Jessica Howell contributed to the study design, laboratory experiments (performance and design), acquisition of data, analysis and interpretation of data, statistical analysis, and manuscript writing. Rohit Sawhney contributed to the laboratory experiments, acquisition of data, and manuscript drafting. Adam Testro contributed to the laboratory protocol and provided assistance with the experimental design. Narelle Skinner provided laboratory expertise and assistance with the experimental design. Paul Gow contributed to the study supervision, study design, and manuscript drafting. Peter Angus contributed to the study supervision, study design, and manuscript drafting. Dilip Ratnam provided assistance with the experimental design. Kumar Visvanathan contributed to the study supervision, study design, and manuscript drafting and provided laboratory expertise and assistance with the experimental design.
None of the authors have any disclosures to make relevant to this study or article.
Funding for this study was provided by nonspecific Innate Immune Laboratory educational funds, which are supported by Monash University (Melbourne Australia). Jessica Howell received scholarship funds for a stipend from the Gastroenterological Society of Australia.
The innate immune system is now recognized to play a key role in various aspects of transplantation, including ischemia/reperfusion injury, tolerance, graft rejection, and infection.[1-4] Toll-like receptors (TLRs) form the cornerstone of the innate immune system and represent the first line of defense against infection and inflammation. They can be stimulated by both pathogen-associated molecular patterns expressed by invading pathogens and damage-associated molecular patterns released by cells damaged during the transplantation process.[3, 4] When TLRs bind with their corresponding ligands, they initiate a downstream signaling cascade, which culminates in the production of type I interferon (IFN) and proinflammatory cytokines that are crucial to transplantation biology. TLRs also have a proven role in many liver disease processes that can recur after transplantation, such as hepatitis C infection.
Traditionally, it has been assumed that unlike adaptive immunity, the innate immune system is unaffected by immunosuppression. However, emerging evidence suggests that immunosuppressive agents may have important effects on innate immunity. To date, the effect of calcineurin inhibitors on TLR function after liver transplantation is unknown.
Given the importance of TLR function to clinical outcomes such as infection, graft rejection, and disease recurrence after liver transplantation, we wished to determine whether TLR function is affected by calcineurin inhibitors, the most commonly used immunosuppressant agents after liver transplantation. In this study, we compare TLR function in patients on cyclosporine A (CYA), patients on tacrolimus (TAC), and healthy controls. We also determine the relationships between TLR function and serum calcineurin inhibitor levels after liver transplantation. We demonstrate novel effects of calcineurin inhibitors on TLR function that have significant implications for management after liver transplantation.
PATIENTS AND METHODS
This cross-sectional study was conducted at a single adult transplant center. Patients were recruited between June 6, 2009 and July 31, 2011. Follow-up continued until death, retransplantation, or July 31, 2011. To be included in the study, patients had to be older than 18 years, be at least 6 months after liver transplantation, and have a minimum follow-up period of 6 months. To prevent potential confounding effects on TLR function measurements, blood samples were taken when patients were clinically stable and were not suffering from intercurrent autoimmune hepatitis, bacterial infections, 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.
One hundred thirteen post–liver transplant patients and 13 healthy controls (anonymous blood donors) were recruited for this study. For a minimum of 3 months at the time of the study, 60 patients had been on a stable dose of CYA, and 53 patients had been on a stable dose of TAC.
Data Collection and Analysis
Patient data were prospectively recorded in the Victorian Liver Transplant Unit database. The recorded data included donor and recipient demographics and clinical variables, such as medication and serum immunosuppression levels. TAC levels in serum were measured before dosing, and CYA levels were measured 2 hours after dosing with the Roche Covasc B 502 instrument.
Isolation of Peripheral Blood Mononuclear Cells (PBMCs)
Whole blood (35 mL) was obtained by venipuncture and collected in heparinized tubes. PBMCs were extracted with the Ficoll-Paque density centrifugation method (Ficoll-Paque PLUS solution, GE Healthcare, United Kingdom) and stored in complete Roswell Park Memorial Institute 1640 medium (Sigma Lifesciences, United States) with 1% penicillin, 1% l-glutamine, and 5% fetal calf serum with dimethyl sulfoxide in liquid nitrogen.
