Comparison of Endothelial Cell Proliferation in Normal Liver and Adipose Tissue in B6C3F1 Mice,F344 Rats,and Humans

Abstract

Peroxisome proliferator-activated receptor gamma (PPARγ) and dual PPARα and γ agonists have been developed for use in the treatment of diabetes and hyperlipidemias. Vascular tumors were increased in mice treated with some PPAR agonists, but not in rats. Spontaneous hemangiosarcomas are common in several strains of mice, uncommon in rats, and rarely occur in humans. The objective of this study was to determine the endothelial cell proliferation rate in liver and adipose tissue of B6C3F1 mice, F344 rats, and humans to aid in investigations of the genesis and development of hemangiosarcoma formation, and to determine the relevance of the increased endothelial cell proliferation rate in drug-treated rodents in assessing the risk of these drugs in humans. We determined the endothelial cell labeling index (LI) in untreated mice, rats, and humans, in normal liver, brown fat (rats and mice only) and white fat by dual immunohistochemistry of CD31 and Ki-67. The LI, highest in mice and lowest in humans, was statistically significantly greater in the mouse compared to the human and rat. The increased rate of spontaneous or PPAR agonist-induced hemangiosarcoma formation in mice may be related to the higher background endothelial cell proliferation rate compared to rats and humans.

Introduction

Hemangiosarcomas (angiosarcomas) are rare, highly malignant tumors with a generally poor prognosis in humans (Fletcher et al., 1991; Mark et al., 1996), accounting for less than 1% of all sarcomas (Enzinger and Weiss, 1995). They arise from the endothelial cells of small blood vessels and commonly occur in the skin, soft tissue and liver in humans (Maddox et al., 1981; Fineberg et al., 1994; Neshiwat et al., 1992; Smith et al., 1985). In contrast, hemangiosarcomas occur spontaneously at relatively high rates in many strains of mice, particularly in lifetime (2 year) studies. The incidence of spontaneous hemangiosarcomas in all organs of B6C3F1 mice is 5.3% in the male and 2.8% in the female (ranges of 0–14% and 0–16%, respectively) with a high incidence in all organs, but especially liver, spleen, and bone marrow (National Toxicology Program, 2000). In the liver of B6C3F1 mice, the mean incidence of spontaneous hemangiosarcomas is 2.48% in the male and 0.80% in the female, with ranges of 0–8% and 0–4%, respectively. In rats, vascular tumors, especially sarcomas, are much less common, but occur at rates considerably higher than in humans (Herman et al., 2002). The incidence of hemangiosarcomas in all organs of the F344 rat is 0.40% in the male and 0.28% in the female (ranges of 0–2% and 0–2%, respectively). In the liver of F344 rats, the incidence of spontaneous hemangiosarcomas is 0.07% in the male and 0.00% in the female (National Toxicology Program, 2000).

When troglitazone, a peroxisome proliferator-activated receptor gamma (PPARγ) agonist, was administered to B6C3F1 mice, the incidence of hemangiosarcomas in the liver was increased (Herman et al., 2002). It has also been reported that the BrdU labeling index of endothelial cells in the brown adipose tissue of PPARγ agonist-treated B6C3F1/CrlBR mice was increased by troglitazone administration (Breider et al., 1999). Hemangiosarcomas in mice have been frequently observed in studies with several PPARγ and dual PPARα and γ agonists (El Hage, 2005), as well as in studies with chemicals in other pharmacologic and chemical classes (Hong et al., 2000).

Little is known about the mechanism of induction of hemangiosarcomas except when induced by genotoxic agents like vinyl chloride and thorotrast. Many of the hemangiosarcomas produced by various chemicals in mice occur at the sites with high spontaneous rates of endothelial cell proliferation, such as liver, spleen, and bone marrow. For PPAR agonists, a significant incidence frequently occurs in the subcutaneous adipose tissue. Since the mechanism of hemangiosarcoma induction in mice by nongenotoxic chemicals is not well understood, assessing the risk to humans is difficult. To begin, there is a paucity of comparative data between species. It is important for investigating the genesis and development of the hemangiosarcomas that comparative data regarding proliferation of endothelial cells in the different species be developed. The objective of this study was to determine the endothelial cell proliferation rate in normal liver and adipose tissue of untreated B6C3F1 mice, F344 rats, and humans to aid in the investigation of the genesis and development of hemangiosarcomas.

