Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Six hundred million dollars a year are spent on workplace drug testing. Thirty three million full-time employees are subject to workplace testing in the United States and drug testing in the workplace appears to be an effective deterrent of worker drug use (Hoffman et al., 1997). Hair testing, a growing tool in drug testing, could offer numerous advantages over urine and serum such as reduced cost of sample collection, storage, and shipment, as well as reduced psychological stress for the testing subject. In addition, the longer window of drug use history and resistance to tampering available from hair are definite advantages over urine or blood. Hair testing also has applications in therapeutic drug monitoring, environmental exposure monitoring, and postmortem toxicology.

However the general acceptance of hair testing reliability has been awaiting answers to questions of mechanisms of drug deposition (Heustis, 1996). Questions currently debated include the efficiency of drug extraction, the differentiation of drugs deposited from the systemic circulation from external contamination, the role of sweat and sebum in contributing to hair drug levels (Kidwell and Blank, 1996), and the role of pigment in drug deposition in hair (Cone and Joseph, 1996).

If drugs are in fact deposited onto the hair from sweat or sebum, the deposition would be difficult to distinguish from external contamination. The role of pigment in deposition has raised questions of differential detection sensitivity in various ethnicities (Cone and Joseph, 1996) or of possible gender bias where hair treatments such as coloring and permanent treatments may possibly remove deposited drugs. Several studies have found that drugs associate with synthetic melanin or melanin obtained from cuttle fish (Sepia officinalis) in vitro (Shimada et al., 1976; Potsch et al., 1997). Additional studies have found differences in drug concentrations between differently pigmented hairs (Shimada et al., 1976; Gygi et al., 1995; Cone and Joseph, 1996; Potsch et al., 1997; Joseph et al., 1997; Knorle et al., 1998; Rothe et al., 1997; Slawson et al. 1998). Emmerich et al. (1998) found that melanocytes in culture accumulated more cocaine than did keratinocytes. We have demonstrated pigment-dependent differences in deposition of radiolabeled serum constituents in vivo (Stout et al., 1998).

By way of differentiating drug deposition in hair following systemic or external drug exposure, we have shown histologically that rhodamine was deposited externally on the hair with a completely different pattern of accumulation than systemically administered rhodamine (Stout and Ruth, 1998). Internal deposition was similar in both C57 mice and Balb/C mice and appeared to be by association with forming keratin fibrils in the hair bulb.

To further determine the subcellular location of drugs deposited in the hair from the circulation, we examined three drugs of forensic importance. [³H]Cocaine, [³H]flunitrazepam (a "date-rape" drug), and [³H]nicotine (of growing importance to insurance companies trying to determine smoking behaviors; Kintz et al., 1998), were administered to pigmented and nonpigmented animals and then the distribution of these compounds into the hair was examined by autoradiography of skin sections containing developing hairs.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

The drugs [benzoyl-3,4-³H(N)]cocaine, [methyl-³H]flunitrazepam, and [N-methyl-³H]nicotine were obtained from New England Nuclear Corp. (Boston, MA) Cocaine and nicotine were prepared in 0.9% saline and flunitrazepam was prepared in water/propylene glycol/dimethyl sulfoxide (49.5:49.5:1% v/v/v). All solutions were administered i.p. in a volume of 0.01 ml/g b.wt. [³H]Cocaine was administered at 10 mg/kg (Ruth et al., 1988). [³H]Nicotine was administered at 4 mg/kg (Grun et al., 1992). [³H]Flunitrazepam was administered at 1 mg/kg (Vatashsky and Aronson, 1983). Tracers were combined with unlabeled drug to achieve dosable concentrations and specific activities of 555 dpm/ng for nicotine, 2220 dpm/ng for flunitrazepam, and 139 dpm/ng for cocaine. For deposition and retention experiments, mice were dosed daily for 3 days and then allowed to grow for an additional 21 days.

