Microcirculatory disturbances and cellular changes during progression of hepatic steatosis to liver tumors

Murine steatosis-inflammation-tumor model

The steatosis-inflammation-tumor model was performed as described by Fujii et al.¹¹ The model comprises two independent hits that lead to NAFLD onset and progression. The first hit is an elevated blood glucose level and the second hit comprises a HFD. Hyperglycemia was induced in male C57BL/6 mice (Charles River, Sulzfeld, Germany) through a single intraperitoneal injection of 200 µg STZ (Sigma-Aldrich, St. Louis, Missouri, USA) in 10 µl of 0.05 M trisodium citrate (pH 4.5; Merck, Darmstadt, Germany) at day two postnatal. Starting at day 28 of age, the mice were fed a continuous HFD (fat content: 60 kJ%; D12492(II) modified experimental diet; Ssniff, Soest, Germany). The mice were kept on water and HFD ad libitum at a 12 h light/dark cycle. The general state of health was monitored daily and blood glucose levels and body weight were measured weekly. Additionally, EDTA plasma was collected fortnightly. Animals exhibiting normal blood glucose levels were excluded from the experiment (19 of 151). The ages of 6, 8, 12, and 20 weeks served as final observation time points. For each time point, a set of mice was used for intravital microscopic analysis (n=8–11 per time point). Additionally, mice were sacrificed at the named time points (n=7–9 per time point). EDTA plasma was collected and the liver was excised and weighed. For histological analyses, the left lateral liver lobe was fixed in formalin and half of the medial liver lobe was embedded in Tissue-Tek® (Sakura, Tokyo, Japan), subsequently snap frozen in liquid nitrogen, and stored at −20°C. The remaining liver tissue was snap frozen in liquid nitrogen and stored at −80°C. Healthy C57BL/6 mice without STZ treatment and HFD served as control mice. All animal experiments were approved by “Landesamt für Landwirtschaft, Lebensmittelsicherheit und Fischerei Mecklenburg-Vorpommern” (7221.3-1-039/14) and were executed in accordance with the German legislation and EU-directive 2010/63/EU.

In vivo microscopy and quantitative off-line analysis

At ages of 6, 8, 12, and 20 weeks, respectively, in vivo microscopy of the liver was performed and analyzed as described previously by our group.^23–25 Under ketamine/xylazine anesthesia (90/7 mg/kg body weight ip) mice were laparotomized, the left lateral liver lobe was exteriorized on a plasticine stage and covered with a glass coverslip. In vivo microscopy with ultraviolet epi-illumination was used to examine hepatic stellate cells (HSCs) by vitamin A autofluorescence. Thereafter, sodium fluorescein (2 µmol/kg), rhodamine 6G (1 µmol/kg), and bisbenzimide (10 µmol/kg) were injected via the vena jugularis. Sodium fluorescein fluorescence was assessed using blue light epi-illumination. Through enhancement of tissue contrast vessel diameter, blood velocity, sinusoidal density, and sinusoidal perfusion were analyzed. Direct staining of leukocytes by rhodamine 6G facilitated analysis of venular leukocyte rolling and adherence, and sinusoidal leukocyte stasis using green light epi-illumination. Apoptotic cell death was analyzed by means of ultraviolet epi-illumination and bisbenzimide staining of hepatocellular nuclei. A computer-assisted image analysis system (CapImage, Zeintl, Heidelberg, Germany) was used for quantitative off-line analysis of the video files. Volumetric blood flow VQ (sinusoidal and venular, equation (1)) and shear stress τ (sinusoidal and venular, equation (2)) were calculated from blood velocity (vRBC) and vessel radius (r) as below.

