Liver Sinusoidal Endothelial Cells Escape Senescence by Loss of p19ARF

Liver sinusoidal endothelial cells (LSECs) represent a highly differentiated cell type that lines hepatic sinusoids. LSECs form a discontinuous endothelium due to fenestrations under physiological conditions, which are reduced upon chronic liver injury. Cultivation of rodent LSECs associates with a rapid onset of stress-induced senescence a few days post isolation, which limits genetic and biochemical studies ex vivo. Here we show the establishment of LSECs isolated from p19ARF-/- mice which undergo more than 50 cell doublings in the absence of senescence. Isolated p19ARF-/- LSECs display a cobblestone-like morphology and show the ability of tube formation. Analysis of DNA content revealed a stable diploid phenotype after long-term passaging without a gain of aneuploidy. Notably, p19ARF-/- LSECs express the endothelial markers CD31, vascular endothelial growth factor receptor (VEGFR)-2, VE-cadherin, von Willebrand factor, stabilin-2 and CD146 suggesting that these cells harbor and maintain an endothelial phenotype. In line, treatment with small molecule inhibitors against VEGFR-2 caused cell death, demonstrating the sustained ability of p19ARF-/- LSECs to respond to anti-angiogenic therapeutics. From these data we conclude that loss of p19ARF overcomes senescence of LSECs, allowing immortalization of cells without losing endothelial characteristics. Thus, p19ARF-/- LSECs provide a novel cellular model to study endothelial cell biology.


Introduction
Endothelial cells (ECs) represent a unique cell population originating from the mesoderm and lining all vessels in the organism. ECs can be divided into blood and lymphatic ones based on the capillary surface they are forming. From large vessels to small capillaries, ECs form compact monolayers with semi-permeable properties, mediating transport of small metabolites, migration of immune cells and vessel tone [1,2]. Liver sinusoidal endothelial cells (LSECs) represent a subpopulation of non-parenchymal cells in adult liver [3]. In comparison with other ECs, they possess a unique morphology due to multiple membranous pores, called fenestrations, which are organized into sieve plates [4]. Due to their size of about 0.2 μm, fenestrations mediate fast transfer of small molecules, which accelerates metabolic exchange under controlled selection.
Platelet endothelial cell adhesion molecule (PECAM-1), also known as CD31, is considered as golden standard marker of endothelial cells [5,6]. Unfortunately, many conflicting reports refer to CD31 expression in LSECs in vivo and in vitro [7][8][9][10]. Interestingly, March et al. showed that CD31 expression is predominantly high in the central and portal endothelium of rat liver and low expression is detected in liver sinusoids [10]. CD31 expression could be barely detected in freshly isolated LSECs which correlated with in vivo data. However, CD31 expression increased in cell culture at day 3 after defenestration, probably being accompanied by loss of endothelial phenotype. While CD31 expression is contradictorily published on ECs, the expression of vascular endothelial growth factor receptor 2 (VEGFR-2) is almost exclusively restricted to ECs [11]. In liver, VEGFR-2 expression can be exclusively detected on LSECs [12], thus belonging to markers which enable to distinguish ECs from other liver cell populations [13]. In addition, ECs are known to express several types of cadherins, including vascular endothelial (VE)-, P-and N-cadherin that are part of adherens junctions [14], von Willebrand factor (vWF) [15], stabilin-1 (Stab-1) and Stab-2 [16], CD32b [17], CD146 [9], and the lymphatic vessel endothelial hyaluronan receptor (Lyve)-1 [17].
Human umbilical vein endothelial cells (HUVECs) are frequently used as an EC model in vascular biology as these cells proliferate in cell culture with limited cell doublings [18]. However, conclusions drawn from experiments using HUVECs have to be carefully interpreted in organ-specific studies due to the wide heterogeneity among different types of endothelium. For multiple reasons, tissue-specific ECs are beneficial as these cells can be investigated in a homotypic setting, allowing more accurate results [19]. Yet, organ-specific ECs are hardly available. While human hepatic sinusoidal endothelial cells (HSECs) can be propagated in cell culture for 7-8 passages [15], their handling is delicate as they do not overcome several freeze-thaw cycles. Even more challenging, rodent LSECs [10] cannot be propagated after liver perfusion as they die within few days after cultivation.
The lifespan of human primary cells in vitro is affected by telomere length and their shortening during cell division. When telomeres reach critical length, cells enter mitotic crisis and as protection from further division, they undergo replicative senescence. In contrast, murine cells have active telomerase and are protected from the senescence induced by telomere shortening [20], yet their proliferative capacity remains finite due to enhanced expression of negative cell cycle regulators p16 INK4a , p21 Cip1 , p53 or its regulator p19 ARF [21]. The INK4a/ARF locus encodes for the two crucial tumor suppressor proteins p16 INK4a and p19 ARF (p14 ARF in humans) which act upstream of the retinoblastoma and Mdm2/p53 pathways, respectively [22]. p19 ARF binds to Mdm2, a negative regulator of p53, thus stabilizing it and allowing p53 to act as a tumor suppressor responsible for cell cycle arrest and apoptosis [23,24]. Mice lacking p19 ARF are viable and fertile with longer latency for tumor development as compared to p53 -/mice [22]. In contrast to p53 deficiency, loss of p19 ARF is supposed to overcome senescence and allow infinite proliferation without gaining malignant properties [25,26].
In this study we aimed at establishing LSECs from p19 ARF-/mice. LSECs isolated from p19 ARF-/mice, termed mLSECs escape from senescence and are allowed to proliferate without losing overt genetic stability. mLSECs show strong EC characteristics and vascular properties that can be used in homotypic cell-cell interaction studies. made to minimize suffering of animals. In particular, we intraperitoneally injected 100 mg/kg Ketamin and 5 mg/kg Rompun for anesthesia of mice before undergoing liver perfusion. 10 mg/kg Carprofen were subcutaneously injected into mice for analgesia. Prior to euthanasia by cervical dislocation, mice were intraperitoneally injected with 100 mg/kg Ketamin and 5 mg/kg Rompun.

