The induction of angiogenesis is essential for successful engraftment of freely transplanted cells or cellular composites. How to augment angiogenesis to ensure an appropriate viability of the grafts is still under investigation. This study evaluated the proangiogenic capability of different syngeneic free liver transplants and elucidated the origin of the newly formed vascular network via use of an eGFP+/eGFP− (enhanced green fluorescent protein) cross-over design. Using intravital fluorescence microscopy, we found that neonatal and resected murine liver transplants implanted into dorsal skinfold chambers display a significantly enhanced vascularization compared to regular adult transplants. Immunohistochemically, less tissue hypoxia, apoptosis and macrophage infiltration was observed in the neonatal and resected transplants, which is in line with improved vascularization of those grafts. Additionally, electron microscopy revealed morphological hallmarks of liver cells. eGFP+ liver transplants implanted on eGFP− recipients displayed vascular sprouting from the grafts themselves and connection to the recipients` microvasculature, which also undergoes transient proangiogenic response. This process is described as external inosculation, with microvessels exhibiting a chimeric nature of the endothelial lining. These data collectively show that proliferative stimulation is taking effect on angiogenic properties of free transplants and might provide a novel tool for modulating the revascularization of free grafts.
Organ transplantation is still the only curative treatment for end-stage liver diseases. Even though this surgical intervention is performed highly efficient by now, the frequency of liver transplantations is often limited due to donor organ shortage . Besides the high operative risk further obstacles of this approach are enormous medical costs and the need for a life-long immune suppressive therapy. For these reasons, alternative strategies are currently under investigation, either to assure a temporal bridging of patients until transplantation or to restore the complex liver function in a long-term view . Classical nonbiological (artificial) bridging procedures serving for extracorporeal blood purification are plasma exchange, hemodialysis and hemofiltration techniques [3, 4]. The introduction of the molecular adsorbents recirculating system (MARS) and the Prometheus system offered new possibilities in the field of dialysis [5, 6]. However, it is still highly challenging to establish a technology that substitutes all of the liver functions, including various secretory and synthesis processes . Therefore, one major focus in the field of tissue engineering and regenerative medicine is the development of innovative cell-based therapies for the complete restoration of the hepatic function in patients [7-9].
The injection of isolated hepatocytes into the liver or spleen holds great promise as an alternative strategy to organ transplantation [10-14]. However, by this, only partial correction of metabolic liver disorders has been achieved until today, and the degree to which donor hepatocytes restore failing livers has not been satisfying enough to circumvent the need for liver transplantation [13, 15]. Alternatively, liver cell composites may be transplanted to extrahepatic sites . This approach bears the major advantage that the number of transplantable hepatocytes is not limited by potential complications as they are described for intrahepatic administration, like pulmonary embolism and portal vein thrombosis [17, 18]. In the case of extrahepatic hepatocyte transplantation, successful engraftment and environmental integration of the grafts is crucially dependent on the induction of angiogenesis and a rapid vascularization . In the present study the vascularization of “living” liver slices transplanted into dorsal skinfold chambers on mice, allowing for in vivo assessment of vascular network development, was investigated. To determine the graft's impact on the angiogenic response, we compared regular adult, neonatal and remnant proliferating (liver undergoing 70% resection 24 h before transplantation) liver grafts and thus the effect of the proliferative status on the induction of angiogenesis. Liver slices provide an intact microenvironment thereby reflecting the correct multicellular alignment and function existent within the in vivo organ. Though they have been extensively used for toxicological, pharmacological and biochemical studies for decades [20, 21], physiological or functional investigations, i.e. vascularization of such grafts have not been described yet. For the purpose of this study “living” liver slices represent ideal transplants, as they retain the cellular in vivo structure and cohesion, which enables solid transplantation without any matrix or scaffold. The detailed insight how liver tissue proliferation affects the angiogenic response is important for further clinical implementation of extrahepatic liver graft transplantation to more effectively correct liver malfunction or failure.
Material and Methods
Male C57BL/6J Tyr mice ([10-12] weeks old, 28–35 g body weight, Charles River Laboratories, Sulzfeld, Germany) and B6.Cg-Tg(ACTB-EGFP)1Osb/J mice (eGFP+, 12–20 weeks old, 26–34 g body weight, Jackson Laboratory, Bar Harbor, ME, USA), were kept in standard animal facilities with a 12-hour light/dark cycle and received laboratory chow and water ad libitum. The experiments were performed in accordance with the German legislation on protection of animals and the NIH Guide for the Care and Use of Laboratory Animals (Institute of Laboratory Animal Resources, National Research Council).
