Pharmacokinetics, Safety, and Antiviral Effects of Hypericin, a Derivative of St. John's Wort Plant, in Patients with Chronic Hepatitis C Virus Infection

Abstract

Hypericin is a natural derivative of the common St. Johns wort plant, Hypericum perforatum. It has in vitro activity against several viruses, including bovine diarrhea virus, a pestivirus with structural similarities to hepatitis C virus (HCV). We conducted a phase I dose escalation study to determine the safety and antiviral activity of hypericin in patients with chronic HCV infection. The first 12 patients received an 8-week course of 0.05 mg of hypericin per kg of body weight orally once a day; 7 patients received an 8-week course of 0.10 mg/kg orally once a day. At the end of the 8-week period of treatment, no subject had a change of plasma HCV RNA level of more than 1.0 log₁₀. Five of 12 subjects receiving the 0.05-mg/kg/day dosing schedule and 6 of 7 subjects receiving the 0.10-mg/kg/day dosing schedule developed phototoxic reactions. No other serious adverse events associated with hypericin use occurred. The pharmacokinetic data revealed a long elimination half-life (mean values of 36.1 and 33.8 h, respectively, for the doses of 0.05 and 0.1 mg/kg) and mean area under the curve determinations of 1.5 and 3.1 μg/ml × hr, respectively. In sum, hypericin given orally in doses of 0.05 and 0.10 mg/kg/d caused considerable phototoxicity and had no detectable anti-HCV activity in patients with chronic HCV infection.

Hypericin (4,5,7,4′,5′,7′-hexa-hydroxy 2,2′ dimethyl-mesonaphthodianthron) is a natural compound found in the stems and petals of members of the genus Hypericum, including the common St. Johns wort plant Hypericum perforatum (32). It has shown in vitro activity against a variety of viruses, including murine Friend leukemia virus (22), Rauscher leukemia virus (30), equine infectious anemia virus (14), murine immunodeficiency virus (16), murine cytomegalovirus (P. L. Barnard, J. H. Huffmann, and S. G. Wood, Prog. Abstr. 30th Intersci. Conf. Antimicrob. Agents Chemother., abstr. 1093, 1990), influenza virus (30), vesiculostomatitis virus (18), Sendai virus (18), herpes viruses (30), and duck hepatitis B virus (23). Although hypericin has in vitro antiviral activity without light activation, this activity is enhanced by exposure to light (11, 16, 30). Furthermore, hypericin was effective against Friend leukemia virus and herpes simplex virus type 1 in mice (16, 30).

Recently, bovine diarrhea virus (BVDV) was found to be completely inactivated by hypericin in vitro in the presence of light (26). BVDV is a pestivirus that has structural similarities to the hepatitis C virus (HCV) (17, 24). Infection with HCV establishes a persistent infection in up to 90% of cases (1–3, 7). Four million Americans and 100 million people worldwide are chronically infected with HCV (4). Cirrhosis and hepatocellular carcinoma are long-term sequelae of this infection (27–29, 31). Although hypericin's activity against BVDV in vitro was not studied in the absence of light, it was decided to proceed with an exploratory clinical study of its activity against HCV. It was felt that there was no practical, consistent method of generating a stable, photo-activated form of hypericin for administration to patients.

We conducted a phase I dose escalation study to determine the safety and antiviral activity of hypericin in patients with chronic HCV infection.

MATERIALS AND METHODS

Study population.

We enrolled patients in the study if they met the following criteria: an age of 18 to 70 years, evidence of virologically active HCV infection as determined by a positive PCR HCV RNA assay within 60 days prior to study entry, a hemoglobin concentration greater than 11 g/dl, a total white blood cell count greater than 3,000/mm³, an absolute neutrophil count greater than 1,500/mm³ a platelet count greater than 180,000/mm³ cubic millimeter, a serum bilirubin concentration of less than 1.3 mg/dl, a serum alanine aminotransferase (ALT) concentration less than five times the upper limit of normal, a serum albumin concentration greater than 38 g per liter, a serum creatinine concentration less than 1.6 mg/dl, a prothrombin time less than 2 s above control, no more than 30 mg/dl urinary protein or 5 to 25 erythrocytes/μl urinary blood, and negative anti-DNA and anti-smooth muscle antibody assays.