Mice, 23 days of age, were obtained from Jackson Laboratories (Bar Harbor, ME). At this age, mice undergo a synchronized anagen (Green, 1966) ensuring that the hair was growing at the time of dosage. C57 mice are predominately eumelanin pigmented with little pheomelanin (Green, 1966), whereas Balb/C mice are nonpigmented. This allowed for the comparison of pigment effects on deposition.

All animal protocols and housing were Institutional Animal Care and Use Committee approved and performed in the University of Colorado Health Sciences Center Animal Care facility. For the deposition studies, four C57 and four Balb/C mice were used for each tracer (a total of 24). For the pharmacokinetic and autoradiography experiments, two C57 and two Balb/C mice were used for each of five time points (a total of 60). Time points were selected to collect data over multiple half-lives of the administered drugs for the calculation of area under the curve (AUC)¹ values while using a minimum number of animals. For [³H]flunitrazepam, time points were 10 min and 1, 5, 24, and 48 h. For [³H]cocaine, time points were 5 and 15 min and 1, 2.5, and 5 h. For [³H]nicotine, time points were 5 and 15 min and 1, 5, and 8 h. Skin sections for autoradiography were taken from animals at the first three time points for each drug.

To prepare autoradiographic samples, skin was harvested from sacrificed animals used for pharmacokinetic measurements (two C57 and two Balb/C mice for each time point). These samples were then fixed in 10% buffered formalin for 18 h. Skin samples were imbedded in paraffin, sectioned, and affixed to slides. Sections were dehydrated in graded ethanol and defatted in several changes of xylene. Sections were allowed to dry overnight. Slides were dip coated with photographic emulsion (NTB2; Eastman Kodak, Rochester, NY). Once the emulsion was dry, slides were stored in light-tight boxes with desiccant at 4°C for 8 weeks. Exposed slides were developed in half-strength D-19 developer (Kodak) and fixed. Slides were then stained with Mayer's H&E and coverslips mounted on slides with Permount (Fisher Scientific, Pittsburgh, PA). All developing and staining were carried out at 15°C to prevent damage to the emulsion. All slides were examined and photographed using a Nikon Microphot-FX (EPI-FL-59334; Nikon Instruments, Melville, NY) fitted with an Immunogold Staining (IGS) filter block for incident lighting of specimens.

Silver grains were counted for five follicles from both animals in each treatment group (10 total follicles). Only follicle sections that had melanocytes and melanosomes that completely filled the 25- × 25-µm counting square were counted; these follicles were picked randomly as the sections were microscopically examined. Drug-positive sebaceous glands were counted if more than 100 silver grains were present per gland. One hundred follicles per individual animal were counted and the percentage of positive glands was calculated.

To measure blood concentrations of each tracer, animals were anesthetized with pentobarbital (70 mg/kg i.p.), blood was taken from the retro-orbital plexus, and animals were sacrificed after sampling. Blood samples (60 µl) were taken using a heparinized capillary tube and radioactivity measured by liquid scintillation counting. All samples were corrected for color quenching. Pharmacokinetic parameters were determined by a linear regression fit of log transformed data. AUC values were determined by the trapezoidal rule (Table 1). Because no appreciable differences were observed in the pharmacokinetics between Balb/C and C57 mice, AUC and T_1/2 values were determined using all animals.

View this table: [in this window] [in a new window]   TABLE 1 Pharmacokinetic data for flunitrazepam, cocaine, and nicotine AUC and T1/2 values are mean of four animals. Incorporation index is total drug in hair divided by AUC (after Nakahara et al., 1996).