Materials and methods

Human Samples

Slides of routinely formalin-fixed, paraffin-embedded livers and white fat tissues from 10 female and 10 male humans (different patients for liver and fat), 20–50 years old, were obtained from the surgical pathology files of the Nebraska Medical Center. The protocol was approved by the University of Nebraska Medical Center Institutional Review Board. Liver tissue was obtained from biopsies of donor livers used to assess suitability for transplantation and were histopathologically unremarkable. Fat tissue was obtained from pannus specimens which were histologically normal. Normal pannus tissue was readily available with a proliferation rate similar to the liver, and the proliferation rate in pannus tissue would be expected to be similar to the proliferation rate at other subcutaneous sites.

Test Animals

Ten female and 10 male F344 rats and B6C3F1 mice (Charles River Breeding Laboratories, Inc., Kingston, NY and Raleigh, NC respectively), 6 weeks old at the time of arrival were used in the study. The animals were group housed (5/cage) in plastic cages with dry corn-cob bedding. The animal room was on a 12-hour light/dark cycle at a targeted temperature of 22 ± 2°C and humidity of 50 ± 20%. Food (Certified Purina 5002, Dyets, Inc., Bethlehem, PA) and tap water were available ad libitum at all times during the study. The protocol for the study was approved by the University of Nebraska Medical Center Institutional Animal Care and Use Committee, and the level of care provided to the animals met or exceeded the basic requirements outlined in the Guide for the Care and Use of Laboratory Animals (NIH publication #86–23, revised 1986).

Histological Examinations

All animals were sacrificed under anesthesia by Nembutal (i.p., 150 mg/kg body weight) at 8 weeks of age. Liver, dorsal brown fat tissue with adjacent white fat tissue and section of the small intestine of rats (one-half of each piece of tissue) and mice were removed and fixed in 10% phosphate-buffered formalin (PBF), embedded in paraffin, and sectioned for hematoxylin and eosin staining (HE) for histopathological evaluation. These blocks of tissue from mice were used for dual immunostaining for Ki-67 and CD31. The other half of each piece of tissue from rats was fixed in 70% ethanol overnight and 95% ethanol for 48 hours, and then processed for paraffin embedding, sectioned at 4 μm and dual immunostained for Ki-67 and CD31. Ethanol fixation of rat tissue was required for retrieval and staining with the antibodies for Ki-67 and CD31

Immunohistochemistry

Dual Immunostaining for Ki-67/CD31 on Human Tissues

Dual immunostaining of liver and white fat tissues in 10% formalin-fixed specimens was performed for detection of Ki-67 (proliferation marker) and CD31 (endothelial cell marker).

The dual immunostain was performed on the Ventana Benchmark instrument (Tucson, AZ). The tissues on slides underwent heat-induced epitope retrieval with CC1 solution (Ventana) for 30 minutes. The first antibody applied was for Ki-67 (Clone MIB-1, DAKO, Carpinteria, CA) diluted 1:40, for 30 minutes. The chromogenic stain was 3,3′-diaminobenzidine tetrahydrochloride (DAB) (Ventana). The second antibody applied was for CD31 (DAKO), diluted 1:100, for 32 minutes. The chromogenic stain was enhanced V red (Ventana).