Patients were excluded from the study if they were human immunodeficiency virus (HIV) infected; if they were pregnant; if they had a prior history, or current clinical or laboratory evidence, of cirrhosis or hepatic failure; if they were active alcohol or illicit substance users; if they had a prior or current history of a malignancy, with the exception of basal cell carcinoma or in situ carcinoma; if they had other causes of active liver disease; if they had recently acquired (within 6 months) acute hepatitis; if they had evidence of significant cardiovascular, renal, gastrointestinal, or central nervous system disease; if they had an active infection or major surgery within two weeks prior to study entry; if they received treatment with any investigational drug within 30 days prior to study entry; if they received treatment with glucocorticosteroids or other immunosuppressive medications within 14 days prior to study entry; if they were treated with antiviral agents within 14 days prior to study entry; if they were treated with monoamine oxidase inhibitors, cimetidine, ketoconazole, terfenamine, acetaminophen, or other agents known to cause hepatotoxicity within 14 days prior to study entry; if they received treatment with any medications known to cause photosensitization within 30 days prior to study entry; if they were receiving sulfonamides, sulfones, or other agents known to cause a significant incidence of skin rash; if they had a generalized skin rash; and if they were obese (body mass index >35). Patients were encouraged to wear gloves and hats and apply sunscreen protection when outside, and to avoid excessive exposure to the sun. Patients may have been treated with alpha interferon previously, but not within the previous 3 months. No patient was taking, or had a history of taking, St. Johns wort.

The study was approved by the institutional review boards of the Bronx Veterans' Affairs Medical Center and the Mount Sinai School of Medicine. The patients gave written informed consent to participate. Females of childbearing potential were required to have negative pregnancy tests (serum) within 7 days prior to study entry. All female patients needed to agree to practice barrier methods of contraception or abstinence for the duration of study participation.

Treatment regimens.

The first 12 patients enrolled in the study were to receive an 8-week course of 0.05 mg of hypericin per kg of body weight in liquid form orally once a day (the study medication was kindly provided by VimRx Pharmaceuticals, Wilmington, Del.). The next 12 patients were to receive an 8-week course of 0.1 mg of hypericin per kg orally once a day. The drug was taken in the morning without regard to food. The subjects were instructed to return the empty drug vials to the research nurses after they were used. This was to monitor adherence to the drug regimen. Patients were also instructed to keep a medication log. In addition, dosing in the clinic at the time of pharmacokinetic blood sampling was observed by the research nurses.

Pharmacologic studies.

Hypericin was reconstituted from a lyophilized 40-mg vial to a concentration of 1 mg/ml in 1.9% benzyl alcohol–4.5% dextrose and given orally at doses of 0.05 and 0.10 mg/kg. Blood was drawn into heparinized tubes at predetermined times as follows: week 0, 0 and 6 h; week 1, 0 and 6 h; week 2, 0 h; and week 4, 0 h. At week 8, detailed pharmacokinetic samples were taken at 0, 6, 9, 12, 24, and 48 h. The plasma was centrifuged at 1,000 × g for 10 min, aliquoted in cryovials, and stored at −20°C.

Hypericin levels in plasma were analyzed using high-performance liquid chromatography methodology previously described (19) with a series of modifications to increase the assay sensitivity to 10 ng/ml. These included the use of 0.5-ml aliquots of the plasma samples that were processed through two consecutive extraction cycles with additions of 0.125 ml of dimethyl sulfoxide and 0.625 ml of a mixture consisting of acetonitrile–2-butoxyethanol (90:10, vol/vol). The extractions were followed by centrifugation, removal of the supernatant, and bringing the combined extraction volume up to 2.0 ml. The plasma standards used in the assay ranged from 10 to 120 ng/ml, while precision samples (20, 40, and 120 ng/ml) chosen to span the range of steady-state trough and peak levels were interdispersed among the analysis samples. The amount of the plasma extracts injected onto the analysis column was 200 μl. The detection of the chromatographically separated components was at 590 nm, the visible chromatophore absorption maximum of hypericin. The intraday coefficient of variation (CV) of the assay in terms of the precision samples of 20, 60, and 120 ng/ml was, respectively, 5.7, 6.8, and 7.4%, while the interday CV was somewhat better at 5.1, 6.1, and 6.2%. The assay was sensitive down to 5 ng/ml, but the lower limit of quantification was 10 ng/ml.

