Somatostatin Therapy Protects Porcine Livers in Small-for-Size Liver Transplantation

Authors

  • A. J. Hessheimer,

    1. Department of Surgery, Institut de Malalties Digestives I Metabòliques (IMDiM), Hospital Clínic, CIBERehd, IDIBAPS, University of Barcelona, Barcelona, Spain
    Search for more papers by this author
  • B. Escobar,

    1. Department of Anesthesia, Hospital Clínic, University of Barcelona, Barcelona, Spain
    Search for more papers by this author
  • J. Muñoz,

    1. Department of Surgery, Institut de Malalties Digestives I Metabòliques (IMDiM), Hospital Clínic, CIBERehd, IDIBAPS, University of Barcelona, Barcelona, Spain
    Search for more papers by this author
  • E. Flores,

    1. Department of Surgery, Institut de Malalties Digestives I Metabòliques (IMDiM), Hospital Clínic, CIBERehd, IDIBAPS, University of Barcelona, Barcelona, Spain
    Search for more papers by this author
  • J. Gracia-Sancho,

    1. Liver Unit, Institut de Malalties Digestives I Metabòliques (IMDiM), Hospital Clínic, CIBERehd, IDIBAPS, University of Barcelona, Barcelona, Spain
    Search for more papers by this author
  • P. Taurá,

    1. Department of Anesthesia, Hospital Clínic, University of Barcelona, Barcelona, Spain
    Search for more papers by this author
  • J. Fuster,

    1. Department of Surgery, Institut de Malalties Digestives I Metabòliques (IMDiM), Hospital Clínic, CIBERehd, IDIBAPS, University of Barcelona, Barcelona, Spain
    Search for more papers by this author
  • A. Rimola,

    1. Liver Unit, Institut de Malalties Digestives I Metabòliques (IMDiM), Hospital Clínic, CIBERehd, IDIBAPS, University of Barcelona, Barcelona, Spain
    Search for more papers by this author
  • J. C. García-Valdecasas,

    1. Department of Surgery, Institut de Malalties Digestives I Metabòliques (IMDiM), Hospital Clínic, CIBERehd, IDIBAPS, University of Barcelona, Barcelona, Spain
    Search for more papers by this author
  • C. Fondevila

    Corresponding author
    1. Department of Surgery, Institut de Malalties Digestives I Metabòliques (IMDiM), Hospital Clínic, CIBERehd, IDIBAPS, University of Barcelona, Barcelona, Spain
    Search for more papers by this author

Abstract

Small-for-size (SFS) injury occurs in partial liver transplantation due to several factors, including excessive portal inflow and insufficient intragraft responses. We aim to determine the role somatostatin plays in reducing portal hyperperfusion and preventing the cascade of deleterious events produced in small grafts. A porcine model of 20% liver transplantation is performed. Perioperatively treated recipients receive somatostatin and untreated controls standard intravenous fluids. Recipients are followed for up to 5 days. In vitro studies are also performed to determine direct protective effects of somatostatin on hepatic stellate cells (HSC) and sinusoidal endothelial cells (SEC). At reperfusion, portal vein flow (PVF) per gram of tissue increased fourfold in untreated animals versus approximately threefold among treated recipients (p = 0.033). Postoperatively, markers of hepatocellular, SEC and HSC injury were improved among treated animals. Hepatic regeneration occurred in a slower but more orderly fashion among treated grafts; functional recovery was also significantly better. In vitro studies revealed that somatostatin directly reduces HSC activation, though no direct effect on SEC was found. In SFS transplantation, somatostatin reduces PVF and protects SEC in the critical postreperfusion period. Somatostatin also exerts a direct cytoprotective effect on HSC, independent of changes in PVF.

