SEARCH

SEARCH BY CITATION

Keywords:

  • Innate immunity;
  • liver fibrosis;
  • viral hepatitis

Abstract

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

Recurrence of hepatitis C (HCV) postliver transplant is universal, with a subgroup developing rapid hepatic fibrosis. Toll-like receptors (TLRs) are critical to innate antiviral responses and HCV alters TLR function to evade immune clearance. Whether TLRs play a role in rapid HCV recurrence posttransplant is unknown. We stimulated peripheral blood mononuclear cells (PBMCs) from 70 patients with HCV postliver transplant with TLR subclass-specific ligands and measured cytokine production, TLR expression and NK cell function. Rate of fibrosis progression was calculated using posttransplant liver biopsies graded by Metavir scoring (F0–4; R = fibrosis stage/year posttransplant; rapid fibrosis defined as >0.4 units/year). Thirty of 70 (43%) patients had rapid fibrosis progression. PBMCs from HCV rapid-fibrosers produced less IFNα with TLR7/8 stimulation (p = 0.039), less IL-6 at baseline (p = 0.027) and with TLR3 stimulation (p = 0.008) and had lower TLR3-mediated monocyte IL-6 production (p = 0.028) compared with HCV slow fibrosers. TLR7/8-mediated NKCD56 dim cell secretion of IFNγ was impaired in HCV rapid fibrosis (p = 0.006) independently of IFNα secretion and TLR7/8 expression, while cytotoxicity remained preserved. Impaired TLR3 and TLR7/8-mediated cytokine responses may contribute to aggressive HCV recurrence postliver transplantation through impaired immune control of HCV and subsequent activation of fibrogenesis.

Abbreviations
GMF

geometric mean fluorescence

HCV

hepatitis C virus

HSC

hepatic stellate cell

IFNα

interferon alpha

IFNγ

interferon gamma

IL-6

interleukin-6

mDC

myeloid dendritic cell

NK cell

natural killer cell

PAMPs

pathogen associated molecular patterns

PBMCs

peripheral blood mononuclear cells

pDC

plasmacytoid dendritic cell

TLR

toll-like receptor

TNFα

tumor necrosis factor alpha

Hepatitis C infection (HCV) is the most common indication for liver transplantation globally, accounting for over 50% of liver transplants performed annually [1]. However, HCV recurrence after transplantation is universal and follows an accelerated course compared with the natural history of HCV infection prior to transplantation [2]. Clinical outcomes are highly variable, with around 10–30% developing aggressive HCV recurrence and rapid fibrosis progression, culminating in graft cirrhosis, liver failure and either death or retransplantation [3]. Why some patients develop rapid fibrosis progression posttransplant remains poorly understood and there is currently no reliable method for predicting those patients most at risk of aggressive HCV recurrence.

Toll-like receptors (TLRs) form the cornerstone of the innate immune system and orchestrate the body's initial immune response against infection. TLRs bind to pathogen associated molecular patterns, or PAMPs, which are highly conserved molecules expressed by invading pathogens [4]. When TLRs bind with their corresponding ligands, they initiate a downstream signaling cascade, culminating in production of type I interferons, antiviral proteins including interferon-stimulated genes and proinflammatory cytokines that have been shown to be critical in initiating antiviral immune responses, including those against hepatitis C infection [5].

There is a large body of evidence demonstrating the importance of TLRs in HCV infection. HCV core and nonstructural proteins are recognized by TLRs in both peripheral immune and liver cells [6]. HCV is able to activate TLRs 2, 3, 4 and 9 [7, 8] and this triggers an inflammatory cascade leading to chronic inflammation and hepatic fibrosis. In particular, TLR stimulation of hepatocytes and hepatic stellate cells (HSCs) by HCV viral proteins has been shown to increase inflammatory cytokine release and therefore HSC activation, matrix production and collagen deposition leading to hepatic fibrosis and cirrhosis [9]. HCV has also been shown to evade immune clearance through specifically targeting and impairing TLR 2, 3, 4, 7/8 and 9 signaling in T cells and plasmacytoid dendritic cells (pDCs) [10, 11]. Natural killer cells (NK cells) are also increasingly recognized to be essential in immune defenses against HCV infection [12, 13] and have been shown to express TLR3 and TLR7/8 functional receptors [14]. NK cells have also been shown to have antifibrotic effects in HCV infection via inhibition of HSCs both directly through apoptosis or indirectly via IFNγ, which inhibits HSC activation [15]. Therefore, both recipient peripheral immune cells and donor liver immune cells have key roles to play in HCV recurrence postliver transplantation.

Given the importance of TLRs in early viral infection (such as occurs in the grafted liver after transplantation), coupled with the proven importance of peripheral blood immune cell TLR function in chronic HCV infection, we hypothesized that abnormalities of TLR function may contribute to the development of HCV-mediated rapid fibrosis progression postliver transplantation. We therefore examined TLR function in peripheral blood mononuclear cells (PBMCs) in patients with hepatitis C posttransplant, comparing those with rapidly progressive and slowly progressive graft fibrosis. We identified several novel changes in TLR3 and TLR7/8 function that may contribute to more aggressive recurrence of HCV-mediated liver injury.