Preparation of the dorsal skinfold chamber
The chamber technique and its implantation procedure have previously been described in detail . Details are provided in Supporting Information.
Preparation and transplantation of “living” liver slices
At the day of transplantation, donor mice (either C57BL/6J Tyr mice or B6.Cg-Tg(ACTB-EGFP)1Osb/J mice (eGFP+)) were anesthetized and underwent liver explantation. The explanted tissue was cut into ∼50-μm-thin liver slices (diameter = 3 mm) using a vibratome (Leica VT1200S, Leica, Wetzlar, Germany). The liver slices were kept in Hanks balanced salt solution (HBSS) at 4°C until subsequent syngeneic transplantation into dorsal skinfold chambers of recipient mice (C57BL/6J Tyr). For transplantation, the chamber-equipped animals were anesthetized and placed on a heating pad for maintenance of body temperature at 37°C. The cover glass of the dorsal skinfold chamber was carefully removed and one liver slice was placed directly on the striated muscle in the center of each chamber, taking care to avoid mechanical irritation or damage of the chamber tissue.
Experimental groups and protocol
Donor mice (C57BL/6J Tyr mice or B6.Cg-Tg(ACTB-EGFP)1Osb/J mice) for the preparation of “living” liver slices were randomized into three groups: (i) regular adult (n = 8), (ii) neonatal (n = 8) and (iii) remnant proliferating (24 hours after 70% liver resection, n = 8) liver tissue. The liver slices were transplanted into dorsal skinfold chambers, taking care to avoid contamination, mechanical irritation, or damage of the chamber tissue. The macroscopic appearance of the skinfold chamber was documented daily. Intravital fluorescence microscopic assessment of angiogenesis and microhemodynamic parameters was performed 20 min as well as 3, 5, 7, 10 and 14 days after transplantation. At the end of the last in vivo investigation the animals were sacrificed by an overdose of ketamine, and the chamber tissue bearing the transplanted grafts was processed for immunohistochemistry.
In an additional, equally performed set of experiments, grafts of regular adult, neonatal and remnant proliferating liver tissue of eGFP+ mice were transplanted onto eGFP− recipient mice (n = 3 per group) to better elucidate whether the newly formed microvascular networks originated from the host tissue or the grafts themselves.
The neonatal mice, from which the liver tissue was used for liver slice production, were two days old. The employed tissue masses were comparable: 0.33 ± 0.04 mg (neonatal liver slices), 0.53 ± 0.1 mg (70% resected liver tissue) and 0.40 ± 0.04 mg (regular adult liver tissue).
Intravital fluorescence microscopy
The in vivo microscopy is described in more detail in Supporting Information.
Quantitative off-line analysis of the microscopic images was performed by means of the computer-assisted image analysis system CapImage (Zeintl, Heidelberg, Germany) as described in Supporting Information.
Histology and immunhistochemistry
These techniques are described in detail in Supporting Information.
This method is described detailed in Supporting Information.
All data are expressed as means ± SEM. After testing for normality and equal variance across the groups, differences between the three transplant groups and time points were assessed by two-way ANOVA followed by the appropriate post hoc comparison. For reasons of clarity and comprehensiveness, only statistically significant differences between groups are indicated in the figures and tables. Statistical significance was set at p < 0.05. Statistics were performed using the software package Sigma-Stat (Jandel Corporation, San Rafael, CA, USA).
Vascularization of free liver transplants
In this study, “living” liver slices from different donor mice were syngeneically transplanted into dorsal skinfold chambers of recipient animals (Figure 1). This approach allowed for the first time the repetitive in vivo imaging of angiogenesis and vascular network formation of free liver transplants by means of intravital fluorescence microscopy.
Of interest we found that liver grafts from regular adult mice showed first characteristic signs of angiogenesis, i.e. bud and sprout formation, at day 5–7 after transplantation, whereas grafts from 70% liver resected and neonatal mice already vascularized at day 3 (Figure 2). Moreover, the take rate of the transplants was only 87.5% for the adult transplant group, whereas the take rate of the 70% resected and neonatal grafts was 100%.
Representative intravital microscopic images at day 14 after transplantation clearly revealed a higher microvessel network density in the 70% resected and neonatal transplant group when compared to the regular adult transplant group (Figures 3A-C). The quantitative analysis indicated in more detail that the morphological microvessel density of the neonatal transplants was significantly higher over the whole observation period of 14 days in comparison to both other experimental groups (Figure 3DD). At day 5 after transplantation, the neonatal graft exhibited a microvessel density of ∼210 cm/cm2, compared to ∼155 cm/cm2 in 70% resected and ∼40 cm/cm2 in adult grafts. At day 14 after transplantation grafts of the 70% resected transplant group (267 ± 14 cm/cm2) reached the density values of the neonatal grafts (259 ± 36 cm/cm2; Figure 3DD), while regular adult grafts exhibited a significantly lower microvessel density not exceeding 150 cm/cm2.
