Impact of oral silymarin on virus- and non-virus-specific T-cells responses in chronic hepatitis C infection

Abstract

Silymarin displays anti-inflammatory effects on T-lymphocytes in vitro. The immunomodulatory properties of oral silymarin in vivo in humans with chronic hepatitis C have not previously been characterized. We hypothesized that silymarin would suppress T-cell proliferation and pro-inflammatory cytokine production of virus- and non-virus-specific T-cells while increasing anti-inflammatory IL-10 production in vivo. Patients from one site of the SyNCH-HCV double-masked, placebo-controlled study of oral silymarin in prior interferon non-responders with chronic hepatitis C provided blood samples at baseline and treatment week 20. Mononuclear cells were stimulated with recombinant HCV proteins and controls in ³H-thymidine proliferation assays, IFNγ Elispot and IL-10 Elispot. The frequency of CD4⁺CD25^hi and CD4⁺foxp3⁺ regulatory T-cells, serum cytokine levels, serum IP-10 and lymphocyte interferon-stimulated gene expression were also quantified at baseline and week 20. Thirty-two patients were recruited (10; placebo, 11; 420mg three times a day, 11; 700mg three times a day). Serum ALT and HCV RNA titers did not change in any group. HCV-specific CD4⁺ T-cell proliferation and the frequency of IFNγ– and IL-10-producing T-cells were not significantly changed in silymarin-treated subjects. However, C. albicans-induced T-cell IFNγ and phytohemagglutinin-induced T-cell proliferation were suppressed by silymarin therapy. A trend towards augmentation of interferon-induced ISG15 expression was present in the high-dose silymarin group. While no effect on HCV-specific T-cells was identified, these data confirm that high-dose oral silymarin exerts modest non-specific immunomodulatory effects in vivo. The impact of this anti-inflammatory effect on long-term liver health in chronic hepatitis C merits future clinical investigation.

INTRODUCTION

T-cells play a critical role in the resolution of hepatitis C virus (HCV) infection. In early infection, HCV-specific T-cell responses are characterized by robust IFNγ production and proliferation in patients who resolve their infections compared to weak and dysfunctional responses in those who progress to chronic infection (1–4). In chronic infection, some but not all studies suggest that treatment-related resolution of chronic hepatitis C is also in part mediated by antiviral T-cells (5–9). Furthermore, various regulatory T-cell populations have been associated with protection from chronic liver injury in chronic infection (10–12).

Silymarin, an extract of milk thistle (Silybum marianum; plant extract of Asteraceae family), has been used for two millennia to treat various liver disorders. In the modern era, salutary effects in alcoholic liver disease and chronic viral hepatitis have been ascribed to orally administered silymarin in the absence of rigorous data supporting this practice (13). Modulation of T-cells has been one postulated explanation for the putative anti-inflammatory properties of silymarin. In murine models, silymarin has been variably demonstrated to have both pro- (14) and anti-inflammatory effects on T-lymphocytes (15)(16). Recent human clinical trials somewhat unexpectedly demonstrated that high dose intravenous silibinin exhibits direct antiviral properties in chronic hepatitis C in vivo (17, 18), suppressing HCV replication primarily by inhibiting viral entry and transmission (19). In vitro, the silymarin preparation MK001 also appeared to suppress virus- and non-virus-specific CD4⁺ T-cell proliferation, as well as both pro- (TNFα, IFNγ) and anti-inflammatory (IL-10) cytokine production (20) likely due to blockade of STAT1-induced NFκB activation (20, 21). However, the effects of silymarin on human antiviral T-cells in vivo have not previously been investigated.

