Article Text

Original article
KLF2 exerts antifibrotic and vasoprotective effects in cirrhotic rat livers: behind the molecular mechanisms of statins
  1. Giusi Marrone1,
  2. Raquel Maeso-Díaz1,
  3. Guillermo García-Cardena2,
  4. Juan G Abraldes1,
  5. Juan Carlos García-Pagán1,
  6. Jaime Bosch1,
  7. Jordi Gracia-Sancho1
  1. 1Barcelona Hepatic Hemodynamic Laboratory, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Hospital Clínic de Barcelona, Centro de Investigaciones Biomédicas en Red en Enfermedades Hepáticas y Digestivas (CIBEREHD), Barcelona, Spain
  2. 2Departments of Pathology, Brigham and Women's Hospital & Harvard Medical School, Boston, Massachusetts, USA
  1. Correspondence to Dr Jordi Gracia-Sancho, Barcelona Hepatic Hemodynamic Laboratory, Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Rosselló 149, 4th floor, 08036, Barcelona, Spain; jgracia{at}


Objective In the liver, the transcription factor, Kruppel-like factor 2 (KLF2), is induced early during progression of cirrhosis to lessen the development of vascular dysfunction; nevertheless, its endogenous expression results insufficient to attenuate establishment of portal hypertension and aggravation of cirrhosis. Herein, we aimed to explore the effects and the underlying mechanisms of hepatic KLF2 overexpression in in vitro and in vivo models of liver cirrhosis.

Design Activation phenotype was evaluated in human and rat cirrhotic hepatic stellate cells (HSC) treated with the pharmacological inductor of KLF2 simvastatin, with adenovirus codifying for this transcription factor (Ad-KLF2), or vehicle, in presence/absence of inhibitors of KLF2. Possible paracrine interactions between parenchymal and non-parenchymal cells overexpressing KLF2 were studied. Effects of in vivo hepatic KLF2 overexpression on liver fibrosis and systemic and hepatic haemodynamics were assessed in cirrhotic rats.

Results KLF2 upregulation profoundly ameliorated HSC phenotype (reduced α-smooth muscle actin, procollagen I and oxidative stress) partly via the activation of the nuclear factor (NF)-E2-related factor 2 (Nrf2). Coculture experiments showed that improvement in HSC phenotype paracrinally ameliorated liver sinusoidal endothelial cells probably through a vascular endothelial growth factor-mediated mechanism. No paracrine interactions between hepatocytes and HSC were observed. Cirrhotic rats treated with simvastatin or Ad-KLF2 showed hepatic upregulation in the KLF2-Nrf2 pathway, deactivation of HSC and prominent reduction in liver fibrosis. Hepatic KLF2 overexpression was associated with lower portal pressure (–15%) due to both attenuations in the increased portal blood flow and hepatic vascular resistance, together with a significant improvement in hepatic endothelial dysfunction.

Conclusions Exogenous hepatic KLF2 upregulation improves liver fibrosis, endothelial dysfunction and portal hypertension in cirrhosis.


Statistics from

Significance of this study

What is already known on this subject?

  • The transcription factor Kruppel-like factor 2 (KLF2) confers protection to endothelial cells through the induction of vasoprotective genes.

  • In cirrhosis, KLF2 is overexpressed early during the progression of the disease; nevertheless, it is not enough to slow down the development of vascular dysfunction.

What are the new findings?

  • Pharmacological or adenoviral upregulation of hepatic KLF2 expression induces a profound amelioration in portal hypertension and cirrhosis.

  • KLF2 beneficial effects are mainly due to hepatic stellate cells inactivation and apoptosis, together with reduction in hepatic oxidative stress and improvement in endothelial function.

How might it impact on clinical practice in the foreseeable future?

  • Specific induction of hepatic KLF2 within the liver represents an easy and highly effective strategy to promote regression of liver cirrhosis and amelioration of portal hypertension.


Maintaining normal liver function requires a perfect sinusoidal environment in which hepatic cells, regularly communicating with each other, control liver metabolism, homeostasis and intrahepatic vascular tone. Modifications in liver circulation and deregulations in the phenotype of hepatic cells lead to parenchymal and non-parenchymal dysfunction, characteristic of chronic liver diseases.

