Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Hair offers several potential advantages, compared with serum and urine, as a substrate for analysis of drugs of abuse. This has been demonstrated in terms of the ease of sample collection, a longer window of detection, and increased analyte stability. Because hair is a complex protein matrix, it is potentially capable of multiple types of chemical interactions with drugs of abuse. 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.

Incorporation of drugs and other compounds from the circulation might involve ionic interactions with ionizable, polar, functional groups, such as amino and carboxylic acid groups, within the hair that could function as binding sites. Neutral species might associate with the hair matrix by hydrogen bonding or other noncovalent interactions. Compounds with reactive chemical functionalities, such as cocaine or the amino acid benzoylecgonine, might be covalently linked to the hair matrix. Diffusional barriers, drug lipophilicities, and hair pigmentation might further complicate the mechanisms of drug deposition into hair.

Kidwell and Blank (1996) have proposed a model for drug incorporation in which hair functions as an ion-exchange membrane. They suggest that cationic compounds have greater affinity for hair than do anionic compounds and that this interaction is generally with carboxyl groups within the hair. To examine this, they esterified carboxylic acid groups within the hair with methanolic HCl and then examined the binding of cocaine to the hair. They indicated that the binding of cocaine would be reduced because of the treatment with acid even without esterification, and they attempted to control for this effect by soaking the hair in an aqueous buffer solution for 5 days before the binding experiments. They found that less cocaine could bind to the treated hair, suggesting that carboxylic acid groups participated in the binding of cocaine to hair.

Joseph et al. (1997) suggested that drugs are incorporated into hair not through an ion-exchange process but as ligands binding to specific sites within the protein matrix of the hair. They contend that hair has sites that can act as cocaine receptors, because they observed cocaine binding to be reversible, stereoselective, and saturable. Additionally, they suggested that this binding site for cocaine is melanin. However, melanin may not be the principle binding site for all drugs; Ishiyama et al. (1983) found that methamphetamine bound similarly to both light and dark hair, suggesting that drug binding may be drug specific and for some drugs may involve protein binding sites.

In an effort to elucidate the mechanisms of drug accumulation in hair, we examined the deposition of several normal serum constituents that were unlikely to be structural components of hair, in an albino mouse strain (BALB/c) and a pigmented strain (C57). If drugs are deposited in hair from the circulation, is the same true for constituents present in much higher concentrations and in equilibrium with interstitial fluid? Are all components of the serum deposited to the same extent in the hair matrix? To answer these questions and provide a comparative framework, the degree of in vivo incorporation of a model of cation interaction (Ca²⁺) and the incorporation of a model of anion interaction (Cl) were examined after chronic administration of Na³⁶Cl or ⁴⁵CaCl₂. The contents of other endogenous ions (Mg²⁺, Na⁺, and K⁺) allowed for assessment of the ability of hair to participate in ionic interactions. A neutral serum constituent ([¹⁴C]urea) was examined to determine the accumulation of a neutral constituent in the hair. We also characterized the deposition of [³⁵S]cysteine in the hair matrix. Cysteine is a structural component of hair and thus acted as a model of a covalently bound molecule. We characterized potential compartments within the hair that were capable of ionic or neutral interactions by examining the stability of these constituents to treatment of the hair with mildly acidic aqueous conditions, methanol, or a solution containing an organic cation (choline).

Lastly, we compared serum concentrations of these constituents and their concentrations in hair with AUC values and hair concentrations reported in the literature for several drugs and metabolites. By comparing the ratios of serum AUC values to hair concentrations, the relative inclusion or exclusion of each compound in the hair matrix could be compared.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Tracer Solutions. Four radiolabeled tracer solutions were used in this study. Stock solutions of [³⁵S]cysteine (1075 Ci/mmol) and [¹⁴C]urea (55 mCi/mmol) were purchased from ICN Pharmaceuticals Inc. (Costa Mesa, CA). Stock solutions of ⁴⁵CaCl₂ (0.0167 mCi/µg of Ca²⁺) and Na³⁶Cl (0.124 mCi/mg of Cl) were purchased from Amersham (Arlington Heights, IL).