Isolation of cells
Mouse liver sinusoidal endothelial cells (mLSECs) were isolated from livers of 10-14-weeksold female p19 ARF-/and C57/BL6 wt mice by in situ liver perfusion via the intrahepatic vena cava as described previously [25]. Human telomerase reverse transcriptase (TERT)-immortalized blood endothelial cells (BECs; [27] were cultivated on collagen-coated plates in ECM. MIM1-4 cells were cultivated as described previously [25]. All cell lines were cultured at 37°C and 5% CO 2 and routinely screened for the absence of mycoplasma. Proliferation kinetics 5 x 10 3 mLSECs were seeded in triplicates on collagen-coated 24-well plates and allowed to proliferate for 12 days until cells undergo growth arrest due to high cell density. The number of cells in the corresponding cell populations was determined in a multichannel cell analyzer (Casy Cell Counter, Schärfe Systems, Germany). In addition, a cumulative growth analysis which allows monitoring of proliferation kinetics without an arrest by cell density was performed as described previously [28]. Briefly, 1 x 10 5 cells were seeded on collagen-coated 6 well-plates. The number of cells was determined periodically in a multichannel cell analyzer (Casy Cell Counter, Schärfe Systems, Germany) until day 12. Cumulative cell numbers were calculated from the cell counts plus dilution factors.
Quantitative reverse-transcriptase polymerase chain reaction (qRT-PCR) Total RNA was extracted, treated with DNaseI and reverse transcribed using a RNA isolation and cDNA synthesis kit as recommended by the manufacturer (Quiagen, Hilden, Germany). Wound healing 1 x 10 5 cells were seeded on collagen-coated 6-well plates and grown until confluency. A blue pipette tip was used to generate scratches. Images were taken after cultivation for 20 hours in the corresponding medium and analyzed with ImageJ.

Statistics
Data were expressed as means ± standard deviation. The statistical significance of differences was evaluated using a paired, non-parametric Student's t-test. Significant differences between experimental groups were Ã p<0.05, ÃÃ p<0.01 or ÃÃÃ p<0.005.