In cirrhosis, increased intrahepatic resistance to portal blood flow (PBF) is, in part, due to structural abnormalities and, in part, due to alterations in the liver sinusoidal milieu, primarily contributing to development of portal hypertension.1 Among the non-parenchymal cell alterations, liver sinusoidal endothelial cells (LSEC) become dysfunctional and lose their unique characteristics to acquire a vasoconstrictor, prothrombotic and proinflammatory phenotypes,2 while hepatic stellate cells (HSC), which modulate liver contractility and extracellular matrix deposition, transdifferentiate and proliferate, becoming activated and phenotypically different from the normal quiescent HSC.3 ,4 Therefore, an efficient modulation of the injured hepatic microvascular phenotype, especially the activated HSC and the dysfunctional LSEC, may lead to a significant amelioration of liver cirrhosis. In fact, most of the reported antifibrotic treatments were aimed at inhibiting HSC transdifferentiation and proliferation, or at regulating HSC apoptosis.5–7 On the other hand, therapeutic approaches to improve LSEC phenotype and liver circulation in cirrhosis focused on the benefit of statins8 ,9 and antioxidant therapies.10

Kruppel-like transcription factors (KLF) are zinc finger proteins that act as activators or repressors of the expression of genes involved in cell proliferation and differentiation.11 We recently identified KLF2 as a key component of the hepatic endothelium, where it may act as a defence mechanism in response to damage that occurs during the progression of the disease.12 Simvastatin, through the activation of the endothelial KLF2-nitric oxide pathway, is the most effective statin protecting the hepatic endothelium in cirrhosis.13 Moreover, previous reports suggested that statins may improve HSC phenotype,6 ,14 ,15 however, the possible direct role of KLF2 improving HSC function and liver fibrosis has never been investigated.

Considering this background, we aimed at studying the effects and the underlying molecular mechanisms of hepatic KLF2 induction using pharmacological and adenoviral approaches in in vitro and in vivo models of liver cirrhosis.

Materials and methods

A complete description of materials and methods can be found in the online supplementary material.

Animal models of liver cirrhosis—KLF2 overexpression

Induction of cirrhosis by carbon tetrachloride

Male Wistar rats (50–75 g) underwent inhalation exposure to carbon tetrachloride, and received phenobarbital (0.3 g/L) in the drinking water. After 10 weeks of liver injury, administration of toxicants was stopped and experiments were performed 1 week later. Hepatic KLF2 overexpression was achieved by administrating a unique dose of 1011 adenoviral particles codifying murine KLF2 (Ad-KLF2)16 or control (Ad-CTRL) through the penile vein. Three days after intravenous injection of the adenoviral constructs, cells isolation, haemodynamic studies and molecular analysis were performed. Adenoviruses were prepared by a third person and kept under a code, therefore, the investigators performing the experiments were not aware of the treatment received by the rats. Animals were kept in environmentally controlled animal facilities. All procedures were approved by the Laboratory Animal Care and Use Committee of the University of Barcelona and were conducted in accordance with the European Community guidelines for the protection of animals used for experimental and other scientific purposes (EEC Directive 86/609).

Induction of cirrhosis by common bile duct ligation

Secondary biliary cirrhosis with intrahepatic portal hypertension was induced in male Sprague–Dawley rats (250–275 g) by common bile duct ligation (CBDL). After 4 weeks of CBDL, animals received either activated simvastatin (25 mg/kg/day orally) or its vehicle (n=6 per group), for 3 days. Liver histology and molecular analysis were performed as described below.

HSC isolation, culture and treatments

Isolation and culture of HSC

rHSC were isolated from cirrhotic rats.13 Results using rHSC were derived from at least three independent isolations and three replicates.

Immortalised human-activated HSC LX-2, kindly provided by Dr Bataller, were cultured as described.13 Results were derived from at least three replicates per group.

Simvastatin treatment

LX-2 cells were treated with activated simvastatin (Calbiochem) or its vehicle, 0.05% dimethyl sulfoxide, at different concentrations (0.1 μM, 1 μM and 10 μM) for 24 h and 72 h. Additionally, time-course experiments (8 h, 24 h and 72 h) were performed in primary cultured HSC incubated with 10 μM simvastatin.

KLF2 inhibition

HSC-KLF2 knock-down was performed using siRNA and pharmacological approaches as described.13 ,17

KLF2 upregulation using adenoviral vectors

In vitro: LX-2 were infected with Ad-KLF2 or Ad-CTRL at 10 multiplicity of infection (MOI) for 2 h, rinsed with phosphate buffered saline and then incubated for an additional 22 h or 70 h. Infection was estimated visually, evaluating adenoviral-encoded green fluorescent protein (GFP) expression by fluorescence microscopy, resulting in ≈90% positive cells.

In vivo: KLF2 overexpressing HSCs were isolated from cirrhotic rats previously infected with Ad-KLF2 or Ad-CTRL.

Intracellular superoxide (O2) content

O2 levels in HSC were assessed with the oxidative fluorescent dye dihydroethidium as described.18

HSC apoptosis and necrosis

HSC were stained with acridine orange and propidium iodide to detect apoptosis and necrosis, respectively, as described in the online supplementary methods.

LSEC isolation and coculture with HSC

LSEC were isolated and cultured as described.19 After 3 days, dedifferentiated LSEC were cocultured for 24 h with LX-2 overexpressing KLF2 due to adenoviral or simvastatin preadministration.13 Vascular endothelial growth factor (VEGF) was analysed in culture media supernatants using a commercially available EIA kit (R&D systems) following the manufacturer's instructions.