Estimation of In Vivo Specific Activity. Because radioisotopes of endogenously present constituents were used in this study, isotopic dilution of our tracers in the serum of mice had to be estimated for calculation of specific activity. To do this, we measured the radioactivity values in blood for each tracer after a single dose, at six time points, i.e. 15 min, 1 hr, 2 hr, 4 hr, and, depending on the rate of clearance, 8 and 12 hr or 24 and 48 hr. BALB/c mice (21 days of age) were obtained from The Jackson Laboratory (Bar Harbor, ME). Groups of three mice for each time point (72 total mice) were injected ip with each of the tracers described above. Mice in groups of three were then anesthetized ip with 70 mg/kg pentobarbital; at each time point, two 60-µl aliquots of blood from each mouse were drawn from the retro-orbital sinus, using heparinized microcapillary tubes. The mice were then sacrificed by cervical dislocation. Blood samples were combined with 4 ml of Scintisafe Plus 50% scintillation cocktail (Fisher, Pittsburgh, PA) and counted using a Packard model 1600 TR liquid scintillation counter. Samples were corrected for color quenching from blood. Plots of radioactivity levels vs. time (fig. 1) were constructed from the data for each of the tracers, and time-weighted averages of radioactivity levels (dpm per milliliter of blood) were calculated for each tracer (table 1).

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Radioactivity levels in the blood of mice after a single injection of each radiolabeled tracer. Blood samples were taken from the retro-orbital plexus at times after injection and were assayed by liquid scintillation counting. Points, means (N = 3); error bars, ±1 SD. These results were corrected for quenching effects. A, radioactivity levels for the 45Ca2+ tracer; calcium undergoes rapid turnover in the bloodstream, with consequent rapid loss of radioactivity. B, radioactivity levels for [14C]urea, also with rapid turnover and loss of radioactivity. C, radioactivity levels for 36Cl, with a much slower decline, indicating a slower turnover of Cl in the blood. D, radioactivity levels for [35S]cysteine, with a similar pattern of slow turnover demonstrated by a slow decline in radioactivity.

Studies of Tracers in Hair. Twenty-one-day-old C57 black and BALB/c mice were obtained from The Jackson Laboratory. For each tracer experiment, five C57 and five BALB/c mice (total of 40 mice for four tracer experiments) were injected ip daily, for 4 days, with the tracer solutions described above. Mice were maintained for the 21-day period of their second synchronized anagen. At the end of the growth period, the mice were sacrificed using CO₂, and the hair from the back was harvested with electric clippers to within 0.01 inch of the skin. This hair was then stored at 20°C until further treatment and analysis.

ICP-AES¹ Analysis of Additional Cations. To facilitate the analysis of cations in the hair that could not be administered as radiotracers, hair from undosed mice was analyzed by ICP-AES by Trace Minerals Inc. (Boulder, CO). Thirty milligrams of hair were combined with 1 ml of concentrated nitric acid and digested for 15 min. Samples were then heated to 85°C and vortex-mixed for 1 hr. After digestion, 0.25 ml of 30% hydrogen peroxide and then, after 5 min, another 0.25 ml of 30% hydrogen peroxide were added, followed by an additional 15 min of heating. Samples were then diluted to 5 ml with distilled deionized water. Samples were analyzed with an ARL model 3560 ICP-AES instrument (Busch B, Trace Minerals Inc., personal communication).