Abbreviations
α-SMA

alpha-smooth muscle actin

AST

aspartate aminotransferase

ET-1

endothelin-1

HAF

hepatic artery flow

HSC

hepatic stellate cells

HVP

hepatic venous pressure

ICG

indocyanine green

PCS

portocaval shunt

PDR

plasma disappearance rate

PVF

portal vein flow

PVP

portal venous pressure

PVPG

portal venous pressure gradient

PVR

portal vein resistance

QPT

Quick prothrombin time

SEC

sinusoidal endothelial cells

SFS

small-for-size

SFSS

small-for-size syndrome

SLV

standard liver volume

SSTR

somatostatin receptor

TK

thymidine kinase

Introduction

Partial liver transplantation represents an important potential means to expand the donor pool and increase the applicability of liver transplantation as the sole treatment for end-stage liver disease. Nevertheless, a major obstacle preventing broader use of partial grafts, including splitting a liver for two adults, is the fact that it may result in insufficient hepatic mass for the recipient, leading to the development of small-for-size syndrome (SFSS). SFSS is diagnosed clinically when cholestasis, coagulopathy, encephalopathy, ascites, gastrointestinal bleeding and/or renal failure arise in a liver representing <0.8% of the recipient's body weight or <40% of the recipient's standard liver volume (SLV), at the exclusion of another cause [1, 2]. The clinical presentation of SFSS is typically subacute; when it does occur, however, SFSS leads to graft failure within 2–3 months in up to 50% of recipients [3].

The pathogenesis of SFSS has been attributed to not only small functional liver mass but also excessive portal inflow and insufficient venous outflow and intragraft responses [2]. We studied the acute phase of small-for-size (SFS) injury in a porcine model of 20% liver transplantation and found that portal vein flow (PVF) approximately fourfold higher than baseline and a significantly elevated portal venous pressure gradient (PVPG) were present upon reperfusion. These hemodynamic alterations caused direct mechanical injury to the sinusoidal endothelial cells (SEC) and were the chief stimulus for a subsequent series of intragraft events leading to acute failure [4]. Using the same model, we demonstrated that a portocaval shunt (PCS) placed prior to reperfusion and calibrated to maintain twice-baseline PVF decompressed the splanchnic bed, prevented SEC injury and resulted in excellent recipient survival [5].

Though a calibrated PCS is effective at preventing acute SFS injury, it is a relatively permanent solution for a temporary problem, as the critical period of portal hyperperfusion-induced injury is limited [6]. Hence, it is of interest to develop reversible means for reducing PVF in the immediate postreperfusion period while the partial graft is still adapting to the altered physiology of the recipient's body. Pharmacological portal inflow modification with short-acting drugs, such as somatostatin, is one such option.

Somatostatin is a naturally occurring tetradecapeptide that decreases portal venous pressure (PVP) and splanchnic blood flow in a dose-dependent manner [7, 8]. It induces vasoconstriction in the mesenteric arteries and portocollateral veins [9, 10] and inhibits the secretion of gut-derived vasodilatory peptides, such as glucagon, vasoactive intestinal peptide and substance P [11-13]. While the effects of somatostatin on the portomesenteric vasculature are well established, less is known about the role that somatostatin plays locally in the hepatic microvasculature, in particular among transplanted livers in the acute phase of injury. In the normal innervated liver, somatostatin-containing axons are present in the space of Disse and exert paracrine effects on nearby cells, depending on the expression of the five somatostatin receptor (SSTR) subtypes [14]. Receptor expression is limited to cholangiocytes and a small number of endothelial cells lining the hepatic arterioles in the normal liver; in the cirrhotic liver, on the other hand, both hepatocytes and hepatic stellate cells (HSC) may express all five subtypes [15-17], indicating potential sites of action in chronic liver disease.

The hypothesis of the present work is that somatostatin is capable of reducing PVF in SFS liver transplantation, indirectly leading to less SEC injury and endothelin-1 (ET-1)-mediated activation of HSC and improved graft survival. Additionally, we hypothesize that somatostatin may have direct protective effects in certain key cells in the acute phase of injury in the partial liver, as well. To test these hypotheses, we perform a porcine model of 20% liver transplantation in which somatostatin-treated recipients are compared with untreated controls.

Experimental Procedures

Animals

Male weanling pigs were used as donors and recipients. Animals were cared for according to the guidelines of the University of Barcelona Committee on Ethics in Animal Experimentation and the Catalan Department of the Environment Commission on Animal Experimentation.