HSC and LSEC coculture in a sinusoidal-like environment

Activated human HSC and dedifferentiated LSEC were cocultured using a cell culture chamber with microfluidics that mimics the liver sinusoidal architecture recently developed and validated by our group.20 Briefly, the chamber contains up to three cell culture compartments with porous membranes separating each of them, thus allowing paracrine interactions. In the top compartment, LSEC are cultured under continuous and homogeneous laminar shear stress (5 dyn/cm2) generated by a microfluidic system. In a bottom compartment, HSC are statically cultured. Cells are maintained in 5% CO2 and at 37°C during all the experiments.

Hepatocytes isolation and coculture with HSC

Hepatocytes were isolated from cirrhotic rat livers by collagenase digestion, cultured on collagen-coated plates and infected with Ad-KLF2 or Ad-CTRL (10 MOI) for 2 h. After 22 h, the cells were washed and cocultured with LX-2 for an additional 24 h. In a different set of experiments, hepatocytes received 10 μM simvastatin for 24 h, and afterwards were cocultured for 24 h with LX-2.

Characterisation of cirrhotic animals with hepatic KLF2 overexpression

Hepatic adenoviral localisation

Adenoviral infection was localised in liver tissue and hepatic cells from cirrhotic animals that received Ad-KLF2 taking advantage of its GFP-derived fluorescence.

Hepatic fibrosis and phenotype of HSC and KC

Liver fibrosis was assessed in cirrhotic rats infected with Ad-KLF2 or Ad-CTRL (n=10 per group), and in cirrhotic rats receiving simvastatin or vehicle (n=6 per group) using Sirius Red staining and computerised analysis.21 HSC and Kupffer cell (KC) phenotypes were analysed through the analysis of α-smooth muscle actin (α-SMA) and desmin, and CD68 and CD163, respectively, in paraffin-embedded liver sections as previously described.22 Matrix metalloproteinase (MMP)-2 and MMP-9 activity was measured using zymography.23

In vivo haemodynamic studies

Mean arterial pressure (MAP), portal pressure (PP), PBF and superior mesenteric artery blood flow (SMABF) were measured in Ad-KLF2 and Ad-CTRL-treated cirrhotic rats using microcatheters and flow probes24 (n=10 per group).

Evaluation of hepatic endothelial function

After in vivo haemodynamic measurements, livers were quickly isolated and perfused. Liver endothelial function was determined as response to incremental doses of the endothelium-dependent vasodilator acetylcholine.25

Statistical analysis

Statistical analysis was performed with the SPSS V.19.0 for Windows statistical package (IBM, Armonk, New York, USA). All results are expressed as mean±SE of the mean. Comparisons between groups were performed with the Student t test or analysis of variance, followed by a posthoc test when adequate. Differences were considered significant at a p value <0.05.


Simvastatin induces KLF2 expression in activated HSC and concomitantly improves their phenotype

Human-activated HSC treated for 24 h with increasing concentrations of simvastatin exhibited an induction in the expression of the transcription factor, KLF2, at a concentration of 10 μM (figure 1A), which was associated with a marked downregulation in the activation marker α-SMA both at mRNA and protein levels (figure 1A). Simvastatin effects on KLF2 and α-SMA were maintained, or even increased, after 3 days of treatment (see online supplementary figure S1A). Primary cirrhotic rHSC incubated with simvastatin during 24 h also showed a significant upregulation of KLF2 mRNA expression (figure 1B) and reduction in α-SMA and procollagen I levels (figure 1B). Time-course experiments showed a marked and time-dependent α-SMA and procollagen I downregulation up to 3 days of treatment (see online supplementary figure S1B).

Figure 1

Simvastatin ameliorates hepatic stellate cells (HSC) phenotype. Expression of Kruppel-like factor 2 (KLF2) and HSC activation markers (α-smooth muscle actin (α-SMA) and procollagen I) in response to 24 h of simvastatin in LX-2 (A) and primary cirrhotic rat HSC (B). Left y-axis indicates KLF2-fold induction. Right y-axis indicates mRNA changes of HSC activation markers. Inserts show representative western blots of the corresponding protein. Percentage under the inserts refers to protein increase (↑) or decrease (↓) due to simvastatin treatment, normalised to glyceraldehyde 3-phosphate dehydrogenase (GAPDH), and compared with vehicle cells; n=3 per condition; *p<0.01 versus vehicle cells.