Hematological measurements and plasma analyses

Alanine aminotransferase (ALT) and glutamate dehydrogenase (GLDH) activities in EDTA plasma were determined spectrophotometrically using cobas c 111 analyzer (Roche Diagnostics; Rotkreuz, Switzerland) according to the manufacturer’s instructions. Blood glucose levels were measured using a Contour blood glucose meter (Bayer Vital, Leverkusen, Germany) with one drop of blood from the tail vein. For analysis of plasma triglyceride levels, a triglyceride colorimetric assay kit (Cayman Chemical Company, MI, USA; 10010303) was used according to manufacturer’s instructions. Malondialdehyde (MDA) in plasma was measured using the MDA-586-method according to manufacturer’s instructions (OxisResearchTM, Portland, OR). For assessment of non-fasting plasma insulin concentration, the ultra-sensitive mouse insulin ELISA (Crystal Chem, Zaandam, Netherlands) was conducted according to manufacturer’s instructions.

Histology, immunohistochemistry, and image analysis

Liver tissue was fixed in 4% phosphate-buffered formalin (Grimm med. Logistik, Torgelow, Germany) for two to three days, embedded in paraffin, and cut into 5 µm thick sections. Assessment of all histological stainings was performed in a blinded manner. Hematoxylin and eosin (H&E) staining was used to determine NAFLD activity score (NAS). The total NAS (score 0–8) was calculated as the sum of three different scores, namely steatosis (score 0–3), lobular inflammation (score 0–3), and hepatocellular ballooning (score 0–2).²⁶ For scoring of lobular inflammation, we defined an inflammatory focus as a group of at least five inflammatory cells that were not aligned in a row. All scores were determined by three to four independent observers in a blinded manner.

For additional assessment of inflammatory processes, namely tissue infiltration of granulocytes, sectioned paraffin-embedded liver tissue was stained for CAE (chloracetate esterase) with Naphthol AS-D chloroacetate (Sigma-Aldrich, St. Louis, Missouri, USA) and counterstained with hematoxylin (Merck, Darmstadt, Germany). For quantification, CAE positive cells were counted in 30 consecutive high-power fields (HPFs).

For Oil Red O staining of lipids, the liver tissue was embedded in Tissue-Tek®, snap frozen in liquid nitrogen, and stored at −20°C. The tissue was cut in 8 µm thick sections, air dried, and fixed in paraformaldehyde (ChemCruz, Dallas, TX). The staining was performed using Oil Red O (Sigma-Aldrich, St. Louis, Missouri) and counterstained with hematoxylin. Ten to 20 photographs were taken per sample in a 200× magnification. Quantitative analysis of the red stained area was conducted using Adobe Photoshop CS5 analyzing the percentage of red pixels per image.

Liver cell death was detected by ApopTag® plus peroxidase in situ apoptosis kit (Merck, Darmstadt, Germany) according to the manufacturer’s instructions. For quantitative analysis, positive cells were counted in 14 consecutive HPFs.

To assess fibrosis, immunohistochemical staining of collagen 1α in liver sections was performed as follows: Paraffin sections were incubated at 4°C overnight with a primary antibody against collagen 1α (ab34710; Abcam, Cambridge, UK) and subsequently incubated with a horseradish peroxidase-(HRP) conjugated secondary antibody (P0448; Dako, Hamburg, Germany) at room temperature for 1 h. The antibody signals were detected with 3,3′-Diaminobenzidine (Dako, Hamburg, Germany) followed by counterstaining with hematoxylin. For quantitative analysis of the staining, 10 to 20 photographs in a 200× magnification were taken per object. The positive stained area was analyzed as the percentage of brown pixels per image with Adobe Photoshop CS5.

Liver macrophages were stained immunohistochemically with F4/80 specific antibody (MCA-497, 1:10, Serotec, Oxford, UK) and matching secondary antibody (sc-2021, 1:200, Santa-Cruz) as described previously by our group.²⁷