Methods

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

Study design

This cross-sectional study was conducted at a single adult transplant centre. Patients were recruited between June 6, 2009 and 31 July 2011. Follow-up was until death, retransplantation or July 31, 2011. To be included in the study, patients had to be older than 18 years, at least 6 months postliver transplant and have a minimum follow-up period of 6 months. In order to avoid potential confounding effects on TLR function measurement, patient blood samples were taken when patients were clinically stable, without intercurrent autoimmune hepatitis, bacterial infection or acute cellular rejection. Any blood sample inadvertently taken at the time of occult sepsis or rejection was excluded from the study. Ethical approval for the study was provided by the institutional ethics committee.

Study population

Seventy-seven patients with HCV postliver transplant were recruited for the study, of whom 70 had liver biopsies suitable for calculation of fibrosis rate. There were no recipients of donation after cardiac death grafts included in this study. No patients in this study had received HCV therapy postliver transplant prior to or during the study period.

Data collection and analysis

Patient data were prospectively recorded in the Liver Transplant Unit of Victoria database. Data recorded included donor and recipient demographics and clinical variables, including medication and HCV viral load. Posttransplant HCV viral load was measured using the BDNA Bayer Version 3 instrument. Clinical endpoints including annual fibrosis stage, development of cirrhosis, death and retransplantation were also recorded prospectively.

Measurement of fibrosis progression using liver biopsy

In all patients, protocol liver biopsies were performed annually, with additional liver biopsies when clinically indicated. Liver biopsies were assessed for HCV-related fibrosis stage and grade by a single experienced liver pathologist who was blinded to both clinical details and the results of the TLR function studies (PC). We used the Metavir scoring system for liver fibrosis measurement [16]. Liver biopsies with evidence of mixed pathology or in those where it was unclear whether hepatitis C was the sole cause of pathological findings, were excluded from the study. This resulted in 7 of 77 patients being excluded from the study. For the remaining 70 patients, biopsies without copathology present on histology were available to determine fibrosis progression. A total of 283 liver biopsies were performed on the study cohort. A total of 59 liver biopsies were excluded due to the presence of pathology other than hepatitis C infection and nine were excluded for being of insufficient size (less than eight portal triads present).

Definition of rapid fibrosis progression

Fibrosis rate was calculated by dividing the Metavir fibrosis stage of the most recent liver biopsy by the number of years posttransplant. Patients with rapid fibrosis were then defined as those with a fibrosis rate greater than 0.4 fibrosis units/year, dichotomized about the median observed rate in our cohort, as described by others[17].

Using the definition of rapid fibrosis outlined above, 40 patients (57%) had slow fibrosis progression and 30 patients (43%) had rapid fibrosis progression. The median fibrosis rate was 0.4 fibrosis units/year (IQR 0.19–0.64). No patients died from or were retransplanted for liver failure secondary to HCV-mediated cirrhosis during the course of the study.

Isolation of PBMCs

Whole blood (35 mL) was obtained by venepuncture and collected in heparinized tubes. PBMCs were extracted using the Ficoll–Paque density centrifugation method (Ficoll Paque Plus Solution, GE Healthcare, UK) and stored in complete RPMI-1640 media (Sigma Lifesciences, St. Louis, MO, USA) with 1% penicillin, 1% L-glutamine and 5% fetal calf serum with dimethyl sulfoxide in liquid nitrogen.

Measurement of IL-6, TNFα and IFNα by ELISA

PBMCs from subjects were rapidly thawed, then cultured for 24 h at 37° in complete RPMI media with 1% penicillin, 1% L-glutamine and 5% fetal calf serum with TLR ligands at a concentration of 1×106 cells/mL. TLR subclass-specific ligands and their concentrations used for cell stimulation were P3C [100ng/mL] (TLR2), PIC [25μg/mL] (TLR3; Invitrogen, Carlsbad, CA, USA), LPS [100ng/mL] (TLR4), R848 [5μg/mL] (TLR7/8; Invitrogen) and CpG (TLR9; CpG 2006 [3M] for monocytes, CpG 2216 [5μg/mL] for plasmacytoid dendritic cells; both from Geneworks, Hindmarsh, SA, Australia). Il-6, TNFα and IFNα produced by cell stimulation were then quantified in the cell culture supernatants using ELISA according to the manufacturer's instructions (BD Biosciences, San Jose, CA, USA and R&D Systems, Minneapolis, MN USA). PBMCs cultured in media alone were used as controls. Concentrations of TLR ligand and duration of incubation period were chosen in order to maximize cytokine production based on results from optimization experiments.