Amelioration of HSC activation phenotype due to simvastatin is mediated by KLF2

To investigate whether the α-SMA and procollagen I reduction observed in simvastatin-treated HSC was dependent on KLF2 expression, we performed KLF2 knock-down experiments through two approaches. First, siRNA against KLF2 was used. Surprisingly, these experiments showed that abrogation of simvastatin-derived KLF2 upregulation due to siRNA treatment was not associated with an inhibition in simvastatin-derived improvement of HSC phenotype (figure 2A, B). However, a concomitant induction in KLF4 and KLF6 expression was observed in simvastatin-treated cirrhotic siKLF2 HSC, in comparison with simvastatin-treated siCTRL HSC (25% and 68%, respectively). KLF4 and KLF6 induction upon KLF2 silencing was corroborated using a different siRNA construct (60% increase in KLF4; 89% in KLF6).

Figure 2

Effects of Kruppel-like factor 2 (KLF2) modulation on hepatic stellate cells (HSC) phenotype. Effects of KLF2 silencing using siKLF2 (LX-2, A; HSC, B), mevalonate (LX-2, C) or geranylgeranyl-pyrophosphate (LX-2, D) on HSC phenotype in response to simvastatin. Left y-axis indicates KLF2-fold induction. Right y-axis indicates mRNA changes of the activation markers; n=3 per condition, *p<0.01 versus vehicle cells, #p<0.01 versus simvastatin. (E) KLF2 and α-smooth muscle actin (α-SMA) expression in LX-2 after 24 h or 72 h of incubation with adenovirus codifying for KLF2 (k) or control adenovirus (Ad-CTRL) (c). Inserts show a representative western blot with its quantification normalised to glyceraldehyde 3-phosphate dehydrogenase (GAPDH); n=3 per condition; *p<0.01 versus Ad-CTRL. (F) KLF2, α-SMA and procollagen I expression in HSC isolated from cirrhotic rats previously infected with Ad-KLF2 or Ad-CTRL; n=3 per condition; *p<0.01 versus Ad-CTRL.

As a second approach, LX-2 incubated with simvastatin or vehicle were cotreated with two inhibitors of KLF2 expression, namely mevalonate and geranylgeranyl-pyrophosphate (GGPP). As demonstrated in figure 2C, D (LX-2), and online supplementary figure S2 (rHSC), isoprenoids attenuated the KLF2 induction mediated by simvastatin, which was associated with limited HSC deactivation in response to the drug.

To further understand the role of KLF2 on HSC improvement, we decided to analyse HSC phenotype in cells selectively overexpressing KLF2 due to adenoviral administration. LX-2 infected with Ad-KLF2 exhibited markedly increased expression of KLF2 that came along with a time-dependent decrease in the expression of α-SMA, in comparison with Ad-CTRL cells (figure 2E). Additionally, primary HSC isolated from cirrhotic rats previously infected with Ad-KLF2 displayed a significant HSC activation reversal, compared with cells from Ad-CTRL animals (figure 2F).

KLF2 upregulation decreases intrahepatic fibrosis in cirrhotic rats both by HSC deactivation and apoptosis

The potential effects of KLF2 upregulation as a new antifibrotic strategy were analysed in cirrhotic animals. After 3 days of adenoviral administration, exogenous Ad-KLF2-GFP was mainly localised in fibrotic septa and sinusoids, with marginal detection within the parenchyma (see online supplementary figure S3A, B), resulting in a significant increase in hepatic KLF2 expression in comparison with Ad-CTRL rats (see online supplementary figure S3C, D). The increment in KLF2 levels was associated with a 41% reduction in intrahepatic fibrosis, as proved by Sirius Red staining (figure 3A), as well as a decrease in procollagen I and α-SMA expression (figure 3A, C). Moreover, a profound decline in desmin protein expression (figure 3A) was observed in Ad-KLF2 cirrhotic animals, suggesting that decrease in HSC activation markers may be mostly due to both apoptosis of activated HSC and due to their deactivation. Indeed, we observed augmentation of the apoptosis marker-cleaved caspase 3 and a reduction of phosphorylated-Bad, and a significant decrease in HSC activation measured as Rho kinase activity (figure 3D). Analysis of hepatic MMPs activity revealed a slight diminution in Ad-KLF2 animals, thus suggesting that the maximal peak of fibrinolysis had already occurred (figure 3E). In vitro studies confirmed that KLF2 upregulation using simvastatin, or Ad-KLF2, diminishes HSC proliferation (43.7% and 17.1% reduction after 24 h of simvastatin, or Ad-KLF2, vs their vehicles), and promotes their death (28% and 9.1% increase in non-viable cells after treatment with simvastatin or Ad-KLF2), most probably through a KLF2-dependent apoptotic mechanism (see online supplementary figure S4).

