and stained with standard hematoxylin and eosin (Leeson et al., 1988) for normal light microscopic observation of the cellular structure. To ensure that diffusional processes were not contributing to the results at the 5- and 20-min time points, additional skin samples were cryosectioned. Skin samples were embedded in Cryo-Gel (Instrumedics, Hackensack, NJ) and snap-frozen in dry ice in hexane. Multiple sections were cut from each sample and sections mounted in Permount. In Vitro Study. Hair from untreated Balb/C and C57 mice was clipped to within 1/100 in. of the skin, using electric clippers. Aliquots of hair were soaked in buffer solutions of either 100 mm ph 3 sodium tartrate, 100 mm ph 6 phosphate, 100 mm ph 9 Tris, or in methanol. Hairs were soaked for 1, 5, or 12 hr in solutions containing 0.1 mg/ml of either rhodamine or fluorescein. Samples were continuously agitated during soaking. After soaking, hairs were separated from the soaking solutions and washed briefly with distilled, deionized water to remove any excess solution. Hair from each of these treatments was allowed to air-dry and was then mounted whole in Permount. Aliquots of dried hair from each of the soaking treatments were subjected to soaking in either methanol at 25 C for 24 hr or in 100 mm phosphate at 37 C for 24 hr. Hair was removed from these solutions, washed briefly, air-dried, and mounted whole in Permount. Fluorescence microscopy utilized a Nikon (Melville, NY) Microphot-FX (EPI-FL-59334) scope. For rhodamine, 546 nm was used for excitation, with a barrier filter of 520 nm for observation. For fluorescein, 495 nm was used for excitation wavelength, with a barrier filter of 590 nm for observation. The intensity of the fluorescence observed increased with the concentration of the fluorescent tracer. Results In Vivo Study. Fig. 2 shows the results of in vivo dye administration. Within 5 min of administration, prominent concentration of rhodamine in the hair bulb above the surrounding skin was apparent, which was still pronounced after 2 hr (fig. 2a). More rhodamine deposition was apparent in the medulla and cortex of the forming hair shaft and Henle s layer (fig. 2, b and c). Notably, the forming cuticle, Huxley s layer, and ERS did not demonstrate rhodamine accumulation (fig. 2b). Deposition of rhodamine had a fibrous appearance near the top of the root bulb and more basal bulb cells. Papillary cells did not display as much rhodamine deposition as did the forming hair shaft and IRS. Melanocytes are dendritic cells that are located at the crest of the papilla (fig. 2d). These cells demonstrate dark granules in the hematoxylin and eosin stained section (fig. 2d), and the dark granules are evident in the unstained sections used for fluorescence. Thus the criteria for identification of melanocytes were cells with obvious dark granules along the crest of the hair papilla. No fluorescence was evident in the cytoplasm of melanocytes in fig. 2c.

IN VIVO AND IN VITRO TRACER DEPOSITION IN HAIR 945 FIG. 2. In vivo deposition of rhodamine and fluorescein. Bold-faced symbols refer to structures within each panel. No distinguishable differences were observed between animals in each treatment group. Representative photomicrographs from various individuals are used. (a) Low magnification of skin cross-section from a Balb/C 2 hr after dose; epidermis is toward the left of the frame (bar represents 1 mm). The hair bulbs are clearly more intense than the surrounding dermal tissue. (b) High magnification of C57 hair follicle 20 min after dose (bar represents 50 m). Rhodamine deposited in the cortex and medulla (asterisk) and Henle s layer (double arrowhead) but not in the ERS (e), Huxley s layer (h), or cuticle (single arrowhead). No increased fluorescence is evident in the papilla, and fluorescence increases toward the top of the hair bulb, where IF formation increases. (c) High magnification of the crest of a hair bulb from C57 2 hr after dosing (bar represents 50 m). Association of rhodamine with forming fibers (closed arrows) within the cortex (c) and medulla (m) is evident, as is the absence of fluorescence in the cuticle (single arrowhead) and Huxley s layer. Also evident is the lack of fluorescence of melanocytes (asterisks), indicating that no rhodamine is incorporated into the cytoplasm of the melanocytes. (d) Hemotoxylin and eosin stained longitudinal section from a C57 hair (bar represents 50 m). This section is analogous to those shown in B and C, showing the staining of the various layers and hair papilla (Hp). (e) Fluorescence cross-section of a C57 hair from the top of the isthmus of the hair 20 hr after dose, at high magnification. Medulla (m) is mostly air space in the pictured section. (f) An section analogous to panel e, stained with hemotoxylin and eosin (bar represents 50 m). At this point, the hair shaft is the only structure with significant fluorescence. No fluorescence is evident in the cuticle (single arrowhead), and the cortex has a homogenous fluorescence even though considerable cellular structures still stain. (g) Bands from three daily doses evident along the length of a mature hair (bar represents 1 mm). (h) Higher magnification of one of the bands reveals smooth edges and no evidence of surface structures, such as junctions of cuticle cells, that would indicate incorporation of rhodamine into the cuticle of the hair.