Surgical procedures, intraoperative monitoring and postoperative management

SFS liver transplants were performed as described previously [4, 18]. Briefly, donor pigs (15–20 kg) underwent extended hepatectomy. Partial liver grafts were implanted into larger recipients (30–35 kg), representing approximately 20% of their SLV. Implants were performed without venovenous bypass, vasopressors or inotropes. Animals were divided between three groups. In the first, no treatment other than standard intravenous fluids was given (SFS group, N = 19). In the other two, somatostatin was administered as a bolus during the anhepatic phase (15 μg/kg, intravenously), followed by continuous infusion of either 8 (SST-8, N = 6) or 15 μg/kg/h (SST-15, N = 6) over the next 5 days. (Dosing schemes were chosen based on a set of preliminary experiments performed in anesthetized pigs—see Supplementary Material.) Since this is such a difficult and time-consuming model, the majority (N = 15) of experiments on control animals included were performed in the context of previous SFS studies [4, 5]; however, an additional four control experiments were performed over the course of this study. Experiments on new control animals as well as the animals receiving both doses of somatostatin were performed in random fashion. Since this was not a clinical study and there were hard end points (e.g. mortality) and we, furthermore, had to mix the drugs immediately prior to administering them, we were not blinded to the treatment that each animal received.

Intraoperatively, PVF, PVP, hepatic venous pressure (HVP) and hepatic artery flow (HAF) were monitored [5]. In SST-8 and SST-15 recipients, a Surefuser Plus intravenous infusion pump was implanted (Nipro Europe, Madrid, Spain). The pump consisted of a reservoir, which contained a solution mixed to a predetermined concentration in order to continuously provide the specified amount of somatostatin. The reservoir was placed in the right iliac fossa and the catheter to which it was connected tunneled subcutaneously and connected to the right external jugular vein. Since many of the experiments on control animals were performed prior to undertaking this particular study, they were not implanted with an intravenous infusion pump as a control measure.

Postoperatively, recipient pigs were cared for intensively for 5 days [5]. Animals that survived until the 5th day were anesthetized and reopened and hemodynamic parameters reevaluated; liver grafts were removed and sampled and the animals euthanized. Nonsurvivors were autopsied.

Markers of hepatocellular injury, metabolism and regeneration

In blood samples collected serially during follow-up, aspartate aminotransferase (AST), total bilirubin, creatinine and the Quick prothrombin time (QPT) were determined using the Advia 1650 automatic analyzer (Bayer, Tarrytown, NY).

Hepatic metabolic function was quantified according to the plasma disappearance rate (PDR) of indocyanine green (ICG) at baseline, after reperfusion and on day 5 among surviving recipients [4].

Thymidine kinase (TK), a phosphotransferase up-regulated by actively dividing cells, provides an index of hepatic regeneration [19]. TK activity was measured in serial serum samples [5].

In order to evaluate sinusoidal spatial integrity during the regenerative phase, tissue samples taken at euthanasia in surviving recipients were stained with porcine anti-CD31 antibody [4].

Levels of somatostatin and of C-peptide were determined in serial serum samples (see Figures S1 and S2).

Sinusoidal endothelial cell injury

ET-1 is a potent vasoconstrictive peptide up-regulated in situations of endothelial cell stress; in the liver, it is produced by endothelial cells and metabolized by hepatocytes. Serum levels of ET-1 were measured and cryosections of hepatic tissue immunostained with porcine antibody against ET-1 [5].

Hepatic stellate cell activation

Alpha-smooth muscle actin (α-SMA), a major constituent of the cellular contractile apparatus, is expressed in the liver by activated HSC. Levels of α-SMA protein expression in hepatic tissue were determined via Western blot and immunohistochemistry (see Figure S1).

In vitro studies

Activated human HSC were cultured in DMEM as described previously [20]. Cultured cells were counted and evenly distributed in a six-well plate and incubated with either 4- or 10-nM somatostatin (doses chosen to reflect circulating levels of somatostatin measured in the in vivo experiments) or vehicle. Incubations were performed for 1, 12 or 24 h. Given its short half-life [21], somatostatin was added to the medium every 6 h, with or without an additional dose given at 1 h prior to the end of the experiment in 12- and 24-h incubations. Cells were lysed for α-SMA protein analysis.

Rat liver SEC were isolated [22]; cultured cells were counted and evenly distributed. ET-1 mRNA expression was assessed in freshly isolated cells and in SEC incubated for 24 h with either vehicle or somatostatin (10 nM) or co-cultured with HSC and incubated with somatostatin (10 nM). Doses of somatostatin were given every 6 h, with an additional dose given 1 h prior to the end of the experiment. cDNA templates were amplified by real-time quantitative polymerase chain reaction using fluorescent TaqMan technology (Applied Biosystems, Foster City, CA) on an ABI Prism 7900 sequence Detection System (Applied Biosystems). Quantification of rat ET-1 and the endogenous control GAPDH was performed using predesigned gene expression assays obtained from Applied Biosystems according to the manufacturer's protocol.