Figure 3

KLF2 upregulation decreases intrahepatic fibrosis in cirrhotic rats. (A) Top, representative images of liver fibrosis assessment using Sirius Red staining (×10), and α-SMA and desmin immunohistochemistry (×20) in Ad-KLF2 (k) and Ad-CTRL(c) cirrhotic animals. Bottom, corresponding quantifications (n=10 per group). (B) Representative western blot and densitometric quantification of α-SMA determined in livers from rats described in A. (C) Hepatic procollagen I mRNA expression in rats described in A. (D) Representative western blots and densitometric quantifications of depicted proteins determined in livers from rats described in A; *p<0.05 versus Ad-CTRL. (E) Representative zymography and quantifications of active MMP-2 and MMP-9 in livers described in A. (F) Representative images of Sirius Red staining, and α-SMA and desmin immunohistochemistry in livers from cirrhotic rats treated with simvastatin or its vehicle for 3 days, and corresponding quantifications; n=6 per group; *p<0.05 versus vehicle. α-SMA, α-smooth muscle actin; Ad-CTRL, adenovirus control; Ad-KLF2, adenovirus codifying for KLF2; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; KLF2, Kruppel-like factor 2; MMP, matrix metalloproteinase; p-bad, phosphorylated-Bad.

Additionally, the antifibrotic effects of KLF2 were evaluated in a group of cirrhotic animals where hepatic KLF2 expression was induced using a pharmacological approach. Cirrhotic rats treated during 3 days with simvastatin exhibited upregulation in KLF2 expression (fourfold higher vs vehicle rats), diminution in fibrosis and HSC deactivation (figure 3F).

KLF2 upregulation decreases HSC oxidative stress, probably via NF-E2-related factor 2 mediated pathway

In vitro studies showed that the improvement in HSC activation phenotype due to KLF2 upregulation was accompanied by a significant increment in the expression of the antioxidant and detoxifying genes heme oxygenase-1 (HO-1) and NAD(P)H dehydrogenase quinone 1 (NQO1) (see figure 4A and online supplementary figure S5A), overall reducing the intracellular levels of O2, through a KLF2-dependent mechanism (figure 4B). Similarly, induction of hepatic KLF2 in cirrhotic animals due to simvastatin or adenoviral administration was associated with increased hepatic expression of HO-1 and NQO1, which may be derived from increased nuclear levels of the KLF2-derived antioxidant transcription factor NF-E2-related factor 2 (Nrf2) (see figure 4C and online supplementary figure S5B).

Figure 4

Kruppel-like factor 2 (KLF2) reduces hepatic stellate cells (HSC) oxidative stress through NF-E2 related factor 2 (Nrf2). (A) Relative expression of Nrf2 and its target genes, heme oxygenase-1 (HO-1) and NAD(P)H dehydrogenase quinone 1 (NQO1), in simvastatin-treated LX-2 cells in the presence of the KLF2 inhibitor, geranylgeranyl-pyrophosphate (GGPP). (B) Top, representative fluorescence images of superoxide (O2) in LX-2 and rHSC incubated as depicted. Bottom, quantitative analysis of the fluorescent intensity normalised to the number of cells. Data obtained in three independent experiments; *p<0.05 versus vehicle; #p<0.05 versus simvastatin. (C) Expression of Nrf2 (left) and its target genes HO-1 and NQO1 (right) in livers from cirrhotic rats receiving Ad-KLF2 (k) or adenovirus control (Ad-CTRL) (c) *p<0.05 versus Ad-CTRL.

KLF2 over-expressing HSC paracrinally improve LSEC phenotype

Primary cultured dedifferentiated LSEC exhibited a dysfunctional phenotype defined as diminished expression of endothelial nitric oxide synthase (eNOS) and upregulated endothelin-1, in comparison with healthy LSEC (see online supplementary figure S6).26 Dedifferentiated LSEC cocultured with LX-2 overexpressing KLF2 (due to previous treatment with simvastatin or Ad-KLF2) significantly improved their phenotype, as demonstrated by an increment in eNOS and reduction in endothelin-1 expression, in comparison with LSEC cocultured with vehicle-LX-2 (figure 5A, B). This amelioration in LSEC phenotype was probably due, at least in part, to an increment in VEGF release from KLF2 overexpressing LX-2 (3.2-fold induction in simvastatin pretreated cells and twofold induction in Ad-KLF2 cells; p<0.05).

Figure 5

Hepatic stellate cells and liver sinusoidal endothelial cells (LSEC) paracrinally improve their phenotype. Relative expression of endothelial nitric oxide synthase (eNOS) (A) and endothelin-1 (B) in dedifferentiated LSEC cocultured with LX-2 cells previously treated with simvastatin, adenovirus codifying for Kruppel-like factor 2 (Ad-KLF2), or vehicle; n=3 per condition; *p<0.05 versus vehicle. (C) Expression of KLF2 in dedifferentiated LSEC cocultured with LX-2 under physiological conditions (shear stress), or static, and treated with simvastatin or vehicle. (D) Relative α-smooth muscle actin (α-SMA) expression in LX-2 described in C; n=3 per condition; *p<0.05 versus vehicle-static. #p<0.05 versus simvastatin-static; †p<0.05 versus vehicle-shear stress.