Statistical Analysis. Data sets were analyzed by one-way analysis of variance, using Statgraphics 6.0 software (Manugistics, Rockville, MD). Data sets were also tested by a least-significant difference multiple-range test. Significance was assumed at the  = 0.01 level for all analyses.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

The time course of serum radioactivity levels was initially assessed for each of the four analytes used, so that an average specific activity for each tracer could be determined. The data in fig. 1 illustrate the time courses of radioactivity levels in the blood of mice after a single dose of each of the tracer solutions. Fig. 1A presents the data for radioactivity levels in blood from the ⁴⁵Ca²⁺-dosed animals; these data indicate a rapid turnover of Ca²⁺ in the blood and thus a rapid loss of radioactivity over 12 hr. Fig. 1B presents the data for [¹⁴C]urea in blood, with a comparable rapid turnover of urea and loss of radioactivity. The radioactivity data for ³⁶Cl (fig. 1C) show much slower turnover of the Cl pool and thus a slow loss of radioactivity. Lastly, the results for [³⁵S]cysteine (fig. 1D) display a slow turnover of the cysteine pool and a slow loss of radioactivity.

Because normal serum constituents such as Ca²⁺, Cl, urea, and cysteine are maintained at constant levels by the animals (Green, 1966) but turnover and recycling of the pools of these constituents occur, the specific activities of the tracers injected into the animals change with time, as shown in fig. 1. Therefore, to calculate the concentration of each constituent in the hair, an estimate of the mean specific activity in the serum was needed. The mean specific activities in table 1 were then used to calculate the concentrations of each constituent in the hair (fig. 2).

View larger version (37K): [in this window] [in a new window]   Fig. 2.   Comparisons of hair concentrations of serum constituents estimated by radiotracer experiments. Bars, means of three samples/animal and five animals/group; upper error bars, 1 SD for wash concentrations; lower error bars, 1 SD for hair concentrations after various treatments. A, calcium results were significantly ( = 0.01) different between C57 black and BALB/c white hair. In all cases, relatively little 45Ca2+ could be removed by any treatment. B, [14C]urea results showed no significant differences in hair concentrations or degree of washout among either strains or treatments. Approximately 50% of the total urea could be removed by each treatment. C, C57 black hair contained significantly ( = 0.01) more 36Cl than did BALB/c white hair. The phosphate extraction was able to remove almost all 36Cl, whereas methanol removed far less of the total 36Cl present. D, C57 black hair contained significantly more [35S]cysteine than did BALB/c white hair. Almost no [35S]cysteine could be removed from either strain by any of the treatments.

After the average specific activities were assessed, in vivo experiments were performed. Hair samples were collected, briefly washed, and then subjected to extensive phosphate extraction, to determine the accumulation and extraction of radioactivity in the hair. Fig. 2A presents the concentrations of ⁴⁵Ca²⁺ in hair and solution after a brief wash with water, after a 24-hr extraction in phosphate buffer, and after a 24-hr extraction in methanol. The concentrations of ⁴⁵Ca²⁺ lost to the solutions in each of these treatments are represented as nanograms per milligram of hair. C57 black hair demonstrated significantly higher ⁴⁵Ca²⁺ concentrations than did BALB/c hair. Additionally, only 10-15% of the deposited ⁴⁵Ca²⁺ could be removed from the C57 hair by any treatment, whereas approximately 25% of the ⁴⁵Ca²⁺ deposited in BALB/c hair could be removed by any treatment.

Fig. 2B illustrates the concentrations of [¹⁴C]urea in hair (as nanograms per milligram of hair) after the same treatments as described above. Concentrations of [¹⁴C]urea in the hair far exceeded the concentrations of the other constituents examined. No significant differences between strains were noted. Also, no significant differences were noted in the amounts of urea removed by the treatments; approximately 50% of the [¹⁴C]urea deposited was removed by each of the treatments.

Fig. 2C presents the ³⁶Cl concentrations in hair (as nanograms per milligram of hair) after the same treatments as described above. Less ³⁶Cl was deposited into the hair of both C57 and BALB/c mice, compared with the other constituents examined. C57 black mice demonstrated significantly higher concentrations of ³⁶Cl than did BALB/c white mice. Of the ³⁶Cl deposited in hair, 85% was removed from the black hair by a 24-hr extraction in phosphate buffer and 81% was removed from the BALB/c hair by the same treatment. Less ³⁶Cl was removed by a 24-hr extraction in methanol (28% from black hair and 40% from white hair).