In the NIH-sponsored Phase II multicenter, randomized, placebo-controlled, double-masked, prospective clinical trial completed by the Silymarin in Non-alcoholic Steatohepatitis and C Hepatitis (SyNCH) Study Group, two doses of Legalon^® 140, a standardized preparation of milk thistle extract, or identical sugar-containing placebo were administered to patients with chronic hepatitis C with previous non-response to interferon-based antiviral therapy. In the clinical trial, there was no statistically significant difference in the proportion of patients in placebo or silymarin arms that achieved serum ALT level of ≤ 45 IU/L or ≥ 50% reduction of ALT from baseline level to < 65 IU/L at the end of the 24-week treatment period (22). We present results of a prospective, masked immunological substudy designed to determine the impact, if any, of orally-administered silymarin on hepatitis C-specific CD4⁺ T-cell proliferation and pro-inflammatory cytokine responses in vivo.

MATERIALS AND METHODS

Subjects and Samples

Patients from the clinical trial at one site who gave informed consent under a separate IRB-approved protocol provided 40ml of peripheral blood once prior to initiation of study medication or placebo and once at 20 ± 2 weeks during active treatment or placebo.

Peripheral blood mononuclear cells (PBMC)

PBMC were isolated from blood using Histopaque-1077 density centrifugation (Sigma-Aldrich, St. Louis MO) (23, 24). PBMC were suspended in RPMI 1640 containing L-glutamine, 10% human AB serum, 25mM HEPES (Gibco/Invitrogen), and penicillin/streptomycin (Gibco/Fisher scientific).

Antigens and reagents

Recombinant genotype 1a-derived HCV core (SOD-c22, aa 2-120), NS3/4 (SOD-c200, aa 1192-1931) NS5 (SOD-NS5, aa 2054-2995), and control superoxide dismutase (SOD) proteins were generously provided by Novartis Diagnostics (Emeryville, CA) (1, 25–27). Tetanus toxoid 0.5ug/ul (Sigma) and Candida albicans 20ug/ul (Sigma) were utilized as antigen positive controls.

T-cell proliferative response

CD4⁺ T-cell proliferation assay was performed as previously described (1, 23, 24). Briefly, PBMC (2x10⁵ cells/well) were stimulated for 7 days with SOD-c22, SOD-c200, and SOD-NS5 and controls and harvested after 16 hours of ³H-thymidine incorporation (1μCi/well) (Dupont NEN, Boston, MA). As a positive control, PBMC were stimulated with a T-cell mitogen phytohemagglutinin (PHA) at 2 μg/mL and harvested on day 4. The results were expressed as a stimulation index (SI) with mean counts per minute (cpm) in stimulated wells divided by the mean cpm in control wells. SI greater than 3 was considered positive (24). Testing of in vitro T-cell suppression was performed using PBMC (1–2x10⁵ cells/well) with and without 2 μg/mL PHA in presence of silymarin (Sigma) at concentrations ranging from 0.1 to 100 μg/ml.

T-cell IFNγ and IL-10 secretion

The frequency of IFNγ and IL-10 producing T-cells were quantified by Elispot assay, performed using 2 x 10⁵ PBMC/well in triplicates with and without HCV and control antigens as described previously (9, 11, 24). HCV-specific CD4⁺T cell IFNγ (Th1) or IL-10 (Tr1) responses in whole PBMC was examined using recombinant HCV and control proteins (10μg/ml). Spot forming units (SFU) were counted using IFL04 Elispot reader (AID, Hamburg, Germany). HCV-specific IFNγ or IL-10 T-cell frequency was calculated by subtracting the mean SFU in negative control wells from mean SFU in antigen-stimulated wells and expressed as HCV-specific SFU/10⁶ PBMC.

Quantification of regulatory T-cells

2 x 10⁵ PBMC were stained with 7-actinomycin-D/Viaprobe, CD69 FITC, CD25 PE, CD45RO APC, CD4 PE-Cy7, CD4 APC (all BD Bioscience), CD4 APC-Alexa Fluor 750 (eBioscience), foxp3 FITC (eBioscience) as previously described (11). All samples where acquired using FACS Diva software on a FACSCanto, and analyzed using FlowJo (Treestar Inc., Ashland OR).