Simvastatin enhances the KLF2-derived protective phenotype in HSC and LSEC cultured in an in vitro sinusoidal-like environment

Primary cultured dedifferentiated LSEC were cocultured with activated HSC in a three-dimensional cell culture chamber with microfluidics that mimics the sinusoidal architecture. Cells were simultaneously treated with simvastatin and LSEC exposed to physiological shear stress stimuli. After 24 h of coculture and biomechanical stimulation, LSEC exhibited a significant improvement in their phenotype defined by upregulation in KLF2 (figure 5C) and HSC showed a further downregulation of α-SMA (figure 5D), in comparison with simvastatin-treated sinusoidal cells cultured under static conditions, or cells cultured within the sinusoidal microchamber but treated with vehicle.

Hepatocyte KLF2 overexpression does not influence HSC phenotype

KLF2 overexpressing hepatocytes, by Ad-KLF2 transfection or simvastatin pretreatment, did not paracrinally modify the expression of the activation marker α-SMA in HSC (see online supplementary figure S7).

Hepatic KLF2 overexpression improves portal hypertension

When compared with control rats, cirrhotic rats transfected with Ad-CTRL exhibited arterial hypotension and portal hypertension derived from both increased PBF and intrahepatic vascular resistance (IHVR) (data not shown). Cirrhotic rats overexpressing hepatic KLF2 due to adenoviral administration showed a significant improvement in PP (−15%), due to both an improvement in PBF and IHVR, in comparison with cirrhotic Ad-CTRL rats (figure 6). No differences in SMABF (2.1±0.3 vs 2.2±0.2 mL/min×100 g bw; p=0.7) or MAP were observed when comparing groups.

Figure 6

Hepatic Kruppel-like factor 2 (KLF2) upregulation improves portal hypertension. (A) Mean arterial pressure (MAP), (B) portal pressure (PP), (C) portal blood flow (PBF) and (D) intrahepatic vascular resistance (IHVR) determined in cirrhotic rats overexpressing hepatic KLF2 due to adenovirus administration (Ad-KLF2) compared with cirrhotic rats receiving control adenovirus (Ad-CTRL); n=10 per group; *p<0.05 versus Ad-CTRL.

Hepatic KLF2 overexpression improves liver endothelial dysfunction of cirrhotic rats

To further characterise the effects of KLF2 upregulation on liver vasculature, livers from cirrhotic rats treated with Ad-KLF2 or its control, were isolated and perfused. Basal ex vivo IHVR was significantly lower in cirrhotic rats receiving Ad-KLF2 than Ad-CTRL animals (1.05±0.09 vs 1.37±0.08 mm Hg×min/mL×g; p=0.02). Endothelial function evaluation revealed that livers from cirrhotic animals treated with Ad-CTRL showed endothelial dysfunction, defined as a deficient response to the vasodilator acetylcholine in comparison with control rats, which was not observed in cirrhotic animals treated with Ad-KLF2 (figure 7A). Endothelial function amelioration was accompanied by an increase in the expression of eNOS and its phosphorylated form (figure 7B).

Figure 7

KLF2 upregulation restores hepatic endothelial function in cirrhotic rats. (A) Endothelial function evaluation in livers from control rats (CT), cirrhotic rats infected with Ad-KLF2 (k) or Ad-CTRL (c). (B) Representative western blots and densitometric quantifications of hepatic eNOS (top) and p-eNOS (bottom) in cirrhotic rats described in A; n=10 per group; *p<0.05 versus Ad-CTRL. Ad-CTRL, adenovirus control; Ad-KLF2, adenovirus codifying for KLF2; eNOS, endothelial nitric oxide synthase; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; KLF2, Kruppel-like factor 2.


Cirrhosis is the principal cause of intrahepatic portal hypertension, a deleterious syndrome derived from increments in IHVR and PBF.1 Pharmacological treatments currently available for portal hypertension, non-selective β blockers, cause splanchnic vasoconstriction reducing PBF, but they are not liver-specific, and less than half the patients under β blockade are protected from the portal hypertension-derived complications.27 Strategies that may improve portal hypertension through a reduction in IHVR are very much necessary.

We recently demonstrated that the transcription factor KLF2 is highly expressed in the liver of cirrhotic rats, particularly in the hepatic endothelium.12 However, this endogenous protective mechanism is insufficient at preventing sinusoidal endothelial dysfunction and HSC activation in cirrhotic animals. Thus, considering that KLF2 confers endothelial protection against inflammation, thrombosis and vasoconstriction, and it is a regulator of haemodynamics, we hypothesised that increasing the expression of this transcription factor could be beneficial in cirrhosis.

In the present study, we demonstrate for the first time that hepatic KLF2 overexpression in cirrhotic animals deactivates HSC and ameliorates the dysfunctional hepatic endothelium, which leads to a significant improvement in liver cirrhosis and portal hypertension. KLF2 expression was exogenously induced through two approaches: using simvastatin, the generic statin with major vasoprotective effects in the liver13 and specifically by the administration of an adenoviral construct codifying for KLF2.