Hypericin plasma levels were modeled using WINONLIN (version 1.5; Pharsight Corporation, Mountain View, Calif.) using a one-compartment oral absorption model. Although it is known that the hypericin pharmacokinetic decay can be fitted to a two-compartment model (12; V. McAuliffe, R. Gulick, H. Hochster, L. Liebes, J. Vaccariello, S. Hussey, Y. Bassiakos, H. Balfour, D. Stein, C. Crumpacker, and F. Valentine, 1st Natl. Conf. Hum. Retrovir. Relat. Infect, abstr. 159, 1993), a one-compartment model was used to allow all of the data sets to be fitted with the same model. This compromise was due to the fact that some data sets had only five decay time points (with the 48-h sampling time missing). In addition, it was deemed not practical to require the subjects to have an overnight stay on two nights to obtain the 18- and 36-h sampling time points that would have allowed the data to be fitted to a two-compartment model. The decision was made in light of the additional visits already required for the steady-state samplings on days 7, 14, and 28. The one-compartment model was able to fit the data well and yielded CVs ranging from 11 to 30% for the primary and secondary parameter estimates.

Criteria for response.

The primary end point of the study was a change in plasma HCV RNA level of more than 1.0 log₁₀ from baseline to week 8.

Evaluation of patients and follow-up.

After the screening and baseline evaluations, the patients were seen weekly for the first 2 weeks and then every 2 weeks until week 10 (2 weeks after the end of study treatment). At each of these visits, an interval clinical history was obtained, a physical examination was performed, and blood specimens were obtained for complete blood cell counts, serum chemistries, and hepatic and renal function. Two plasma samples for the measurement of HCV RNA were obtained at baseline; the first one was taken within 7 days before starting study treatment, and the second was taken on the day of starting treatment. Follow-up samples were obtained every 2 weeks until week 10. In addition, blood specimens were obtained for measuring hypericin levels at baseline.

The degree of fatigue and abdominal pain was assessed by the subjective reports of patients. In the event of intolerable photosensitization or other adverse events of grade 3 or higher (according to National Institute of Allergy and Infectious Diseases adverse events criteria), the study medication was permanently discontinued. Toxicity was assessed by reviewing patient symptoms, physical findings on examination, and the results of laboratory tests. The higher-dose cohort was not initiated until the first four patients of the lower dose cohort were treated for 4 weeks without the occurrence of significant phototoxicity or other serious adverse effects. There were no predetermined guidelines for stopping a dosing cohort based on toxicity. Instead, the study team regularly reviewed the toxicity data.

Statistics.

Response was defined as a reduction in HCV RNA levels of at least 1.0 log₁₀. The study was designed to have 80% power and a type I error rate of 0.05 to detect at least a 10% response rate (2 or more responders out of 10 subjects in each group) if the true response rate were 50%. To allow for a 20% dropout rate, 12 subjects were accrued to each dose arm of the study.

The null hypothesis that the probability of a drop of at least 1.0 log₁₀ in HCV RNA levels is less than or equal to 10% was tested with an exact 0.05 level two-sided test based on the binomial distribution. The null hypothesis that the median change in plasma HCV RNA levels from baseline to week 8 was zero within a single group was tested with a two-sided 0.05-level Wilcoxon sign-rank test. The Wilcoxon sign-rank test was also used to evaluate the median changes in serum ALT levels from baseline to week 8.

RESULTS

Study population.

Between January 1997 and October 1997, 19 patients were enrolled; 12 received the 0.05-mg/kg daily dose of hypericin and seven received 0.10 mg/kg/day. The baseline characteristics of the study subjects are shown in Table 1.

TABLE 1.

Baseline characteristics

Characteristic	Result in hypericin dosage group
	0.05 mg/kg (n = 12)	0.1 mg/kg (n = 7)
Mean age ± SD (yr)	48 ± 7	45 ± 4
Gender (no.)
Male	10	5
Female	2	2
Race or ethnic group (no.)
Black	4	4
Caucasian	6	2
Hispanic	2	1
Mean serum ALT level ± SD (U/liter)ab	73 ± 34	75 ± 42
Mean plasma HCV RNA level ± SD (log10 RNA molecules/ml)b	7.17 ± 0.84	7.60 ± 0.47

^aNormal laboratory value = 0 to 45 U/liter. 

^bSerum ALT results as well as plasma HCV RNA results obtained at the prescreening visit (within 7 days of baseline) and on the day of treatment initiation were averaged in order to obtain the baseline value. 