Flow cytometry analysis of cell surface markers and intracellular cytokine

A smaller subset of 34 HCV patients posttransplant were selected from the larger cohort for flow cytometry. Patients were selected by time of recruitment and subjects in this cohort were matched between rapid and slow fibrosis groups for age, gender, time posttransplant, immunosuppression and donor age. PBMCs were rapidly thawed and cultured for 6 h with TLR 7/8 ligand R848 [5 μg/mL], TLR7 ligand loxoribine [1 mM] and TLR3 ligand PIC [25 μg/mL] at a cell concentration of 1×106 cells/mL (Invitrogen). The Golgi transport inhibitor Brefeldin A (Golgiplug, Becton Dickinson, Franklin Lakes, NJ, USA) was added for the final 4 h of culture at a concentration of 1 μg/mL. Cells were then stained with fluorochrome-conjugated antibodies for cell surface markers CD3-Pacific Blue (Clone UCHT1, BD Biosciences), CD56-PE-Cy7 (Clone B159, BD Biosciences), CD16-APC (Clone 3G8, BD Biosciences), CD14-APCCy7 (Clone MΦPG, BD Biosciences) and CD69-PE (BD Biosciences). Following this, cells were permeabilized (Cytofix/Cytoperm solution, BD Biosciences) and stained for intracellular cytokines IL-6-PE, TNFα-APC and IFNγ-FITC (all by BD Biosciences). Separate cell aliquots (1×106 cells/mL) not stimulated with TLR ligands were also stained for surface expression of CD14-APC-Cy7, CD3-Pacific Blue, CD56-PE-Cy7 and CD107a-PE (BD Biosciences) to assess NK cell degranulation and then permeabilized and stained with TLR7-CSF (R&D Systems) and TLR8-PE (Clone 44C143, Imgenex, San Diego, CA, USA) fluorochrome antibodies to measure intracellular TLR7 and TLR8 expression. Cells cultured with media alone were used as controls. Cells were run on a FACSCaliburTM flow cytometer (BD Biosciences) and results analyzed using Flowjo software (Flowjo 9.2, Treestar, Ashland, OR, USA).

Co-culture PBMCs with TLR ligand and exogenous IFNα

PBMCs at a concentration of 1×106 cells/mL were cultured for 6 h with TLR7/8 ligand R848, with and without either recombinant IFNα-2a 1000 IU/mL (rIFNα-2a; Roche, Australia Pty Ltd) or mouse antihuman neutralizing interferon-α antibody 1 μg/mL (Becton Dickinson). Cells were stained for CD56-PE-Cy7 and CD3-Pacific Blue, then permeabilized and stained for detection of intracellular TNFα-APC and IFNγ-FITC using flow cytometry.

Statistics

Statistical analysis was performed using SPPS (SPSS Inc., Version 19 software, Chicago, IL, USA) and Prism 5.0c for Macintosh software (Graphpad Software Inc., La Jolla, CA, USA). Clinical parameters were compared between groups using the Student t-test for continuous variables and the chi-square test for categorical variables. Nonparametric statistical tests were used for comparing group data: Mann–Whitney for two groups and Kruskall–Wallis with Dunn's posttest for multiple groups. Correlation between fibrosis rate and TLR function was determined using Spearman correlation. A twosided p-value of 0.05 was considered to be statistically significant. Multivariate analysis was performed using binary logistic regression and backward stepwise elimination procedure and the model was built using variables with significance <0.1.

Results

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

HCV patients with rapid fibrosis progression have impaired TLR7/8-induced type I interferon and impaired TLR3 and TLR9-mediated proinflammatory cytokine production compared with those with slow fibrosis.

We compared type I interferon and proinflammatory cytokine production by PBMCs in response to TLR ligand stimulation in HCV patients with rapid fibrosis (n = 30) and slow fibrosis progression (n = 40). IFNα production was expressed as mean concentration, rather than a fold increase above baseline, as all IFNα production at baseline (unstimulated with TLR ligands) was zero pg/mL.

HCV patients with rapid fibrosis progression had impaired IFNα production in response to TLR7/8 stimulation with R848 compared with patients with slow fibrosis progression (p = 0.039, Figure 1).

image

Figure 1. Mean concentration of IFNα secretion (pg/mL) by PBMCs in response to TLR3, TLR7/8 and TLR9 stimulation in HCV patients with rapid fibrosis (n = 30) compared with slow fibrosis progression (n = 40) postliver transplant. (A) IFNα from TLR3 stimulation (PIC), p = 0.18. (B) IFNα from TLR7/8 stimulation (R848), p = 0.039. (C) IFNα from TLR9 stimulation (CpG), p = 0.23. *p-value < 0.05.