946 STOUT AND RUTH Fluorescence was only observable in the hair shaft at the upper portion of the follicle isthmus by 20 hr (fig. 2e). After 20 hr, fluorescence was not observable in Henle s layer. The medulla and cortex had a homogenous appearance of fluorescence, even though obvious cellular structures stained in the cortex and medulla (fig. 2f). The cuticle did not observably fluoresce. At the 44- and 168-hour sampling points, fluorescence was only observable in hair that was above the skin s surface. Once the hair had matured and grown above the skin s surface, rhodamine persisted in the hair as localized bands, with each day s dose easily distinguishable approximately 1 mm apart (fig. 2g). This indicates a growth rate of approximately 1 mm/day, putting approximately 4 mm between the final dose of 1 week and the first dose of the following week. The first week s dose was apparent at the tip of the hair, indicating that, approximately, a 3-week record was present in the hair. The pattern of in vivo deposition in the mature hair shaft appeared to be primarily in the medulla and cortex and excluded the cuticle (fig. 2h). This is supported by both the lack of fluorescence in the forming cuticle (fig. 2c) and the lack of any discernible surface structures, such as the appearance of a serrated edge or staining of scale cell junctions in the mature hair (fig. 2h). Both C57 and Balb/C mice demonstrated fluorescent staining of forming fibrils within the hair bulb. Additionally, no observable interindividual differences were seen between mice receiving the same treatments. Fluorescein deposition appeared to have a similar pattern to rhodamine within 5 min after dosing, namely in the cortex and medulla, with some faint staining of Henle s layer. However, after 2 hr, almost no fluorescence was observable in fluorescein-dosed animals. No significant deposition of fluorescein was observed in the hair. Rhodamine and fluorescein deposited in the hair did not appear to be significantly available to extraction either by 24-hr aqueous phosphate buffer extraction or 24-hr methanol extraction. No noticeable changes in either the intensity of the fluorescence or the pattern of deposition were observed after either treatment. In Vitro Results. The pattern of rhodamine and fluorescein deposition in vitro was markedly different from the in vivo results. As shown in fig. 3, externally applied rhodamine deposited along the FIG. 2.Continued. junctions of cuticle scales without significant deposition into the medulla except at the cut ends of hair where the soaking solution could move into the interior of the hair by bulk flow. The right side of the frame shows distinct staining of the junctions of cuticle cells with no observable medullary structures, while the left side of the frame shows the cut end with the distinctive pattern of medullary air spaces containing rhodamine. No observable difference in the intensity of rhodamine fluorescence was seen between the different ph groups. Additionally, the pattern of cuticular deposition without medullar deposition observed after 5 hr (fig. 3a) was conserved even after 12 hr of exposure to rhodamine solutions (fig. 3b). Soaking the hair in the methanol solution yielded a pattern of medullar deposition identical to the pattern observed for fluorescein (fig. 3c). No noticeable diminution of fluorescence was observed in the hair soaked in aqueous rhodamine solutions after hair was washed in either phosphate buffer or methanol. However, the fluorescence was markedly decreased by both the phosphate buffer and methanol washes of the hair treated with the methanolic rhodamine solution. Fluorescein deposition demonstrated a dramatic ph dependence. When soaked in ph 6 phosphate buffer, a pattern of cuticular deposition was observed. Hair soaked in ph 9 Tris buffer did not demonstrate appreciable deposition. The pattern of deposition appeared to be cuticular but was sparse and of low fluorescence intensity (the exposure time for hair shown in fig. 3d was three times as long as that for hair in fig. 3c). Hair soaked in the methanolic