Virologic data.

The median plasma HCV RNA levels at baseline were 7.3 log₁₀ RNA molecules/ml for the 0.05-mg/kg dosage group and 7.6 log₁₀ RNA molecules/ml for the 0.1-mg/kg dosage group (Table 1). No subjects had a change of plasma HCV RNA levels of more than 1.0 log₁₀ (Table 2). With the 11 evaluable subjects in the 0.05-mg/kg dose cohort and 7 in the 0.10-mg/kg dose cohort, the study had 89 and 77% power, respectively, to detect a response rate of more than 10%. There were no significant changes from baseline to week 8 in plasma HCV RNA levels with either dose of study treatment (Table 2; Fig. 1). At the end of the 8-week study treatment period, the median plasma HCV RNA level was 7.27 log₁₀ RNA molecules/ml for the 0.05-mg/kg/day dosage group and 7.35 log₁₀ RNA molecules/ml for the 0.10-mg/kg/day dosage group.

TABLE 2.

Plasma HCV RNA values for subjects in this study

Patient no.	Dose (mg/kg)	Plasma HCV RNA level (log10 RNA molecules/ml) at:
		Baselinea	Day 7	Day 14	Day 28	Day 42	Day 56
1	0.05	7.62	7.68	7.73	7.79	7.65	7.27
2	0.05	7.29	7.73	7.83	7.43	7.62	7.62
3	0.05	8.30	8.20	8.21	8.50	8.61	8.48
4	0.05	6.99	7.19	7.46	7.67	7.40	7.24
5	0.05	7.68	7.50	7.60	8.42	8.10	6.33b
6	0.05	6.35	6.55	7.11	6.42	6.57	6.98
7	0.05	7.12	7.18	7.03	7.38	7.13	7.35
8	0.05	5.17	5.55	5.31	5.31	5.36	5.38
9c	0.05	7.31	7.03	7.02	7.45c
10	0.05	7.85	7.98	7.74	7.99	8.20	7.71
11	0.05	7.87	7.51	8.03	8.31	8.03	8.30
12	0.05	6.53	7.51	7.01	7.24		7.24
Mean		7.17	7.30	7.34	7.49	7.40	7.26
SD		0.84	0.70	0.76	0.90	0.91	0.86
13	0.10	7.60	7.52	7.42	7.72	7.72	7.84
14	0.10	8.04	8.09	8.16	7.75	7.60	7.05
15	0.10	8.07	8.05	8.36	7.25	7.78	8.48
16	0.10	7.45	7.58	7.51	7.19	7.75	7.51
17	0.10	8.05	7.60	7.36	7.40	7.61	7.35d
18	0.10	6.97	6.88	7.11	6.99	7.33	6.76
19	0.10	7.05	6.25	7.28	6.67	6.40	6.50
Mean		7.60	7.42	7.60	7.28	7.46	7.36
SD		0.47	0.66	0.47	0.39	0.49	0.67

^aPlasma HCV RNA results obtained at the prescreening visit (within 7 days of baseline) and on the day of treatment initiation were averaged for the baseline plasma HCV RNA value. 

^bPatient 5 did not complete 56 days of hypericin. The last day of treatment was day 42. 

^cPatient 9 did not complete 56 days of hypericin. The last day of treatment was day 11. 

^dPatient 17 did not complete 56 days of hypericin. The last day of treatment was day 30, and the value shown here is the day 28 value. 

FIG. 1.

Median plasma HCV RNA levels in patient cohorts treated with hypericin (either 0.05 or 0.10 mg/kg/day) for 56 days. Inf and −Inf, positive and negative infinity, respectively; error bars, 95% confidence intervals.

Safety data.

Seven of 12 subjects treated with the 0.05-mg/kg dosage and all 7 treated with the 0.1-mg/kg dosage had a photosensitivity reaction(s) while taking hypericin. There were four different types of photosensitivity reactions identified—paresthesias, dermatitis, darkened coloration of exposed skin, and pruritic nodules (Table 3). All of the photosensitivity reactions were judged to be probably related to hypericin.

TABLE 3.