Serum cytokine analysis

Freshly isolated plasma from whole blood were aliquoted and stored at −80c for cytokine analysis of IL-4, IL-6, IL-8, IL-10, IL-12, IL-17, TNFα, TNFβ, and IFNγ using Milliplex MAP Kit (Millipore, Billerica, MA) on a Luminex 200 system (Luminex Corporation, Austin, TX) using Masterplex QT software (Hitachi/MiraiBio, South San Francisco, CA) and IP-10 by ELISA (R&D Systems).

Interferon-stimulated gene expression

Cryopreserved PBMC from pre- and on-treatment samples from 6 patients from each treatment assignment were thawed and cultured overnight in RPMI 10%ABS media with or without stimulation using rhIFNα (0.1ng/ml) or PHA (2 μg/mL). Using Taqman Gene expression kits and Cells-to-CT kit (Applied Bioscience), mRNA levels for 18S RNA, GAPDH, IRF7, Mx1, PP2A, ISG15, USP18 and STAT1 were measured on ABI 7000 rtPCR system (Applied Bioscience) and analyzed by ΔΔCt methodology.

Statistical analysis

The median values for clinical and immunologic parameters at individual time points were compared using the nonparametric Kruskal-Wallis ANOVA, Wilcoxon Rank Sum or Mann-Whitney U test. MANOVA for repeat measures was utilized to determine the significance of the difference of pre-and on-treatment immunological results across treatment assignment. Statistical analysis was performed using JMP 9 (SAS Institute Inc, Cary NC). A p-value < 0.05 was considered significant.

RESULTS

Baseline characteristics and clinical outcomes

Thirty-two patients consented to participate in the immunological substudy. The placebo and two treatment arms were similar in terms of age and gender, but there was a modest over-representation of black patients in the silymarin 700mg arm (Table 1). Similar to overall trial results, substudy patients demonstrated no significant on-treatment change in ALT or HCV RNA regardless of treatment assignment.

Table 1.

Baseline patient characteristics and clinical outcome

	Treatment Assignment
	Placebo TID	420mg TID	700mg TID
At Randomization				p
N	10	11	11
Median Age (IQR)	57 (52–59)	56 (50–60)	55 (48–57)	0.42
Gender (M/F)	9/1	9/2	6/5	0.14
Race (W/B)	10/0	10/1	6/5	0.01
Median ALT (IQR)	128 (106–187)	109 (90–181)	110 (81–165)	0.58
Median log HCV RNA (IQR)	6.1 (5.8–6.6)	6.4 (5.9–6.7)	6.3 (5.9–6.5)	0.69
During Treatment
Median ALT at week 20 (IQR)	125 (99–189)	104 (73–161)	118 (100–157)	0.35*
Median log HCV RNA at week 12 (IQR)	6.3 (5.7–6.4)	6.4 (5.9–6.7)	6.2 (5.7–6.4)	0.46*
Median log HCV RNA at week 24 (IQR)	6.3 (5.8–6.6)	6.4 (6.1–6.5)	6.2 (6.0–6.4)	0.83*

^*comparison across groups for difference from values at randomization

Oral administration of silymarin has no significant impact on HCV-specific CD4⁺T-cell proliferation in chronic hepatitis C

As shown in Figure 1A–C, neither silymarin treatment arm nor the placebo group demonstrated any significant change in CD4⁺ T-cell proliferation in response to HCV Core, NS3/4 or NS5 proteins from baseline to on-treatment (summed HCV-specific SI 3.7 to 3.2 in placebo arms, 2.6 to 3.0 in 420mg, and 3.8 to 4.0 in the 700mg arm, p=0.54). C. albicans-specific T-cell proliferation responses were at baseline approximately 1-log greater than HCV-specific responses (median 70 SI units, Figure 1D) but similarly were unchanged between baseline and week 20. By contrast, proliferation responses induced by the T-cell mitogen PHA decreased significantly in silymarin-treated patients (mean 205 to 77 SI units, p=0.014 combining 420mg and 700mg arms, Figure 1F) but not in placebo-treated subjects (mean 122 to 144 SI units, p=0.51; p=0.035 for difference of differences).