The concentrations of [³⁵S]cysteine in hair are shown in fig. 2D. The hair from C57 black mice contained significantly more [³⁵S]cysteine per milligram of hair than did the hair of BALB/c mice ( = 0.01). In all cases, <2% of the deposited [³⁵S]cysteine could be removed from the hair by any treatment.

To complete the assessment of endogenous cation contents of mouse hair, total concentrations of several ions were determined in control hair by ICP-AES. The concentrations of unlabeled Ca²⁺ in untreated, water-extracted, and choline-extracted hair, as determined by ICP-AES, are shown in fig. 3A. As with the ⁴⁵Ca²⁺ concentrations, the C57 hair contained significantly more Ca²⁺ than did the BALB/c hair. Only 100 mM choline was able to displace a significant amount (10% of the original amount) of Ca²⁺ from the C57 hair. Of the Ca²⁺ deposited in BALB/c hair, 50% could be removed from the hair by the water extraction and 83% by the choline extraction; these differences were significant. Ca²⁺ concentration estimates in hair obtained by radiotracer analysis were approximately 30% of the unlabeled concentration determined by ICP/AES for both BALB/c and C57 mice.

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Comparisons of hair concentrations of Ca2+, Mg2+, K+, and Na+ measured by ICP-AES after various treatments. Bars, means of three animals; error bars, 1 SD. A, calcium concentrations measured by ICP-AES were approximately 3 times the values estimated by the radiotracer method, demonstrating that approximately one third of the Ca2+ pool was effectively labeled by 4 days after injection. Again, black hair contained significantly more Ca2+ than did white hair, and only choline could displace a small amount of the Ca2+ present. B, magnesium data were similar to those for Ca2+, and black hair contained significantly more Mg2+ than did white hair. As with Ca2+, only a small portion of the Mg2+ could be removed from the hair. C, C57 black hair contained a slightly but significantly greater concentration of K+ than did white hair. Most of the K+ could be removed from the hair by both water and choline extractions. D, no significant differences in Na+ concentrations were observed between black and white hair. Both the water extraction and the choline extraction were able to remove almost all of the Na+ present.

Concentrations of Mg²⁺ in hair, as determined by ICP-AES, are shown in fig. 3B. As with Ca²⁺, Mg²⁺ concentrations in C57 black hair were significantly greater than those in BALB/c white hair, although Mg²⁺ concentrations were considerably lower than Ca²⁺ concentrations The water extraction removed 15% of the original concentration of Mg²⁺, and the choline extraction removed 26% from the C57 hair. Both of these differences were significant. Essentially all of the Mg²⁺ could be removed from the BALB/c white hair by both the water extraction and the choline extraction.

Potassium concentrations determined by ICP-AES are presented in fig. 3C. The K⁺ concentrations were comparable to Ca²⁺ concentrations in C57 black hair, but K⁺ concentrations were much greater than Ca²⁺ concentrations in BALB/c white hair. C57 hair contained significantly more K⁺ than did BALB/c hair. Water extraction removed 90% and choline extraction 95% of the K⁺ originally measured in the black hair. Both the water and choline extractions removed >98% of the original K⁺ concentration in BALB/c white hair.

Fig. 3D presents the Na⁺ concentrations in hair. No significant differences in Na⁺ concentrations were noted between the two strains. Hair Na⁺ concentrations for both strains were comparable to hair Mg²⁺ concentrations in C57 mice but were considerably lower than hair K⁺ concentrations or Ca²⁺ concentrations in black hair. Sodium concentrations were below the detection limit (0.0307 ng/mg of hair) in BALB/c white hair extracted with both water and choline. The water extraction removed 95% of the original Na⁺ concentration, and the choline extraction lowered Na⁺ concentrations below the detection limit (0.0211 ng/mg of hair).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The use of an animal model in this study offered several advantages in the investigation of deposition of compounds into hair. Inbred strains have less interindividual genetic variation, thus facilitating statistical analysis of the results. The use of C57 black mice, an almost exclusively eumelanin-pigmented strain, and BALB/c mice, a nonpigmented strain (Hamilton et al., 1974), allowed us to assess the effects of pigment on deposition. The pigment in the hair of C57 mice is predominantly eumelanin, except for a few hairs around the mammae, ears, and perineum (Hamilton et al., 1974), which were not harvested. Mice also undergo a period of synchronized anagen starting at 23 days of age (Green, 1966). This ensured that the hair was growing during the dosage period. Mouse hair, unlike human hair, does not have the potential to be damaged by harsh detergents, UV radiation, or cosmetic agents. Therefore, variables that could confound the examination of deposition interactions could be tightly controlled.