Photosensitivity reactions experienced by subjects by dosage group^a

Photosensitivity reaction	No. of subjects in dosing group
	0.05 mg/kg (n = 12)	0.1 mg/kg (n = 7)
Paresthesias
Mild (grade 1)	5	3
Moderate (grade 2)	1	2
Severe (grade 3)	2b	0
Dermatitis
Mild (grade 1)	1	2
Moderate (grade 2)	1	0
Severe (grade 3)	0	0
Darkened coloration of exposed skin
Mild (grade 1)	0	3
Moderate (grade 2)	0	0
Severe (grade 3)	0	0
Pruritic nodules
Mild (grade 1)	0	1
Moderate (grade 2)	0	0
Severe (grade 3)	0	0

^aNote: All the photosensitivity reactions listed in Table 3 were judged to be probably related to hypericin. 

^bThe study medications were discontinued in these two patients. 

Paresthesias (reported as a burning and/or tingling sensation in the skin after sun exposure) were the most common photosensitivity reactions experienced by subjects in both dosage groups. Seven subjects treated with the 0.05-mg/kg dose and five subjects treated with the 0.1-mg/kg dose had mild (grade 1) paresthesias. One subject treated at the 0.05-mg/kg level initially experienced grade 1 paresthesias, which after 6 days increased to grade 2. Five of the subjects treated at the 0.05-mg/kg dose were able to complete the 56 days of treatment despite the paresthesias. Two subjects who experienced paresthesias at this dosage level terminated the study treatment early, finding the sensation too painful to continue treatment (grade 3). Neither of these two subjects took analgesia for the paresthesias, preferring to wait for the effect to end on its own. All five subjects treated with the 0.1-mg/kg dosage who experienced paresthesias were able to complete the 56 days of treatment. The paresthesias resolved without sequelae in all subjects who experienced them after the subjects stopped taking hypericin.

Dermatitis was experienced by two subjects treated with the 0.05-mg/kg dosage and two subjects treated with the 0.1-mg/kg dosage. The dermatitis was mild (grade 1) for one subject treated with the 0.05-mg/kg dose and for both subjects treated with the 0.1-mg/kg dose and consisted of erythema of the skin after sun exposure. One subject treated with the 0.05-mg/kg dose experienced a moderate (grade 2) dermatitis, dry desquamation on the tips of her fingers.

Darkened coloration of exposed skin and pruritic nodules, were experienced by subjects treated with the 0.1-mg/kg dose but not by subjects treated with the 0.05-mg/kg dose. Of the seven subjects treated with the 0.1-mg/kg dose, three experienced darkened coloration of exposed skin. One of these three subjects also experienced pruritic nodules on the fingers of both hands. Of note was that all three subjects were black men. No treatment was required for these reactions. For two subjects who experienced darkened coloration of skin, the adverse event was ongoing at day 70 but resolved by the 6-month follow-up visit. The third subject had to discontinue treatment early due to another adverse event which was not related to study treatment (migraine-related amaurosis fugax). His pruritic nodules and darkened coloration of skin resolved after treatment was discontinued.

Because uncomfortable photosensitivity reactions occurred at both the 0.05- and 0.1-mg/kg dosage levels and no antiviral effect was detected in either dose cohort, the decision was made to stop enrollment early. As a result, only seven subjects were entered into the 0.1-mg/kg dosage group.

There were no laboratory-related grade 3 (severe) adverse events for either dosage group, and no subject had to withdraw because of laboratory abnormalities.

Other adverse events were uncommon. One subject treated with the 0.01-mg/kg dose experienced both dry mouth and angular cheilosis, which resolved 2 days after the completion of the 8-week course of hypericin. One subject experienced headache, dizziness, and amaurosis fugax diagnosed by a neurologist after physical examination, electroencephalogram, and brain magnetic imaging as a migraine headache. Although it was judged to be unrelated to hypericin use, because the amaurosis fugax was a grade 3 (severe) adverse event, this patient was discontinued from study treatment. All symptoms resolved without treatment and without sequelae after hypericin was discontinued.

Other clinical data.

Most subjects who entered the study had no symptoms of chronic illness. Five subjects complained of chronic fatigue at baseline: two in the 0.05-mg/kg dosage group and three in the 0.1-mg/kg dosage group. Four subjects, two in each dosage group, had intermittent right upper quadrant abdominal pain at baseline. The degree of fatigue improved during study treatment for only one of the five subjects who initially reported this symptom. In the other four subjects, the severity of the symptom remained approximately the same. In the four subjects with abdominal pain, no significant change occurred while on study treatment for three of them. Only one subject in the 0.1-mg/kg group reported a slight decrease in abdominal pain. All but two subjects had mild-to-moderate elevations of serum ALT levels at baseline (median, 70 U/liter; range, 39 to 158 U/liter). There were no significant changes in these values with either dose of hypericin given (Table 4).