Western blot

Snap-frozen liver tissue was homogenized, lysed in lysis buffer (50 mM Tris-HCl, 150 mM NaCl, 1% NP-40, 0.5% sodium deoxycholate, 0.1% SDS, 10× protease inhibitor) for 30 min on ice and centrifuged for 10 min at 10,000 × g. The concentration of extracted protein was determined using Pierce BCA protein assay kit (Thermo Fisher Scientific, Waltham, Massachusetts) with BSA serving as a standard. SDS gel electrophoresis was conducted with 40 µg extracted protein per lane on a 12% SDS gel. The separated proteins were transferred to nitrocellulose membranes (Amersham Biosciences, Freiburg, Germany) and stained for α smooth muscle actin (αSMA) and GAPDH. Primary antibodies against αSMA (1:1.000, A2547, Sigma-Aldrich, St. Louis, Missouri) and GAPDH (1:20.000, MAB374, Merck Millipore, Darmstadt, Germany) were incubated overnight at 4°C in 2.5% BSA. An antibody conjugated to HRP was used as secondary antibody (1:20.000 for αSMA and 1:40.000 for GAPDH, A9044, Sigma-Aldrich, St. Louis, Missouri; 1 h room temperature) and was detected using ECL Western blot reagents (Amersham Biosciences, Amersham, UK). The signals were quantitatively analyzed with Quantitiy One 1D analysis software (Bio-Rad, Hercules, California) and αSMA signals were normalized to GAPDH signals.

Statistical analysis

Data are presented as mean±s.e.m. if normally distributed and as median if not normally distributed. In normally distributed data sets with more than nine data points, outliers were identified using Grubb’s test and were excluded if significance was p<0.05. Statistical analysis was performed using Sigma-Plot 12.0 (Systat Software Inc., Erkrath, Germany). After testing for normality and equal variance across groups, differences between the groups were assessed by one-way ANOVA, followed by the appropriate post hoc comparison test. If the data were not normally distributed, one-way ANOVA on Ranks was performed. Statistical significance was set at p<0.05. For details see figure legends.

General aspects

Treatment of mice with STZ led to increased blood glucose levels throughout the entire observation period with blood glucose levels of about 20 mmol/L (Figure 1(a)). Elevated blood glucose levels were accompanied by slightly (at 12 and 20 weeks significantly) reduced non-fasting plasma insulin levels compared to healthy mice (Figure 1(b)). Feeding of a HFD resulted in continuously elevated plasma triglyceride levels (data not shown). Body weight of STZ/HFD-treated mice increased slowly over the period of 20 weeks (wks), but was generally 10–20% lower when compared to healthy mice of the same age (Figure 1(c)). However, liver weight increased disproportionately compared to body weight, leading to a continuously increasing liver/body weight index over time (Figure 1(d) and (e)). This model was associated with a mortality rate of about 28% (37 of 132) over the period of 20 weeks.

Hepatic microcirculation and microhemodynamics

Microcirculatory changes during NAFLD progression were examined by means of in vivo fluorescence microscopy. Hepatic microvasculature in STZ/HFD-treated mice was characterized by an impairment of nutritive perfusion with a high number of non-perfused sinusoids. This resulted in a reduced sinusoidal perfusion rate at all examined time points, with 8 and 12 weeks being significantly diminished by 7% and 11%, compared to healthy mice (Figure 2(a)). Simultaneously, morphological density of sinusoids was significantly reduced at all time points compared to control mice (Figure 2(b)). Further analysis even showed lipid-laden hepatocytes, resulting in tortuous appearance of sinusoids varying from small to large diameters. Nevertheless, slightly increasing sinusoidal diameters were observed over time (Table 1). In contrast, the venular diameters were marginally reduced at six weeks (Table 1). In correlation with disease severity, normal vascular architecture was more and more lost and changed to an irregular and distorted sinusoidal network (Figure 2). Particularly for the tumor state (20 weeks), numerous superficial vessels and highly vascularized zones were striking, which resulted in an overall higher sinusoidal density compared to the other disease stages (Figure 2(b)). Quantitative analysis of hepatic microhemodynamics revealed no significant changes of venular and sinusoidal blood velocity in STZ/HFD-treated mice compared to control animals. Consequently, volumetric blood flow in venules and sinusoids was found to be almost unchanged. Calculation of the shear stress revealed marginally reduced values in venules and sinusoids at late time points (Table 1).