fluorescein solution demonstrated a medullar pattern of deposition without noticeable deposition in the cuticle or cortex (fig. 3e). The intensity of fluorescence was diminished by both the methanol and phosphate buffer extractions. However, significant fluorescence was still observable, indicating that not all of the fluorescein could be removed from hair soaked either in methanolic or ph 9 Tris buffer solutions of fluorescein. Discussion It is clear from the evidence shown in fig. 2g that multiple daily rhodamine doses deposit in discrete bands within mature hair. The pattern of in vivo deposition within the cortex and medulla of the hair appears to exclude the cuticle (fig. 2, b, c, and h). The forming hair

IN VIVO AND IN VITRO TRACER DEPOSITION IN HAIR 947 FIG. 3. In vitro deposition of rhodamine and fluorescein. (a) Cut end of a C57 hair soaked in 0.1 mg/ml rhodamine in 100 mm ph 6 phosphate buffer for 5 hr (bar represents 50 m). Bulk movement of the solution into the hair from the cut end (left of frame) is evidenced by the staining of air spaces within the hair. Beyond the reach of bulk solution flow (right of frame), the junctions (CMC) between cuticle cells are stained. (b) C57 hair soaked in 0.1 mg/ml rhodamine in 100 mm ph 6 phosphate buffer for 12 hr (bar represents 50 m). The same pattern of cuticle cell junction staining is still evident, with some diffusion of the borders. ph of the buffer appeared to have no effect on the intensity or pattern of rhodamine fluorescence. (c) C57 hair soaked in 0.1 mg/ml fluorescein in 100 mm ph 6 phosphate buffer for 12 hr (bar represents 50 m). Fluoroscein presents a similar pattern of deposition to rhodamine. (d) Balb/C hair soaked in 0.1 mg/ml fluorescein in 100 mm ph 9 Tris buffer for 12 hr (bar represents 50 m). This exposure time was three times that of the section shown in panel c and indicates that nominal fluorescein deposition occurred at ph 9. (e) Balb/C hair soaked in 0.1 mg/ml fluorescein in methanol for 12 hr (bar represents 50 m). Medullary structures are stained without staining of the cortex or cuticle. This is evident from the lack of obvious surface features such as serrated edges due to staining of cuticle cell junctions or the patterns of cuticle cell junction staining evident in panels a, b, and c. Additionally, staining of the medullary structures are not contiguous at the edges, indicating that the cortex, which is contiguous with the medulla at the edges, is not stained. This may be due to increased permeability of the cuticle by methanol. Rhodamine demonstrated a similar pattern. shaft at the top of the hair bulb (fig. 2c) and the upper isthmus (fig. 2e) also demonstrates little rhodamine deposition in the cuticle. These results, coupled with the likelihood that all of the components of the forming hair are exposed to a similar concentration of rhodamine, suggest that the cuticle has a lower affinity for rhodamine during systemic deposition than do the cortex and medulla. The rapid appearance of rhodamine through the length of the forming hair shaft and Henle s layer, coupled with the lack of fluorescence in the hair papilla, suggests that diffusion of small molecules may occur more prominently from the capillaries surrounding the hair follicle than from the papilla. The blood stream appears to be the major route of deposition of rhodamine and fluorescein from the systemic circulation, as mice do not have significant sweat glands in the skin of their backs (Green, 1966). Heid et al. (1988) found that the cuticle had epithelial keratins, whereas the cortex and medulla had only keratins unique to trichocytes. This would suggest that the trichocyte-specific keratins have a greater affinity for compounds such as rhodamine. These keratins have a higher sulfur content than do epithelial keratins (Bertolino et al., 1990), which may result in these trichocyte keratins having more capacity for interaction or occlusion of compounds. The higher degree of disulfide crosslinking may result in increased occlusion of rhodamine within the matrix. Rhodamine also appears to preferentially associate with keratin

