General information

Introduction

The visibility, accessibility, and rapid developmental growth of the chorioallantoic membrane (CAM) offer clear advantages to study and manipulate vascular functions. The first use of the system dates of 1911 when Rous and Murphy demonstrated the growth of chicken sarcoma tumors transplanted onto the CAM [1]. One year later, Murphy published successful heterologous transplantations of tumors in this model [2].

In 1930s the CAM was first exploited for cultivation of viruses and bacteria [3–5]. Its application in vascular-related questions occurred only several decades later, but it quickly gained recognition as an excellent model system for studies associated with vascular responses.

The initial vascular-related experiments employed grafts from tumors [1, 2] and later methylcellulose discs in which compounds were encapsulated to determine induction of vessel growth or inhibition by morphological criteria [6–8]. These discs provided a venue for the slow release of pro- or anti- angiogenic factors and thus, a “high” throughput platform for therapeutic exploration emerged. Problems with this method were associated with the difficulty in distinguishing the existing from the newly-formed (or induced) capillaries. In addition, the methylcellulose discs by itself could, on occasion, induce angiogenesis even in the absence of growth factors, particularly if crystals of salt were present. Thus, the approach was subsequently replaced by 3D models of vascular invasion and other alternatives [9, 10].

In addition to angiogenesis screens, the CAM has been used to explore hemodynamics, immune cell trafficking, transplantation and responses to therapy. In this review we will provide a comprehensive summary of the use of this model to study vascular-related questions bringing into this description a rigorous evaluation of the pros and cons.

Structure and development of the CAM

The CAM develops in a similar manner as the allantois in mammals as it extends extra embryonically from the ventral wall of the endodermal hind-gut (Fig. 1). However, in birds it also fuses with the chorion to form the chorioallantoic membrane. Initially avascular, it quickly gains a rich vascular plexus (Fig. 2a, b) that rapidly acquires hierarchic complexity with the emergence of arteries and veins (Fig. 2c). The growth of the CAM occurs from embryonic development day 3 (stage 18 of Hamburger and Hamilton, [11] ) and it is completed by day 10 (Fig. 3) however, it is only fully differentiated by day 13. This growth requires impressive rates of cell division with short cell cycles. While this has not been significantly explored, bromodeoxyuridine (BrdU) incorporation revealed rapid cell cycles and extensive proliferation (Fig. 4).

Fig. 1
figure 1

Chicken embryo and associated extra-embryonic structures. a Chicken embryo at day 4 post-fertilization. Note the highly vascularized chorioallantoic membrane (CAM) that expands from the hind-gut. The yolk sac membrane (YSM), also highly vascularized, is seen in the background; b Embryo at day 12 post-fertilization into a 10 cm petri dish. Note the significant growth of the CAM that now exceeds the area covered by the yolk. The YSM is also present, but under the embryo and only associated with the yolk

Full size image
Fig. 2
figure 2

Images illustrating the maturation of the CAM vasculature. a Primitive vascular plexus at day 3 post-fertilization; b Vascular remodeling, growth and anastomosis can be seen at day 7 post-fertilization; c Hierarchic vascular structures and fully differentiated vessels are noted at embryo development day 10

Full size image
Fig. 3
figure 3

Expansion of the CAM was measured daily. Graph shows the average of 5–7 CAMs per time point after plating the fertilized eggs into a 10 cm petri dish (±SD). Note the exponential growth between days 3 and 10

Full size image
Fig. 4
figure 4

Proliferation kinetics of CAM vessels. a Quantification of BrdU incorporation in CAM vessels at the indicated days post-fertilization. Counts per time point were the average of positive nuclei in the vessels from 3 to 4 CAMs per time point; b Example of images used to obtain the measurements. The upper panel represents embryo development at day 4 and lower panel is an example of embryo development at day 6

Full size image

The CAM functions as the respiratory organ of avian embryos. Although most of attention is given to the vascular system, it is important to emphasize that the CAM has a fully developed lymphatic system that holds remarkable functional and molecular similarities to mammalian lymphatics [12]. A more detailed description of lymphatic development in the CAM will be presented later in the review.

Histologically, the CAM consists of two epithelial sheets that limit a thin layer of stroma (Fig. 5). The upper epithelium is of ectodermal origin, while the stroma and the lower epithelium are of mesodermal and endodermal origin respectively. It is within the stroma that the blood vasculature and lymphatics reside. In this manner, it is important to realize that any compound delivered on the surface of the CAM has to pass across the surface epithelium and reach the vessels in the stroma. Poor humidity conditions when the CAMs are cultured ex ovo induce significant cell division and keratinization of this upper, and normally single, epithelial layer; making it even more difficult to deliver soluble molecules to the underlying vascular tree. Therefore the humidity status is critical during the experimental use of this system.