Our data demonstrate that cationic, anionic, and neutral species are incorporated into the hair matrix from the circulation. The data clearly show significant strain differences in the deposition of ionic molecules. The greater deposition of Ca²⁺, Cl, Mg²⁺, and K⁺ in the C57 black hair, compared with the BALB/c white hair, suggests that these ions may be associated with eumelanin. This is consistent with observations of melanin having a high affinity for metal ions (6 × 10²⁰ binding sites/g of melanin) (Enochs et al., 1997). Numerous authors have also reported the association of natural melanin with a variety of bound metals (Potts and Au, 1976; Froncisz et al., 1980; Sarna et al., 1976, 1980; Lyden et al. 1984; Sky-Peck, 1990). Additionally, Potts and Au (1976) reported that higher-atomic weight cations have a higher affinity for melanin, which is also consistent with our observation of relative hair concentrations of Ca²⁺ > Mg²⁺ > Na⁺.

Calcium, which was one of the most abundant cations we measured in hair (figs. 2A and 3A) and which was also reported as the most abundant by Sky-Peck (1990), demonstrated an affinity for the hair matrix more closely resembling that of [³⁵S]cysteine (fig. 2D), a compound that is covalently incorporated into keratin, than that of more exchangeable ions such as Na⁺ and K⁺ (fig. 3, C and D). Only a small percentage of the Ca²⁺ measured in hair could be displaced by another cation such as choline. This suggests that anionic sites paired with Ca²⁺ would not be readily available for ion exchange with cationic drug species. The most displaceable of the cations, Na⁺, constituted the smallest portion of the total cationic contribution observed, suggesting that the easily displaceable cation-exchange capacity of the hair is limited.

Potts and Au (1976) reported that Na⁺ and K⁺ have little or no affinity for melanin; this is consistent with our results showing no significant differences in Na⁺ concentrations and small differences in K⁺ concentrations between black and white hair. This observation, coupled with the large color-related differences in Ca²⁺ and Mg²⁺ concentrations, suggests that the primary ions associated with melanin are Ca²⁺ and Mg²⁺. However, these ions were not readily displaceable, making it unlikely that drugs could compete for these ionic binding sites. This suggests that ionic associations of drugs with the hair matrix would be more easily made with sites that are associated with Na⁺ and K⁺. Sarna et al. (1980) found that the polymeric hydroxyindoles of natural melanin are 50% (by weight) associated with protein. Therefore, drugs may have both ionic associations directly with the carboxylic or phenolic hydroxyl groups of melanin (Froncisz et al., 1980; Sarna et al., 1980), requiring displacement of Ca²⁺ and Mg²⁺, and ionic associations with melanin-associated protein, requiring displacement of Na⁺ and K⁺. Melanin binding observed by Joseph et al. (1997) and Cone and Joseph (1996) may involve these two binding compartments, i.e. one compartment of less-available melanin sites and another compartment of more-available melanin-associated protein sites. Both of these compartments would be larger with increasing quantities of melanin.

Additionally, the observation that calcium was not displaced by a variety of treatments suggests that drug species ionically associated with melanin in the hair may be significantly occluded from extraction. Such interactions would be consistent with observations by Kidwell and Blank (1) that externally applied cocaine was retained by the hair after extraction. This suggests that only small fractions of Ca²⁺ and Mg²⁺ can displaced from the hair matrix by organic cations such as choline or cocaine. Such an interaction suggests that noncovalently bound serum constituents can be significantly occluded from extraction.