Immunostaining for CD31 on 10% Formalin-Fixed Rat and Mouse Tissues

After deparaffinization of the sections, the slides were exposed to 3% H₂O₂ for 5 minutes to quench endogenous peroxidase, and epitope retrieval was performed by exposure to Proteinase K (P-6556, Sigma-Aldrich, St. Louis, MO) for 90 minutes at 37°C. After washing with tap water and 1% non-fat milk used for blocking non-specific reactions, the sections were incubated overnight at approximately 4°C with primary antibody (monoclonal rat anti-mouse CD31, clone MEC13.3, Fitzgerald Inc., Concord, MA or monoclonal mouse anti-rat CD31, Clone TLD-3A12, Serotec Ltd., Oxford, UK) diluted 1:25. The avidin-biotin-peroxidase complex (ABC) method was used for visualization of antigen location (Vectastain Elite ABC kit, Vector Laboratories, Burlingame, CA). Positive reactions for endothelium resulted in purple cytoplasmic staining with the VIP substrate kit (Vector Laboratories).

Dual Immunostaining for Ki-67/CD31 on Mouse Tissues

Dual immunostaining for Ki-67 and CD31of liver and brown and white fat was performed on 10% formalin-fixed mouse tissues. After deparaffinization of the sections and exposure of the slides to 3% H₂O₂ for 5 minutes to quench endogenous peroxidase, heat-induced epitope retrieval was performed in 10 mM citrate buffer, pH 6.0. After incubation with 1% non-fat milk used for blocking nonspecific reactions, the sections were immunostained using monoclonal rat anti-mouse Ki-67 (TEC-3; DAKO), diluted 1:25, for 1 hour at room temperature. The sections were incubated using the ABC method. Positive reactions resulted in brown nuclear staining with the DAB substrate kit (Vector Laboratories). After washing with tap water and 1% non-fat milk again, the sections were incubated overnight at approximately 4°C with monoclonal mouse anti-human CD31 (Clone; JC70A, DAKO) diluted 1:25. The ABC method was used for visualization of antigen location. Positive reactions for endothelial cells resulted in purple cytoplastic staining with the VIP substrate kit (Vector Laboratories).

Dual Immunostaining for CD31/Ki-67 for Rat Tissues

Dual immunostaining of liver, brown, and white fat tissues in ethanol-fixed rat tissues was performed for detection of CD31 and Ki-67. After deparaffinization of the sections, the slides were exposed to 3% H₂O₂ for 5 minutes to quench endogenous peroxidase activity and 1% nonfat milk for blocking nonspecific reactions, the sections were reacted overnight at 4°C with mouse anti-rat CD31 monoclonal antibody diluted 1:25 (Serotec Ltd., Oxford, UK) without epitope retrieval. The sections were incubated using the ABC method. Positive reactions resulted in brown cytoplasmic staining with the DAB substrate kit. Heat-induced epitope retrieval was performed in Tris-EDTA buffer, pH 9.0, for 8 minutes at 96–98°C. After exposure to 1% non-fat milk for blocking of non-specific reactions, the sections were immunostained using monoclonal mouse anti-rat Ki-67 antigen (MIB-5; DAKO) diluted 1:25 for 1 hour at room temperature. The ABC method was used for visualization of antigen location. Positive reactions for endothelial cells resulted in purple nuclear staining with the VIP substrate kit.

Statistical Analysis

Data are presented as means ± SD. Ki-67/CD31 labeling indices were analyzed by 2-tailed Students t-test using Microsoft Excel. Values represent the results of the t-test, with values of p < 0.05 considered statistically significant.

Results

Labeling Index of Endothelial Cells in Humans, Rats, and Mice

We determined the labeling index (LI) in mice, rats and humans of endothelial cells in normal liver, brown fat (rats and mice only) and white fat by dual immunohistochemical staining for Ki-67 and CD31 (Table 1, Figure 1). The LI in the male and female mouse liver was significantly higher (p < 0.01) compared to the LI in the male and female rat and human liver, and the LI in the male and female rat liver was significantly higher (p < 0.05) than the LI in the human liver. The LI in the male and female mouse brown fat was significantly higher (p < 0.01) than the LI in the male and female rat brown fat. The LI in the male mouse white fat was significantly higher (p < 0.01) compared to the LI in the male rat and human white fat. The LI in the female mouse white fat was significantly higher compared to the female rat (p < 0.05) and the female human (p < 0.01). Otherwise, the LI was similar in males and females although the LI in the male mouse tissues was consistently higher than in the female mouse tissues.