Herein, we demonstrate that simvastatin induces KLF2 expression both in human and rat-activated HSC, which is accompanied by a time-dependent decrease in α-SMA and procollagen I expression. The role of KLF2 in HSC deactivation was first studied using siRNA-KLF2, however, this methodology failed in demonstrating whether simvastatin effects on HSC were solely KLF2-dependent, most likely because of a compensatory increment in the expression of other members of the KLF family with antifibrogenic properties, specifically KLF4 and KLF6.28–30 Unfortunately, the compensatory mechanisms of other KLFs could not be herein confirmed, since simultaneous silencing of KLF2, KLF4 and KLF6 resulted as being lethal to HSC (over 90% cell death). As a second approach, pharmacological inhibition of the KLF2 pathway using mevalonate and GGPP resulted in a significant blockade of the simvastatin-derived HSC amelioration (see online supplementary figure S2). Considering that these inhibitors may affect molecular pathways different from KLF2, we do not discard that other mechanisms may contribute to improve HSC phenotype in response to statins.31 ,32

To definitely confirm the importance of KLF2-modulating HSC phenotype, we decided to study the effects of overexpressing KLF2 in activated HSC through adenoviral technology. Specific KLF2 upregulation reversed HSC activation, both in vitro and in vivo. This amelioration could be due to the fact that KLF2 suppresses transforming growth factor (TGF)-β signaling,33 in fact, KLF2 overexpressing HSC showed reduced TGF-β mRNA (data not shown).

To validate the KLF2-mediated antifibrotic effects in the cirrhotic liver, hepatic KLF2 expression was augmented in cirrhotic animals via adenoviral administration. KLF2 upregulation led to a significant regression in fibrosis, probably derived from both HSC phenotype improvement and apoptosis. The definition of KLF2 as a proapoptotic factor in certain situations is of interest. Indeed, our findings reveal that KLF2 per se induces HSC apoptosis (in vivo and in vitro) and that simvastatin-derived HSC apoptosis would be due to KLF2 upregulation (in vitro). Considering our results and previously published data,17 ,34 ,35 it can be proposed that KLF2 may have a dual role in terms of apoptosis, depending on the cell type and the basal phenotype of such cell. In the particular case of activated HSC, apoptosis might be mediated by a KLF2-derived down-regulation in nuclear factor (NF)-kB (data not shown), which tightly regulates the expression of different antiapoptotic genes.36 Additionally, our findings suggest that the proapoptosis and antifibrotic properties of other KLF2 inducers, such as resveratrol22 and curcumin,37 may also be mediated by KLF2.

In parallel experiments, hepatic KLF2 upregulation in simvastatin-treated cirrhotic rats was associated with a decline in liver fibrosis (reduced α-SMA and collagen amount), but no differences in desmin were observed, suggesting that a 3-day pharmacological treatment would mainly modulate HSC phenotype. Interestingly, and although not significant, hepatic Rho kinase activity was also diminished in simvastatin-treated cirrhotic animals (data not shown), a result that has been previously pointed out using atorvastatin.31 We, therefore, hypothesise a possible link between KLF2, Rho kinase and fibrosis resolution.

To explore if the antifibrotic effects of in vivo KLF2 overexpression might be partly mediated by infected hepatocytes, and although their in vivo infection rate was relatively low, we performed coculture experiments with KLF2-overexpressing hepatocytes and activated HSC. These studies showed no direct influence of parenchymal cells over HSC, confirming lack of paracrine effects of hepatocytes over sinusoidal cells.38 Nevertheless, we cannot discard that possible hepatocyte phenotype amelioration due to KLF2 upregulation in vivo might contribute to reduce the hepatic damage, and indirectly deactivate HSC. Moreover, improvement in KC phenotype in KLF2 overexpressing livers may further contribute to liver fibrosis resolution (see online supplementary figure S8).39 Taken together, our data define hepatic KLF2 upregulation as an efficient mechanism to reduce fibrosis by improvement in HSC phenotype and promoting their apoptosis.

One of the mechanisms by which simvastatin may directly improve HSC phenotype might be related to an amelioration in intracellular oxidative stress. Indeed, simvastatin triggers nuclear translocation of the antioxidant transcription factor Nrf2,40 which has been shown to play a critical role attenuating liver fibrosis.41 Under basal conditions, Nrf2 is retained in the cytoplasm bound to Keap1 that promotes its proteasomal degradation. However, upon stimulation, Nrf2 is released and translocates to the nucleus where it binds to the antioxidant-responsive elements of cytoprotective genes such as NQO1 and HO-1, promoting their transcription. Herein, we demonstrate that HSC activation reversal due to KLF2 upregulation was accompanied by increments in the expression of the Nrf2-targets HO-1 and NQO1, altogether promoting a marked decline in intracellular oxidative stress. It has been reported that KLF2 enhances the antioxidant activity of Nrf2 by increasing its nuclear localisation and activity42 ,43 and that a specific Nrf2-activating stimulus, apart from KLF2, is required for full transcriptome effects.43 This could explain the increased hepatic nuclear accumulation of Nrf2, with a concomitant activation of its pathway, observed in Ad-KLF2 and simvastatin-treated cirrhotic animals in which the activating stimulus may probably be the elevated oxidative stress characteristic of cirrhosis.44 Taken together, these results suggest that KLF2 improves HSC phenotype by reduction in intracellular oxidative stress through Nrf2 activation.