The total concentration of each tracer in mature hair was measured by first digesting 10 mg of hair (n = 4) in 1 ml of 1 M NaOH for 24 h at 37°C. This solution was then neutralized to pH 7 with HCl, Fisher Scintisafe 50% Plus scintillation cocktail (Fisher Scientific) added and the sample counted using a Packard liquid scintillation counter. Additional 10-mg aliquots of hair (n = 4) were subjected to several extraction procedures. For an aqueous extraction, hair was washed briefly in 1 ml of methanol and then suspended in 1 ml of 100 mM pH 6 phosphate buffer at 37°C for 24 h. Tracer concentrations were measured in both the methanol wash and the phosphate buffer extraction. For an enzymatic digestion of hair, 10-mg aliquots of hair (n = 4) were suspended in1 ml of a Proteinase K (Sigma Chemical Co., St. Louis, MO) digestion solution described in Nakahara et al. (1992). This solution was then vortexed with 2 ml of ethyl acetate for 5 min, the organic phase decanted, and the tracer concentration measured in the organic phase by liquid scintillation counting. To measure the tracer recoverable from NaOH-digested hair by a liquid/liquid extraction, 10-mg aliquots of hair (n = 4) were digested in 1 ml of 1 M NaOH as above. The solution was then vortexed with 2 ml of ethyl acetate for 5 min, the organic phase decanted, and the tracer concentration measured by liquid scintillation counting. Lastly, to measure the tracer in untreated hair, 10-mg aliqouts of hair were suspended in scintillation cocktail and the tracer concentration measured by liquid scintillation counting.

Incorporation indices were calculated by the method described in Nakahara and Kikura (1996) by dividing the concentration of tracer in the hair by the serum AUC for each tracer. This gives an estimate of the incorporation rate for each drug relative to the serum levels of each tracer. All data were analyzed by a one-way crossed ANOVA or two-way crossed ANOVA followed by a least-significant difference means test using Statgraphics 6.0 (Manugistics, Rockville, MD). For all data, significance was assumed at the  = 0.01 level.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Table 1 shows the pharmacokinetic data for the elimination of systemically administered [³H]flunitrazepam, [³H]nicotine, and [³H]cocaine. Because no significant differences were seen between drug time courses in C57 and Balb/C mice, AUC and T_1/2 values are the mean of both C57 and Balb/C mice. Significantly greater total amounts of each of the three drugs were present in C57 hair compared with Balb/C hair. Incorporation indices calculated were significantly greater in C57 mice than in Balb/C mice.

Skin sections taken from animals at time points early in the metabolism of each drug (10-15 min after dosage) show rapid association of each drug with melanin within the hair bulb of C57 mice (Fig. 1, A and B). Figure 1, A and B are of a C57 mouse 1 h after dosage with flunitrazepam. In Fig. 1A, a H&E- stained section demonstrates histological detail of a C57 follicle. Fig. 1B shows the same section with reflected incident lighting demonstrating silver grains. Silver grain deposition (indicating the presence of the tracer) is clearly over the eumelanin melanosomes of the forming medulla and cortex in the hair bulb without any deposition over the other cellular structures or other parts of the hair bulb. In particular, no silver deposition is evident outside of the forming cortex and medulla. Also, little silver grain deposition was evident in Balb/C mice treated in a similar manner (Fig. 1, C and D). The same pattern of deposition was observed in each strain following the systemic administration of [³H]cocaine and [³H]nicotine. Figure 1E demonstrates the background silver grain deposition in an untreated C57 mouse.

View larger version (188K): [in this window] [in a new window]   Fig. 1.   A, H&E-stained section of a C57 follicle 1 h after a 1 mg/kg i.p. dose of flunitrazepam (original magnification, 1200×; scale bar, 50 µm). Melanosomes (*) in the cortex (Cx) and medulla (M) are evident with transmitted light. B, incident light reflects from silver grains deposited over melanin in section indicating flunitrazepam present in melanosomes. Visibly more silver deposition in C57 mice indicates more drug deposition in melanin in black hair than in nonpigmented hair. This pattern was the same for nicotine and cocaine. C, transmitted lighting of a Balb/C hair follicle, with hair papilla (Hp) evident, also taken 1 h after i.p. flunitrazepam dose (original magnification, 1200×; scale bar, 50 µm). D, incident lighting reflects off diffuse silver grains in maturing keratinocytes. Again this pattern of deposition was the same for both cocaine and nicotine. E, background sliver grain reflectance in an untreated C57 follicle with incident lighting (original magnification, 600×; scale bar, 25 µm). Fewer than five grains per 625-µm2 field was considered acceptable background. F, a positive sebaceous gland with incident lighting from a Balb/C 10 min after flunitrazepam dose (original magnification, 200×; scale bar, 50 µm). Reflective structures above and below hair shaft in middle are sebaceous gland tissue with heavy silver deposition. This deposition was typical of positive sebaceous glands.