Data and statistical analysis

All quantitative values are expressed as the median and 25–75% interquartile range. Differences between groups were compared using the Mann–Whitney test for two samples and the Kruskal–Wallis one-way analysis of variance for multiple samples. Survival was analyzed according to the method of Kaplan–Meier, and comparisons between groups were made using the Mantel–Cox log rank test. p < 0.05 was considered significant. Calculations were performed using Predictive Analytics SoftWare Statistics version 18.0 (IBM, Somers, NY).

Results

Transplant characteristics

Among the SFS, SST-8 and SST-15 groups, graft weight-to-body weight ratios were 0.58 (0.54–0.66), 0.58 (0.47–0.69) and 0.59% (0.56–0.63); SLV 23.0 (19.4–25.5), 22.5 (20.0–25.0) and 21.0% (17.4–24.5); and cold ischemic times 322 (283–340), 308 (269–359) and 329 min (322–335), respectively.

Pre- and postreperfusion hepatic hemodynamics

PVF and HAF values are expressed in mL/min/kg hepatic tissue. The PVPG was calculated as the difference between PVP and HVP. Portal vein resistance (PVR) was calculated as described previously [4].

There were no significant differences between the groups in the parameters measured at baseline. Upon reperfusion, PVF, PVPG and PVR were all significantly higher in the SFS group versus the other two. Compared with baseline, PVF underwent 4.0-fold change (2.9–4.9) in the SFS group versus 3.0-fold change (2.9–3.1) in SST-8 and 2.7-fold change (2.2–3.3) in SST-15 (p = 0.033). PVF, PVPG and PVR were all significantly higher in the SFS group after arterial reperfusion, as well, which occurred approximately 40 min after portal reperfusion. Also after arterial reperfusion, HAF was lower in the SFS group versus the other two, although the difference did not reach statistical significance (p = 0.078) (Table 1).

Table 1. Pre- and postreperfusion hepatic hemodynamic parameters
 SFS (N = 19)SST-8 (N = 6)SST-15 (N = 6)p1
  • This table depicts the evolution of hemodynamic parameters measured at baseline in the donor, after portal and arterial reperfusion (approx. 40 min after portal reperfusion) in the recipient, and at 5 days in the surviving animals. AR, arterial reperfusion; BAS, baseline; EUT, euthanasia; HAF, hepatic artery flow; PR, portal reperfusion; PVF, portal vein flow; PVPG, portal venous pressure gradient; PVR, portal vein resistance; SFS, small-for-size; SST, somatostatin.
  • 1Calculated according to the Kruskal–Wallis one-way analysis of variance for multiple samples (BAS, PR, AR) and the Mann–Whitney test for two samples (EUT).
  • 2For both SFS and SST-8, N = 5 at euthanasia.
PVF (mL/min/kg)
BAS996 (702–1284)1052 (886–1074)1032 (928–1392)0.789
PR3797 (3089–5309)2884 (2687–3167)2593 (2360–3565)0.042
AR3418 (2578–3783)2488 (2310–2959)2304 (1974–2826)0.021
EUT21556 (1278–1816)1817 (1678–2130)0.730
PVPG (mmHg)
BAS22 (1–2)2 (1–2)0.232
PR11 (9–13)8 (7–9)8 (7–9)0.027
AR9 (8–10)6 (5–7)6 (5–6)0.001
EUT26 (5–7)5 (5–6)0.556
PVR (dyn s m2/cm5)
BAS333 (262–457)200 (182–300)291 (234–385)0.202
PR1224 (828–1686)767 (624–1011)795 (632–1008)0.028
AR1181 (976–1714)703 (583–769)787 (506–800)0.002
EUT2583 (466–753)381 (377–447)0.857
HAF (mL/min/kg)
BAS333 (277–547)381 (287–475)422 (225–561)0.799
AR189 (124–259)264 (230–356)282 (192–304)0.078
EUT2128 (104–135)155 (144–175)0.111

Postoperative evolution

Hepatocellular injury

Upon reperfusion, serum AST, total bilirubin and creatinine increased and QPT decreased in all of the recipient animals. Comparing the SFS group with SST-8, AST was significantly higher at 2, 3, 4 and 5 days (p = 0.039, 0.028, 0.014 and 0.025, respectively) and creatinine was significantly higher at 2 days (p = 0.047). Bilirubin was significantly higher in SFS versus SST-15 at 6 h (p = 0.007), while QPT was significantly lower at 3 h (p = 0.001). Finally, bilirubin was significantly higher in SST-8 versus SST-15 at 6 and 12 h (p = 0.017 and 0.042, respectively) (Figure 1).