Discussion

A new class of drugs called PPARγ and dual PPARα and γ agonists has been developed for use in the treatment of diabetes and hyperlipidemias (Cantello et al., 1994; Lehman et al., 1995). PPARγ is expressed in adipose tissue, endothelial cells, urothelium and intestine (Marx et al., 1999; Yki-Jarvinen., 2004). Vascular tumors were increased in mice treated with some of these drugs. In a 2-year bioassay, troglitazone, a PPARγ agonist, caused an increased incidence of hemangiosarcomas of the liver in male and female mice but not in rats (Herman et al., 2002). Rodents treated with these drugs also had increased interscapular brown fat. It has also been reported that the bromodeoxyuridine (BrdU) labeling index of endothelial cells in the brown adipose tissue of PPARγ agonist-treated B6C3F1 mice was increased (Breider et al., 1999).

Ras mutation and p53 inactivation in B6C3F1 mice did not play a role in vascular tumorigenesis in a long-term study with troglitazone (Duddy et al., 1999a, 1999b). From the viewpoint of our data, the species-specific high incidences of hemangioma and hemangiosarcoma may be the result of increased endothelial cell proliferation on a background of an already high proliferation rate in B6C3F1 mice. The 8-week old rats and mice were in similar growth phases and the background cell proliferation rates in the liver and adipose tissue would be expected to be at adult rates by these ages. The high background proliferation rate of endothelial cells in mice might also be a significant basis for the high rate of spontaneous endothelial cell tumors in mice compared to other species, such as the rat and human. Increased numbers of cell divisions (DNA replications) have been theorized to provide the basis for an increased risk of carcinogenesis based on the spontaneous rates of mutation in somatic cells (Cohen et al., 1990, 1991; Cohen, 2005; Dominick et al., 2006).

Platelet endothelial cell adhesion molecule-1 (PECAM-1, CD31) belongs to the immunoglobulin family of cell adhesion molecules (Newman et al., 1990). Anti-CD31 antibody primarily labels endothelial cells, and is useful for the identification of vascular lesions, including heangiosarcomas (Newman et al., 1990; Parums et al., 1990). In addition, anti-CD31 antibody is valuable for determining angiogenesis in several types of tumors (DeYoung et al., 1995; Engel et al., 1996). An anti-human CD31 antibody was used for labeling paraffin-embedded human tissue sections fixed in formalin with heat-induced epitope retrieval (Giatromanolaki et al., 1996).

An anti-human CD31 antibody that was cross-reactive with mouse tissue was used in this study (Figure 1C–E), for the dual labeling procedure for paraffin-embedded mouse tissue sections fixed in formalin. Immunohistochemical studies using anti-rat and anti-mouse CD31 antibody are useful for tissues dehydrated with ethanol or in frozen sections (Carrithers et al., 2005; Zacchigna et al., 2005). Anti-rat CD31 antibody (clone number: TLD-3A12) and anti-mouse CD31 antibody (clone number: MEC13.3) may be used for labeling paraffin-embedded mouse tissue sections fixed in formalin with proteinase K for epitope retrieval (Figure 2). We chose Ki-67 as an indicator of cell proliferation so that the same marker could be used for human and rodent tissue. Obviously, BrdU labeling is not possible for examination of human tissues since BrdU cannot be routinely administered to humans. Using Ki-67 and CD31 permitted use of already available normal human tissues.

In conclusion, the increased rate of spontaneous hemangiosarcoma formation in mice may be related to the increased proliferation rate of endothelial cells normally present in this strain of mouse compared to rats and humans. Additional increases in endothelial cell proliferation might explain the sensitivity of mice to hemangiosarcomagenesis by various agents such as PPARγ and dual PPAR α/ γ agonists.