Measurements of Interleukin-6 (IL-6), Tumor Necrosis Factor α (TNFα), and IFNα by Enzyme-Linked Immunosorbent Assay (ELISA)
PBMCs from subjects were rapidly thawed and then cultured for 24 hours at 37°C in complete Roswell Park Memorial Institute medium with 1% penicillin, 1% l-glutamine, and 5% fetal calf serum with TLR ligands at a concentration of 1 × 106 cells/mL. The TLR subclass–specific ligands used for cell stimulation were Pan-3-cys (P3C; TLR2) (TLR2), Poly-IC (PIC, TLR3; InvivoGen, San Diego, CA), lipopolysaccharide (LPS; TLR4), R848 (TLR7/8; InvivoGen), and CpG (TLR9; CpG 2006 for monocytes and CpG 2216 for plasmacytoid dendritic cells; GeneWorks, United States). IL-6, TNFα, and IFNα, produced by cell stimulation, were then quantified in cell culture supernatants with ELISA according to the manufacturers' instructions (BD Biosciences, San Jose, CA, and R&D Systems, Minneapolis, MN). PBMCs cultured in media alone were used as controls.
Flow Cytometry Analysis of Cell Surface Markers and Intracellular Cytokines
Fifty-nine posttransplant patients were selected from the larger cohort for flow cytometry. PBMCs were rapidly thawed and cultured for 6 hours with TLR7/8 ligand R848, TLR7 ligand loxoribine, and TLR3 ligand PIC at a cell concentration of 1 × 106 cells/mL (InvivoGen). The short incubation for the PBMCs allowed us to measure the early innate immune contribution to cytokine production. The Golgi transport inhibitor brefeldin A (GolgiPlug, Becton Dickinson, San Jose, CA) was added for the final 4 hours of culturing at a concentration of 1 μg/mL. The cells were then stained with fluorochrome-conjugated antibodies for cell surface markers CD3–Pacific Blue (clone UCHT1, BD Biosciences), CD56–phycoerythrin (PE)–cyanine 7 (Cy7; clone B159, BD Biosciences), CD16-allophycocyanin (APC; clone 3G8, BD Biosciences), CD14-APC-Cy7 (clone MΦPG, BD Biosciences), and CD69-PE (BD Biosciences). After this, the cells were permeabilized (Cytofix/Cytoperm solution, BD Biosciences) and stained for intracellular cytokines IL-6–PE, TNFα-APC, and IFNγ–fluorescein isothiocyanate (all from BD Biosciences). Separate cell aliquots (1 × 106 cells/mL) not stimulated with TLR ligands were also stained for the surface expression of CD14-APC-Cy7, CD3–Pacific Blue, CD56-PE-Cy7, and CD107a-PE (BD Biosciences) to assess natural killer (NK) cell degranulation, and then they were permeabilized and stained with TLR7-CSF (R&D Systems) and TLR8-PE fluorochrome antibodies (clone 44C143, Imgenex, United States) to measure intracellular TLR7 and TLR8 expression. Cells cultured with media alone were used as controls. Cells were run on a FACSCalibur flow cytometer (BD Biosciences), and the results were analyzed with FlowJo 9.2 (TreeStar, Ashland, OR).
Assessment of the Effect of Calcineurin Inhibitors on TLR Function in Healthy Controls
In order to determine whether calcineurin inhibitors affected TLR function in PBMCs from healthy controls, PBMCs from 5 healthy controls were cultured with media, TAC (at a therapeutic concentration of 10 ng/mL), or CYA (at a therapeutic concentration of 1000 ng/mL) for 2 hours. At 2 hours, a TLR stimulant (LPS, P3C, or R848) or media (controls) were added, and protocols for cell stimulation and subsequent intracellular flow cytometry or cell supernatant ELISA were followed as outlined previously.