Download figure to PowerPoint

HCV patients with rapid fibrosis also had impaired IL-6 production in PBMCs at baseline (p = 0.027) and in response to TLR3 stimulation with PIC (p = 0.008, Figure 2) and TLR9 stimulation with CpG (p = 0.038, Figure 3) compared to those with slow fibrosis. They also had reduced TNFα production at baseline (p = 0.060) and in response to TLR3 stimulation with PIC (p = 0.040, Figure 2). There was no significant difference in the fold increase above baseline production of these cytokines, suggesting that while there was impaired cytokine production at baseline and with TLR3 and TLR9 stimulation in HCV patients with rapid fibrosis compared with slow fibrosis, the incremental increase in cytokine with stimulation in patients with rapid fibrosis was not significantly different to that in patients with slow fibrosis.

image

Figure 2. Mean concentration (pg/mL) and fold increase over baseline of IL-6 and TNFα secretion by PBMCs in response to TLR3 stimulation in HCV patients with rapid fibrosis (n = 30) compared with slow fibrosis progression (n = 40) posttransplant. (A) Baseline unstimulated IL-6 production, p = 0.027. (B) IL-6 with TLR3 stimulation (PIC), p = 0.003. (C) Fold increase in IL-6 with TLR3 stimulation, p = 0.940. (D) Baseline unstimulated TNFα production, p = 0.06. (E) TNFα with TLR3 stimulation (PIC), p = 0.040. (F) Fold increase TNFα with TLR3 stimulation, p = 0.920. *p-value < 0.05.

Download figure to PowerPoint

image

Figure 3. Mean concentration (pg/mL) and fold increase over baseline of IL-6 secretion by PBMCs in response to TLR9 stimulation in HCV patients with rapid fibrosis (n = 30) compared with slow fibrosis progression (n = 40) posttransplant. (A) Baseline unstimulated IL-6 production, p = 0.027. (B) IL-6 with TLR9 stimulation (CpG), p = 0.038. (C) Fold increase in IL-6 with TLR9 stimulation, p = 0.810. *p-value < 0.05.

Download figure to PowerPoint

There were no other observed differences in cytokine production between HCV patients with rapid and slow fibrosis progression for any other TLR ligand.

These findings were confirmed in a subset of HCV patients evaluated at 1 year posttransplant

In order to determine whether our findings applied to the early posttransplant period and to avoid the potential confounding variable of the time posttransplant at which the blood samples were donated, we measured cytokine responses to TLR stimulation in a subgroup of 16 HCV patients early posttransplant. These patients all had PBMCs recovered and a liver biopsy to determine fibrosis stage between 6 and 12 months posttransplant. Rapid fibrosis progression in this smaller cohort was defined as the presence of the Metavir F1 stage or greater fibrosis on the year 1 liver biopsy, as the fibrosis rate equals the fibrosis stage measured at 1 year (fibrosis stage being divided by 1 year). Using this definition, nine HCV patients had slow fibrosis and seven patients had rapid fibrosis.

In this cohort, HCV patients with early rapid fibrosis also had impaired IFNα production in response to TLR7/8 stimulation (p = 0.010) and impaired a fold increase in IL-6 (p = 0.050) and TNFα production (p = 0.009) in response to TLR3 stimulation compared with HCV patients with slow fibrosis, as demonstrated in the larger cohort. However, there was no difference in TLR9-mediated IL-6 production between HCV patients with rapid fibrosis and slow fibrosis progression (data not shown). This suggests that early posttransplant there is impairment in cytokine responses to TLR3 and TLR7/8 stimulation in HCV patients with rapid fibrosis.

HCV patients with rapid fibrosis posttransplant have impaired TLR7/8-mediated IFNγ production by NK cells and impaired TLR3-mediated IL-6 and TNFα production by monocytes

In order to determine which cell types may be responsible for the differences in TLR3 and TLR7/8 function we had demonstrated in PBMCs, we next examined TLR3 and TLR7/8-mediated cytokine responses in individual cell types using flow cytometry.

In HCV patients with rapid fibrosis progression, NK CD56dim cells had significantly impaired IFNγ secretion in response to TLR7/8 stimulation with R848 (geometric mean fluorescence (GMF) of IFNγ p = 0.0057, fold increase in IFNγ GMF over baseline p = 0.036) compared with those with slow fibrosis progression). There was a similar trend also seen in NK CD56 bright cells in HCV patients with rapid fibrosis, which also had impaired GMF of IFNγ compared with those with slow fibrosis (p = 0.053, Figure 4). Furthermore, there was a significant correlation between fibrosis rate and NK CD56dim cell IFNγ secretion in response to TLR7/8 stimulation (p = 0.027, r = −0.4).

image

Figure 4. Geometric mean fluorescence and fold increase in fluorescence over baseline of IFNγ production with TLR7/8 stimulation (R848) in CD56dim and CD56bright NK cells, comparing HCV patients with rapid fibrosis (n = 16) and slow fibrosis progression (n = 16) posttransplant. (A) Baseline unstimulated IFNγ production by NK CD56dim cells, p = 0.214. (B) IFNγ with TLR7/8 stimulation in NK CD56dim cells, p = 0.006. (C) Fold increase IFNγ with TLR7/8 stimulation in NK CD56dim cells, p = 0.036. (D) Baseline unstimulated IFNγ production in NK CD56bright cells, p = 0.720. (E) IFNγ with TLR7/8 stimulation in NK CD56 bright cells, p = 0.053. (F) Fold increase IFNγ with TLR7/8 stimulation in NK CD56bright cells, p = 0.213. *p < 0.05.

Download figure to PowerPoint

Monocytes from HCV patients with rapid fibrosis progression also produced less IL-6 with TLR3 stimulation with PIC (GMF IL-6 p = 0.028, Figure 5), with a trend to impaired IL-6 production by monocytes at baseline also in HCV patients with rapid fibrosis compared with slow fibrosis (GMF IL-6 p = 0.060).

image

Figure 5. Geometric mean fluorescence and fold increase in fluorescence over baseline of IL-6 production with TLR3 stimulation (PIC) in monocytes, comparing HCV patients with rapid fibrosis (n = 13) and slow fibrosis progression (n = 9) posttransplant. (A) Baseline unstimulated IL-6 production in monocytes, p = 0.06. (B) IL-6 with TLR3 stimulation in monocytes, p = 0.028. (C) Fold increase in IL-6 with TLR3 stimulation in monocytes, p = 0.616. *p-value < 0.05.

Download figure to PowerPoint

NK CD56dim cell TLR7/8-mediated IFNγ secretion occurs through an IFNα-independent pathway in HCV rapid fibrosis

Type I interferon secretion by plasmacytoid dendritic cells (pDCs) is an important mechanism for NK cell activation and subsequent IFNγ secretion and cytotoxicity [18]. We therefore wished to determine whether the impaired NK CD56dim cell TLR7/8-mediated IFNγ secretion in HCV rapid fibrosis we had identified using flow cytometry (Figure 3) was the downstream effect of the impaired IFNα secretion by PBMCs in response to TLR7/8 stimulation we had shown by ELISA (Figure 1). To do this, we cocultured PBMCs from patients with HCV patients with rapid fibrosis (n = 10) with TLR7/8 ligand R848, in the presence or absence of IFNα or an anti-IFNα neutralizing antibody. We then measured GMF of IFNγ produced by NK CD56dim cells using flow cytometry.

The addition of IFNα to PBMCs from HCV patients with rapid fibrosis did not restore NK CD56 dim cell IFNγ secretion in response to TLR7/8 stimulation to levels comparable to HCV patients with slow fibrosis (p = 0.89, Figure 6). Similarly, concomitant use of an anti-IFNα blocking antibody and TLR7/8 stimulation with R848 did not alter NK CD56dim cell IFNγ secretion.

image

Figure 6. Geometric mean fluorescence and fold increase in fluorescence over baseline of IFNγ production by NK CD56dim cells in HCV patients (n = 10), comparing stimulation with TLR7-specific agonist (loxoribine) and TLR7/8 agonist (R848), with and without IFNα. (A) IFNγ at baseline, with TLR7 agonist and TLR7/8 agonist in NK CD56dim cells, p = 0.049. (B) Fold increase in IFNγ at baseline, with TLR7 agonist and TLR7/8 agonist in NK CD56 dim cells, p = 0.009. (C) Geometric mean fluorescence of IFNγ with TLR7/8 stimulation in NK CD56dim cells, with and without the addition of IFNα, p = 0.89. (D) Fold increase in fluorescence of IFNγ with TLR7/8 stimulation in NK CD56dim cells, with and without addition of IFNα, p = 0.661. *p-value < 0.05.

Download figure to PowerPoint

We then went on to determine whether NK cells express TLR7 and TLR8. We found that NK cells (including CD56 bright and dim subsets) express both TLR7 and TLR8 intracellular receptors. However, there was no difference in TLR7 or TLR8 expression on NK cells between HCV patients with rapid and slow fibrosis to explain the differences in IFNγ secretion (data not shown).

Impairment in IFNγ secretion by NK CD56dim cells in HCV rapid fibrosis is mediated through TLR8 rather than TLR7

In our experiments we had used a TLR7/8 agonist R848 as a reliable and powerful stimulant of cytokines through TLR7/8-mediated pathways. We therefore did not know whether the impairment in IFNγ secretion in NK CD56dim cells was mediated by TLR7 or TLR8. We next ascertained whether TLR7 or TLR8 were mediating the functional impairment we had demonstrated in NK cell secretion of IFNγ by comparing the effects of a TLR7-specific agonist loxoribine and TLR7/8 agonist R848 on NK CD56dim cell IFNγ production. In NK CD56dim cells from both HCV posttransplant patients and healthy controls, TLR7/8 ligand R848 produced a significantly greater fold increase in IFNγ production compared with the TLR7-specific agonist loxoribine. Furthermore, the IFNγ production ratio to baseline was clustered around 1.0 with loxoribine, suggesting that there was minimal TLR7 contribution to cytokine production in NK cells (Figure 6).