TABLE 4.

ALT values for subjects in this study

Patient no.	Dose (mg/kg)	Serum ALT value (U/liter) at:
		Baselinea	Day 7	Day 14	Day 28	Day 42	Day 56
1	0.05	67	66	73	65	61	66
2	0.05	54	58	56	38	31	40
3	0.05	88	39	78	66	82	88
4	0.05	40	40	37	38	31	35
5	0.05	91	98	102	91	86	—b
6	0.05	57	61	95	100	89	95
7	0.05	46	44	44	41	43	43
8	0.05	53	56	46	52	64	65
9	0.05	155	131	—	—	—	—
10	0.05	96	119	99	75	84	68
11	0.05	64	80	58	54	65	58
12	0.05	79	81	80	78	80	98
Mean		74	73	70	63	65	66
Median		66	64	73	65	65	66
SD		31	30	23	21	22	23
13	0.10	39	38	35	37	41	31
14	0.10	158	172	137	122	135	135
15	0.10	45	37	48	49	51	45
16	0.10	53	58	59	42	44	36
17	0.10	76	83	87	93	—	—
18	0.10	75	79	70	72	69	69
19	0.10	81	90	95	137	94	122
Mean		75	80	76	79	72	73
Median		75	79	70	72	60	57
SD		40	46	34	40	36	45

^aSerum ALT results as well as plasma HCV RNA results obtained at the prescreening visit (within 7 days of baseline) and on the day of treatment initiation were averaged in order to obtain the baseline value. 

^b—, Patient discontinued study medication. 

Pharmacology.

Blood specimens were analyzed initially in terms of the peak and trough levels. Averages of peak levels were derived from the day 14 and day 56 samples, while the trough averages were from the day 7, day 14, and day 28 samples (Table 5). A second set of trough values (minimum drug concentration if serum [C_min] at the end of the treatment cycle) which were derived from the measurements at day 56 and day 57 are also summarized. These data show good consistency over these different time points and do not indicate any effects of chronic dosing over 58 days. The mean values of the minimum drug concentration in serum (C_min) and the maximum drug concentration in serum (C_max) are dose proportional. Figure 2 shows a comparison of the mean ± standard deviation (SD) values from plasma decays obtained from those subjects receiving doses of 0.05 and 0.10 mg/kg who had detailed pharmacokinetic samplings at the end of the dosing cycle. These plots show dose proportionality, and the pharmacokinetic estimates from those subjects who had evaluable data are summarized in Table 6. The plasma hypericin decays were found to have a mean ±SD elimination half-life of 36.1 ± 22.6 and 33.8 ± 18.8 h, respectively, for the 0.01- and 0.05-mg/kg doses. The mean area under the curve (AUC) determinations for the two hypericin doses were, respectively, 1.5 and 3.1 μg/ml × hr, which again showed dose proportionality. The time to C_max, 4.4 ± 2.5 h, along with C_max values from the earlier study monitoring (Table 5) comparable with those at the end of the study (Table 6) showed enough overlap and confirmed that the monitoring for C_max at 6 h was adequate. The volume of distribution and clearance were greater at the 0.10-mg/kg dose. This possibly reflects a secondary tissue absorption that is more apparent at the higher concentration.

TABLE 5.

Summary of hypericin peak and trough concentrations

Patient no.	Dose (mg/kg)	Mean Cmaxa ± SD (ng/ml)	Mean Cminb ± SD (ng/ml)	Mean Cmin, endc ± SD (ng/ml)
1	0.05	32.7	21.3 ± 4.0	21.4 ± 0.6
2	0.05	16.7	10.9
3	0.05	17.5	12.7 ± 0.7	13.6
4	0.05	39.2	25.6 ± 1.4	24.6 ± 2.6
5	0.05	21.1	16.7 ± 4.3
6	0.05	54.9	18.9 ± 4.0	20.4
7	0.05	19.6	11.1
8	0.05	18.4	9.9
9	0.05	35.3 ± 4.1	21.5 ± 7.8	11.6
10	0.05	25.6	13.0 ± 2.9	20.3 ± 1.2
11	0.05	19.0	9.2
12	0.05	36.8 ± 0.1	13.9 ± 4.5	14.1 ± 0.3
Mean		28.1	15.4	18.0
Median		4.0	13.5	20.3
SD		11.8	5.3	4.9
13	0.10	79.2	28.5 ± 8.4	25.5 ± 6.9
14	0.10	66.3	39.4 ± 3.8	42.2 ± 5.1
15	0.10	78.5 ± 11.1	40.8 ± 7.3	47.0 ± 1.6
16	0.10	30.4	19.7 ± 5.3	21.2 ± 2.5
17	0.10	66.6	41.7 ± 5.8
18	0.10	112.4	47.7 ± 6.0	48.8 ± 2.8
19	0.10	120.9 ± 10.4	71.9 ± 14.5	85
Mean		79.2	41.7	45.0
Median		78.5	40.8	44.6
SD		30.4	15.9	22.7

^aAverage of values from days 7 and 14, when available. 

^bAverage of values from days 7, 14, and 28, when available. 

^cC_{min, end}, C_min at end of treatment cycle. 

FIG. 2.

Summary of mean values ± SD (error bars) for hypericin decay plots from patients receiving hypericin at 0.05 (n = 7) and 0.1 (n = 4) mg/kg who agreed to undergo detailed pharmacokinetic sampling. The decays were fitted with one-compartment model with a first-order absorption (nonlinear fit) and showed dose proportionality for the AUC determinations (1.5 versus 3.1 μg/ml × h, respectively, for 0.05- and 0.1-mg/kg dose levels).

TABLE 6.

Hypericin pharmacokinetic parameter estimates^a

Patient no.	Dose (mg/kg)	Tmax	Cmax	AUC (μg/ml × h)	Elimination t1/2 (h)b	V/F (liter)c	CLTB (ml/ min/m2)d
1	0.05	2.6	46.4	1.75	24.1	14.0	4.7
2	0.05	8.1	13.3	1.03	36.8	36.0	7.1
3	0.05	3.7	18.1	1.37	65.7	48.4	5.3
4	0.05	3.1	29.3	2.87	65.8	22.5	2.2
6	0.05	2.7	24.9	1.32	34.8	28.2	4.5
10	0.05	2.2	42.6	1.02	19.3	24.0	8.1
12	0.05	8.7	39.3	0.93	6.3	13.9	8.9
Mean		4.4	30.6	1.5	36.1	26.7	5.8
Median		3.1	29.3	1.3	34.8	24.0	5.3
SD		2.7	12.6	0.7	22.6	12.3	2.3
13	0.10	5.9	42.1	1.82	22.6	109.0	22.8
15	0.10	6.9	69.4	4.86	37.4	71.9	9.8
16	0.10	3.3	30.0	2.64	58.6	112.1	11.9
18	0.10	1.6	118.0	2.99	16.4	44.0	10.5
Mean		4.4	64.9	3.1	33.8	84.3	13.4
Median		4.6	55.8	2.8	30.0	90.5	11.2
SD		2.4	39.1	1.3	18.8	32.5	6.4

^aT_max, time to C_max. 

^bt_1/2, half-life. 

^cV/F, volume of distribution/bioavailability. 

^dCL_TB, total body clearance. 

The gradient high-performance liquid chromatography methodology employed for the analysis of hypericin was adequate to allow quantitation of any hypericin-derived metabolites, as the assay used the visible absorption maximum of the hypericin molecule. However, no discernible peaks other than the hypericin molecule were detected over all of the study monitoring time points.

DISCUSSION

In the doses studied hypericin demonstrated no detectable anti-HCV activity. No subjects treated with an 8-week course of hypericin (either 0.05 or 0.10 mg/kg/day) had a change in plasma level of HCV RNA of more than 1.0 log₁₀, the natural variability of the laboratory assay. Twelve subjects receiving the 0.05-mg/kg/day dose and seven receiving the 0.10-mg/kg/day dose were studied. These results provide significant evidence that hypericin in these doses is not effective in lowering plasma HCV levels in HCV-infected patients. In addition, hypericin had no effect on improving elevated serum liver enzyme levels in the patients studied.

Although immune responses to the infecting organism are likely to play a role in disease pathogenesis (13), elimination of detectable virus from the blood is closely associated with the successful treatment of chronic HCV infection (15, 20). In all clinical trials involving interferon preparations, long-term histologic and clinical outcome in a patient was determined by the sustained (>6 months after treatment ended) suppression of detectable HCV activity in plasma (15, 20). Thus, the lack of an effect of hypericin on plasma HCV levels strongly suggests that hypericin as a single agent has no role in the treatment of chronic hepatitis C.

Five of 12 subjects receiving the 0.05-mg/kg/day dosing schedule and almost all (6 of 7) subjects receiving the 0.10-mg/kg/day dosing schedule developed uncomfortable phototoxicity. This necessitated treatment discontinuation in only two subjects in the 0.05-mg/kg/day treatment group and none in the 0.10-mg/kg/day treatment group. Nevertheless, it is likely that the phototoxic reactions would be even more severe at higher doses. Without even a hint of activity at the doses studied, it would be difficult to justify evaluating higher doses of hypericin for hepatitis C. Similar phototoxicity was reported recently in a study of hypericin given intravenously in doses of 0.25 and 0.50 mg/kg given two or three times weekly for HIV infection (8). No antiviral activity was detected in that study either (8). No other serious adverse events associated with hypericin use occurred during our study.

Hypericin is being evaluated for other antiviral and antineoplastic activities (6, 8–10, 33). The plant from which it is derived, St. Johns wort, has wide use for the treatment of depression (5). Thus, the safety and pharmacokinetic data reported here have usefulness for the application of hypericin to these indications.

The pharmacokinetic data from the two dose levels of hypericin examined in this study were consistent with respect to dose proportionality in the steady-state levels observed as well as for the pharmacokinetic parameter estimates yielding the AUC calculations. These data are also comparable in terms of the long elimination half-life (mean values of 36.1 and 33.8 h, respectively, for doses of 0.05 and 0.10 mg/kg/day) with the data derived from the phase I studies of hypericin in HIV-infected adults, in which an elimination half-life of 35.3 ± 9 h was observed (McAuliff et al., 1st Natl. Conf. Hum. Retrovir. Relat. Infect). A study with plant-derived hypericin preparations administered to healthy volunteers at hypericin levels 5- to 10-fold lower than those used in the present study also showed an elimination rate of 43 h (12). The AUC calculation from the 0.05-mg/kg dosing group of 3.1 ± 1.3 μg/ml × h is in good agreement with the 22% bioavailability that was determined from the phase I HIV-infected-adult study, in which a theoretical extrapolation from the intravenous AUC determined from the 0.5-mg/kg dose level of 53.7 μg/ml × h corrected for the ∼20% bioavailability gives an expected AUC for 0.05-mg/kg dosing of 2.39 μg/ml × hr. There was no discernible drug accumulation from the daily dosing, as evidenced over time.

If one uses a multidosing simulation of the single-dosing pharmacokinetic parameter estimates (McAuliffe et al., 1st Natl. Conf. Hum. Retrovir. Relat. Infect.), steady-state levels should be reached by 10 days with a twofold peak/trough ratio. This was observed in our study, in which there was good consistency for the day 14, 28, and 56 trough levels as well as for the peak level measurements obtained at study days 14 and 56 (data not shown). While the daily dosing employed in this study appears to take close to 2 weeks to achieve steady state, it still appears to be preferable not to dose more frequently given the considerable toxicity with daily dosing at the 0.5-mg/kg level in this study as well as in the phase I HIV study (8).

It is worthwhile to comment on the lack of any detectable hypericin metabolites, even in any of the detailed pharmacokinetic samples taken on day 42 of the study. Our findings extend the observations of Kerb and coworkers (12), who found little difference between the multidose and the single dose AUC from 0 h to infinity. They also found no change in the plot of trough levels over their 14 days of continuous dosing. It is possible that the recent suggested perturbations imparted by St. Johns wort on the pharmacokinetics of cyclosporin and protease inhibitors (25) could be related to the use of additional plant-derived components and the concentrations used. A recent report by Markowitz and coworkers (21) found little perturbation of cytochrome P-450 and 3A4 activity in healthy volunteers taking St. Johns wort preparations (H. perforatum) at recommended doses for depression.

In sum, hypericin given orally in tolerable doses had no detectable anti-HCV activity in patients with chronic HCV infection. It has considerable phototoxicity when administered in doses of 0.05 mg/kg/day or more. Nevertheless, hypericin is being evaluated for other antimicrobial, antineoplastic, and antidepressant indications, and the pharmacokinetic and safety data reported here can be used to guide further study of this compound.