Figure 1.

Postoperative evolutions of serum aspartate aminotransferase (AST) (A), total bilirubin (B), creatinine (C) and Quick prothrombin time (QPT) (D) among somatostatin-treated (SST-8, SST-15) and small-for-size untreated recipients (SFS). Dots indicate the mean and whiskers the standard error of the mean; tables depict the number of surviving animals in each group at each time point. p < 0.05 for SFS versus SST-8 (*), SST-8 versus SST-15 (**) and SFS versus SST-15 (***).

Survival

Survival was 26% in the SFS group, 83% in SST-8 and 17% in SST-15 (p = 0.019 and 0.014 for SST-8 vs. SFS and SST-15); five of the six animals in the SST-15 group suffered cardiorespiratory arrest between 12 and 18 h after reperfusion (Figure 2). All nonsurviving recipients were subjected to autopsy, and no apparent cause of death (including, but not limited to, vascular thrombosis, gastrointestinal perforation or hemorrhage) could be found. Seven animals in the SFS group had moderate amounts of ascites (one at euthanasia, six at autopsy), while none of the animals receiving somatostatin had ascites.

Figure 2.

Survival curves for somatostatin-treated and untreated recipients. Five-day survival was 83%, 26% and 17% in the SST-8, SFS and SST-15 groups, respectively.

Hepatic regeneration and function during follow-up

Graft mass increased during the follow-up period. Because survival varied, the increase in mass was divided by survival in hours and multiplied by 24 h/day. Growth per day was higher in the SFS group, 114 g/day (72–181), versus SST-8, 69 g/day (65–90), though the difference did not reach statistical significance (p = 0.065).

Among the recipients of untreated SFS grafts, TK activity initially peaked at 12 h and was significantly higher in the SFS group versus SST-8 (p = 0.039). Among recipients treated with somatostatin, however, TK activity remained relatively stable throughout the entire follow-up period (Figure 3).

Figure 3.

(A) Postoperative evolution of serum thymidine kinase (TK) activity, a marker of cellular proliferation. There was a dramatic increase in TK activity after reperfusion among the SFS grafts. Among recipients treated with somatostatin, however, TK activity remained relatively stable throughout the entire follow-up period. Dots indicate the mean and whiskers the standard error of the mean; tables depict the number of surviving animals in each group at each time point. *p < 0.05. (B) Anti-CD31 immunohistochemical staining of tissue samples taken at the end of follow-up. Among surviving recipients in the SFS group, the hepatic sinusoidal spaces were markedly compressed by the widened hepatocyte cords. In grafts treated with somatostatin, however, the hepatocyte cords were narrower, and the normal sinusoidal microarchitecture was better preserved.

Tissue sampled on the 5th day in survivors revealed compressed sinusoidal spaces and widened hepatocyte plates among untreated grafts. In grafts treated with somatostatin, however, hepatocyte plate widths were narrower and the integrity of the sinusoidal spaces was better maintained (Figure 3).

The PDR of ICG was similar among all recipients at baseline: SFS 10.2 (6.6–10.5), SST-8 10.6 (6.6–12.1) and SST-15 12.3 (8.9–15.1). Upon graft reperfusion, it declined to a similar extent among all three groups: SFS 6.0 (5.5–6.4), SST-8 5.5 (5.4–5.6) and SST-15 5.7 (5.4–5.9). On the 5th day, however, significant differences between the SFS and SST-8 groups were detected, with improved function observed in the latter versus the former: 4.2 (3.7–4.7) versus 8.0 (6.6–13.5), respectively (p = 0.014).

Sinusoidal endothelial cell injury

Histology

After reperfusion, there was extensive destruction of SEC lining the portal venous branches, with significant periportal as well as intraparenchymal hemorrhage in the SFS group; these findings were minimal-to-absent in the SST groups (Figure 4).

Figure 4.