Drug deposition occurred within 10 to 15 min of the dosage (Table 2) Significantly (p < .0001) more silver grains were evident in C57 mice than in Balb/C mice for all three drugs. Table 2 also shows the appearance of each of the tracers in sebaceous glands. [³H]Flunitrazepam appeared in the sebaceous glands of Balb/C mice but not C57 mice. [³H]Cocaine appeared in both C57 and Balb/C mice without a significant difference in the number of positive glands. [³H]Nicotine did not appear in either C57 or Balb/C sebaceous glands. Significantly (p < .0001) more silver grains were evident in C57 mice 10 to 15 min after dosage of all three drugs (Table 2) when compared with control animals (Fig. 1E). Even though the three drugs are labeled in such a way that the major metabolites of these drugs are also radiolabeled, the time course of the deposition of drugs into melanin, when compared with the half-lives of the drugs (Table 1) suggests that parent drug is predominantly deposited in the hair.

Substantially more of each of the drugs was present in the hair than could be recovered by a variety of extraction methods (Fig. 2). Significantly more of each drug was present in the hair (total NaOH digested) than could be recovered in an enzyme digest or phosphate buffer extraction (Fig. 2). When the hair was digested in NaOH and this digest extracted with ethyl acetate, no more than 53% of the drug present could be recovered. For nicotine and flunitrazepam, significantly more drug was recovered by an organic extract of the NaOH-digested hair than from the enzyme digested hair (Fig. 2; nicotine p < .0001; flunitrazepam p = .0001). This suggests that the drug present may be tightly occluded or covalently linked to hair components and that the enzyme digest does not effectively liberate the drug present. Significantly less cocaine was recovered by an organic extraction of the NaOH-digested hair than from the enzyme digest (p = .0001), however this is attributable to the instability of cocaine in strong base. Additionally, when undigested hair was counted, significantly less drug was measured when compared with NaOH-digested hair (p = .0001).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Deposition and retention of each of the three drugs in mature hair. Bars represent mean of four animals and error bars represent 1 S.D. Numbers indicate drug recovered by each treatment as a percentage of total drug present in hair (total drug in hair represents 100%). For each of the drugs no treatment was more than 50% effective at recovering drug present in hair and in most cases treatments were no more than 20% efficient. Significantly (flunitrazepam, p = .0005; cocaine, p < .0001; nicotine, p = .0001) more of all three drugs were present in C57 hair than in Balb/C hair. This is consistent with pattern observed histologically. Significantly less of all three drugs was measured in undigested hair than in digested hair (cocaine, p < .0001; nicotine, p = .0005; flunitrazepam, p = .0001). *Difference between C57 and Balb/C values was significant by ANOVA assuming significance at p < .01. Treatment method values were significantly different than total digested values by ANOVA assuming significance at p < .01.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

These results are consistent with the results of other authors who indicate that unmetabolized cocaine and nicotine are the dominant species deposited in the hair following systemic administration and analysis by gas chromatography-mass spectrometry (Kintz, 1996; Pichini et al., 1997). However this is inconsistent with Cirimele et al. (1997) who indicated that the metabolite 7-aminoflunitrazepam is present in higher concentrations than flunitrazepam. Their results were determined by gas chromatography-mass spectrometry rather than radiotracer and may reflect a differential extent of parent drug versus metabolite extraction from hair.

