Objectives: The liver's capability to completely regenerate after injury is a unique phenomenon in which cytokines are of particular interest. Here, we aimed to assess the release patterns and prognostic relevance of liver regeneration-related cytokines in the setting of living-donor liver transplant.
Materials and Methods: Eleven cytokines related to liver regeneration (hepatocyte growth factor, interleukin 6, insulin-like growth factor-1, tumor necrosis factor alpha, transforming growth factor beta, granulocyte colony-stimulating factor, stem cell factor, chemokine (C-X-C motif) ligand 12, angiogenin, fibroblast growth factor-2, and vascular endothelial growth factor) were compared in 13 living-donor liver transplant recipients and their corresponding donors before and daily (10 days) after transplant. Patients and donors were stratified by clinical outcomes (early graft loss within 4 weeks after transplant vs beneficial outcome).
Results: Most cytokines tested (especially tumor necrosis factor alpha and stem cell factor) were elevated in recipients versus donors. Many cytokines were also increased in recipients with graft loss (especially CXCL12) and in donors of recipients with beneficial outcomes (especially fibroblast growth factor 2). Fibroblast growth factor 2 levels were also correlated positively with serum gamma-glutamyltransferase, and higher preoperative concentrations in donors were associated with recipients having beneficial outcomes, indicating an improved regenerative capacity. In contrast, elevated CXCL12 levels in recipients before and after LDLT predicted graft loss and were linked to ongoing liver damage.
Conclusions: In living-donor liver transplant, there are distinct differences between donors and recipients regarding the release of liver regeneration-related cytokines. Moreover, fibroblast growth factor 2 and CXCL12 may be of diagnostic value in a complementary way to describe or even predict the possible outcomes after transplant. These results may be of clinical interest not only for living-donor liver transplant but also for acute liver failure.
Key words : Chemokine (C-X-C motif) ligand 12; Fibroblast growth factor 2, Stem cells
The liver is the only organ with a high capacity of regeneration. Indeed, acute liver failure not only is associated with mortality but also with restitutio ad integrum when the injury stops. After living-donor liver transplant (LDLT), which has become an important alternative to liver transplant with deceased donors for overcoming the increasing problem of organ shortage, efficient regeneration of the liver is essential for outcomes of donors and especially recipients.1,2 Remarkably, in the large majority of both donors and recipients, liver volume restores within 4 weeks after LDLT.2-6 However, in some patients, the liver grafts exhibit impaired regeneration, resulting in liver failure, and the identification of those livers with sufficient regenerative capacity before LDLT still remains challenging.7
The capability of the liver to completely regenerate after injury is a unique phenomenon in which cytokines are of particular interest. Within a selection of cytokines classically associated with liver regeneration are tumor necrosis factor alpha (TNF-α; an important regulator of the regeneration priming phase and a marker of cytotoxicity), interleukin 6 (IL-6; also a pivotal primer of regeneration enhanced by TNF-α), hepatocyte growth factor (HGF; a major hepatocyte mitogen activated by IL-6), and insulin-like growth factor-1 (IGF-1; another potent paracrine stimulator of hepatocyte mitogenesis).8-10 Many of these cytokines have also been shown to be associated with liver regeneration in acute liver failure.11 As a potent growth inhibitor mediating the termination of excessive liver regeneration, transforming growth factor beta (TGF-β or TGF-β1) was also included into this study.8,9
The role of stem cell biology for liver regeneration is being increasingly recognized.12 Among the stem cell types that have been investigated as promising sources of liver regeneration are CD34/CD133-expressing bone marrow-derived hematopoietic stem cells (BMHSC), which can be mobilized to the regenerating liver by various cytokines.12-14 Granulocyte colony-stimulating factor (G-CSF, or colony stimulating factor 3) is a strong BMHSC mobilizer also involved in stem cell differentiation and proliferation.15,16 Stem cell factor (SCF or c-kit ligand, another key stimulator of BMHSC mobilization, proliferation, and differentiation)17-19 and chemokine (C-X-C motif) ligand 12 (CXCL12 or stromal cell-derived factor-1, a G-CSF- and SCF-modulated BMHSC chemoattractant and mediator of inflammation that acts via its cognate receptor chemokine receptor 4 [CXCR4])18,20-23 were also assessed in the present study.