NK CD56dim cell degranulation is preserved in HCV rapid fibrosis despite impaired TLR7/8-mediated IFNγ secretion

Finally, we wished to determine whether NK CD56dim cell cytotoxicity was also affected given the impairment in NK CD56dim cell secretion of IFNγ we had demonstrated in HCV patients with rapid fibrosis. CD107a is a marker of NK cell degranulation, binding to the cell surface once degranulation has occurred. We therefore measured ex vivo NK CD56dim cell surface expression of NK degranulation marker CD107a via flow cytometry as a marker of cell cytotoxic degranulation that had occurred in-vivo, comparing HCV patients with rapid fibrosis to those with slow fibrosis.

We found that there was no difference in CD107a expression between HCV patients with rapid fibrosis progression in NK cells (p = 0.39), including NK CD56dim subsets (0.57). Interestingly, there was a trend toward increased CD107a expression in CD56bright NK cells in HCV patients with rapid fibrosis compared with slow fibrosis (p = 0.060), but this did not reach statistical significance (data not shown).

Time posttransplant was the only significant clinical factor that differed between HCV patients with rapid fibrosis and slow fibrosis

As there are well-recognized clinical factors which may influence both the risk of HCV rapid fibrosis, such as advanced donor age and HCV viral load, as well as factors that may theoretically influence TLR function posttransplant, such as immunosuppression level, we compared clinical characteristics of our HCV rapid and slow fibrosis cohorts (Table 1). We found that on univariate analysis, donor age (p = 0.003) and time posttransplant (p < 0.001) were the only significant differences between the two groups, which were otherwise well matched for clinical variables which may affect fibrosis outcomes. Of note, the number of patients with cirrhosis, HCV viral load and serum calcineurin inhibitor levels were similar between groups. Immunosuppressive regimen was also not statistically different between groups.

Table 1. Comparison of clinical factors in HCV patients posttransplant, comparing those with rapid fibrosis to those with slow fibrosis progression
 Slow fibrosis (n = 40)Rapid fibrosis (n = 30)p-Value
  1. ETOH = alcohol; SEM = standard error of the mean.

Univariate analysis   
Age at transplant49.1 + /− 1.151.2 + /−1.00.156
(mean +/−SEM)   
Current age55.0 + /− 1.054.0 + /− 0.90.474
(mean +/−SEM)   
Donor age33.9 + /− 2.545.1 + /− 2.70.003
(mean +/−SEM)   
Gender (M:F)37:530:20.433
Race (Caucasian)38230.062
ETOH as a cofactor at time of transplant1291.00
HCV viral load at time of experiments (IU/mL) (mean+/−SEM)1.79 × 1061.00 × 1060.399
Cold ischemia time (min, mean +/−SEM)501.6 + /− 22.7441.7 + /− 24.70.078
Cyclosporine level at time of experiments (mean +/−SEM)404.6 + /− 42.7449.4 + /− 55.00.520
Tacrolimus level at time of experiments6.9+/−1.07.0+/−4.70.850
(mean +/−SEM)   
F3 or F4 Metavir stage fibrosis within 5 years posttransplant220<0.0001
Presence of cirrhosis4110.085
Time posttransplant (years)8.0+/−0.74.0+/−0.5<0.0001
Multivariate analysis   
Time posttransplant  0.002

When we performed a multi-variate analysis, time posttransplant was the only significant difference between the two groups (p = 0.002, Table 1), with donor age no longer being significant.

Discussion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

Despite rapid developments in our understanding of TLR function and its role in HCV infection, little is known of the role of TLR function in HCV recurrence posttransplantation. TLRs are the premier sentinel immune sensors in viral infection and have been linked with fibrosis development [6]; thus their role in HCV reinfection posttransplantation is theoretically of considerable importance to study. This work represents the first systematic exploration of TLR function in peripheral immune cells in hepatitis C recurrence postliver transplantation.

In patients with rapid fibrosis due to recurrent HCV posttransplant, we found evidence of impaired IFNα secretion by PBMCs in response to TLR7/8 stimulation. We also identified impaired secretion by PBMCs of TNFα and IL-6 at baseline and with TLR3 stimulation in those with rapid fibrosis progression. These findings were confirmed in a subgroup with liver biopsy and TLR function studies performed early posttransplant. For this reason, we chose to focus on TLR3 and TLR7/8-mediated cytokine production by cell subtypes using flow cytometry in order to determine which cell subsets may be responsible for these TLR functional changes.