We also found that an anion, Cl, had significant access to the hair matrix and also demonstrated a pigment-dependent relationship. Unlike calcium, Cl was easily removed from the hair by an aqueous buffer, suggesting that ionic sites paired with Cl would be available for ion exchange with negatively charged drugs. This is not consistent with reports of little incorporation of negatively charged drugs such as ⁹-tetrahydrocannabinol (Cone and Joseph, 1996; Uhl, 1997). This observation suggests that, even though ionic sites within the hair matrix are available for ion exchange with anions, these drugs are excluded from the hair matrix by other mechanisms.

Our data also indicated that a neutral organic species, i.e. urea, had significant access to the hair matrix and was incorporated similarly in pigmented and nonpigmented hair. Although the presence of urea in the hair could indicate urine contamination, the dramatically inverted Na⁺/K⁺ ratio measured in hair, relative to the ratio in urine, argues against significant urine contamination. The K⁺ and Na⁺ concentrations in hair presented in fig. 3, C and D, demonstrate that the K⁺/Na⁺ ratio in hair (K⁺/Na⁺ = 6.25) is the inverse of this ratio in urine (K⁺/Na⁺ = 0.74) (Green, 1966), indicating that urine contamination of the hair was minimal. Additionally, the control experiments using Cl demonstrated that undosed mice housed with dosed mice experienced minimal contamination (<10% of the dpm per milligram of hair values, compared with the dosed mice); this confirms that the hair was not contaminated by urine. These results are consistent with the presence of a neutral compartment, independent of pigment, in the hair matrix with a potentially large storage capacity.

One half of the pool of urea in the hair was removed by aqueous extraction or methanolic extraction. This indicates that the pool is externally accessible and that a portion of a neutral species associated with hair may be significantly occluded from extraction and tightly bound to the hair. The observation of nondisplaceable pools of constituents is significant.

Table 1 compares serum concentrations with hair concentrations for each constituent examined in this study. This comparison shows that large portions of these constituents are incorporated into hair, relative to serum. The results for [³⁵S]cysteine are consistent with those for a compound that is actively incorporated into the hair matrix. These results suggest that the capacity of hair to store neutral compounds is greater than that for ionic compounds and that this association is independent of color. The relative extent of deposition of these constituents was cysteine > urea > Ca²⁺ > Cl, an order that does not parallel relative serum concentrations. This also supports the existence of a large neutral storage compartment in hair.

A comparison between the serum analytes and several drugs and metabolites shows primarily that drug concentrations in hair are substantially lower than the serum analyte concentrations. Although these compounds are organic cations (cocaine, codeine, and morphine) or a zwitterion (benzoylecgonine) capable of ionic interactions, they are relatively precluded from the hair matrix, compared with the endogenous constituents. This suggests that these ions cannot effectively compete with Ca²⁺ and Mg²⁺ for ionic binding sites on melanin.

Table 2 presents a comparison of serum AUC values for several drugs and metabolites with hair concentrations of each compound. From the serum AUC/hair ratios, it is evident that deposition of the drugs into hair varies widely from compound to compound. The more water-soluble compounds benzoylecgonine and ecgonine methyl ester demonstrate less accumulation in hair than do more lipid-soluble compounds.

However, none of the drugs demonstrate as much incorporation as urea. This supports the existence of a nondisplaceable, pigment-independent compartment of accumulation within the hair matrix. The greater association of a neutral compound with hair suggests that ion exchange and ionic association cannot fully account for chemical modes of drug deposition.

The very low measured concentrations of drugs in hair might also be in part attributable to very strong chemical interactions between drugs and hair matrix components. The digestion of hair either by strong base or by proteolytic enzymes might reduce the length of the proteins but leave protein-drug interactions intact, limiting the concentrations of drug that could be dissociated from the hair.