Masson's trichrome, H&E and anti-endothelin-1 (ET-1) immunohistochemical staining of tissue samples taken 1 h after portal reperfusion. In the SFS group, there was significant endothelial denudation in the medium-sized portal vein branches and hemorrhage into perivenular connective tissue (D), which extended into the hepatic parenchyma (E). Treatment with somatostatin led to better preservation of the sinusoidal endothelial and the periportal connective tissue (G), and intraparenchymal hemorrhage was absent (H). Anti-ET 1 immunostaining demonstrated considerable endothelial cell injury among SFS grafts, with a significant upregulation in expression and subsequent uptake of ET-1 in the distorted hepatocyte cords (F). Somatostatin-treated grafts, on the other hand, were notable for very minimal anti-ET 1 staining (I). Levels of ET-1 measured in serum confirmed the immunohistochemical findings (J). Dots indicate the mean and whiskers the standard error of the mean; tables depict the number of surviving animals in each group at each time point. p < 0.05 for SFS versus SST-8 (*), SST-8 versus SST-15 (**) and SFS versus SST-15 (***).

Endothelin-1 immunodetection

After reperfusion in untreated grafts, there was considerable uptake of ET-1 into distorted hepatocyte cords. In the SST groups, however, ET-1 expression was significantly less, and the hepatocyte cords remained well preserved (Figure 4).

Serum endothelin-1 levels

Serum levels of ET-1 were similar at baseline but were significantly higher in the SFS group versus animals receiving somatostatin regardless of the dose at 3 and 12 h postreperfusion (p = 0.024 and 0.014 for SFS vs. SST-8 and 0.024 and 0.007 for SFS vs. SST-15 at 3 and 12 h, respectively) (Figure 4).

Hepatic stellate cell activation

After reperfusion, there was notably more immunostaining of activated HSC in grafts in the SFS group versus those treated with somatostatin. As well, on Western blot analysis, expression of α-SMA protein in hepatic tissue was significantly up-regulated among grafts in the SFS group versus those treated with SST (Figure 5).

Figure 5.

Measurement of alpha-smooth muscle actin (α-SMA) expression as a marker of hepatic stellate cell activation after graft reperfusion. Anti-α SMA immunostaining was more notable in untreated grafts in the SFS group (C) versus those treated with somatostatin (B). Western blot analysis confirmed significantly greater expression of α-SMA protein in untreated samples (E, F). Data are expressed as the median, with the 25–75% percentiles in boxes and the 5–95% percentiles as whiskers. *p > 0.05.

In vitro studies

Hepatic stellate cells

For incubations lasting 1 h, levels of α-SMA were significantly lower in cells incubated with 10 nM somatostatin versus vehicle (p = 0.01). Levels also tended to be lower in cells incubated with 10 nM versus 4 nM somatostatin and in those incubated with 4 nM somatostatin versus vehicle, although these differences did not reach statistical significance (p = 0.109 and 0.127, respectively) (Figure 6). For incubations lasting longer than 1 h, no differences were observed with respect to vehicle when the last dose of somatostatin was given 6 h prior to the end of the experiment; however, the addition of a final dose of somatostatin 1 h prior to the end of the experiment led to results comparable to those observed in the 1 h incubations.

Figure 6.

In vitro studies. (A) Activated hepatic stellate cells (HSC) were incubated for 1, 12 and 24 h with either 4 or 10 nM somatostatin. Alpha-smooth muscle actin expression measured by Western blot analysis was significantly reduced in activated HSC incubated with somatostatin for periods as short as 1 h, indicating a direct and immediate effect in these cells. *p < 0.05 versus vehicle. (B) Sinusoidal endothelial cells (SEC) were cultured for 24 h in order to develop an injured phenotype. After 24 h in culture, levels of endothelin-1 expression determined by real-time quantitative polymerase chain reaction were significantly up-regulated with respect to freshly isolated SEC, indicating that the injury phenotype was achieved. Endothelin-1 expression did not change, however, when injured SEC were treated with 10-nM somatostatin, either alone or in co-culture with HSC. Data are expressed as the median, with the 25–75% percentiles in boxes and the 5–95% percentiles as whiskers. **p < 0.05 versus 0 h (freshly isolated SEC).