Results
Loss of p19 ARF allows proliferation of mLSECs mLSECs were isolated from livers of female p19 ARF-/mice. In the presence of the endothelial cell growth supplement (ECGS) medium, cells started to spontaneously proliferate within one week in culture. Phase contrast microscopy showed mLSECs displaying the typical cobblestone-like morphology ( Fig 1A). Interestingly, mLSECs showed slightly faster proliferation kinetics at later passages (passage 20) as compared to mLSECs at early passage 5 (Fig 1B). At day 12 of cultivation, mLSECs exhibited proliferation arrest due to high cell density. Similar results were obtained from cumulative growth analyses of mLSECs at early and late passage without proliferation arrest (Fig 1B). Immunolocalization of the tight junction constituent ZO-1 and the adherens junction component β-catenin showed staining at cell boundaries in both early and late passaged mLSECs, indicating cells of epithelial origin (Fig 1D). Similarly, N-cadherin which is required for the interaction among ECs and pericytes was detected at cell borders [30]. In addition, both early and late passaged mLSECs displayed the expression of the intermediate filament protein vimentin which plays an important role in the sprouting of ECs during angiogenesis [31]. Together, these data show that loss of p19 ARF equips mLSECs with the ability to proliferate by concomitantly displaying an endothelial morphology.

mLSECs display endothelial properties in culture
Next we analyzed more closely the EC phenotype of mLSECs. Immunoblotting showed that mLSECs express CD31, VE-cadherin and VEGFR-2 similarly to hHSECs and the human TERT-immortalized blood endothelial cells (BECs, Fig 2A). Interestingly, mLSECs retained expression of these markers during cell passaging. Furthermore, we analyzed other published markers for their expression in mLSECs. vWF, Stab-2 and CD146 which are reported as characteristic markers of rodent liver endothelium were detected in both, early and late passaged cells. In contrast, Stab-1, Lyve-1 and CD32b expression were not increased in comparison to immortalized MIM1-4 hepatocytes (Fig 2B). In addition, endothelial cells are endowed with a high migratory capacity which precedes angiogenesis in vivo and which could be observed by wound healing assays upon stimulation with ECGS. Notably, mLSECs displayed a capability of wound closure which is comparable to established cultures of BECs and hHSECs (Fig 2C). To further demonstrate the endothelial integrity of isolated mLSECs, we performed tube formation assays on growth factor-reduced Matrigel. Formation of vessel-like structures upon stimulation with pro-angiogenic mitogens is a particular feature of all ECs [32][33][34]. Isolated mLSECs generated tube-like structures upon stimulation with VEGF-A, ECGS or a combination of both, whereas no tubes were detected without stimulation (Fig 2D). From these data we concluded that mLSECs exhibit characteristics that are typical for ECs.

mLSECs overcome cellular senescence
As mentioned, cultivation of rodent LSECs is accompanied by a rapid loss of the endothelial phenotype and onset of cellular senescence several days after isolation [10,35,36]. Accordingly, we isolated wild type (wt) LSECs from C57/BL6 mice and introduced them into cell culture. After several days in culture, p19 ARF-/-mLSECs started to proliferate (Fig 1A), whereas wt mLSECs did not exhibit any detectable outgrowth under the same culture conditions. Remarkably, cell cycle analyses revealed that 89.5% of freshly isolated wt mLSECs remained quiescent in the G1/G0 phase of the cell cycle at day 1 of cultivation with a low proportion of cell death (7.8%; Fig 3A) indicating an arrest in proliferation. However, the portion of dead mLSECs increased to 48.9% on day 4 after isolation, suggesting that mLSECs evince a clear blockade in cell division followed by cell death (Fig 3A). The analysis of DNA content by flow cytometry revealed that cultured mLSECs represent a diploid cell population. Higher ploidy levels would point to a non-stable DNA content associated with a malignant phenotype or a contamination with hepatocytes. Interestingly, the cell cycle distribution among different phases of the cell cycle was stable between early (p10) and late (p20) passage numbers, pointing to genomic stability (Fig 3B).

Loss of p19 ARF does not influence susceptibility to anti-angiogenic agents
We next analyzed whether mLSECs respond to anti-angiogenic drugs such as sorafenib and sunitinib which block platelet-derived growth factor receptor (PDGFR) and VEGFR. Cell survival assays showed similar inhibitory concentration (IC)50 levels of both drugs, namely 8.0 μM for sunitinib and 7.3 μM for sorafenib which are comparable to those detected in BECs (4.33 μM sorafenib; 8.28 μM sunitinib) and hHSECs (2.88 μM sorafenib; 5.67 μM sunitinib), respectively (Fig 4A-4F). To analyze whether the effect of drugs is rather cytostatic or cytotoxic, we performed flow cytometry analysis using DAPI staining. Treatment of mLSECs with sunitinib at the IC50 concentration resulted in 40% of dead cells after 72 hours, thus closely correlating with results from the cell survival assay, while administration with sorafenib led to a cytostatic effect without induction of cell death (Fig 5A-5C). Similar results were obtained by flow cytometry analysis of hHSECs and BECs. These data show that mLSECs respond to antiangiogenic drugs by inhibition of cellular proliferation and/or cell death.