Steatosis

Fat deposition in the liver was assessed by quantitative analysis of Oil Red O-stained frozen liver sections. While healthy animals revealed lipid contents of about 8% (red line), Oil Red O positive area of the liver in STZ/HFD-treated mice was markedly increased at all time points, reaching values of 20–30% (Figure 3(a)). Oil Red O (Figure 3(a)) but also H&E staining (Figure 6(e)) showed both micro- and macrovesicular steatosis being present independent of the time point. Particularly in livers of 20 weeks old mice, in accordance with tumor development, fat was rather heterogeneously distributed. As lipid peroxidation is used as an indicator of oxidative stress, the lipid peroxide MDA was measured systemically (Figure 3(b)). Compared to healthy animals, plasma MDA levels were significantly elevated during almost the entire observation period of 20 weeks.

Inflammation

In vivo fluorescence microscopic analysis of leukocyte interaction with the microvascular endothelium revealed massively and significantly increased leukocyte adherence in postsinusoidal venules at all time points of STZ/HFD-treated mice compared to control mice (Figure 4(a) and (c)). This observation was accompanied by continuously increasing sinusoidal leukostasis (Figure 4(d)). Elevated adherence of leukocytes in venules leads to leukocyte migration into the liver parenchyma and thus reflects ongoing inflammatory processes. To further investigate inflammation, the number of granulocytes in the liver tissue was histologically assessed by quantitative analysis of CAE staining (Figure 4(b) and (c)). During NAFLD progression, numbers of infiltrated granulocytes increased, reaching a peak at 12 weeks (Figure 4(b) and (c)). At the same time, the number of F4/80-stained cells, namely Kupffer cells and monocyte derived macrophages, remained unchanged during NAFLD progression (data not shown).

Fibrosis/end-stage liver disease

Steatohepatitis is continuously challenging the liver tissue which in turn can lead to remodeling of extracellular matrix and fibrosis. HSCs play a crucial role in fibrosis development. HSCs exhibit a transition from a quiescent to a proliferative fibroblast-like phenotype that is accompanied by loss of intracellular vitamin A. Evaluation of vitamin A-positive spots by means of IVM revealed a decreased number in STZ/HFD-treated mice at all time points compared to healthy animals, being significantly reduced at eight weeks (Figure 5(a)). Concomitantly, activation of HSCs is characterized by increased expression of αSMA. Indeed, we observed elevated αSMA protein expression in STZ/HFD-treated mice, particularly at an age of 12 weeks (Figure 5(b)). Consequently, HSC activation leads to increased collagen 1α deposition. Quantitative assessment of collagen 1α via immunohistochemical staining revealed highest collagen 1α content of about 17% at 12 weeks of age, being significantly increased compared to 8 weeks (Figure 5(c)). The observed slight reduction of fibrotic markers (collagen 1α deposition and αSMA expression) at 20 weeks was explained by the highly heterogeneous appearance of livers at this time point with tumors often being non-fibrotic. Chronic hyperglycemia and HFD induced tissue damage and functional impairment of the liver. Initially, plasma activities of ALT and GLDH were only slightly elevated, but increased continuously with severity of the liver disease (Figure 6(a)). In accordance with that observation, in vivo analysis of bisbenzimide-stained hepatocytes unveiled elevated numbers of apoptotic cells at all observed time points, with 20 weeks being significantly increased compared to untreated mice (Figure 6(c)). This was also verified by immunohistochemical assessment of cell death in TdT-mediated dUTP-biotin nick end labeling (TUNEL)-stained liver sections ( Figure 6(d)). Even though that NAFLD activity score is not suitable for tumor assessment, it differentiates NAFLD severity grades and progression stages. In STZ/HFD-treated mice, NAS increased with age and disease severity, confirming progression of NAFLD (Figure 6(b)). At late time points, different types of tumors were observed, reaching from neoplasia with high-fat content but intact vascular architecture, unstructured tumors displaying a loss of vascular architecture, to highly differentiated HCC of pseudoglandular and trabecular type (Figure 6(e)). While no tumors were observed at the time points 6 and 8 weeks, 41% of 12 weeks old (n=17) and 84% of 20 weeks old mice (n=19) displayed tumors (Figure 1(e)).