Given that angiogenic factors play key roles in liver regeneration, the 3 cytokines angiogenin, vascular endothelial growth factor (VEGF or VEGF-D), and fibroblast growth factor-2 (FGF-2) were also included in this investigation.8,24,25 It is noted that the grouping into classic regenerative, BMHSC-associated, and angiogenesis-related cytokines is a simplification for the purpose of this study and does not reflect the complex signaling networks during liver regeneration. For example, FGF-2 not only is a potent driver of angiogenesis26 but also stimulates hepatocyte mitogenesis8 and facilitates stem cell transdifferentiation toward a hepatic lineage.27,28
In this pilot study, which included 13 LDLTs, our objective was to assess the levels of the above-mentioned cytokines in donors and recipients before and after LDLT. Specifically, we aimed to identify cytokine profile differences in donors and their corresponding recipients (ie, regeneration of liver tissue originating from the same organ under different conditions regarding preservation injury, immunosuppression, and so forth) and investigate predictive markers obtained noninvasively that are associated with graft loss or beneficial outcome. For this purpose, both donors and recipients were divided into 2 subgroups (recipients with transplant loss occurring in the early postoperative period [4 weeks] and recipients with beneficial outcomes, as well as donors of recipients with and without early graft loss).
Materials and Methods
Patients and donors
In the present study, 13 consecutive living donors of a right liver lobe (segment 5-8) and their recipients were included, with all LDLT procedures performed at the Department of General, Visceral, and Transplantation Surgery of the University Hospital Essen, Germany. All patients provided written informed consent according to the ethical guidelines of the 2000 Declaration of Helsinki and the 2008 Declaration of Istanbul. All donors were healthy and eligible for LDLT in accordance with the guidelines for selection of living related liver donors at the University Hospital Essen.29 Demographics of donors and recipients and indications for LDLT are displayed in Table 1. The recipients received between 54.7% and 66.4% of the donor's liver volume (median of 60.6%; mean of 60.4%). Four recipients required retransplant due to organ failure within the first 4 weeks after transplant.
Enzyme-linked immunosorbent assays
Serum and plasma samples of donors and recipients were obtained as a baseline (postoperative day [POD] 0) before LDLT. Serum and plasma samples were obtained daily (7:30 AM ± 15 min) from POD1 to POD10. Specimens were immediately stored on ice, centrifuged, and stored at -20°C. Concentrations of HGF, IL-6, IGF-1, TNF-α, TGF-β1, G-CSF, SCF, CXCL12, angiogenin, FGF-2, and VEGF were measured with the use of commercially available enzyme-linked immunosorbent assay kits (Quantikin, R&D Systems, Minneapolis, MN, USA) according to the suppliers' protocols. The internal controls were within the sensitivity of the tests.
Data were analyzed for the time courses after LDLT and are expressed as means and standard deviation (SD) unless stated otherwise. For statistical analyses, the two-sided t test for unrelated groups was used. Correlations were calculated with Spearman rank correlation as nonparametric test. Receiver operating characteristic (ROC) analyses were performed to characterize potential predictive markers. Differences were considered as significant at levels of P < .05. Statistical analyses were performed using GraphPad Prism version 4.00 (GraphPad Software Inc., La Jolla, CA, USA) and a Web-based calculator for ROC curves (Eng J. ROC analysis, Johns Hopkins University, Baltimore, MD, USA).
General differences in cytokine profiles of donors and recipients
Before LDLT (POD0), higher average concentrations of HGF (P < .05), IL-6 (P < .05), TNF-α (not significant), and CXCL12 (P < .05) were found in recipients, whereas donors exhibited increased levels of IGF-1 and TGF-β1 compared with recipients (both P < .05; Figures 1 and 2C, left panels). All other cytokine concentrations were similar in donors and recipients on POD0.
Early after LDLT (POD1), concentrations of HGF, IGF-1, and CXCL12 converged, whereas serum levels of TNF-α and TGF-β1 remained significantly different (Figure 1A and 1C-E and Figure 2C, left panels). In donors and recipients, IL-6 concentrations peaked on POD1, with the donors showing markedly higher IL-6 increases (Figure 1B, left). Levels of G-CSF also peaked on POD1, reaching almost preoperative values on POD9; however, there were no relevant differences between donors and recipients (Figure 2A, left).