948 STOUT AND RUTH fibers that have elongated. As can be seen in fig. 2c, less fluorescence is associated with cells in and around the papilla than in the forming hair shaft near the top of the hair bulb. As can be seen in the higher magnification field (fig. 2c), the deposition pattern in the forming hair shaft is fibrous. Heid et al. (1988) noted that little intermediate filament formation occurs in the lower cells of the hair bulb, and Potsch et al. (1997) describes fiber elongation in the upper portion of the bulb. These descriptions are consistent with rhodamine associated with or occluded in maturing keratin. For rhodamine, its association with protein fibers appears to dominate over its association with melanin granules. Melanocytes and melanin granules are evident in fig. 2c and do not demonstrate any fluorescence or fluorescent halos that would indicate rhodamine deposition associated with the granule. The lack of fluorescence could be due to quenching of either the excitation energy or the emission energy. However, in C57 mice, the pigment is almost exclusively eumelanin (Green, 1966), and Ikejima and Takeuchi (1978) reported that eumelanin has an excitation maximum at 376 nm, with nominal absorption above 400 nm. Because we utilized 546 nm for excitation and 590nm for emission, it is unlikely that melanin quenching was involved. Additionally, the pattern of melanin deposition in mouse medulla is in clumps (Hojiro, 1972); thus, if rhodamine associated with melanin more than protein fibers, a periodic pattern of fluorescence would be expected in the medulla. This was not observed in the present study. Additionally, a similar pattern of deposition along fibers of the forming Balb/C hair indicate that association of rhodamine with forming keratin fibrils occurs in a non-pigmented strain as well as in a pigmented strain. This is inconsistent with reports by Joseph et al. (1997) and Potsch et al. (1997) that melanin is a significant site for the binding of small molecules other than rhodamine and fluorescein in the hair. This suggests that melanin binding may be compound-specific. Fluoroscein did have some faint banding apparent in the hair but was not as prominent as rhodamine. Fluoroscein was rapidly cleared from the forming hairs within 1 hr of the dose. Though fluorescein can interact with the hair as demonstrated by the in vitro results, the in vivo results suggest that fluorescein is cleared from the tissue faster than it can be trapped by the compression and crosslinking of the forming hair matrix. The rapid disappearance of fluorescein could also be due to a lack of suitable binding sites within the forming hair and the diffusion of fluorescein back into the blood stream. Thus other anionic compounds, such as 9 -tetrahydrocannabinol carboxylic acid, may display low hair concentrations in part because they are cleared from the hair follicle rapidly, and not because of a lack of potential binding sites. External rhodamine appeared to associate only with the CMC of the cuticle cells without staining the medulla. This is consistent with the findings by Potsch and Moeller (1996), who found that rhodamine appeared to gain access to human hair from external solution through the CMC of cuticle cells. Rhodamine and fluorescein deposited from a methanol solution stained the air spaces of the medulla without staining the CMC of the cuticle cells. This may be due to an increased permeability of the cuticle by methanol. Externally deposited rhodamine and fluorescein appear to be strongly associated with the hair fiber; neither was completely removed by either aqueous extraction or a methanolic extraction. This is consistent with findings by Kidwell and Blank (1996), who found that externally applied cocaine was significantly occluded from recovery. We did not find that rhodamine deposited from a non-aqueous solution could be removed from the hair, as Potsch and Moeller (1996) found. This may be due to differences in the rhodamine solutions used. We used pure methanol with rhodamine, whereas Potsch and Moeller (1996) used 1:9 v/v methanol to ethanol. A pure methanol solution may have resulted in greater access to the matrix and deposition of rhodamine more resistant to extraction. The results of this study indicate that rhodamine and fluorescein can both deposit into the hair from the systemic circulation. This deposition appears to be by association of the dye with forming trichocytic-specific keratins in the hair follicle. This pattern was similar in both Balb/C and C57 mice. The presence of this pattern in both a pigmented and non-pigmented strain supports the association of rhodamine with forming fibrils. This association is very rapid, occurring within 5 min of dosage. This pattern of deposition is markedly different from dye deposited in vitro, which deposits along the CMC of the cuticle cells without staining the cortex and medulla. Both pathways result in dye deposition that is resistant to removal by either aqueous or methanolic extraction. Because mice have similar follicle morphology to humans, these results suggest that endogenous and exogenous deposition of drugs may present different patterns in humans. However, distinguishing these patterns by extraction procedures may be difficult. Rhodamine, though visible different in external deposition from endogenous deposition, was not easily removed in either case. Thus drugs deposited from the environment may be difficult to distinguish by extraction and washing procedures. Additionally, it is clear in mice that rhodamine deposition can occur by rapid association with the forming keratin fibers before the hair is exposed to sweat or sebum. Thus, in humans, it is possible that the deposition of short-lived species, such as cocaine, could occur by incorporation into the forming hair rather than though deposition onto the mature hair from the sweat or sebum. References Bertolino AP, Checkla DM, Heitner S, Freedberg IM and Yu D (1990) Differential expression of type I hair keratins. J Invest Dermatol 94:297 303. Bertolino AP, Checkla DM, Notterman R, Sklaver I, Schiff TA, Freedberg IM and DiDona GJ (1988) Cloning and characterization of a mouse type I hair keratin cdna. J Invest Dermatol 91:541 546. Cone EJ and Joseph RE (1996) The potential for bias in hair testing for drugs of abuse, in Drug Testing in Hair (Kintz P ed) pp 69 93, CRC Press, New York. Green E (1966) The Biology of the Laboratory Mouse, 2nd Ed., McGraw-Hill, New York. Heid HW, Moll I and Franke WW (1988) Patterns of expression of trichocytic and epithelial cytokeratins in mammalian tissues. I. Human and bovine hair follicles. Differentiation 37: 137 157. Heid HW, Werner E and Franke WW (1986) The complement of native -keratin polypeptides of hair forming cells: A subset of eight polypeptides that differ from epithelial cytokeratins. Differentiation 32:101 119. Hirobe T (1995) Structure and function of melanocytes: Microscopic morphology and cell biology of mouse melanocytes in the epidermis and hair follicle. Histol Histopathol 10:223 237. Hojiro O (1972) Fine structure of the mouse hair follicle. J Electron Microsc 21:127 138. Ikejima T and Takeuchi T (1978) Fluorescence spectrophotometric analysis of melanins in the house mouse. Biochem Genet 16:673 679. Joseph RE, Tsai W, Tsao L, Su T and Cone EJ (1997) In vitro characterization of cocaine binding sites in human hair. J Pharmacol Exp Ther 282:1228 1241. Kidwell DA and Blank DL (1996) Environmental exposure the stumbling block of hair testing, in Drug Testing in Hair (Kintz P ed) pp 17 68, CRC Press, New York. Leeson TS, Leeson CR and Paparo AA (1988) Atlas of Histology. W.B. Saunders, Philadelphia. Potsch L and Moeller MR (1996) On pathways for small molecules into and out of human hair fibers. J Forensic Sci 42:121 125. Potsch L, Skopp G and Moeller MR (1997) Biochemical approach on the conservation of drug molecules during hair fiber formation. Forensic Sci Int 84:25 35. Stout PR, Claffey DJ and Ruth JR (1998) Fentanyl in hair: chemical factors involved in accumulation and retention of fentanyl in hair after external exposure or in vivo deposition. Drug Metab Dispos 26:689 700.