Sinusoidal endothelial cells

Levels of ET-1 were significantly higher in cultured versus freshly isolated SEC, regardless of somatostatin treatment (p < 0.05 for all comparisons). However, no differences were observed between cultured SEC treated with vehicle versus somatostatin, either alone or in co-culture with HSC (p = 0.827 and 0.513, respectively) (Figure 6).

Discussion

The porcine model is the most relevant means available to study SFSS outside the clinical setting. While rat and mouse models require relatively little material, personnel and space to perform, it is hard to perform hemodynamic studies in this context. Arterialization of the liver is difficult and often omitted in the case of rats, and the cuffed portal vein anastomosis may artificially decrease PVF [23-25]. Such a finding is not in any way representative of the clinical situation with partial liver transplantation and raises concern as to whether such a small-cuffed anastomosis truly allows unimpeded laminar flow through the portal vein. Large animal models, on the other hand, offer size and anatomy similar to humans. Vascular anastomoses may be performed end-to-end, as they are clinically, and are significantly less prone to artificial narrowing.

Using our porcine SFS liver transplant model, our group previously demonstrated that postreperfusion PVF increases considerably (fourfold with respect to baseline) in extremely small grafts [4]. Herein, we demonstrate that somatostatin administered prior to and after reperfusion can significantly reduce this flow, improving posttransplantation outcomes. By decreasing portal flow and pressure, somatostatin has an indirect protective effect on the partial liver. However, the reduction in PVF and the PVPG produced by somatostatin was not as immediate or as important as that we achieved with calibrated PCS in our previous study, so we looked for other means by which somatostatin was able to offer hepatoprotection in this setting. For this reason, we performed in vitro studies, which demonstrated that somatostatin acts directly on and reduces the activation of HSC, as well. We also determined for the first time the effect of somatostatin on isolated, phenotypically injured SEC.

Another group looked at the effects of somatostatin on activated HSC in vitro and found significant decreases in α-SMA gene expression and de novo protein synthesis but no differences in the cellular contractile apparatus or overall levels of α-SMA protein [26]. When we re-dosed somatostatin every 6 h, as was done in the previous study, we, too, observed no differences in α-SMA protein expression versus vehicle. Only when somatostatin was given 1 h before HSC were lysed for protein extraction did we observe a significant decrease in α-SMA protein expression, regardless of the overall duration of incubation. This is an important finding that supports the results of our porcine studies by confirming the fact that the effect of somatostatin on the HSC contractile apparatus is immediate and offers further proof that somatostatin may play a key role in overcoming early SFS liver injury through a direct hepatoprotective mechanism, independent of the reduction in PVF.

In addition to its direct effect on HSC, we have been able to demonstrate a protective effect for SEC in grafts treated with somatostatin in vivo, as reflected by decreased immunostaining and circulating levels of ET-1. Other authors have similarly observed down-regulated expression of ET-1 in partial grafts treated with somatostatin in both rodents and humans, although they did not investigate the specific mechanism through which this down-regulation occurred [27, 28]. Given the results of the in vitro studies in which we were unable to demonstrate a reduction in ET-1 expression in phenotypically injured SEC incubated with somatostatin, both in isolation and co-cultured with similarly treated HSC, it appears that the reduction in ET-1 expression in vivo is an effect of the splanchnic vasoconstriction and decrease in PVF. That is to say, there does not appear to be any direct protective effect of somatostatin on SEC. No other studies have looked at the direct effects of somatostatin on hepatic SEC in culture, so there is no other data with which to compare our results.

Postoperatively, biochemical parameters were improved in somatostatin-treated animals and tended to be the best among those receiving 15 μg/kg/h, indicating that liver failure was not the cause of the death among these animals. An effect of the learning curve was not to blame, either, as our group had extensive experience in this model when this study was undertaken. The lower dose of somatostatin we used was chosen based on how the drug is typically administered clinically in the context of bleeding esophageal varices: an initial 250–500 μg bolus (the equivalent of about 4–8 μg/kg, based on a 60–70 kg adult) followed by 250–500 μg/h over the course of 5 days [29]. With respect to our choice of 15 μg/kg, we performed a preliminary set of experiments in anesthetized pigs and observed a more immediate response in portal hemodynamic parameters with an initial bolus of 15 µg/kg versus 8 μg/kg, although some slowing of the heart rate with a 15 μg/kg/h infusion was also observed. At standard doses somatostatin is considered to be a very safe drug [30, 31], but high doses may provoke bradypnea, bradycardia, hypotension and even apnea [32-35]. Thus, while it was effective at decreasing splanchnic blood flow and PVP and even led to some improvements over the more standard dose in terms of the hepatic biochemical profile, we believe that the high-dose somatostatin infusion we used in the SST-15 group may have had the adverse effect in this model of provoking terminal cardiorespiratory events.