Figure 1. CD4⁺T-cell proliferation responses to HCV- and non-HCV-specific stimulation at baseline and during silymarin therapy.

Stimulation index for PBMC stimulated with A. c22 (Core), B. c200 (NS3/4), C. NS5 (aa2504-2995), D. C. albicans, E. sum of HCV, and F. phytohemagglutinin. Pie charts show fraction of patients with 50% increase to SI greater than 3 (black), 50% decrease from SI greater than 3 to below 3 (white), or no significant change (grey). Active treatment patients demonstrated a statistically significant decrease in PHA-induced T-cell proliferation (mean −128.1 SI units, p=0.0135 when 420mg and 700mg groups combined) and the decrease was statistically different from the difference in placebo patients (+22.5 SI units, p=0.035).

HCV-specific T-cell cytokine responses were unchanged by silymarin therapy

To examine the impact of oral silymarin administration on T-cell cytokine production, we performed IFNγ and IL-10 Elispot with freshly isolated PBMC. As shown in Figure 2 and Supplemental Figure 1, there were no significant changes in the frequency of HCV-specific CD4⁺T-cells producing IFNγ or IL-10 between baseline and week 20 of treatment. Interestingly, the frequency of C. albicans-specific IFNγ-producing T-cells decreased in both silymarin treatment arms (Figure 2F) without any consistent effect observed on C. albicans-specific IL-10 T-cell frequency. Combining the silymarin treatment arms, Candida-specific T-cell IFNγ decreased from a mean 256 to 114 SFU/10⁶ PBMC (p=0.0066) compared to no change (116 to 107 SFU/10⁶ PBMC, p=0.85) in the placebo arm with a strong trend towards a difference in these differences by MANOVA (p=0.076). Anecdotally, the single individual in the silymarin 700mg cohort with detectable HCV-specific cytokine and proliferation responses at baseline manifested parallel reductions in HCV- and Candida-specific T-cell proliferation and IFNγ production with increased IL-10 responses. While overall HCV-specific T-cell proliferation and cytokine production was not altered by silymarin therapy, we observed evidence that silymarin suppresses mitogen- and recall-antigen-induced proliferation and cytokine responses, respectively, in vivo.

Figure 2. Frequency of HCV- and non-HCV-specific IFNγ-producing T-cells at baseline and during silymarin therapy.

IFNγ spot-forming units per 10⁶ PBMC are shown for PBMC stimulated with A. c22 (Core), B. c200 (NS3/4), C. NS5 (aa2504-2995), D. summed HCV-specific responses, E. Tetanus toxoid, and F. C. albicans. Pie charts show fraction of patients with ≥50 SFU/10⁶ increase to SFU/10⁶ greater than 50 (black), 50 SFU/10⁶ decrease from SFU/10⁶ greater than 50 to below 50 (white), or no significant change (grey). Active treatment (420mg plus 700mg) patients demonstrated a statistically significant decrease in frequency of C. albicans-induced IFNγ production (mean −142 SFU/10⁶, p=0.007). There was a trend for the difference in C. albicans response in active treatment patients to be statistically different from the difference in placebo patients (−9 SFU/10⁶, p=0.076). G. Composite of changes of ALT, HCV RNA, HCV- and control-specific SI, IFNγ and IL-10 responses for patient in 700mg TID group with detectable HCV-specific T-cell responses at baseline.

Silymarin did not alter regulatory T-cell frequency or global cytokine levels

We then asked if other non-specific signs of reduced inflammatory markers were present in silymarin-exposed subjects. As shown in Table 2, we did not observe any significant reduction in serum pro-inflammatory (IL-6, IL-12p40, IL-17, TNFα, IFNγ) or anti-inflammatory (IL-8, IL-10). Serum IP-10 chemokine levels, a marker of interferon-responsiveness (when low)(28), were similarly unchanged by silymarin 700mg thrice daily. No change in the frequency of CD4⁺foxp3⁺ or CD4⁺CD25^hi regulatory T-cells nor activated CD4⁺CD69⁺ or CD8⁺CD69⁺ activated T-cells were identified between baseline and week 20 in either treatment or placebo group (Figure 3C–F).

Table 2.

Serum cytokine levels at baseline and during silymarin administration

	Placebo TID		420mg TID		700mg TID		p*
	N=6		N=6		N=7
	Pre	Post	Pre	Post	Pre	Post
IFNγ(pg/ml)	7.3 ± 12.5	23.7 ± 37.5	< LLD	< LLD	2.8 ± 2.7	3.0 ± 3.5	0.11
IL-4 (pg/ml)	< LLD	4.5 ± 6.9	< LLD	< LLD	< LLD	< LLD	0.36
IL-6 (pg/ml)	4.4 ± 4.0	8.9 ± 10.6	5.5 ± 7.3	< LLD	4.4 ± 3.3	4.2 ± 3.5	0.17
IL-8 (pg/ml)	26.5 ± 32.5	40.6 ± 48.7	8.3 ± 5.0	8.8 ± 6.1	30.1 ± 58.7	5.0 ± 6.6	0.20
IL-10 (pg/ml)	5.7 ± 5.8	10.7 ± 17.6	7.2 ± 11.1	3.6 ± 14.2	4.1 ± 2.9	4.8 ± 4.7	0.69
IL-12p40 (pg/ml)	27.5 ± 44.8	102.2 ± 135.6	39.7 ± 86.4	44.8 ± 97.1	10.9 ± 11.4	5.4 ± 6.3	0.04
IL-17 (pg/ml)	51.5 ± 111.6	54.1 ± 104.9	< LLD	< LLD	37.7 ± 79.8	< LLD	0.42
TNFα(pg/ml)	10.4 ± 3.2	12.9 ± 5.9	6.3 ± 2.4	8.0 ± 4.2	4.8 ± 1.1	7.2 ± 2.1	0.59
TNFβ(pg/ml)	9.5 ± 18.0	21.5 ± 446.3	5.0 ± 7.1	4.1 ± 10.0	< LLD	< LLD	0.34
IP-10 (pg/ml)	500.7 ± 173.5	456.8 ± 205.0	ND	ND	500.7 ± 187.9	533.0 ± 220.9	0.06

^*p-value for differences in mean change across groups

Figure 3. Frequency of activated and regulatory T-cells at baseline and during silymarin therapy.

A. Gating strategy for identifying CD4⁺CD25^hi regulatory T-cells. B. Examples of pre- and post-treatment determination of CD4⁺foxp3⁺ frequency. C. Pre- and on-treatment CD4⁺CD25^hi T-cells. D. Pre- and on-treatment CD4⁺foxp3⁺ T-cells. E. Pre- and on-treatment %CD69⁺/CD4⁺ T-cells. F. Pre- and on-treatment %CD69⁺/CD8⁺ T-cells. None of specified populations significantly changed frequency in any group from baseline to on-treatment.

Subtle alterations of lymphocyte interferon-stimulated gene expression were observed related to silymarin therapy

To determine if silymarin might alter interferon-associated or interferon-stimulated gene expression in a manner that might suggest a restoration of interferon-sensitivity, we measured baseline and week 20 lymphocyte expression of STAT1, Mx1, ISG15, IRF7, USP18 and PP2A in lymphocytes from six subjects in each group. USP18 and PP2A did not amplify in the majority of samples from all three groups of patients, and therefore these data were excluded. As shown in Table 3, there was no change in expression of STAT1, Mx1, ISG15 and IRF7 in ex vivo unstimulated lymphocytes in any arm. There was a trend for a greater increase in interferon-induced ISG15 mRNA expression in the 700mg group (median 4.7 fold versus 2.1 fold, p=0.12; 4/6 patients in the 700mg group exhibited a ≥ 3-fold increase in interferon-induced ISG15 compared to 0/6 in the placebo arm p=0.03, Supplemental Figure 2A). PHA-stimulation led to greater than 5-fold increases in STAT1 and Mx1 in 3/6 (50%) of placebo patients but 0/10 (0%) silymarin-exposed patients (p=0.035, Supplemental Figure 2B). Thus, while the highest dose silymarin group exhibited improvement in IFNα-sensitivity, silymarin suppressed PHA-induced lymphocyte activation.

Table 3.

Fold-change of interferon-stimulated gene expression on-treatment relative to pre-treatment

		Group
Condition/Gene		Placebo	420mg	700mg
Media-stimulated
IRF7	N	6	4	6
	Median (Min-Max)	0.8 (0.1–83.9)	1.1 (0.1–3.7)	1.8 (0.4–17.8)
Mx1	N	6	4	6
	Median (Min-Max)	1.5 (0.1–109.1)	0.5 (0.0–3.4)	1.6 (0.6–10.2)
ISG15	N	1	2	5
	Median (Min-Max)	25.1 (25.1–25.1)	143.2 (48.3–238.0)	13.1 (3.6–156.0)
STAT1	N	6	4	6
	Median (Min-Max)	0.6 (0.1–94.7)	0.7 (0.1–4.6)	1.0 (0.2–7.5)
IFNα-stimulated
IRF7	N	6	4	6
	Median (Min-Max)	0.9 (0.1–34.3)	0.5 (0.3–9.9)	1.7 (1.2–11.0)
Mx1	N	6	4	6
	Median (Min-Max)	1.6 (0.1–168.9)	0.5 (0.3–6.9)	2.0 (0.9–6.0)
ISG15	N	5	4	6
	Median (Min-Max)	2.1 (0.1–2.6)	0.3 (0.0–8.8)	4.7 (0.5–13.5) †
STAT1	N	6	4	6
	Median (Min-Max)	1.4 (0.2–119.4)	0.6 (0.3–5.4)	2.0 (0.9–6.5)
PHA-stimulated
IRF7	N	6	4	6
	Median (Min-Max)	4.2 (0.4–112.2)	0.7 (0.2–2.3)	1.3 (0.5–2.6)
Mx1	N	6	4	6
	Median (Min-Max)	4.8 (0.5–117.4)	1.0 (0.1–2.1)	1.6 (0.4–2.7)
ISG15	N	2	3	5
	Median (Min-Max)	19.1 (9.1–29.1)	0.4 (0.1–3.1)	1.0 (0.5–1.7)
STAT1	N	6	4	6
	Median (Min-Max)	3.7 (0.3–147.5)	1.3 (0.5–4.5)	1.3 (0.2–2.1)

^† p=0.12 for difference between placebo and 700mg group

Dose responsiveness of silymarin-induced T-cell proliferation

To address whether or not silymarin at concentrations obtainable by oral administration could potentially suppress in vivo lymphocyte responses, we tested PHA-induced lymphocyte proliferation over a broad concentration range. As shown in Figure 4A, no significant suppression of T-cell proliferation developed at silymarin concentrations ≤ 1 μg/dl. The IC₅₀ for silymarin appeared variable among subjects within the range of 10–50 μg/dl (Figure 4B–D).

Figure 4. In vitro sensitivity of T-cells to proliferation suppression by silymarin.

A. Percentage of maximal proliferation after PHA-stimulation of PBMC from 3 baseline samples in the presence of silymarin 0.1–100μg/ml. B–D. Percentage of maximal proliferation after PHA stimulation in PBMC from 3 individual subjects in the presence of silymarin 1–100μg/ml.

*p<0.05 **p<0.001 compared to maximal proliferation condition

DISCUSSION

Silymarin, a mixture of 6 major flavonolignans and multiple flavonoids (29), remains the most widely used herbal supplement utilized for chronic liver diseases, including chronic hepatitis C. In addition to pleiomorphic effects on hepatocytes (29), silymarin has been demonstrated to have anti-inflammatory immunomodulatory effects both in vitro (20, 21, 30) and in multiple rodent hepatitis models (14, 16). In both normal humans and mice, silymarin at concentrations 50–100 μM in vitro has been shown to suppress mitogen and anti-CD3-mediated T-cell proliferation, IFNγ, IL-2 and TNFα production, and NFκB nuclear translocation (21). Similarly in the concanavalin A T-cell mediated hepatitis model, intraperitoneal administration of silibinin suppressed T-cell pro-inflammatory cytokine production (IFNγ, TNFα, IL-2), increased T-cell production of anti-inflammatory IL-10, and reduced intrahepatic expression of iNOS (16). More recently a direct antiviral effect of silymarin at concentrations of 1–20 μM has been appreciated in vitro (19, 31), an effect clinically observed after administration of intravenous silibinin to patients with chronic hepatitis C infection (17, 18). While short term courses of oral silymarin has never consistently shown salutary effects in alcoholic or viral liver disease (13), long-term silymarin use in the HALT-C trial was associated with slower fibrosis progression (32) suggesting that oral silymarin might indeed have anti-inflammatory effects in vivo in humans.

The present study reflects the first systematic, prospective assessment of the immunological effects of a well-characterized orally-administered silymarin preparation on human T-cell responses in chronic hepatitis C in vivo. The parent clinical trial identified no significant effect of silymarin on alanine aminotransferase or HCV RNA titers. Overall, we did not identify any significant impact of silymarin on HCV-specific T-cell proliferation, IFNγ production, or IL-10 secretion. Similarly, there was no global alteration of serum cytokines, liver-related chemokines (IP-10), or circulating regulatory or activated T-cell frequencies. The lack of suppression of HCV-specific responses in our study is discrepant with the in vitro findings of Morishima et al. (20) who clearly demonstrated that a distinct preparation of silymarin, MK001, at 20 μg/ml suppressed HCV-specific CD4⁺ and CD8⁺ T -cell proliferative responses from lymphocytes isolated from HCV-infected patients in vitro. In that study, similar to the present study, the suppressive effect of silymarin was more convincingly demonstrated with PHA, anti-CD3, C. albicans or tetanus toxoid stimulation, due to low frequency and weak HCV-specific responses typically found in chronic hepatitis C patients (9). Indeed, in the single patient in the 700mg silymarin group with detectable HCV-specific responses, suppression of virus-specific proliferation and IFNγ production with increased IL-10 was observed. Combined with the rarity of detectable HCV-specific immune responses in all cohorts, that the plasma concentrations of flavonolignans such as silybin A and silybin B achieved from oral administration were approximately 1–2 log lower than concentrations utilized in vitro (33) likely contributed to difficulties in detecting changes in virus-specific T-cell responses. Our in vitro PHA proliferation results suggest an IC₅₀ of silymarin in the 10–50μg/ml, a concentration not currently attainable with oral silymarin. We cannot discount the possibility that a type II error explains our negative findings, as the sample size of 10–11 per group was below our initial targets of 20 patients per group. Our ability to sample patients only at one timepoint during the study could also have impaired our ability to detect dynamic changes in T-cells responses particularly early after medication initiation. Furthermore, the initial starting population of prior interferon non-responders could also potentially bias against detection of changes in T-cell responses.

Despite these potential limitations, consistent with in vitro findings of Morishima et al., we did observe significant suppression in vivo of Candida-specific T-cell IFNγ production as well as PHA-induced T-cell proliferation. Notably, suppression of mitogenic or anti-CD3-mediated T-cell activation and proliferation has been consistently demonstrated with various preparations of silymarin in vitro (21, 34, 35). One may speculate that this non-specific suppression of T-cell pro-inflammatory effects might explain the association of silymarin exposure and reduced hepatic fibrosis progression observed in the large, long-term HALT-C cohort study (32).

The possibility that silymarin might restore interferon-responsiveness via its oxidant properties was the initial rationale for the human clinical trials that ultimately demonstrated a direct antiviral effect of intravenous silibinin (17). In our laboratory’s previous work examining the role of T-cells in the outcome of IFNα/ribavirin therapy for chronic HCV infection, early virological response was paradoxically associated with baseline suppressed PHA-induced proliferation responses (9). Similarly, both suppressed serum levels of the ISG IP-10 (28) and suppressed pre-treatment intrahepatic ISG mRNA (36) have been associated with improved virological responses suggesting that basal ISG hyperactivation is associated with reduced interferon sensitivity. Similar to previous in vitro studies (20, 34, 35), we observed significant reductions of mitogen-induced T-cell proliferation in vivo and in vitro that might be a signature of the restoration of interferon-sensitivity. Focusing on IRF7, ISG15, Mx1 and STAT1 whose intrahepatic expression strongly correlates with interferon-responsiveness (37), we did not observe any significant alteration in basal lymphocyte ISG mRNA expression from silymarin. However, a strong trend toward increased interferon-induced ISG15 expression was observed in patients in the high-dose silymarin group with the caveat that the extent to which lymphocyte ISG induction correlates with intrahepatic ISG induction is unknown. Furthermore, these data suffer from a significant type II error due to limitations on paired samples available to perform these analyses post hoc. Our data and a small clinical trial by Rutter et al. (38) that raised possibility that intravenous silibinin improved interferon sensitivity in vivo, provide rationale for a formal prospective evaluation of silymarin’s impact on intrahepatic ISG expression and a potential interferon-sensitizing role.

In conclusion, oral silymarin administered at higher than usual dosages had no measurable impact on hepatitis C-specific T-cell proliferation, IFNγ or IL-10 responses in vivo. Silymarin also had no effect on global serum cytokines, IP-10, lymphocyte ISG expression or regulatory T-cell frequency. However, silymarin did significantly reduce the pro-inflammatory cytokine response to Candida albicans and PHA lectin-induced T-cell proliferation, similar to previous in vitro findings, confirming that silymarin does exert an immunomodulatory effect in vivo. The impact of this anti-inflammatory effect on long-term liver health in chronic hepatitis C should be examined in future clinical investigation.

Supplementary Material

Supp Fig S1. Supplemental Figure 1. Frequency of HCV- and non-HCV-specific IL-10-producing T-cells at baseline and during silymarin therapy.

IL-10 spot-forming units per 10⁶ PBMC are shown for PBMC stimulated with A. c22 (aa 2-120 HCV Core), B. c200 (aa 1192-1931 NS3/4), C. NS5 (aa2504-2995), D. summed HCV-specific responses, E. C. albicans, and F. Tetanus toxoid. No antigen or group showed any statistically significant changes.

Supp Fig S2. Supplemental Figure 2. Interferon-related and interferon-stimulated gene expression prior to and during silymarin or placebo treatment.

A. Relative upregulation of STAT1, ISG15, Mx1 and IRF7 of – pre- and on-treatment PBMC after stimulation with interferon-alpha. B. Relative upregulation of STAT1, ISG15, Mx1 and IRF7 of –pre- and on-treatment PBMC after stimulation with phytohemagglutinin.

Abbreviations

HCV

hepatitis C virus

IFNγ

interferon-gamma

IL

interleukin

IP-10

Interferon gamma-induced protein 10

IRF7

interferon-response factor 7

ISG

interferon-stimulated gene

Mx1

Myxovirus resistance 1

NFκB

nuclear factor kappa B

PBMC

peripheral blood mononuclear cells

PHA

phytohemagglutinin

SFU

spot-forming unit

SI

stimulation index

SOD

superoxide dismutase

STAT1

signal transducer and activator of transcription 1

TNFα

tumor necrosis factor-alpha

TNFβ

tumor necrosis factor-beta