Within the subsequent postoperative course (POD2-POD10), most cytokines, that is, IL-6 (POD5-POD10 and reciprocally to POD1), IGF-1 (POD3-POD10 and reciprocally to preoperative values on POD0), CXCL12, angiogenin (POD2-POD6), and especially TNF-α and SCF were found to be significantly elevated (but not at every POD) in recipients versus donors (Figure 1B-D, Figure 2B-C, and Figure 3A, left panels). Only levels of TGF-β1 were significantly higher in donors (Figure 1E, left). Regarding concentrations of HGF, G-CSF, FGF-2 (only slightly increased in donors during POD5-POD9), and VEGF (only slightly increased in donors during POD5-POD10), no relevant differences were detected between POD2 and POD10 (Figure 1A, Figure 2A, and Figure 3B and 3C, left panels). Thus, most cytokines tested (especially TNF-α and SCF) were found to be elevated in recipients versus donors postoperatively.
Comparison of cytokine profiles regarding living-donor liver transplant
outcomes in recipients
Figures 1 to 3 (middle panels) depict cytokine profiles of recipient subgroups according to the outcomes after LDLT (9 recipients had beneficial outcome and 4 recipients had graft loss within 4 weeks after transplant). Preoperatively (POD0), only CXCL12 levels were significantly increased (P < .05) in recipients with graft loss compared with recipients with beneficial outcomes (Figure 2C, middle panel).
After transplant (POD1-POD10), average concentrations of HGF (POD3-POD6) and especially CXCL12 (although CXCL12 levels decreased after LDLT) were significantly higher (P < .05, but not at every POD) in recipients with graft loss versus recipients with beneficial outcomes (Figures 1A and 2C, middle panels). Interleukin 6 (POD5, POD9, POD10, with attenuated peak on POD1), TNF-α, SCF (except on POD4 and POD7), and G-CSF (retained peak on POD1) were also found to be (slightly) increased, although no significant differences were shown on any POD (Figure 1B and 1D, Figure 2A and 2B, middle panels). In contrast, serum concentrations of angiogenin (POD7-POD9 and reciprocally to POD1-POD5; P < .05 on POD3) and IGF-1 (POD5-POD10; not significant on every POD) were higher in recipients with beneficial outcome than in recipients with graft loss (Figures 1C and 3A, middle panels). Levels of TGF-β1, FGF-2, and VEGF (except on POD7) were quite similar in both recipient subgroups (Figures 1E, 3B, and 3C, middle panels). Thus, many of the cytokines tested (especially CXCL12) were found to be elevated in recipients with graft loss versus recipients with beneficial outcomes.
Comparison of cytokine profiles regarding living-donor liver transplant
outcomes in donors
Figures 1 to 3 (right panels) also depict cytokine profiles of donors stratified by LDLT outcomes (9 donors of recipients with beneficial outcome vs 4 donors of recipients with graft loss within 4 weeks after transplant). Preoperatively (POD0), only FGF-2 serum concentrations were elevated (P < .05) in donors of recipients with beneficial outcome versus donors of recipients with graft loss (Figure 3B, right panel).
After transplant (POD1-POD10), average concentrations of TGF-β1 (POD7, POD9, POD10), angiogenin (POD1-POD6), and especially FGF-2 (POD2-POD10) were significantly elevated (P < .05, but not at every POD) in donors of recipients with beneficial outcome versus donors of recipients with graft loss (Figure 3A and 3B, right). Interleukin 6 (reciprocally to its profile in the recipient subgroups [Figure 1B, middle panel]), G-CSF (POD1-POD6 and POD10; reciprocally to its profile in the recipient subgroups [Figure 2A, middle panel]), and VEGF (POD9 and POD10) were also found to be (slightly) increased; however, results were not statistically significant on every POD (Figures 1B and 2A, right panels). Interestingly, none of the cytokines tested was significantly increased in donors of recipients with graft loss compared with donors of recipients with beneficial outcome (only HGF on POD2, POD9, and POD10 and CXCL12 were slightly elevated, although not significantly; Figures 1A and 2C, right panels). Concentrations of IGF-1 (except on POD2), SCF, and especially TNF-α (in contrast to its other profiles in Figure 1D, left and middle panels) were quite similar in both donor subgroups (Figures 1C, 1D, and 2B, right panels). Thus, many of the cytokines tested (especially the angiogenic cytokine FGF-2, as well as angiogenin) were found to be elevated in donors of recipients with beneficial outcome compared with donors of recipients with graft loss.