Given the adverse effects of high-dose somatostatin, one might wonder whether it could potentially be administered through the portal vein or its venous tributaries and act on the liver directly while avoiding some of its systemic effects due to first-pass metabolism. Administration of the drug in this manner, however, means that it would have much less of a vasoconstrictive effect on mesenteric arterial vasculature, and the reduction in portal inflow would be less. Similarly, the administration of longer-acting somatostatin analogs such as octreotide and lanreotide, which are more resistant to endogenous peptidases, might be appealing. Native somatostatin and its analogs, however, have different affinities for the five SSTR subtypes: the former binds all five, while octreotide binds only SSTR2 and SSTR5 [36]. Previous studies have shown that relaxation of cultured HSC is mediated by SSTR1 [15, 17, 37], indicating that octreotide would perhaps not offer a direct hepatoprotective effect in the setting of acute SFS injury.

Based on the fact that it lowers PVF and suppresses the release and action of several hepatotrophic factors, somatostatin is known to suppress hepatic regeneration [38-40]. While some authors argue that SFSS arises primarily due to the failure of a partial liver to regenerate [41, 42], other groups have made observations to the contrary. In the clinical setting, good restoration of hepatic size on computer tomographic volumetry is observed in patients who simultaneously demonstrate clinical signs of liver insufficiency [43, 44], and small grafts have been shown to have greater regenerative capacity compared with larger ones [45]. In our own preclinical experience, we have consistently observed a significantly greater increase in mass and TK activity among failing SFS grafts versus those with initial good function, in particular those treated with measures to reduce PVF and suppress regeneration (e.g. an appropriately calibrated portocaval flow derivation or, in the present case, somatostatin) [4, 5].

Events occurring at the cellular level help explain why livers that regenerate rapidly might still be SFS. After major hepatic resection, entry into the cell cycle and DNA synthesis occur much earlier in hepatocytes versus nonparechymal cells. In the rat, DNA synthesis in hepatocytes begins to increase after 12 h and peaks at 24 h, whereas it begins around 48 h for Kupffer cells and cholangiocytes and 96 h for SEC [46, 47]. This order of events results in the formation of avascular clusters of 10–14 hepatocytes devoid of extracellular matrix, the so-called “hepatocyte islands” [48-50], which persist until SEC are replicated and the normal extracellular matrix restored. Consequently, reformation of the normal liver microarchitecture occurs only after the original mass has been replaced. Hepatocyte islands are considered less functional than normal single-cell-wide hepatocyte plates due to the fact that there are fewer sinusoids and bile canaliculi for each hepatocyte. Therefore, it follows that the smaller the graft, the greater is the proportion of hepatocytes entering the replicative process to form these islands, decreasing the amount of functional liver tissue available even as gross mass is restored [44]. This helps explain why multiple studies, including our own, have reported that functional regeneration does not correlate with volumetric regeneration during the early stages of liver regrowth [4, 5, 51-54] and that strategies that decelerate regeneration can actually lead to better preservation of normal hepatic microarchitecture and improved survival during the regenerative process [43].

Based on the results of this study, we conclude that the perioperative administration of somatostatin in SFS liver transplantation offers a reversible means to reduce portal inflow and protect the cells lining the hepatic sinusoid in the critical postreperfusion period, thereby significantly improving posttransplantation outcomes. Additionally, somatostatin exerts a direct and immediate cytoprotective effect on HSC, independent of the reduction in PVF. Taken together, these findings indicate that somatostatin may play an important role in alleviating SFSS and increasing the applicability of partial liver transplantation in the clinical setting.

Acknowledgments

Research was supported by F.I.S. Grant 08/0273 from the Instituto de Salud Carlos III, Ministerio de Ciencia e Innovación, Spain. The authors thank Nipro Europe for providing the Surefuser Plus intravenous infusion pumps. They also thank Jose Manuel Asencio, David Calatayud, Nuria Mestres, Eliano Riani and Germán Sobrino for their participation in and assistance with various aspects of the experimental protocol.

Disclosure

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

Ancillary