Statistical analysis was performed with [SPSS] 19 (SPSS, Inc., Chicago, IL) and Prism 5.0c for Macintosh (GraphPad Software, Inc., La Jolla, CA). Clinical parameters were compared between groups with the Student t test for continuous variables and with the chi-square test for categorical variables. Nonparametric statistical tests were used for comparing group data: the Mann-Whitney test for 2 groups and the Kruskall-Wallis test with Dunn's post test for multiple groups. Correlations between calcineurin inhibitor levels and TLR function were determined via Spearman correlation. A paired t test was used to compare the effects of TLR ligand stimulation in the presence of TAC and CYA on cytokine production with respect to controls. A 2-sided P value of 0.05 was considered to be statistically significant. A multivariate analysis was performed with a binary logistic regression and backward stepwise elimination procedure, and the model was built with variables with significance P < 0.1.
Comparison of TLR Function Between Patients on Calcineurin Inhibitors and Healthy Controls
First, we wished to determine whether peripheral blood immune cell TLR function was impaired in patients on calcineurin inhibitors versus healthy controls. We, therefore, measured cytokine production by PBMCs in response to TLR ligands and compared patients taking either CYA or TAC to healthy controls.
PBMCs From Patients on Calcineurin Inhibitors Had Impaired IL-6 and TNFα Production With TLR2, TLR4, and TLR7/8 Stimulation in Comparison With Healthy Controls
We found that PBMCs from patients on CYA and TAC had impaired fold increases in IL-6 production above the baseline with TLR2 stimulation (P = 0.02; Fig. 1A), TLR4 stimulation (P = 0.02; Fig. 1B), and TLR7/8 stimulation (P = 0.02; Fig. 1C) in comparison with healthy controls. We also found that TNFα production in response to TLR2 stimulation (P = 0.01; Fig. 1D), TLR4 stimulation (P = 0.01; Fig. 1E), and TLR7/8 stimulation (P = 0.02; Fig. 1F) was impaired in patients on CYA and TAC versus healthy controls. In comparison with patients on CYA, patients on TAC appeared to have greater impairment of cytokine production with TLR stimulation; however, there was no statistically significant difference in cytokine production between patients on TAC and patients on CYA.
There was no difference in IFNα production by PBMCs with TLR stimulation between patients on calcineurin inhibitors and healthy controls.
NK Cells From Patients on Calcineurin Inhibitors Had Impaired TLR7/8-Mediated IFNγ Secretion in Comparison With Healthy Controls
Using flow cytometry, we then looked specifically at cytokine production by individual cell subtypes with TLR stimulation.
We found that both CD56bright and CD56dim NK cell subsets from patients on calcineurin inhibitors produced less IFNγ with TLR7/8 stimulation than those from healthy controls (CD56bright NK cells, P = 0.002 for the geometric mean fluorescence, P = 0.06 for the fold increase above the baseline; CD56dim NK cells, P = 0.004 for the geometric mean fluorescence, P = 0.002 for the fold increase above the baseline; Fig. 2).
However, we did not find a specific cell subtype to be responsible for the differences in IL-6 and TNFα production by PBMCs in response to TLR2, TLR4, and TLR7/8 stimulation as demonstrated by ELISA.
Correlations Between TLR Function and Calcineurin Inhibitor Levels
Having demonstrated impairment in TLR function in patients on calcineurin inhibitors versus healthy controls, we next examined whether there was a direct relationship between serum calcineurin inhibitor levels and cytokine production in response to TLR stimulation.
TLR4-Induced IL-6 Production by PBMCs Was Inversely Proportional to Serum TAC Levels
We then examined cytokine production in response to TLR stimulation in PBMCs and the relationship to calcineurin levels. We found that PBMCs produced less TLR4-induced IL-6 with increasing TAC levels (P = 0.03, r = −0.30, 95% CI = −0.528 to −0.022; data not shown). However, there was no relationship between CYA and PBMC cytokine production.
CD56dim NK Cells Produced More TNFα in Response to TLR7/8 Stimulation With Increasing TAC Levels
In contrast to our other findings, TLR7/8-induced TNFα production by NK cells and CD56bright NK cells positively correlated with TAC levels (P = 0.01 for CD56bright NK cells, r = 0.66, 95% CI = 0.162-0.894; data not shown). There was no correlation between CYA levels and TNFα production.