Fibroblast growth factor 2 and CXCL12 might serve as predictors of beneficial
outcomes and graft loss
Before LDLT (on POD0), only FGF-2 and CXCL12 serum concentrations were significantly increased (P < .05) in donors of recipients with beneficial outcomes (Figure 3B, right panel) and recipients who had subsequent graft loss (Figure 2C, middle panel), respectively. Therefore, FGF-2 and CXCL12 might be of use as preoperative predictive markers. Receiver operating characteristic (ROC) curve analysis revealed that higher levels of FGF-2 in donors on POD0 (cut-off value of 5.4 pg/mL) were associated with beneficial outcomes of their recipients (sensitivity of 100%, specificity of 90%, positive predictive value = 90%, negative predictive value = 100%, accuracy of 92.31%; Figure 4A). Regarding CXCL12, higher preoperative serum concentrations in recipients (cut-off value of 4.2 ng/mL) were associated with graft loss within 4 weeks after LDLT (sensitivity of 75%, specificity of 100%, positive predictive value = 100%, negative predictive value = 90%, accuracy of 92.31%; Figure 4B).
To further characterize the cytokine profiles of donors and recipients, we also checked for potential correlations with conventional liver enzyme profiles, including aspartate aminotransferase (AST), alanine aminotransferase (ALT), and gamma-glutamyltransferase (γGT), which were previously described for the same cohort.5 Indeed, nonparametric correlation analyses revealed a positive correlation of the FGF-2 profile with serum γGT profile in donors (r = 0.3500; P < .001) and recipients (r = 0.2402; P = .005). In addition, IL-6 positively correlated with serum AST in donors (r = 0.3328; P < .001), whereas G-CSF positively correlated with serum ALT in recipients (r = 0.2684; P = .002).
Thus, FGF-2 levels correlate positively with serum γGT, and higher preoperative concentrations in donors may serve as a predictor of beneficial outcome in recipients, whereas elevated preoperative CXCL12 levels in recipients might be associated with early graft loss.
The results of this pilot study provide new insights regarding release patterns and prognostic relevance of 11 liver regeneration-related cytokines in donors and recipients undergoing LDLT. These data indicate that (1) many cytokines (especially TNF-α and SCF) are elevated in recipients compared with donors, (2) numerous cytokines are increased in recipients with graft loss (especially CXCL12) and in donors of recipients with beneficial outcome (especially FGF-2), and (3) higher preoperative FGF-2 levels in donors may predict beneficial recipient outcomes, whereas elevated preoperative CXCL12 in recipients might be associated with early graft loss. These findings are discussed in greater detail below.
In the present study, various cytokines, including TNF-α, SCF, and CXCL12, were found to be (slightly) elevated in recipients versus donors, an observation that often was associated with early graft loss. Together, increased values of these cytokines seem to reflect increased cytotoxic and inflammatory processes, resulting in impaired liver regeneration and function. For IL-6, this was true at later time points after an early peak that was more pronounced in donors than in recipients. Additionally, IL-6 (donors) and G-CSF (recipients) levels were positively correlated with peaking AST and ALT levels, respectively, implying a connection between liver damage in general and IL-6/G-CSF release into the serum, as described previously for subtype α of the detoxifying enzyme glutathione-S-transferase (GST-α).5 Granulocyte colony-stimulating factor is widely used as a mobilizing agent in stem cell-based liver therapy approaches with heterogeneous results.12 In the present study, G-CSF was not found to be associated with a favorable outcome, which is consistent with previous findings in liver transplant recipients.30 From all cytokines tested, only TGF-β1 (a strong repressor of liver regeneration) was found to be significantly decreased in recipients. On partial hepatectomy, TGF-β is still expressed, but the responsiveness to this cytokine declines transiently.31 Because liver grafts are also challenged with preservation injury and immunosuppression, downregulation of this growth inhibitor might therefore be an additional mechanism to ensure intact hepatocyte proliferation. The significantly higher CXCL12 levels observed in recipients with early graft loss before and after LDLT indicate a link between this marker and a nonfavorable outcome. Wilson and associates similarly reported that signaling via CXCR4 (the receptor of CXCL12) is detrimental to liver recovery and regeneration after ischemia/reperfusion.32 Moreover, the application of CXCR4 antagonists may even improve hepatic recovery after liver injury, confirming a connection between high CXCL12 levels and impaired liver regeneration.32 In our patient cohort, we found a significant correlation between FGF-2 and γGT. As we have previously shown, γGT is significantly increased in patients who show beneficial outcomes after acute liver failure.33 Thus, a temporary increase of FGF-2 and γGT may imply regenerative activity and thus a favorable outcome after surgery, an assumption that has also been made for GST-α.5 In contrast, continuously raised serum levels are likely to be a marker for ongoing liver damage and/or constant insufficient regeneration. Similarly, higher FGF-2 values in donors (before and early after LDLT) were associated with graft survival in their corresponding recipients, confirming FGF-2 as a potential indicator of improved regenerative liver capacity. However, under different conditions after LDLT (eg, preservation injury, immunosuppression), FGF-2 levels were not found to be similarly elevated in recipients with beneficial outcomes themselves. Thus, livers of donors with elevated FGF-2 levels may already be primed and more suitable for LDLT preoperatively due to one or more of the pleiotropic FGF-2 effects regarding angiogenesis,26 hepatocyte mitogenesis,8 and facilitated hepatic transdifferentiation of (possibly already mobilized) stem cells.27,28 Other angiogenic factors like VEGF have also been described as important drivers of liver regeneration.8,24 Indeed, we also found that higher concentrations of angiogenin (especially in donors) were slightly associated with improved outcome; however, VEGF levels were quite similar in donors and recipients and not linked to the clinical course.
Although ambitious, due to the limited number of patients, this study is preliminary and thus conclusions and ROC analyses of potential predictive cytokines remain speculative. Further studies with larger patient cohorts are warranted to investigate noninvasively obtainable predictive markers. Nevertheless, the present data reveal distinct liver regeneration-related cytokine release patterns in LDLT that are different in donors and recipients. The results also indicate that FGF-2 levels correlate positively with serum γGT, and higher preoperative concentrations in donors may predict an improved regenerative capacity, resulting in a favorable recipient outcome. Furthermore, elevated CXCL12 concentrations in recipients before and after LDLT may serve as a predictor of graft failure and marker of liver damage, respectively, contraindicating LDLT (or deceased-donor liver transplant). These data might also be of interest regarding acute liver failure in which a higher regenerative capacity and decreased cell death is associated with prolonged patient survival.11,34 In conclusion, FGF-2 and CXCL12 may be of diagnostic value in complementary ways to describe or even predict possible outcomes after LDLT.
Volume : 16
Issue : 5
Pages : 554 - 561
DOI : 10.6002/ect.2017.0039
From the 1Department of General, Visceral and Transplantation
Surgery, the 2Department of Gastroenterology and Hepatology,
University Hospital Essen, University Duisburg-Essen, Essen, Germany; the 3Department
of General, Visceral and Transplant Surgery, University Hospital Tuebingen,
Tuebingen, Germany; and the 4Department of Gastroenterology,
Hepatology, Infektiology, University Hospital, University of Magdeburg,
Acknowledgements: This work was supported by the German Research Foundation/Deutsche Forschungsgemeinschaft (DFG, grants 267/11-1 and 267/14-1) and the Wilhelm Laupitz Foundation to AC as well as the grants FI 1630/3-1 (DFG) and D/107-114400 (IFORES) to C.D. Fingas. The authors have no conflicts of interest to declare.
Corresponding author: Ali Canbay, University Hospital Magdeburg, Department of Gastroenterology, Hepatology, Infectiology, Leipziger Str. 44, 39120 Magdeburg, Germany
Phone: +49 391 67 13100
Figure 1. Profiles of Cytokines That Mediate Liver Regeneration in Donors and Recipients After Living-Donor Liver Transplant
Figure 2 Profiles of Cytokines Associated With Hematopoietic Stem Cell Mobilization During Liver Regeneration in Donors and Recipients After Living-Donor Liver Transplant
Figure 3. Profiles of Angiogenesis-Related Cytokines During Liver Regeneration in Donors and Recipients After Living-Donor Liver Transplant
Figure 4 Receiver Operating Characteristic Curves for Single Parameters (Fibroblast Growth Factor 2 and CXCL12)
Table 1. Donor and Recipient Characteristics and Indications for Living-Donor Liver Transplant