Further studies demonstrated the impairment in TLR3-mediated IL-6 and TNFα production by PBMCs associated with rapid fibrosis was likely due to monocytes producing reduced amounts of proinflammatory cytokines in response to TLR3 stimulation. Monocytes have been shown to express TLR3 functional receptors [13] and TLR3 is a potent inducer of primary and secondary antiviral pathways [14]. Monocyte-derived IL-6 and TNFα activate mDCs and pDCs, as well as recruiting inflammatory cells to the liver in HCV infection [19]. Impaired secretion of TLR3-mediated IL-6 by monocytes has been described in HCV-HIV coinfection, a state where accelerated liver fibrosis is also seen [20]. Furthermore, recent evidence suggests that IL-6 can have antifibrotic effects on hepatocytes through promoting hepatic regeneration without fibrosis [21, 22]. It is therefore conceivable that impairment of TLR3-mediated IL-6 and TNFα production by monocytes contributes to inappropriate activation of NK cells and dendritic cells and increased fibrosis.

TLR7/8-mediated IFNα secretion by PBMCs was reduced in HCV patients with rapid fibrosis progression. Almost all type I interferon is secreted by plasmacytoid dendritic cells [23] and these express functional TLR7 receptors [24]. IFNα secreted by pDCs is critical for effective T cell activation and may contribute to poor T cell responses to HCV antigens, leading to reduced viral control and clearance in patients with rapid fibrosis [25]. Our findings correspond well with the known importance of TLR7-mediated IFNα in the immune response to HCV, with substantial evidence for the efficacy of IFNα therapy for clearance of HCV infection and more specifically the efficacy of TLR7 agonist isotoribine in suppressing HCV viremia and achievement of sustained virological response [26]. Further evidence for the importance of TLR7/8-mediated IFNα in HCV-mediated rapid fibrosis progression is that TLR7 polymorphisms have been linked to liver fibrosis progression [27].

Furthermore, we demonstrated impaired TLR7/8-mediated IFNγ secretion by NK CD56dim cells. NK CD56dim cells produce large amounts of antiviral IFNγ and initiate perforin-mediated cytotoxicity of viral-infected cells, but produce less proinflammatory cytokines compared with CD56bright NK cells [28]. As failure of pDCs to secrete IFNα has been shown to reduce NK cell activation [18], we wished to determine whether this was a dependent effect. The addition of exogenous IFNα did not correct the deficit in IFNγ secretion by NK CD56dim cells, demonstrating that impairment in NKCD56 dim cell IFNγ secretion is independent of impaired TLR7/8-mediated IFNα secretion by PBMCs in HCV patients with rapid fibrosis progression.

We then wished to determine what other factors may be responsible for the impairment in TLR7/8-mediated NK CD56dim cell function. We determined that NK cells (including CD56dim) express TLR7 and TLR8 receptors and that NK CD56dim cell TLR7 and TLR8 expression was not different between HCV patients with rapid and slow fibrosis. In addition, we found that the TLR7/8 agonist R848 was more a more potent inducer of NK CD56dim cell IFNγ secretion than TLR7-specific agonist loxoribine, suggesting that NK CD56dim cell secretion of IFNγ is related to TLR8 stimulation, rather than TLR7 stimulation. We also demonstrated that CD107a expression was not different between HCV patients with rapid versus slow fibrosis. Taken together, these results suggest that NK cells are capable of being directly stimulated by TLR7/8 ligands and can produce IFNγ in an IFNα/plasmacytoid dendritic cell-independent manner. They also suggest that IFNγ secretion occurs separately to cytotoxicity. These findings are supported by evidence from other groups showing that NK cells express functional TLR7 and TLR8 receptors [14], NK cell cytotoxicity is more dependent upon pDC derived IFNα than IFNγ secretion [29] and that IFNγ secretion by NK cells can occur independently of IFNα [14]. It has also been shown that CD107a expression is preserved in HCV infection despite reduction in IFNγ [30]. Both NK and NKT cells have been shown to inhibit HCV viremia via both cytolytic and noncytolytic mechanisms [31] and this may explain why there is an effect of impaired NK cell-derived IFNγ without impaired cytotoxicity in patients with HCV-mediated rapid fibrosis posttransplant.

There are several potential mechanisms to explain how our findings may contribute to rapid fibrosis progression in HCV recurrence posttransplant. NK CD56dim cell IFNγ secretion activates dendritic cells and T cells and promotes anti-viral Th1 cytokine responses. Impairments in NK cell IFNγ secretion promote Th2 compared with Th1 cytokine responses [32], leading to impaired activation of T cell immune responses against HCV infection, particularly impairing the protective effect of CD4+ T cells against liver disease progression [25, 33]. Poor immune control of HCV viremia is well described in HCV rapid fibrosis progression [34] and leads to greater proinflammatory cytokine production and greater activation of profibrogenic pathways [35, 36]. NK CD56dim cells also have important antifibrotic effects via inhibition of hepatic stellate cells, either directly through induction of stellate cell apoptosis or indirectly through production of IFNγ, which inhibits hepatic stellate cell activation [15]. Impaired NK cell IFNγ secretion as we have demonstrated leads to uncontrolled activation of hepatic stellate cells and increased fibrogenesis. In addition, the impairment in IFNγ secretion by NK CD56dim cells and subsequent promotion of Th2 cytokine responses in HCV infection leads to greater TGF-β secretion and IL-10 secretion [37], which both promote hepatic stellate cell survival through downregulation of CD95Ligand [38] and upregulation of liver fibrogenesis.