Exposure of Healthy PBMCs to Calcineurin Inhibitors Impaired TLR Function
In order to confirm that calcineurin inhibitors alter TLR function in PBMCs and to ensure that our previous findings were not due to some other factor related to the post–liver transplant state, we incubated PBMCs from healthy controls with therapeutic doses of TAC, CYA, or media (controls). We then compared cytokine production in response to TLR2, TLR4, and TLR7/8 stimulation in the presence and absence of calcineurin inhibitors with both flow cytometry and ELISA. The ratio of cytokine production with a TLR ligand in the presence of TAC or CYA versus the TLR ligand alone was calculated.
We found that the ratio of TLR4-mediated TNFα production by PBMCs above the baseline was lower in the presence of TAC and CYA versus the TLR4 ligand alone (P = 0.009 for TAC, P = 0.03 for CYA; Fig. 3E). We also found that the ratio of TLR4-mediated IL-6 production by PBMCs above the baseline was reduced in the presence of TAC and CYA versus controls (P = 0.02 for TAC, P = 0.03 for CYA; Fig. 3B). The ratio of TLR7/8-mediated TNFα production by PBMCs versus controls was lower in the presence of CYA versus TAC (P = 0.12 for TAC, P = 0.01 for CYA; Fig. 3F). Similarly, the ratio of TLR7/8-mediated IL-6 production by PBMCs was lower in the presence of CYA but not TAC (P = 0.02 for CYA, P = 0.79 for TAC; Fig. 3C). There was no significant difference in TLR2-mediated TNFα production in the presence or absence of calcineurin inhibitors (P = 0.80 for TAC, P = 0.29 for CYA; Fig. 3D). However, there was significantly less TLR2-mediated IL-6 produced by PBMCs in the presence of CYA versus TAC (P = 0.005 for CYA, P = 0.65 for TAC; Fig. 3A).
Collectively, these data support our finding of impaired TLR2-, TLR4-, and TLR7/8-mediated cytokine production by PBMCs in the presence of calcineurin inhibitors after liver transplantation, with CYA being observed to have the strongest effect.
Using flow cytometry, we then assessed the effect of therapeutic doses of TAC and CYA on cytokine production in individual cell subtypes from healthy subjects (Fig. 4). We found that IFNγ production by CD56bright and CD56dim NK cells was reduced in the presence of calcineurin inhibitors. CD56dim NK cells produced less TLR7/8-mediated IFNγ with CYA versus TAC or controls (P = 0.02 for CYA, P = 0.56 for TAC), whereas CD56bright NK cells produced less TLR7/8-mediated IFNγ with TAC versus CYA or controls (P = 0.03 for TAC, P = 0.17 for CYA).
These data also concur with our previously reported finding of impaired TLR7/8-mediated IFNγ production by CD56bright and CD56dim NK cells in the presence of calcineurin inhibitors.
Difference in Cytokine Responses to TLR Stimulation in the Presence of Calcineurin Inhibitors Was Independent of Drug Toxicity and Cell Frequency
In previous experiments, we determined that ex vivo PBMC toxicity from therapeutic levels of TAC and CYA is minimal (A. Testro, unpublished data). Furthermore, in this experiment, we compared the cell frequencies in the TAC, CYA, and control groups and found no significant differences to account for the differences in cytokine production demonstrated (P = 0.87 for monocytes, P = 0.83 for CD56bright NK cells, P = 0.99 for CD56dim NK cells; data not shown).
There Was No Significant Clinical Difference Between Patients on CYA and TAC to Account for Differences in TLR Function
In order to ensure that the differences in TLR function that we had demonstrated between patients on CYA and patients on TAC were not due to confounding clinical factors, we compared the 2 cohorts for clinical characteristics that might affect TLR function (Table 1). There were no significant differences between the 2 groups that could account for the demonstrated differences in TLR function. There was a higher number of patients with hepatitis C infections in the CYA cohort versus the TAC cohort; however, this did not reach statistical significance (P = 0.085).
Table 1. Clinical Characteristics of Post–Liver Transplant Patients: CYA Therapy Versus TAC Therapy
CYA (n = 60)
TAC (n = 53)
The data are presented as means and standard errors of the mean.