Other authors have noted that the concentration of parent compound in hair often predominates over metabolites and have suggested this is due to contamination of the hair from drug-containing sweat (Kidwell and Blank, 1996). The rapid association of each of the drugs in this study with the forming hair suggests that the observed predominance of parent drug incorporation can occur from direct incorporation into the forming hair from the circulation. The deposition pattern observed in this study also demonstrates association of the drug with the forming hair below the point at which sweat or sebum have access to the hair.

Sebaceous deposition does not appear to be a significant route of deposition for these three drugs in this study. Although it is clear that sebaceous glands in Balb/C mice were strongly positive for flunitrazepam soon after dosage (Table 2), C57 mice did not demonstrate deposition of tracer in sebaceous glands. For cocaine, the presence of the tracer in sebaceous glands was small in both C57 and Balb/C mice, whereas nicotine was not present in the sebaceous glands of either strain. Yet, significantly more flunitrazepam, cocaine, and nicotine were incorporated into C57 hair than Balb/C hair. Significantly more of each drug was recovered by each treatment from C57 hair than from Balb/C hair (Fig. 2) Additionally, an initial methanol wash of the hair did not remove more than 3.3% of the total incorporated drugs and was not significantly different between Balb/C and C57 mice (Fig. 2). These results demonstrate nominal deposition onto the hair from sebum.

When hair concentrations are indexed to the serum AUC (Table 1), it is apparent that nicotine and cocaine incorporate at a similar rate. However flunitrazepam accumulation was much greater, suggesting that a more lipophilic compound has greater access to the hair. Additionally, the incorporation rate for all three drugs into pigmented hair far exceeded the incorporation rate for each drug into nonpigmented hair. This indicates that for the same AUC far more drug deposits into pigmented hair than nonpigmented hair.

Because tritium is a low-energy emitter, it is easily quenched; the differential between the digested and undigested hair measurements indicates that the location of the tracer was within the hair matrix and not on the surface. This is consistent with the microautoradiographic evidence showing association of each of the drugs with internal structures of the hair.

Several authors have found association of various compounds with melanin in vitro. These studies have used either Sepia melanin or synthetic. This study demonstrates association of flunitrazepam, nicotine, and cocaine with melanosomes in the forming hair shaft after systemic administration. This relationship is confirmed by analysis of mature hair, which shows significantly more incorporation of all three tracers in pigmented hair than nonpigmented hair. Additionally, this incorporation is rapid enough for the incorporation into the forming hair of parent drug to occur directly from the circulation. Sebum is a nominal deposition pathway for these three drugs in this study and the incorporated drug is not effectively recovered from the hair even after the protein matrix of the hair has been digested. This suggests that incorporated drug is either tightly bound or occluded within components of the hair. These results also demonstrate that interpretation of hair drug analyses is complicated by the extent of hair pigmentation. Although some studies have found reasonably complete recovery of externally applied radiolabeled drugs, drugs deposited from the systemic circulation are resistant to recovery from both pigmented and nonpigmented hair. This is consistent with external and internal deposition exhibiting different patterns (Stout and Ruth, 1998). The results demonstrate that systemically administered drugs are strongly retained in a unique manner of deposition in forming hair. Thus, deposition from the systemic circulation and environmental contamination may be distinguishable.

The strong association of many drugs with melanin in vitro seen in other studies (Shimada et al., 1976; Gygi et al., 1995; Cone and Joseph, 1996; Potsch et al., 1997; Joseph et al., 1997; Knorle et al., 1998; Rothe et al., 1997; Slawson et al., 1998) and the association of these three drugs with melanin in this study in vivo suggests that darker-haired individuals with more eumelanin are likely to accumulate and retain more drugs in their hair. Thus the interpretation of human results must be done carefully to avoid potential bias due to hair color.