Discussion
We have established mouse liver sinusoidal endothelial cells from p19 ARF-/mice. Loss of p19 ARF provides the ability to overcome cellular senescence. Isolated mLSECs have a typical cobblestone-like morphology, and in comparison with LSECs isolated from wt mice, mLSECs start to proliferate in vitro a few days after introduction into cell culture. By comparing cells at different passage numbers, a slightly faster growth of cells was observed at later passages. As the analysis of proliferation kinetics require seeding of low numbers of cells, we speculate that cells kept longer in cell culture can more easily overcome stress associated with sub-confluent growth and are better adapted to in vitro conditions. Notably, mLSECs displayed a stable morphology and kept sensitivity to contact inhibition during cultivation up to passage number 50 (data not shown) which corresponds to roughly 75 cell doublings within 5 months.
mLSECs exhibit an endothelial phenotype due to the expression of typical markers mostly cited in the literature, even if some ambiguity exists among their expression levels. Our study shows that mLSEC express low levels of CD31 which is underlined by the fact that streptavidin/biotin amplification was required to detect the protein. Similarly, VE-cadherin was detected upon signal amplification and its level in hepatic endothelial cells is lower than in BECs. These data support recent reports suggesting that the sinusoidal endothelium in normal liver lacks VE-cadherin or expresses it at low levels and its relative lack might be a consequence of the absence of classical adherens junctions [8]. Further analyses revealed mLSECs expression of vWF as well as Stab-2 and CD146 but not Stab-1, CD32b and Lyve-1, the latter being rather a marker of lymphatic endothelium [37]. Furthermore, mLSECs harbor a high migratory capacity as shown by the wound healing assay as well as tube forming abilities on growth factor-reduced matrigel, reflecting important aspects of endothelial integrity and angiogenic capabilities. In addition, loss of p19 ARF does not induce malignancy as shown by subcutaneous injection of mLSECs into immunocompromized mice. None of the animals developed tumors or died 6 months after injection (data not shown), demonstrating that mLSECs fail to exhibit a malignant phenotype.
Cellular senescence is a state of irreversible growth arrest that can be induced in stress response to various cellular stimuli [21]. p19 ARF is a known executor of senescence in cultured cells [38] and its expression level is increasing during aging [39]. Thus, loss of p19 ARF allows overcoming cellular senescence which was described for several cell types in culture [25,26]. The mechanism of cell death of primary rodent wt LSECs is still poorly understood and many authors refer to dedifferentiation and loss of phenotype instead of the induction of cellular senescence. We speculate that the upregulation of p19 ARF involves the activation of p53, causing the induction of the cell cycle regulators p21 Cip1 and p27 Kip1 and thus cell cycle arrest, which subsequently leads to the loss of the EC phenotype and cell death in culture. In accordance with this consideration, loss of p19 ARF allows overcoming cell cycle arrest (Fig 3B), which is typical for many primary cells and precedes cellular senescence. mLSECs lacking p19 ARF reveals a distribution across all cell cycle phases in early and late passaged cells. In contrast, wt LSECs remain quiescent in the first days of cultivation with about 90% of cells arrested in the G 1 /G 0 phase of the cell cycle ( Fig 3A). On day 4 of cultivation, this portion is reduced to 50% of cells with the other 50% being apoptotic. Noteworthy, treatment of mLSECs with either sunitinib or sorafenib showed that mLSECs are able to respond to anti-angiogenic drugs in a similar way as HSECs and BECs as shown by comparable IC50 levels. Loss of p19 ARF still allows mLSECs to respond by cell cycle arrest and cell death that is probably driven by p53 in a p19 ARF -independent way.
Together, our results show that loss of p19 ARF supports the escape of mLSECs from cellular senescence and enables their cultivation in vitro. mLSECs can be immortalized without overt features of dedifferentiation. As compared to previously reported p19 ARF-/hepatocytes [25] and p19 ARF-/stellate cells [26], mLSECs do not evince signs of malignancy despite an "unlimited lifespan" and thus could represent a valuable tool in vascular biology and liver-specific studies.