An important consideration for our data was whether the changes in TLR-mediated immune responses are merely a result of HCV viral load or immunosuppression levels, two key clinical determinants of HCV rapid fibrosis progression [34, 39]. It is well established that patients with rapid HCV fibrosis posttransplant have higher levels of HCV viremia [40]. HCV has been shown to impair both TLR3[10, 11, 41, 42] and TLR7/8 [43] signaling through various mechanisms, which may explain our findings in HCV rapid fibrosers. We demonstrated that there was no significant difference in either serum calcineurin inhibitor level or HCV viral load at the time of PBMC recovery between the HCV rapid and slow fibrosis cohorts (Table 1). Furthermore, we found no correlation between either HCV viral load or serum calcineurin inhibitor level and the TLR changes we identified (data not shown). This suggests that the TLR functional impairment we describe in HCV patients with rapid fibrosis development is independent of both HCV viral load and calcineurin inhibitor level.

Time posttransplant (in years) was the only clinical factor that significantly differed between the rapid fibrosis and slow fibrosis groups on multivariate analysis (Table 1). This likely reflects increasing rates of HCV progression postliver transplant in recent years, in part due to increased use of marginal grafts and greater immunosuppression [39, 44]. We endeavored to reduce the effect of time posttransplant in our study by analyzing a subset of patients who had a liver biopsy and TLR functional studies performed 6–12 months posttransplant and by ensuring time posttransplant was matched between rapid and slow fibrosis groups in the patients in whom flow cytometric analysis was performed (data not shown).

One difficulty with studies evaluating risk associations with rapid fibrosis progression in HCV recurrence postliver transplant is the choice of definition of rapid fibrosis progression, for which there is no current gold standard. In this study, we have used fibrosis rate and fibrosis progression on the first-year liver biopsy posttransplant as these have both been widely used in clinical studies previously [16, 17, 45-52]. Ideally, additional analyses using F2 stage or greater fibrosis on the year 1 liver biopsy and the development of cirrhosis within 5 years posttransplant would also have been performed, however in our study the numbers of patients developing F2 stage fibrosis on the year 1 biopsy or cirrhosis within 5 years posttransplant were small, meaning any analysis of the relationship between liver fibrosis and TLR function would be underpowered to detect a difference.

A key limitation to our study is that it is cross-sectional and retrospective and therefore precludes our ability to definitively establish cause and effect between HCV fibrosis and TLR function. Similarly, it provides no information about what happens to TLR function over time in HCV reinfection. Further prospective studies measuring TLR responses in PBMCs and fibrosis progression on liver biopsy at multiple time points posttransplant would help to answer these questions. However, our study still describes important novel relationships between TLR function and HCV fibrosis progression posttransplant that warrant further investigation.

Conclusion

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

We have demonstrated key differences in TLR function between HCV patients with rapid fibrosis progression compared with slow fibrosis progression posttransplant. TLR3-mediated IL-6 production by monocytes is impaired in HCV rapid fibrosers compared with slow fibrosers. We also demonstrated impaired TLR7/8-mediated IFNα secretion by PBMCs and impaired IFNα-independent, TLR8-mediated NK cell IFNγ secretion without impairment of NK cell cytotoxicity in HCV patients with rapid fibrosis. These novel findings of TLR3 and TLR7/8 functional impairment in patients with aggressive HCV recurrence may contribute to the observed poor immune control of HCV infection posttransplant and subsequent activation of profibrogenic pathways.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

We would like to thank Professor Peter Crowley for kindly staging the degree of fibrosis in all liver biopsies.

Grant Support: Funding was provided for the study by nonspecific Innate Immune Laboratory educational funds, supported by Monash University, Melbourne Australia.

Dr. Jessica Howell received scholarship funds for stipend from the Gastroenterological Society of Australia (GESA).

Author Contributions:

  • Dr Jessica Howell: Study design, laboratory experiments (performance and design), acquisition of data, analysis and interpretation of data, statistical analysis, manuscript writing.
  • Dr Rohit Sawhney: Laboratory experiments, acquisition of data, manuscript drafting.
  • Ms Narelle Skinner: Laboratory experiment expertise and assistance with experiment design.
  • A/ Prof Paul Gow: Study supervision, study design, manuscript drafting.
  • Prof Peter Angus: Study supervision, study design, manuscript drafting.
  • Dr Dilip Ratnam: Assistance with experiment design and expertise.
  • A/ Prof Kumar.
  • Visvanathan: Study supervision, study design, laboratory expertise and experiment design, manuscript drafting.

Disclosure

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References

The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.

References

  1. Top of page
  2. Abstract
  3. Methods
  4. Results
  5. Discussion
  6. Conclusion
  7. Acknowledgments
  8. Disclosure
  9. References