We have recently demonstrated that the preservation of a normal LSEC phenotype, via KLF2 upregulation, maintains HSC in a quiescence status, or even promotes their deactivation.13 However, it is unknown whether therapeutic strategies that deactivate HSC may impact on the phenotype of dedifferentiated LSEC. Herein, we observe that HSC deactivation due to KLF2 overexpression paracrinally improves LSEC. Indeed, dedifferentiated LSEC recover eNOS and lose endothelin-1 mRNA expression (genes profoundly deregulated during in vitro capillarisation26) when cocultured with HSC pretreated with simvastatin or Ad-KLF2. We suggest that this improvement in LSEC phenotype would partly be due to HSC-derived VEGF that may act via a microvascular internal loop mechanism. VEGF might bind to its receptor in LSEC determining amelioration in their phenotype, but it also could induce an attenuation of the contractile properties of HSC via upregulation of VEGFR-1.45 Although these data seem to be of great value, more investigations are required.

To better understand the paracrine interactions between LSEC and HSC in response to simvastatin, we in vitro reproduced the liver sinusoid and analysed the effects of simvastatin on both cell types. Although a limitation of the experiment may be the use of cells from different animal sources, it certainly showed a global improvement in the sinusoidal microenvironment as the consequence of cells’ paracrine communications, thus validating the hypothesis that administration of simvastatin to individuals with sinusoidal microvascular injuries (ie, cirrhosis or ischaemia/reperfusion) exerts strong liver protection.46 ,47

Finally, the effects of KLF2 overexpression on hepatic and systemic haemodynamics were determined. Improvement in liver sinusoidal cells due to KLF2 induction promoted a significant reduction in PP, consequence of both reductions in PBF and IHVR, without changes in systemic haemodynamic parameters. Additionally, KLF2 overexpressing cirrhotic rats exhibited restored liver endothelial function. Thus, the mechanisms by which KLF2 upregulation improves portal hypertension go back to the recovery of sinusoidal function and the restoration of paracrine signalling. Although we did not appreciate the re-establishment of fenestra in these improved LSEC (data not shown) or differences in analytical biochemistry data (see online supplementary table S1), we do not discard that a more prolonged treatment could lead to even better results. Nevertheless, special attention should be paid when considering the use of KLF2-encoding adenoviruses, since liver steatosis may develop.48

In conclusion, this study provides evidence that increasing KLF2 in cirrhotic animals leads to an improvement in liver sinusoidal cells phenotype, deactivating HSC, ameliorating the dysfunctional endothelium and reducing oxidative stress, that turns into regression of cirrhosis and amelioration of portal hypertension. The use of simvastatin or other drugs capable of augmenting KLF2 expression might be an appealing proposition to treat portal hypertension in cirrhosis.


The authors are indebted to Sergi Vila, Héctor García and Montse Monclús for excellent technical assistance, and to Sergi Guixé, Diana Hide and Marina Vilaseca for their experimental expertise. This work was partly carried out at the Centro Esther Koplowitz, Barcelona, Spain.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • GM and RM-D share first authorship.

    JB and JG-S share senior authorship.

  • Contributors GM designed the research, conceived ideas, performed experiments, analysed data and wrote the manuscript. RM-D conceived ideas, performed experiments and analysed data. GG-C and JG-A provided essential materials. JCG-P critically revised the manuscript. JB conceived ideas, critically revised the manuscript, obtained funding and co-directed the study. JG-S designed the research, conceived ideas, wrote the manuscript, obtained funding and co-directed the study. All authors edited and reviewed the final manuscript.

  • Funding This study was funded by grants from the Instituto de Salud Carlos III, Ministerio de Economía y Competitividad (FIS PI11/00235 to JG-S and PS09/01261 and PI13/00341 to JB) and cofunded by FEDER (Fondo Europeo de Desarrollo Regional, Unión Europea. ‘Una manera de hacer Europa’). JG-S has a Ramón y Cajal contract from the Ministerio de Economía y Competitividad, and received funding from the European Community's Seventh Framework Programme (EC FP7/2007–2013, grant agreement 229673), GM has a fellowship from the Instituto de Salud Carlos III (PFIS09/00540). Ciberehd is funded by Instituto de Salud Carlos III.

  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles