High-Performance Liquid Chromatography-Based Determination of Total Isothiocyanate Levels in Human Plasma: Application to Studies with 2-Phenethyl Isothiocyanate

Analytical Biochemistry 291, 279 –289 (2001) doi:10.1006/abio.2001.5030, available online at http://www.idealibrary.com on

High-Performance Liquid Chromatography-Based Determination of Total Isothiocyanate Levels in Human Plasma: Application to Studies with 2-Phenethyl Isothiocyanate 1 Leonard Liebes,* ,2 C. Clifford Conaway,† Howard Hochster,* Sandra Mendoza,* Stephen S. Hecht,‡ James Crowell,§ and Fung-Lung Chung† *Division of Oncology, Department of Medicine, New York University Medical Center, New York, New York 10016; ‡University of Minnesota Cancer Center, Minneapolis, Minnesota 55455; †American Health Foundation, Valhalla, New York 10595; and §Division of Cancer Prevention, National Cancer Institute, Bethesda, Maryland 20892

Received November 27, 2000; published online March 9, 2001

Dietary and pharmacologic isothiocyanates (ITCs) may play a role in reducing the risk of certain cancers. The quantification of ITCs in humans is important both for epidemiological and pharmacokinetic studies. We describe a modification of an HPLC-based assay of urinary ITCs for use with human plasma. The assay utilizes the cyclocondensation reaction of 1,2-benzenedithiol with ITCs present in human plasma, followed by a twostep hexane extraction and analysis by HPLC using UV detection at 365 nm. The method shows linearity and reproducibility with human plasma over a range of 49 – 3003 nM phenethyl isothiocyanate (PEITC) (r 2 ⴝ 0.996 ⴞ 0.003). A similar degree of linearity was seen with two other biologically occurring conjugates of PEITC: PEITC–N-acetylcysteine (PEITC–NAC) and PEITC– glutathione (PEITC–GSH). The recovery of PEITC assessed on multiple days was 96.6 ⴞ 1.5% and was 100% for PEITC–GSH and PEITC–NAC. The reproducibility of the assay on multiday samplings showed a mean %CV of 6.5 ⴞ 0.3% for PEITC, 6.4 ⴞ 4.3 for PEITC–NAC and 12.3 ⴞ 3.9 for PEITC–GSH. In clinical studies, mean plasma ITC level of 413 ⴞ 193 nM PEITC equivalents was determined for a non-dietary-controlled group of 23 subjects. Multiday analysis data from pharmacokinetic plasma sets of 3 subjects taking a single dose of PEITC at 40 mg showed a good CV (range: 16 –21%). The applicability of the methodology to pharmacokinetic studies of PEITC in humans is demonstrated. © 2001 Academic Press 1 Work was supported by NCI-CN-55120, CA-166081-21, CRC-RR00096, and CA 46535. 2 To whom correspondence should be addressed at New York University Medical Center, Kaplan Comprehensive Cancer Center, Bellevue C & D Building, Room 556, 462 First Avenue, New York, NY 10016. Fax: 212-263-8703. E-mail: [email protected].

0003-2697/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.

Key Words: PEITC; PEITC–NAC; PEITC–GSH; phenethyl isothiocyanate; pharmacokinetics; HPLC; plasma assay; isothiocyanates.

Isothiocyanates (ITCs) 3 are effective inhibitors of tumorigenesis induced by polycyclic aromatic hydrocarbons and nitrosamines in laboratory animals (1–5). 2-Phenethyl isothiocyanate (PEITC, Fig. 1), a naturally occurring ITC found as its glucosinolate, gluconasturtiin, in watercress and other cruciferous vegetables (6), is one of the most extensively investigated ITCs. It has long been recognized that ITCs have multifold biological effects due to their chemical reactivity (7, 8). The main protective effect of ITCs against cancer in animals has been attributed to the activity of these compounds on two mechanisms: (i) inhibitory effects on phase I enzymes responsible for the bioactivation of carcinogens such as NNK (3–5) and (ii) activity as inducers of phase II detoxification enzymes (9 –12). In dietary studies in humans given large quantities of cruciferous vegetables in their daily diet, similar effects on both phase I and phase II enzymes have been observed (13–15). The actual definitive role of dietary 3 Abbreviations used: AUC, area under the curve; BDT, 1,2-benzene-dithiol; BITC, benzyl isothiocyanate; CL, clearance; C max, maximum plasma concentration; CV, coefficient of variation; DL; detection limit; HPLC, high-performance liquid chromatography; ITC, isothiocyanates; K10HL, elimination rate half-life, max, maximum; min, minimum; PEITC, phenethyl isothiocyanate; PEITC–NAC, PEITC–N-acetyl-L-cysteine; PEITC–GSH, PEITC– glutathione; QL, quantification limit; T max, time to maximum concentration.

279

280

LIEBES ET AL.

protective effects of ITCs against the development of cancer. We have developed a method for the routine analysis of PEITC and two metabolites, PEITC–NAC and PEITC–GSH (Fig. 1), in human plasma employing a condensation reaction with 1,2-benzene-dithiol (BDT) that yields an assessment of the total ITC content. The assay of the 1,3-benzenedithiol-2-thione reaction product (Fig. 2) is carried out by HPLC with UV detection at 365 nm. The assay was originally developed by Zhang and co-workers (23, 24), for determination of ITCs, although it also reacts with carbamates and dithiocarbamates. The assay was subsequently modified for assay of ITCs in urine by Chung et al. (25); this allowed the assessment of the uptake and excretion of total ITCs in several population studies (25, 26). We have modified the assay to determine total ITCs in plasma. The assay is sensitive, accurate, rapid, and sufficiently reproducible to follow the absorption and excretion of PEITC and metabolites in evolving clinical studies. FIG. 1.

Structures of PEITC, PEITC–NAC, and PEITC–GSH.

MATERIALS AND METHODS

Chemicals ITCs or their glucosinolates in human cancer prevention remains to be adequately defined. There are, however, several studies where epidemiological data have demonstrated that dietary consumption of Brassica vegetables reduces the incidence of cancers at various sites (16 –18). One recent cohort study has shown that intake of ITCs is associated with a reduced risk of lung cancer (19). It is also of interest to note that thiol conjugates of isothicyanates also inhibit phase I enzymes. The L-cysteine, glutathione, and N-acetyl-L-cysteine (NAC) conjugates of PEITC, as well as identical conjugates of 6-phenylhexyl isothiocyanate and benzyl isothiocyanate (BITC), have been found to inhibit the metabolism of NNK by mouse lung microsomes and to inhibit NNK-induced lung tumorigenesis in A/J mice (20). More than half of PEITC and BITC administered to humans are excreted as their respective NAC conjugates (21, 22). An assay is needed to determine the levels of ITCs in human plasma to more adequately assess potential

All reagents were HPLC grade and were obtained from Fisher Scientific, (Springfield, NJ). 1,2-Benzenedithiol was obtained from Lancaster Synthesis, Inc. (Windham, NH). PEITC was manufactured by Aldrich Chemical Company (Milwaukee, WI) and was obtained from the National Cancer Institute, Division of Cancer Prevention and Control drug repository managed by McKesson BioServices (Rockville, MD). PEITC–NAC and PEITC–GSH were synthesized and purified as previously described by Jiao et al. (27). Standard Assay Preparation Basic procedures. Human plasma was obtained from the NYU Medical Center blood bank and was stored in 50-ml aliquots in polypropylene tubes at ⫺20°C. Prior to being used for addition of ITC standards or precision samples, it was processed through the sample preparation and extraction procedure detailed below for assessment of background ITCs and

FIG. 2. Reaction of ITC substrate (RONACAS) in the cyclocondensation reaction with 1,2-benzenedithiol to yield 1,3-benzenedithiol-2thione. PEITC, the metabolites PEITC-NAC or PEITC-GSH, and other ITCs in a human plasma matrix, can serve as substrates in this condensation reaction.

HUMAN PLASMA TOTAL ISOTHIOCYANATE CHROMATOGRAPHIC ASSAY

dithiocarbamates. A background level of 75 nM ITCs was determined to be the upper acceptable limit for a given lot of plasma used in an assay where standards were added; the background ITC value of the plasma was subtracted from all measurements. Deionized/ distilled grade water was obtained using a Milli-Q purification system (Millipore Corp., Milford MA). Calibration curve standards consisted of PEITC concentrations of 49, 98, 251, 496, 999, 1998, and 3002 nM in plasma. Precision samples were prepared at concentrations of 98, 502, and 1998 nM as described below. Preparation of Plasma Standards PEITC was the primary source of ITCs for the plasma standards. Working stock solutions of PEITC in plasma at 200 ␮M (32.6 ␮g/ml) were prepared from a 61 mM (10 mg/ml) master stock in acetonitrile. Serial dilutions were then prepared from this secondary stock to range from 49 to 3002 nM as described above for the standard curves and precision samples. For studies with PEITC–NAC and PEITC–GSH, stock solutions of 153 and 106 ␮M were respectively prepared and working dilutions of 3, 10, and 20 ␮M used for the precision and calibration curves at the concentrations stated above. To assess the effects of different ratios, a series of four mixtures of PEITC, PEITC–NAC, and PEITC– GSH in 1 ml of plasma were prepared. Equimolar 500 ␮M amounts, and 2:1 (1000:500 ␮M) amounts of respectively PEITC–GSH:PEITC–NAC, PEITC–NAC: PEITC, and PEITC:PEITC–GSH were prepared and subsequently analyzed for total ITC levels. Collection of clinical specimens Subjects donated a single blood sample (5 ml) or were sampled serially as part of their enrollment in a National Cancer Institutesponsored phase I clinical trial with NYU Institutional Review Board approval. This study focused on the single dose safety and pharmacokinetics of orally administered PEITC at a concentration of 10 mg/ml in an olive oil suspension, at doses ranging from 5 to 40 mg. Blood was serially collected in sodium-heparinized, blood Vacutainer tubes (Becton–Dickinson, Franklin Lakes, NJ) from three subjects taking a single 40-mg dose of PEITC. For these detailed pharmacokinetic studies, samples were collected at the NYU General Clinical Research Center at the following time points: baseline at screening visit, prior to taking PEITC, and at 7.5, 15, 30, 45 min; 1, 1.5, 2, 3, 4, 6, 8, and 24 h after dosing. The plasma was collected after centrifugation at 1000g for 10 min at 4°C. The supernatant was transferred to cryovials and was immediately stored at ⫺20°C until further analysis. Sample Preparation and Extraction Procedure Plasma (1 ml) was incubated in a 10 ml glass screw top tube at 65°C for 1 h after the addition of 0.5 ml 0.1

281

M potassium phosphate (pH 5.5) and 0.8 ml of 10 mM 1,2-benzenedithiol in isopropanol. After the addition of both of these solutions, they were then mixed thoroughly for 1 min with a vortex mixer. Immediately after the 1-h incubation period, while the sample was still warm, 2.0 ml of n-hexane was added and the sample was vortex mixed in the capped tube for 1 min, followed by centrifugation at 1200g for 8 min. A 2.0-ml portion of the hexane phase was removed and was placed in a 12 ⫻ 75-mm glass tube. A second extraction was carried out by adding 2.0 ml n-hexane and mixing thoroughly for 1 min with a vortex mixer. Centrifugation was repeated as above and the second hexane extract (2.0 ml) was combined with the first extract. The combined extracts were evaporated to dryness using a Speed-Vac (Forma Scientific, Hauphauge, NY). The residue was then dissolved in 0.5 ml of 7:3 (v/v) methanol/H 2O by vortex mixing for 15 s. The methanol/ water-reconstituted sample was centrifuged at 6000g for 4 min and was placed in 1.0-ml autosampler vials for analysis by HPLC. For the case of spectral confirmation of the baseline ITC condensation product, four separate 1-ml plasma extractions from the same subject were combined together. Chromatographic Procedure The chromatographic equipment consisted of a ThermoQuest UV/VIS detector Model 8450 set at 365 nm (ThermoQuest, Austin, TX) along with a Waters Model 6000 pump and 712B auto-sampler (Waters Associates, Milford, MA). The data system/pump controller was an Axxiom 747 multi-system controller/data system (Cole Scientific, Moorepark, CA). For collection of UV spectral scans, a Waters 990 photodiode array detector was employed. A 4.6 ⫻ 250-mm, 5 ␮ Phenosphere ODS2 C18 column (Phenomenex, Torrance, CA) was eluted isocratically at a flow rate of 0.7 ml/min with 85/15 (v/v) methanol/water, with the entire mixture containing 0.1 M triethylamine. The pH of the mobile phase was adjusted to 7.0 with 85% phosphoric acid and then filtered and degassed through a 0.22-␮m, 47-mm diameter nylon filter. Aliquots (200 ␮l) of the cyclocondensation reaction product were injected using the autosampler. Validation and Clinical Analysis Protocols A multiday study of the reproducibility of the assay was carried out with a series of plasma precision and standard curve samples prepared as a batch and stored frozen (⫺20°C) until analysis. This was repeated on multiple days to define the interday variation of the assay. The intraday assay variation was carried out with the analysis of triplicate samples of each of the three levels of precision samples processed and analyzed in one day.

282

LIEBES ET AL.

System Control and Data Reduction A master sequence program was used to control the injection and analysis of up to 50 plasma samples. The ordering of the standard curve, precision, and clinical samples for a given automated analysis cycle was such that there was an equal interspersion of the different samples. Each analysis run was keyed to a unique raw data file with the user-defined baseline and integration parameters for the calculation of the chromatographic peak area. The chromatographic areas were transferred to a Microsoft Excel 98 (Microsoft, Redmond, WA) spreadsheet for calculation of average, standard deviation, and coefficient of variation. A linear regression of the concentration vs area data was prepared using Cricket Graph III (version 1.53, Computer Associates, Garden City, NY). An iterative two-step process was employed for the generation of the regression lines of the standard curves. First, a calculation of the regression line using all of the initial data was carried out. A second regression line was then generated with only those data that provided the best linear fit, using the criteria that at least five data points were used for the generation of the standard curve. The precision values and back calculations were carried out using this “best fit line.” The detection limit (DL) of total ITC levels is based on a 2:1 signal above the plasma (endogenous) baseline level which is subtracted from all analyses specific to this plasma, and the quantification limit (QL) is based on 10:1 signal above this baseline value. The decay of the total ITC levels in plasma following a single 40-mg dose of PEITC were fitted to a single compartment first-order model available in WinNonlin, Standard Version 1.5 (Pharsight Corp., Mountainview, CA). Statistical comparisons between groups used nonparametric tests available in Statview Version 5 (SAS Institute, Research Triangle, NC). The Mann–Whitney U test was used for paired groups and the Wilcoxon signed rank test was utilized to compare unpaired groups. RESULTS

Chromatography Figure 3 presents representative results of the HPLC analysis of the cyclocondensation product for PEITC added to plasma at two concentrations. The typical retention time for the cyclocondensation product was 9.1 ⫾ 0.1 min. As detailed under Materials and Methods, Fig. 3 graphically also shows the small amount of cyclocondensation product peak present in plasma alone that was accounted for by a correction in all calculations. A closer examination of the intersubject variability of this background is described below. Linearity of Plasma Curves on Multiple days Figure 4 shows a plot of four plasma standard curves run on different days. These results demonstrate the

FIG. 3. Chromatographic profile of a human plasma samples: no addition of PEITC (—); 100 nM PEITC added (䡠 䡠 䡠) and 2.04 ␮M PEITC added (- - -). The 1,3-benzenedithiol-2-thione conjugate product elutes at a retention time of 9.1 min, while the peak at 4.0 min is 1,2-benzenedithiol.

linearity of standard curves of the cyclocondensation reaction product prepared from plasma containing PEITC levels over a range of 49 to 3002 nM PEITC equivalents. Table 1 summarizes the descriptive statistics for the standard curves plotted in Fig. 4. The linearity was consistent over the different assay days (r 2 ⫽ 0.996 ⫾ 0.003). The small differences in the y-intercept values (Fig. 4) may reflect variances that occur upon storage of the standard stocks used for varying amounts of time, and differences in the baseline ITC content of the plasma stocks. The range of variance from 0.4 to 15.4%, along with the interday percentage bias range of 0 to 12.1% indicates a good consistency and accuracy for the interday standard curves. Precision A summary of the precision standards for the three possible forms of PEITC present in plasma (PEITC, PEITC–NAC, and PEITC–GSH) is presented in Table 2. The three levels were chosen to define the assay precision at the expected range of subject baseline and peak levels following an oral dose of PEITC ranging from 5 to 40 mg. The assay showed consistent precision for each of the PEITC conjugates at the three concentrations examined over the three different assay days (mean overall %CV ⫽ 6.5 ⫾ 0.3) with individual mean standard deviations (SD) of 6.3, 31.7, and 130 nM, respectively, for the 98, 502, and 1998 nM concentra-

283

HUMAN PLASMA TOTAL ISOTHIOCYANATE CHROMATOGRAPHIC ASSAY

FIG. 4. Representative standard curves for PEITC run on different days. Plasma samples containing PEITC additions over the range of 49 to 3003 nM were subjected to the BDT cyclocondensation reaction. Each plotted point is the average of duplicate injections for the 1,3-benzenedithiol-2-thione cyclocondensation product.

tions of PEITC in plasma. For PEITC–NAC, an overall mean %CV of 6.4 ⫾ 4.3% was calculated with individual mean SDs of 2.8, 62, and 109 nM, respectively, for the same concentrations of PEITC–NAC. With respect to PEITC–GSH, a higher overall mean %CV of 12.3 ⫾ 3.9 was calculated, and higher mean SDs, respectively, of 80 and 158 nM for the 502 and 1998 levels. Data for the low end of the standard curve on day 3 of the PEITC–GSH assessment were found to fall off the linear regression line, and precluded calculation of the 98 nM level.

Recovery The recovery of PEITC, PEITC–NAC, and PEITC– GSH from human plasma employing the two-step extraction protocol was assessed from the data summarized in Table 2. The results show that the mean interday recovery over these three concentration ranges for PEITC, PEITC–NAC, and PEITC–GSH were, respectively, 96.6 ⫾ 1.5, 104.2 ⫾ 7.8, and 105.4 ⫾ 13.3%, in good agreement with the expected concentration values. The data in Table 2 also show that the

TABLE 1

Descriptive Statistics of Plasma Multiday Linearity of PEITC Cyclocondensation Reaction Product (1,3-Benzenedithiol-2-thione) r2 Interday, expected nM Mean, nM SD % SD % Bias

49 56 8.6 15.4% 12.1%

98 97 6.7 6.9% ⫺0.6%

251 236 28.2 11.9% ⫺6.5%

496 552 44.1 8.0% 10.1%

999 953 75.4 7.9% ⫺4.8%

1998 1998 61.9 3.1% 0%

3002 3003 13.5 0.4% 0%

0.996 0.003 0.3%

284

LIEBES ET AL. TABLE 2

Reproducibility of PEITC and PEITC Conjugate Precision Standards 98 nM expected Day

Intra- (mean)

Intra- (SD)

502 nM expected %CV

Intra- (mean)

1998 nM expected

Intra- (SD)

%CV

Intra- (mean)

Intra- (SD)

%CV

32.1 28.2 34.9 31.7

6.4 6.3 7.2 6.6

2026 1827 2039 1964 (⫺1.7%)

31.3 146 182 130

1.5 8.0 8.9 6.2

54.1 9.3 122.6 62.0

10.6 1.7 21.5 11.3

2248 2302 2054 2201 (10.2%)

126.5 119.8 80.4 109

5.6 5.2 3.9 4.9

83.7 74.3 82 80.0

16.5 16.4 14.9 15.9

1838 1917 1966.5 1907 (⫺4.5%)

72.3 285.8 115 157.8

3.9 14.9 5.8 8.2

PEITC 1 2 3 Interday (1–3) Mean (bias)

92.5 93.8 93.8 93.3 (⫺4.8%)

7.4 7.4 4.3 6.3

7.9 7.8 4.6 6.8

506 455 488 483 (3.7%) PEITC-NAC

1 2 3 Interday (1–3) Mean (bias)

76.6 110.2 93.8 93.5 (⫺4.5%)

1.4 2.8 4.3 2.8

1.9 2.6 4.6 3.0

512 531.8 569.3 537.7 (7.1%) PEITC-GSH

1 2 3 Interday (1–3) Mean (bias)

119.9 116.4

(20.6)

15.6 14.8

13.0 12.7

506 453 551 503 (0.3%)

percentage bias of the precision standards ranged from ⫺1.7 to 3.7% for PEITC, ⫺4.5 to 10.2% for PEITC– NAC, and ⫺4.5 to 20.6% for PEITC–GSH. Specificity of Conjugate Product and Linearity with Other PEITC Metabolites The ability of the assay to detect ITC moieties in PEITC metabolites is shown in Fig. 5 for PEITC-NAC and PEITC–GSH compared with PEITC. The slopes of the linear plots obtained for PEITC and each of the conjugates were comparable. An assessment was also made on possible effects of varying amounts of PEITC, PEITC–NAC, and PEITC–GSH at four different mixtures each totaling 1500 ␮M total ITC in plasma. These were designed to mimic the levels expected to occur after subjects had ingested PEITC at amounts of 40 mg. Table 3 summarizes the results of the analysis of a single set of these mixtures carried out on a single day that show no discernable effect from the varying ratios of PEITC moieties. An overall mean of 1559 nM ITC equivalents for these four mixtures was calculated with a range of 1394 to 1682 nM; the values which bracketed the theoretical target value of 1500 nM ITC equivalents. The mean percentage bias was 3.9% for these mixtures (range ⫺7.1 to 12.1%), and was comparable to the data summarized in Table 2 for the individual preparations of PEITC, PEITC–NAC, and PEITC–GSH.

Determination of Endogenous Plasma ITC Levels A limited assessment of the intersubject variation and performance of the assay for baseline ITC levels was carried out. The analysis was conducted on two consecutive days to assess the reproducibility of the assay and to determine possible effects of storage and handling of blood samples from six subjects recruited from laboratory personnel. The detected cyclocondensation product range of 344 nM was from a min–max of 439 to 783 nM PEITC equivalents for the first analysis day, with a mean of 603 ⫾ 125 nM (Fig. 6A). The repeat analysis of the same samples (that had been refrozen and thawed again) resulted in a lower mean (529 ⫾ 84 nM) as well as a lower range of 243 nM in the data (min–max, 434 – 677 nM). However, this effect between the two different analysis days of the same samples did not test to be significantly different (P ⫽ 0.2, Mann– Whitney U test), probably due to the small sample size. As seen in Fig. 6A, a downward trend for each sample is apparent (range of reduction, 1.1–22%), suggesting perturbations on the amount of condensation reaction detected following one cycle of freezing and thawing of plasma. Another baseline ITC assessment was conducted with a larger sample pool comprised of subjects enrolled in our single dose phase IA study of PEITC. We sought to assess the consistency of the assay on repeat plasma samples that were obtained from 17 subjects for baseline ITC levels. The samples were drawn on two separate days, but for consistency were

HUMAN PLASMA TOTAL ISOTHIOCYANATE CHROMATOGRAPHIC ASSAY

285

FIG. 5. Plot of a range of plasma samples containing PEITC (⫹), PEITC–GSH (E), and PEITC–NAC (〫) from 98 to 3002 nM. Each of the samples was analyzed for the cyclocondensation reaction product in duplicate. Slopes were comparable for each of the substrates (range of 8 –10 ⫻ 10 ⫺4 A/nM).

then processed for ITC content on the same assay day. A preliminary examination of the results showed high levels with a range of 1641 nM (min–max, 524 –2165 nM) in the initial screening sample for 6 subjects. However, when the second plasma samples obtained ⬃1 week later were analyzed, markedly lower levels, with a range of 467 nM (min–max, 52–518 nM) were determined. The mean of these two samplings showed a %CV ⬎ 50%. Since the first sample was obtained prior to the time that the study subjects had been counseled to refrain from eating crucerifous vegetables, we only included data from the second plasma collection in the summary plotted in Fig. 6B for this

small subgroup (subjects 1– 6). The data plotted for subjects 7–17 show that the two weekly measurements in the other 11 subjects had a 4 –35% variation between the two measurements and a mean CV of 16%. We used the mean of the two measurements from these 11 subjects and the second sample data from the 6 subjects discussed above to calculate a mean of 346 ⫾ 168 nM for this study group. Although a much wider range of 731 nM in the data was found (min–max, 52–783 nM) compared with the smaller laboratory volunteer group plotted in 6A, the two group means overlap one another (Wilcoxon signed rank test, P ⫽ 0.17). Combining the data from these two groups results in a

TABLE 3

Comparison of the Effects of Mixtures of PEITC, PEITC–NAC, and PEITC–GSH in Plasma Mixture

PEITC (nM)

PEITC–NAC (nM)

PEITC–GSH (nM)

1 2 3 4

500

500 500 1000

500 1000

500 1000

Calculated results (nM)

500 Mean:

a

Based on a theoretical total of 1500 nM.

1575 1584 1394 1682 1558.8

Bias, a % 5.0 5.6 ⫺7.1 12.1 3.9

286

LIEBES ET AL.

FIG. 6. Range of baseline plasma ITC levels from a small group of laboratory volunteers and subjects entered into the phase IA study of PEITC. Scatter plot: (A) Subject numbers from 1 to 6 show the values from one collection day analyzed on two sequential days. (B) Scatter plot: Baseline samples from subject entered into phase I single-dose study. Subjects 1– 6, data are from second plasma sample obtained, data for subjects 7–17 show values at initial screening and 1 week later after dietary restriction.

mean of 413 ⫾ 193 nM (n ⫽ 23). This high %CV of 47% is an indication of the need for careful dietary restrictions when defining a baseline plasma range. A photodiode array scan UV spectrum of the peak at retention time 9.1 min of several of the baseline plasma samples showed a wavelength maximum at 365 nm, identical to that of the BDT cyclocondensation product (data not shown). Sensitivity The data summarized in Fig. 1 and Tables 1 and 2 show that the DL limit is 20 nM, while the QL was 98 nM for PEITC and PEITC–NAC. For PEITC–GSH the QL was 250 nM. Biological Reproducibility A further test of the quality of data generated as part of our phase I clinical study with PEITC was conducted as follows: Three pharmacokinetic plasma sets from the highest PEITC dosing group (40 mg) were analyzed in duplicate. The average %CV for all of the samples in the duplicate clinical sets was 21, 20, and 16% for each of the three subjects with a mean CV for the group of 19%. The 40-mg PEITC plasma elimination data sets were subjected to nonlinear modeling using a single compartment oral absorption model with a time delay parameter of 1 ⫾ 0.8 h that provided the best fit to the

data. Each data set containing the mean values for each of the three subjects was corrected using the average of two separate baseline determinations on two different days, and each was fit well using a one compartment oral absorption nonlinear model (Fig. 7). Table 4 summarizes the mean pharmacokinetic secondary parameters derived from these analyses. With the exception of the half-life estimates, the range found for the standard deviation of the parameter estimates, 26 –36%, was slightly higher, but paralleled that determined for the duplicate data sets described above. DISCUSSION

This assay provides a sensitive and reproducible method for PEITC and PEITC conjugate levels in human plasma. In addition, the assay is amenable to multiple processing of up to 50 samples in one batch. Plasma samples can be stored frozen at ⫺20°C, but repeat freeze–thawing affects the amount of detected cyclocondensation product. The assay is linear over the range of 49 to 3002 nM for PEITC, and it showed the same stoichiometry for PEITC–GSH and PEITC–NAC over the range of 98 to 3002 nM. The recoveries assessed on multiple days ranged from 96.7 to 105% for PEITC and its conjugates consistent with that found for ITC substrates reacting in the cyclocondensation reaction in buffer alone (24). The assay is dependent upon the use of a plasma standard curve that is pre-

287

HUMAN PLASMA TOTAL ISOTHIOCYANATE CHROMATOGRAPHIC ASSAY

FIG. 7. Plot of pharmacokinetic decay of three subjects after a single dose of 40 mg PEITC. The data were fit to a single compartment oral absorption model with a time delay ranging between 0.2 and 1.8 h.

pared along with each plasma set analyzed; the plasma should be screened for a low background of reactive thiols and ITCs. This background can be attributed primarily to ITCs, but could result to a minor extent from other reactive thiocarbamates present in the plasma (24). The assay is consistent on an intraday and interday basis. The mean coefficient of variation of 6.5% for the precision values assayed on three different days was comparable with the daily %CV for each of the three concentration levels evaluated (98, 502, and 1998 nM). This %CV also compared well with the mean CV of 6.4% found with the standard curve data sets generated on multidays. A slightly higher %CV and mean SD was determined for the PETIC–GSH conjugate data, which could possibly be attributed to purity and stability of the PEITC–GSH stock solutions. We observed that normal human plasma contains a considerable range of ITC-derived 1,3-benzenedithiol2-thione cyclocondensation products. Based on a personal communication (28), we were aware of a potential source of contaminant reactive dithiocarbamates present in rubber stoppers of Becton–Dickinson lavender-top blood collection tubes (K 3EDTA). We investigated what effect on the cyclocondensation reaction

could occur from an extreme exposure to contamination from the rubber stopper in the blood collection tubes used in this study (green-top, sodium heparin). For the case of total contact of plasma for 2 h at room temperature to the rubber stopper, we found an approximate doubling of the baseline level of a “tube naive” plasma sample from 43 nM to a range of 74 – 86 nM PEITC equivalents. The mean level of the endogenous plasma levels of ITCs in the combined group of 23 subjects enrolled in our phase I PEITC study and the laboratory cohort was 413 nM PEITC equivalents. We calculated that the possible effects from thiocarbamate exposure from heparinized blood collection tubes is 18 –20% of the mean baseline total ITC. Our routine processing of the collected plasma was generally within an hour after venipuncture, with the blood collection tubes remaining in a vertical position at 4°C, with minimal contact of the blood with the stopper of the tube. Based upon these precautions, we estimate that less than 10% of the baseline variability found in our subject pool could possibly be attributed to reactive thiocarbamates. Thus, other sources of plasma ITCs and related compounds probably contribute to our observed baseline levels of cyclocondensation product.

TABLE 4

Total ITC Pharmacokinetic Secondary Parameter Summary a

a

Parameter/subject

S-26

S-27

S-28

AUC (␮M ⫻ h) CL (ml/m 2 per min) K10-HL, h T max, h C max, ␮M

7.76 310 1.7 3.1 1.40

10.5 201 6.2 5.5 0.64

13.3 198 3.1 5.2 1.07

From a single 40-mg dose of PEITC.

Average

SE

10.5 236 3.7 4.6 1.04

1.61 36.8 1.3 0.7 0.22

288

LIEBES ET AL.

The UV spectrum of the cyclocondensation reaction product generated from a subject’s baseline plasma in our study showed the same spectral characteristics as that with a standard cyclocondensation product (23). Studies of oral administration of 37.3 ␮Ci 14C-labeled PEITC in rats showed that a major amount of the 14 C-labeled PEITC dose was distributed in the liver, kidneys, and intestinal organs (esophagus, stomach, small intestine, cecum, and colon). Elimination halflives were determined to be approximately 22 h for the liver, lungs, and blood, while a faster range of elimination half-lives (4 –10 h) was found with the intestinal organs (29). These data suggest that those tissue compartments that retain PEITC and related metabolites serve as a source for the gradual release of these compounds from protein sulfhydryl conjugates (30, 31). The subgroup of six subjects who had high (⬎50%) CVs in the baseline ITC determinations could reflect an uncontrolled dietary contribution to the ITC plasma content on one of the two sampling days. As part of an initial developmental program, we had previously conducted preliminary studies with plasma spiked with 60 –3000 nM PEITC that showed recoveries of 80 –90% for intact PEITC. This work employed an assay we developed using solid-phase extraction followed by gradient reverse-phase chromatography. We evaluated this assay for its applicability to monitoring PEITC plasma levels from human subjects in two studies: (1) Plasma from volunteers who had ingested 40-g servings of watercress (releasing 6 –12 mg PEITC) with serial blood samples taken up to 4 h postingestion, and (2) plasma from several pharmacokinetic collection sets obtained from subjects enrolled in our phase I PEITC single-dose study at the 40-mg dose level. With most of the samples examined from these two studies, little or no PEITC was detected. The results were similar to those reported by Negrusz and co-workers from analyses of plasma of dogs chronically dosed with PEITC as part of a toxicology study. These workers failed to detect PEITC in dog plasma (32). In humans, intact PEITC was not detected in urine in a previous study of watercress consumption by Chung and coworkers (21). Recent data from our phase IA singledose study (33) showed that substantial amounts of PEITC were detectable as the N-acetylcysteine conjugate 4 – 6 h after ingestion of PEITC, suggesting the rapid conjugation of PEITC to PEITC-NAC. The peak concentrations of total ITCs in the urine occurred at the same time points (34), suggesting that all or most of the administered PEITC was in conjugated forms. We have shown conclusively that the reaction is not dependent on the relative ratio of PEITC present in plasma and show a comparable linear response for each of the PEITC conjugates and with PEITC alone. The pharmacokinetic parameter estimates agree in most aspects with recent data generated from a 14Clabeled PEITC rat tissue distribution study (29). 14C

was measured in rat blood for as long as 48 h after dosing male rats with 37.3 ␮Ci 14C-labeled PEITC. The bulk of orally administered 14C-labeled PEITC in plasma was found by radioflow-HPLC to cochromatograph as peaks corresponding to standards of PEITC (⬃20%) or PEITC–NAC (⬃80%), with little 14C (⬍2%) detected in plasma as PEITC–GSH 0.5– 4 h after dosing; a greater proportion of 14C appeared as PEITC– NAC in plasma samples collected at later time points (29). The time to maximal concentration in whole blood in the rat study was 3 h, a value within the range defined by the 4.6 (⫾ 1.3) h T max observed in human plasma. The sensitivity imparted by the high specific activity of the 14C label in the PEITC side chain permitted the modeling of these data using a two-compartment model. The elimination half-life of 14C-labeled PEITC from the major tissue compartments absorbing the PEITC in the rat ranged from 4 to 22 h. Major contributing components for the blood elimination half-life of 21.7 h calculated for the rat study were long tissue decays. Given the less-sensitive human plasma assay based on measurement of conjugates, we did not feel justified in using a two-compartment model, especially with sampling of blood levels only up to 24 h. A one-compartment model fit our data well. However, given these two different models, it is possible to compare the distribution half-lives. The initial 2.4 h alpha or distribution half-life defined in the rat study was within the range of the 3.7 (⫾ 2.3) K10 half-life (distribution component to the blood compartment) defined for total ITCs following PEITC administration to human subjects. The assay has demonstrated its utility in following the pharmacokinetics of oral dosing of PEITC, and should be useful in monitoring future phase I and II studies. ACKNOWLEDGMENT The authors greatly appreciate the helpful comments and suggestions by V. K. Prasad of Drug Regulatory Services.

REFERENCES 1. Morse, M. A., Wang, C. X., Stoner, G. D., Mandal, S., Conran, P. B., Amin, S. G., Hecht, S. S., and Chung, F. L. (1989) Inhibition of 4-(methynitrosamino)-1-(3-pyridyl)-1-butanone-induced DNA adduct formation and tumorigenicity in lung of F344 rats by dietary phenethyl isothiocyanate. Cancer Res. 49, 549 –553. 2. Morse, M. A., Amin, S. G., Hecht S. S., and Chung, F.-L. (1989) Effects of aromatic isothiocyanates on tumorigenicity, O 6-methylguanine formation, and metabolism of the tobacco-specific nitrosamine 4-(methylnitrosamino)-1-(3-pyridyl-1-butanone in A/J mouse lung. Cancer Res. 49, 2894 –2897. 3. Chung, F.-L. (1992) Chemoprevention of lung carcinogenesis by aromatic isothiocyanates. In Cancer Chemoprevention (Wattenberg, L., et al., Eds.), pp. 227–244. CRC Press, Boca Raton. 4. Hecht, S. (1995) Chemoprevention by isothiocyanates. J. Cell. Biochem. 22(Suppl.), 195–209.

HUMAN PLASMA TOTAL ISOTHIOCYANATE CHROMATOGRAPHIC ASSAY 5. Zhang, Y., and Talalay, P. (1994) Anticarcinogenic activities of organic isothiocyanates: Chemistry and mechanisms. Cancer Res. 54(Suppl.), 1976s–1986s. 6. Wattenberg, L. W. (1977) Inhibition of carcinogenic effects of polycyclic-hydrocarbons by benzyl isothiocyanate and related compounds. J. Natl. Cancer Inst. 58, 397–398. 7. Fenwick, G. R., Heaney, R. K., and Mullin, W. J. (1983) Glucosinolates and their breakdown products in food plants. CRC Crit. Rev. Food Sci. Nutr. 18, 123–201. 8. Nishie, K., and Daxenbichler, M. E. (1980) Toxicology of glucosinolates, related compounds (nitriles, R-goitrin, isothiocyanates) and vitamin U found in cruciferae. Food Cosmet. Toxicol. 18, 1191–1194. 9. Guo, Z., Smith, T., Wang, E., Sadrieh, N., Ma, Q., et al. (1992) Effects of phenethyl isothiocyanate, a carcinogenesis inhibitor, on xenobiotic-metabolizing enzymes and nitrosamine metabolism in rats. Carcinogenesis 13, 2205–2210. 10. Mahe´o, F., Morel, F., Langoue¨t, S., Kramer, H., LeFerrec, E., et al. (1997) Inhibition of cytochromes P-450 and induction of glutathione S-transferases by sulforaphane in primary human and rat hepatocytes. Cancer Res. 57, 3649 –3652. 11. Zhang, Y., Talalay, P., Cho, C.-G., and Posner, G. (1992) A major inducer of anticarcinogenic protective enzymes from broccoli: Isolation and elucidation of structure. Proc. Natl. Acad. Sci. USA 89, 2399 –2403. 12. Sparnins, V. L., Venegas, P., and Wattenberg, L. (1982) Glutathione S-transferase activity: Enhancement by compounds inhibiting chemical carcinogenesis and by dietary constituents. J. Natl. Cancer Inst. 68, 493– 496. 13. Pantuck, E., Pantuck, C., Garland, W., Min, B., Wattenberg, L., et al. (1979) Stimulatory effect of brussels sprouts and cabbage on human drug metabolism. Clin. Pharmacol. Ther. 25, 88 –95. 14. Bogards, J., Verhagen, H., Willems, M., van Poppel, G., and van Bladeren, P. (1994) Consumption of brussels sprouts results in elevated ␣-class glutathione S-transferase levels in human blood plasma. Carcinogenesis 15, 1073–1075. 15. Nijhoff, W., Mulder, T., Verhagen, H., van Poppel, G., and Peters, W. (1995) Effects of consumption of brussels sprouts on plasma and urinary glutathione S-transferase class-␣ and -␲ in humans. Carcinogenesis 16, 955–957. 16. Verhoeven, D., Goldbloom, R., van Poppel, G., Verhagen, H., and Van Den Brandt, P. (1996) Epidemiological studies on Brassica vegetables and cancer risk. Cancer Epidemiol. Biomarkers Prev. 5, 733–751. 17. Lin, H., Probst-Hensch, N., Louie, A., Kau I., Witte, J., et al. (1998) Glutathione transferase null genotype, broccoli, and lower prevalence of colorectal adenomas. Cancer Epidemiol. Biomarkers Prev. 7, 647– 652. 18. Steinmetz, K., and Potter, J. (1991) Vegetables, fruit, and cancer. I. Epidemiology. Cancer Causes Control 2, 325–357. 19. London, S. J., Yuan, J-M, Chung, F-L, Gao, Y-T, Coetzee, GA, et al. (2000) Isothiocyanates, glutathione S-transferase M1 and T1 polymorphisms and lung cancer risk: A prospective study of men in Shanghai, China. Lancet 356, 724 –729. 20. Jiao, D., Smith, T. J., Chung S. Y., Pittman, B., Desai, D., Amin, S., and Chung, F. L. (1997) Chemopreventive activity of thiol conjugates of isothiocyanates for lung tumorigenesis. Carcinogenesis 18, 2143–2147.

289

21. Chung, F.-L., Morse, M. A., Eklind, K. I., and Lewis, J. (1992) Quantitation of human uptake of the anticarcinogen phenethyl isothiocyanate after a watercress meal. Cancer Epidemiol. Biomarkers Prev. 1, 383–388. 22. Mennicke, W. H., Gorler, K., Krumbiegel, G., Lorenz, D., and Rittmann, N. (1988) Studies on the metabolism and excretion of benzyl isothiocyanate in man. Xenobiotica 18, 441– 447. 23. Zhang, Y., Cho, C. C., Posner, G. H., and Talalay, P. (1992) Spectroscopic quantitation of organic isothiocyanates by cyclocondensation with vicinal dithiols. Anal. Biochem. 205, 100 –107. 24. Zhang, Y., Wade, K., Prestera, T., and Talalay, P. (1996) Quantitative determination of isothiocyanates, dithiocarbamates, carbon disulfide and related thiocarbonyl compounds by cyclocondensation with 1,2-benzenedithiol. Anal. Biochem. 239, 160 – 167. 25. Chung, F.-L., Jiao, D., Getahun, S. M., and Yu, M. C. (1998) A urinary biomarker for uptake of dietary isothiocyanates in humans. Cancer Epidemiol. Biomarkers Prev. 7, 103–108. 26. Seow, A., Shi, C.-Y., Chung, F.-L., Jiao, D., Hankin, J. H., Lee, H.-P., Coetzee, G. A., and Yu, M. C. (1998) Urinary total isiothiocyanate (ITC) in a population-based sample of middle-aged and older Chinese in Singapore: Relationship with dietary total ITC and glutathione S-transferase M1/T1P1 genotypes. Cancer Epidemiol. Biomarkers Prev. 7, 775–781. 27. Jiao, D., Conaway, C. C., Yang, M.-H., Koehl, C. S., and Chung, F.-L. (1996) Inhibition of N-nitrosodimethylamine demethylase in rat and human liver microsomes by isothiocyanates and their glutathione, L-cysteine, and N-acetyl-L-cysteine conjugates. Chem. Res. Toxicol. 9, 932–938. 28. Talalay, P. (2000) Personal communication. 29. Conaway, C., Jiao, D., Kohri, T., Liebes, L., and Chung, F.-L. (1999) Disposition and pharmacokinetics of phenethyl isothiocyanate and 6-phenylhexyl isothiocyanate in F344 rats. Drug. Metab. Dispos. 27, 13–20. 30. Drobnica, L., and Augustin, J. (1965) Reaction of isothiocyanates with amino acids, peptides and proteins. I. Kinetics of the reaction of aromatic isothiocyanates with glycine. Collect. Czech. Chem. Commun. 30, 99 –104. 31. Xu, K., and Thornalley, P. J. (2000) Studies on the mechanism of the inhibition of human leukaemia cell growth by dietary isothiocyanates and their cysteine adducts in vitro. Biochem. Pharmacol., in press. 32. Negrusz, A., Moore, C. M., McDonagh, N. S., Woods, E. F., Levine, B. S., and Crowell, J. A. (1996) Determination of phenethylamine, a phenethyl isothiocyanate marker in dog plasma using solid phase extraction and GC–MS with chemical ionization. Clin. Chem. 42, 254. [Abstract 660] 33. Liebes, L., Hochster, H., Chachoua, A., Farrell, K., Rafique., M., Garbers, S., Mendoza, S., Chung, F.-L., Conaway, C. C., Malpas, P., Gehling, A.-M., Crowell, J., and Hecht, S. S. (1999) Phase I clinical and pharmacokinetic (PK) trial of phenethyl isothiocyanate (PEITC) in chronic smokers. Proc. Am. Assoc. Cancer Res. 40, 68. [Abstract 557] 34. Liebes, L., Hochster, H., Chachoua, A., Farrell, K., Garbers, S., Mendoza, S., Chung, F.-L., Conaway, C. C., Crowell, J., Carmella, S., Murphy, S. E., Hecht, S. S. (2000) Phase I pharmacokinetics of single dose administration of PEITC: Effects on nicotine metabolism. Proc. Am. Assoc. Cancer Res. 41, 659. [Abstract 4192]