In accordance with the increased morphological microvessel density, the fraction of functional, i.e. perfused, capillaries was also significantly higher in the neonatal transplants when compared to both other transplant groups (Figure 3EE). Again, at day 14 after transplantation comparable values of functional capillary density could be found in neonatal and remnant proliferating but not in adult transplants (Figure 3EE).
Since vessel maturation is associated with a decrease of vascular diameter, we determined the diameter of the newly formed microvessels in order to define the degree of vessel maturity (Figures 3F and G). Whereas the capillary diameters of the regular adult and the 70% resected transplant group displayed initial values of ∼50 and ∼37 μm, respectively, the microvessel diameters in the neonatal grafts ranged around ∼20 μm (Figure 3FF). The functional microvessels showed a decrease of the diameter from initially ∼17 μm (day 3 after transplantation) to ∼11 μm (day 14) in the neonatal transplant group, compared to ∼19 μm (day 3) and ∼10 μm (day 14) in the 70% resected grafts (Figure 3G). In regular adult grafts no functional microvessels were detected at day 3 after transplantation, and moreover from day 5 until day 14 they displayed diameters with slightly increasing values (9.0–13.6 μm; Figure 3GG). The centerline velocity of microvessels within the neonatal transplant group was significantly increased at day 14 when compared to that of both other transplant groups, whereas the volumetric blood flow in microvessels of all three different grafts was not significantly differing (Table 1).
Table 1. Diameter, velocity and volumetric blood flow of newly formed microvessels 3, 5, 7, 10 and 14 days after transplantation
Centerline Velocity (μm/s)
Volumetric blood flow (pl/s)
All data are given as means ± SEM with n = 7–9 grafts for each time point. A two-way ANOVA followed by Holm-Sidak procedure was performed to proof statistical significance. *p < 0.05 versus adult; #p < 0.05 versus 70% resected at the corresponding time points.
18 ± 6
60 ± 2
89 ± 50
33 ± 20
9 ± 0
17 ± 2
18 ± 1
69 ± 4
134 ± 28
124 ± 37
4 ± 1
31 ± 6
30 ± 11
18 ± 3
17 ± 2
145 ± 37
164 ± 21
181 ± 35
39 ± 10
37 ± 10
35 ± 8
19 ± 1
15 ± 2
13 ± 1
238 ± 60
223 ± 54
244 ± 63
63 ± 10
54 ± 14
36 ± 13
19 ± 2
13 ± 1
14 ± 2
224 ± 42
243 ± 47
346 ± 93*#
32 ± 9
58 ± 17
To better elucidate the origin of the newly formed microvascular network, we employed an eGFP+/eGFP− donor/recipient cross-over design. eGFP+ endothelial cell lining of the newly formed microvascular network indicates that the vessels originate from proliferating cells within the grafts. eGFP− endothelial cell lining indicates that the vessels originate from the host tissue. Therefore we prepared regular adult, remnant proliferating and neonatal “living” liver slices from eGFP+ donor mice. In all transplant groups eGFP+-lined endothelial sprouts could be detected, as shown in Figure 4. These sprouts connected to angiogenic sprouts from the host tissue (eGFP−, black vessels without any fluorescence), and showed to some extent chimeric characteristics, i.e. a discontinuous distribution of eGFP+ endothelial cells (Figures 4B and C), an effect which is also visualized by electron microscopic images showing both, eGFP+ and eGFP− endothelial cell lining within the same vessel (Figure 4D). Besides the chimeric structure only eGFP+ or only eGFP− lined vessels were also detected (Figures 4E and F). The newly formed vessels were connected to the host's microvasculature and perfused as determined by the application of FITC-dextran for contrast enhancement of plasma (Figures 5A-C).