Other immunosuppressants with a calcineurin inhibitor (n)
TLRs have an important role in many aspects of transplantation biology, including infection, graft rejection, and tolerance.[1-4] However, in contrast to the known effects of immunosuppression on adaptive immunity, the influence of immunosuppressive agents on human TLR function has not been fully determined. Our data demonstrate that TLR function is lower in patients on calcineurin inhibitors after liver transplantation versus healthy controls, and CYA and TAC appear to have divergent effects on TLR function.
We first demonstrated that there was impairment in proinflammatory cytokine production by PBMCs with TLR2, TLR4, and TLR7/8 stimulation in patients on TAC and CYA versus healthy controls, and this suggested a class effect for calcineurin inhibitors on TLR2, TLR4, and TLR7/8 function. However, when individual cell subtypes were examined with flow cytometry, we did not find any specific cell subtype responsible for these alterations in TLR function. This suggests that although the overall effect of calcineurin inhibitors appears to be inhibitory on TLR2, TLR4, and TLR7/8 signaling in PBMCs, the effect of calcineurin inhibition in individual cell subtypes is variable.
Kang et al. initially described calcineurin inhibition of TLR signaling through interactions with TLR2, TLR4, myeloid differentiation protein 88, and TIR domain containing adaptor protein-inducing IFNβ (TRIF). They found that mouse peritoneal macrophages cocultured with calcineurin produced greater amounts of TNFα. The same group also reported down-regulation of TLR4 signaling with LPS exposure in the presence of calcineurin inhibitors, with pretreatment of cells with TAC reducing subsequent inflammatory responses to LPS; this suggested induction of LPS tolerance. Collectively, these data suggest that calcineurin inhibitors may have both activating effects (through the inhibition of calcineurin) and inhibitory effects (through the induction of tolerance) on TLR signaling. However, there have been concerns raised about these studies because very high doses of calcineurin inhibitors were used, and this may have resulted in cell toxicity with subsequent inflammatory cytokine release. A further study performed with human PBMCs and neonatal cord blood cells found minimal effects of calcineurin inhibitors on adult monocyte TLR function. However, that study did not look at PBMCs collectively, and the doses of calcineurin inhibitors were different from what are used in clinical practice in liver transplantation.
We also demonstrated impaired IFNγ secretion by both CD56bright and CD56dim NK cell subsets in response to TLR7/8 stimulation. This is a very interesting finding because IFNγ production by NK cells is of great importance to viral infections, including hepatitis C infections.[12-14] This suggests that there may be potential ramifications of calcineurin inhibitor use for hepatitis C virus recurrence after transplantation, which is known to follow a more aggressive course in comparison with the nontransplant setting.[15, 16] We have previously reported that TLR7/8-mediated CD56dim NK cell IFNγ production is lower in patients with hepatitis C virus who develop rapid fibrosis after liver transplantation and is associated with poor graft outcomes. A direct effect of calcineurin inhibitors on CD56dim NK cell function may, therefore, be crucial to the development of rapid hepatitis C virus recurrence after liver transplantation, and this warrants further investigation. To date, there have not been studies looking at the relationship between TLR7/8 function and immunosuppression.
We then explored whether there were direct relationships between serum calcineurin inhibitor levels and TLR function. Only TLR4-mediated IL-6 production by PBMCs was inversely proportional to increasing serum TAC levels but not CYA levels. Interestingly, there were no other direct relationships between serum calcineurin inhibitor levels and TLR ligand–mediated cytokine production. This suggests but does not prove that the relationship between TLR signaling and calcineurin inhibition (other than TLR4-mediated IL-6 production and TAC) is not direct. An implication of this is that serum calcineurin inhibitor levels may not provide a direct measure of TLR signaling in PBMCs. However, this hypothesis cannot be confirmed by our present study and warrants further investigation.