Determination of graft survival
For this purpose we employed histological and immunohistochemical processing of the transplanted grafts as well as electron microscopy after the 14 days of in vivo imaging. Firstly, H&E staining served to assess gross morphology of the grafts (Figure 6AA). By means of electron microscopy the presence of intact hepatocytes with bile canaliculi could be detected (Figure 6BB). To proof that the improved vascularization of the neonatal and 70% resected transplants compared to the regular adult transplants correlates with a better supply of oxygen, the grafts were stained immunohistochemically against HIF1alpha (Figures 7A-C). While in the adult transplant group multiple HIF1alpha positive cells could be detected, significantly less positive cells were present within the 70% resected and neonatal transplant group, indicating a better oxygen supply of the cells in the latter grafts (Figure 7DD). To directly correlate the microcirculatory data at day 14 (i.e. functional capillary density) with tissue viability at day 14 (i.e. HIF1alpha), a linear regression analysis was performed between both parameter, which revealed a correlation coefficient r of –0.69 and a p-value of 0.0006. This indicates that the higher functional capillary density is, the lower tissue hypoxia (HIF1alpha) is. Since tissue hypoxia represents a classical trigger for cellular apoptosis, we stained the tissue specimen against the effector caspase-3. Significantly more cleaved-caspase-3 positive cells could be detected within the adult transplant group, compared to both other transplant groups (Figures 8A-D), pointing towards hypoxia-induced cell apoptosis. The staining against the F4/80 antigen revealed diffuse cellular infiltration of the adult graft with macrophages (Figures 9A-D). In contrast, within the 70% resected and the neonatal grafts, F4/80 positive macrophages were less frequent and rather detected in the border zone of the transplants.
Hepatocyte vitality determined by cytochrome P450 1a1 immunohistochemical staining revealed significantly more active hepatocytes within the 70% resected and neonatal transplant group, when compared to the adult transplant group (Figure 10). When comparing regular adult, 70% resected and neonatal liver tissue before transplantation into the dorsal skinfold chamber, distinct differences could be detected only for the hepatocyte activity (cytochrome P450 1a1 staining) of 70% resected and neonatal liver compared to regular adult tissue, whereas tissue hypoxia, apoptosis and macrophage infiltration did not differ between grafts (Figure 11).
Inflammatory host tissue response
Visualization of rhodamine-stained leukocytes revealed physiologic leukocyte-endothelial cell interaction, i.e. a small fraction of rolling and a low number of firmly adherent cells in the newly formed microvascular networks (data not shown). This underscores the syngeneic transplant regimen with absence of inflammation and, thus, lack of signs of graft rejection.
Under normal physiological conditions angiogenesis in humans is tightly regulated [28, 29]. During free graft transplantation, an appropriate and timely induction of the angiogenic process is essential for an adequate oxygen and nutrient supply and consequently critical for the integration and survival of the grafts. Considerable progress in the field of tissue engineering during the past decades involved several strategies to stimulate vascularization of free transplanted 3D tissue constructs [30, 31]. Approaches reach from the utilization of matrices over chemical modification of scaffold surfaces and the generation of a defined porous architecture  to stimulate prevascularization and connection of cell seeded scaffolds via inosculation [31-34]. Although all techniques aim at a better vascularization of the tissue constructs, an adequate angiogenic response is still absent, but a prerequisite for consideration of an extrahepatic transplantation side as a clinical alternative in the treatment of liver failure or other liver associated metabolic diseases.
In the present study “living” liver slices were utilized as freely transplanted grafts within dorsal skinfold chambers with in vivo analysis of their vascularization by means of intravital fluorescence microscopy. For the study of the angiogenic process, the dorsal skinfold chamber offers a unique tool, because it allows for repetitive observation and determination of microcirculatory parameters over a prolonged time period of at least 2 weeks  as well as the transplantation of various “live” tissues or even foreign engineered tissue constructs. With regard to transplantation, implantation and angiogenesis, the dorsal skinfold chamber model has been already used extensively, proving its enormous suitability for these purposes [36-40].
The use of regular adult, 70% resected and neonatal liver transplants and their comparative investigation in terms of their vascularization characteristics revealed considerable differences of the angiogenic response. The earlier and more pronounced angiogenesis in the resected and the neonatal grafts might thereby rely on different reasons. The induction of the vascular network formation at day 3 after transplantation in the neonatal grafts is most likely related to the greater prevalence of stem cells in developing tissues [41, 42], while slices from liver tissue undergoing resection are exposed to numerous growth factors, like epidermal growth factor, hepatocyte growth factor or TGFα [43, 44], which most probably mediate the rapid microvascular network development [45-49]. In contrast to the other transplants, in the regular adult transplants no particular stimulation of the graft-confined cells might be existent, except of free-transplantation associated graft hypoxia. Thus, timely and adequate vascular connection of the graft to the host tissue with preservation of intact hepatocytes could be observed in neonatal and 70% resected grafts, while vital hepatocytes were less numerous in adult liver tissue grafts.
This study clearly shows that freely transplanted neonatal and resected liver tissues have a significant advantage staying vital for a certain time after transplantation compared to regular adult liver tissue. Therefore, one conclusion for a successful clinical implementation of an extrahepatic transplantation approach would be to utilize neonatal or proliferative stimulated tissue, which, however, is limited to realize due to ethical concerns. Another option is the transformation and stimulation of regular adult and stationary liver tissue into proliferative and versatile liver cell composites. For this purpose an understanding of the underlying signal transduction pathways and deeper knowledge of the presence of classic growth factors involved in angiogenesis is important. One responsible candidate for the more prominent angiogenic response within the neonatal and hepatectomized liver tissue could be, for example, VEGF. VEGF is the most prominent inducer of an angiogenic response and its augmented activity is well described for hepatectomized animals . In particular VEGFA is important for blood vessel formation . Therefore, it might be possible that in neonatal or hepatectomized liver tissue increased VEGFA levels can be detected. Vice versa application of VEGFA might stimulate regular adult liver tissue to enhance vascular development and new blood vessel formation. Furthermore, the stimulation of certain signal transduction pathways might be a possible method to potentiate the angiogenic response in order to improve cellular graft survival. This aspect deserves to be investigated and the approach here applied, i.e. employing in vivo graft observation within dorsal skinfold chambers, would allow for further examination in those directions. Relating to this, important roles for Notch signaling in mediating embryonic vascular development and normal vascular remodeling  as well as for liver regeneration are described . Kohler et al. observed an upregulation of Notch receptor-1 and its ligand Jagged-1 after partial hepatectomy, and also an increased nuclear translocation of the intracellular Notch followed by an upregulation of Notch-related target genes Hes-1 and Hes-5 . Notch signaling pathway is reported to control endothelial cell differentiation, arteriovenous specification and vascular development . Since Notch signaling regulates tip cell formation and function during sprouting angiogenesis , those genes might also be candidates to investigate for playing a role in angiogenic development and regeneration.
Looking at the vascularization characteristics, i.e. the vascular sprouting, network formation and vessel maturation, as it is described for other free transplanted tissues [56-58], neonatal, resected and adult transplants showed the same occurrence order, starting with sprouting of vessels from both, the transplant and the host tissue, subsequent network formation and finally maturation, i.e. perfusion of the newly developed vasculature. The sprouting of vessels from the grafts themselves and connection to the host's vasculature outside the graft is termed external inosculation [31, 59] and accounts probably for a more rapid vascular communication between graft and host tissue. This effect is particularly important in the field of tissue engineering in order to assure adequate angiogenesis. The chimeric nature of eGFP+ and eGFP− endothelial cell lining observed within the newly developed vascular network underscores even more that the vasculature originates from both, the graft and the host tissue. However, specific characteristics of liver vasculature, like fenestration of the endothelial lining and the absence of a basal lamina, even though their existence is discussed controversially in the literature [60, 61], were not observed in the examined grafts during our observation time. In the first instance an appropriate supply of the grafts with oxygen and nutrients was intended by formation of a new vascular structure, a possible specification into liver specific endothelial cell lining might occur at later time points, as it has been described for the vascular development and differentiation during liver organogenesis .
Microhemodynamic parameters of the newly developed vascular networks within the transplants revealed that neonatal grafts were already perfused at the early time point of 3 days after transplantation, whereas microvessels of the resected and adult grafts showed no perfusion. This might be explained on the one hand by the fact that from the beginning more mature vessels, at least indicated by a smaller diameter, are present in the neonatal grafts, which were perfused immediately. On the other hand it might be possible that the neonatal microvessels show no decrease in diameter during vessel maturation or that immature microvessels are perfused at this early time point in the neonatal transplants.
In summary, we could demonstrate that the neonatal origin of the liver transplant is of major impact for the kinetics and extent of revascularization when compared to regular adult or proliferative stimulated (resected) transplants. This is important for preservation of functional transplant tissue in order to assure restoration of organ function with respect to subsequent clinical implementation. To identify more specifically the underlying mechanisms of the induction of the angiogenic process is highly relevant for the interdisciplinary field of tissue engineering, as until now the functionality of free transplanted tissue constructs is a major limiting factor for successful clinical application. Therefore further investigations should be directed towards an enhancement of the angiogenic response in freely transplanted grafts, by activation of particular signaling pathways identified to promote vascularization.
This work was partly funded and supported by the FORUN Project No. 889015 (University of Rostock) and also in part by the Bundesministerium für Bildung und Forschung (BMBF), Germany (01GN0986). The authors kindly thank Prof. Dr. R. Koehling (Department of Physiology, University of Rostock) for providing his vibratome and Dorothea Frenz (Institute for Experimental Surgery) and Ute Schulz (Electron Microscopy Center) for their excellent technical assistance.
The authors of this manuscript have no conflicts of interest to disclose as described by the American Journal of Transplantation.