In contrast, TNFα production by CD56dim NK cells with TLR7/8 stimulation increased with increasing serum TAC levels. This occurred despite the impairment of TLR7/8-mediated TNFα production by PBMCs in patients on calcineurin inhibitors versus healthy controls and likely reflected different effects of calcineurin inhibition in different cell types. TLR7/8 stimulation of CD56dim NK cells results in increased cytotoxicity and proinflammatory cytokine production. Although a role for TLR7/8-mediated cytokine production in posttransplant events such as rejection has yet to be demonstrated, increased cytotoxic responses by peripheral CD56dim NK cells upon recruitment to areas of inflammation in the liver may contribute to the progression of inflammatory hepatic damage in a number of conditions that recur after liver transplantation, including hepatitis C recurrence.[18-20]
An important question is whether the changes in TLR signaling that we have demonstrated in patients after liver transplantation are due to immunosuppression or other factors related to transplantation. To explore this, we cultured PBMCs from healthy controls ex vivo with therapeutic concentrations of both TAC and CYA and then assessed cytokine responses to TLR stimulation. Essentially, we found that healthy PBMCs cultured with TLR ligands in the presence of calcineurin inhibitors also exhibited impaired TLR2-, TLR4-, and TLR7/8-mediated TNFα and IL-6 production in comparison with controls, and this was similar to our findings in posttransplant patients on calcineurin inhibitors. Interestingly, the effect of CYA on cytokine production by PBMCs with TLR2, TLR3, and TLR4 stimulation was greater than the effect of TAC, which significantly reduced only TLR4-mediated cytokine production in comparison with controls. We also demonstrated specific impairment in TLR7/8-mediated IFNγ production by NK cells with calcineurin inhibitors. These data concur with our previous findings in posttransplant patients and suggest that calcineurin inhibitors are the cause of impaired TLR signaling after liver transplantation rather than another posttransplant factor. Furthermore, there was no difference in cell frequency with exposure to calcineurin inhibitors versus controls to explain the reduced cytokine production with TLR stimulation.
Importantly, by using multivariate analysis, we determined that there were no clinical variables that could have contributed to the differences in TLR function between those on TAC and those on CYA. Although some patients were also on other immunosuppressive medications in addition to a calcineurin inhibitor, the numbers were small and were not significantly different between groups.
There are several important limitations to this study. First, the study is cross-sectional, and the numbers are small. Further prospective studies with PBMCs harvested at multiple time points would allow longitudinal confirmation of our findings. Our study also does not provide insight into the mechanism by which calcineurin inhibition reduces TLR signaling in PBMCs, and further studies in animal models or cell culture models would be useful to delineate this with greater control of extraneous variables. A further consideration is the measurement of the effect of calcineurin inhibition on adaptive immune function as a marker for the degree of immunosuppression associated with serum drug levels in study subjects. In this study, we exposed PBMCs from healthy subjects to identical concentrations of calcineurin inhibitors that were equivalent to therapeutic levels used in the Victorian Liver Transplant Unit (Australia). In this way, we sought to remove the potential variability in immunosuppression present in a study cohort of posttransplant patients on variable doses of immunosuppression. However, the direct measurement of T cell activation in the presence of immunosuppression would be an additional way of quantifying the effects of immunosuppression levels on immune function and would be particularly useful for comparing the effects of CYA to the effects of TAC.
These findings collectively demonstrate that calcineurin inhibitors do affect innate immune function and that these agents may have different effects. To date, most studies of calcineurin inhibitor effects on TLR function have used CYA, with TAC and CYA not being directly compared. The difference in the magnitude of the effect of calcineurin inhibitors on innate immune function that we have demonstrated is in keeping with data suggesting that TAC and CYA are associated with qualitatively different outcomes in various diseases that recur after liver transplantation, including hepatitis C virus recurrence.[21-27] Our data shed light on potential reasons that differences in clinical outcomes have been demonstrated.
In conclusion, both CYA and TAC impair TLR function after liver transplantation in comparison with healthy controls, and there are direct correlations between serum levels and TLR function in peripheral immune cells. The impairment of TLR function by these agents has important implications for preventing clinical outcomes such as rejection, infection, and disease recurrence after liver transplantation. Importantly, CYA and TAC appear to have different effects on TLR function. These differences in the inhibition of TLR function need further exploration as potential therapeutic strategies for improving clinical outcomes after liver transplantation.