Fig. 5
figure 5

Immunohistochemical visualization of α-smooth muscle actin (dark brown) in a transversal CAM section. Smaller capillary plexus is more superficial in comparison to the larger vessels located deeper in the CAM. (Color figure online)

Full size image
Fig. 6
figure 6

Angiogram of the CAM lymphatic system. Lymphatic vessels around a large blood vessel were visualized by FITC-dextran fluorescence angiography (25 mg/kg; 20 kDa, λex = 470 ± 20 nm). At a more superficial plane small non-fluorescent blood vessels are also visible. Bar = 20 μm

Full size image
Fig. 7
figure 7

Images of human A2780 ovarian carcinoma grown on the CAM. a Bright field image; b Fluorescence angiography (FITC-dextran, λex = 470 ± 20 nm, λem = 520 nm). Functional vasculature of the CAM is integrated well in the tumor. Bar = 1 mm

Full size image

It is important not to mistake the CAM with the yolk membrane, which is also highly vascularized. This second membrane is associated with the yolk and it has distinct properties. Figure 1 clearly shows both membranes, that are easily distinguishable in the young embryos and less noticeable in an older embryo (Fig. 1). In the egg, the CAM attaches to the surface of the egg shell, removing calcium and obtaining oxygen for the atmosphere across the porous shell. The attachment of the CAM to the shell inner membrane occurs shortly after fertilization (day 4–5) and thus, opening fertilized eggs for cultivation has to be done by day 3, as after this time, rupture of the shell is associated with rupture of the CAM. Also when opening small windows for application of compounds at latter time points, meticulous dissection of the shell and its underlying membrane is necessary to expose the vascular CAM without damaging small vessels.

Advantages and disadvantages of the CAM model

As already stated, the CAM model offers multiple advantages over other in vivo models used to study angiogenesis and vascular biology. The most prominent advantage of the model is its accessibility and rapid growth (see Table 1). The CAM develops in a short time from a small avascular membrane into a structure that covers the entire inner surface of the shell displaying a densely organized vascular network. This process occurs over a period of only 7 days (day 3–10). It is this enormous angiogenic boost provides a wealth of information related to required developmental pathways. In addition, this model can also be used to study pathological processes by simple exposure to cytokines, hormones or drugs, or by transplantation of tissues, isolated cells or materials.

Table 1 Advantages and disadvantages of the CAM model as a screening platform
Full size table

The great accessibility of the CAM and the easy handling of the model for both intervention and imaging of the vasculature have attracted many researchers. In particular, the broad applicability of distinct imaging modalities that range from microscopic- to magnetic resonance- to positron emission tomography imaging (MRI and PET) offer multiple benefits. In addition, due to the transparency of its superficial layers, nearly any wavelength in the visible part of the electromagnetic spectrum can be used for fluorescence imaging. For applications related to transplantation, an important benefit is the slow developmental progression of the immune system that reaches physiological activity only by day 15 post-fertilization [13]. Thus, unwanted immunological reactions are limited, facilitating the transplantation of xenogeneic tissues, e.g. human tumors.

Last but not least, the cost-effectiveness of the model adds to the long-list of benefits. In fact, the price of a fertilized egg is 100 times lower than that of a mouse of a common strain.

Nonetheless there are also limitations. Perhaps the most prominent being the limits on the number of reagents compatible with avian species, including antibodies, cytokines and primers which might not work in chick tissues. Along these lines, genetic manipulation, albeit possible through virus, it is not as “easy” as in mice.

The CAM contains already a well-developed vascular network, which makes it difficult to discriminate between new capillaries and already existing ones. Another limitation is that if experiments extend after 15 days, a non-specific inflammatory reaction can occur that may limit the success of grafting. Such reactions may provoke angiogenic responses, which are difficult to distinguish from an angiogenic activity of grafted material [7].

Cultivation protocols

Several protocols are available for culturing avian embryos. The type of experiment to be conducted directs the choice of the protocol. In many cases, investigators rupture the egg and transfer the embryo to a 10 cm petri dish at day 3 post-fertilization. This ex ovo culture of the embryo, also called shell-less embryo culture offers ample observation and manipulation space, allowing for testing of several samples in a single CAM. For this approach, the eggs are first kept at 38 °C in a humidified incubator (days 0–3) on the side to ensure the position of the embryo is opposite to where the cracking of the shell occurs. Subsequently, a small whole is made on the side of the air chamber to equilibrate pressure and the egg is cracked into a petri dish. Naturally, the exposure of the entire embryo and yolk requires sterile conditions. Also removing the embryo from the shell is associated with a low survival rate, due to the frequent rupture of the yolk membrane either during or after culture. This occurs due to excessive distension of the yolk in the flat dish (Fig. 1b) [6, 8]. Ex ovo survival rates can be drastically improved through reducing the impact during the cracking of the egg shell and placing the embryo in a confined curved space rather than a flat petri dish for culture. Gentle cracking of the egg can be achieved using a circular saw rather than cracking the egg against a sharp edge (Fig. 8a). Alternative cultivating dishes include large weight boats (Fig. 8b) and plastic inserts that can be placed inside 100 mm petri dishes to introduce a curvature to the dish (Fig. 8c). In addition, it is paramount to keep the humidity in the incubator at 98 % or higher and to maintain the culture environment sterile. Using these approaches the viability is around 50 % at day 14. For a demonstration of a refined chick ex ovo culture and CAM assay with survival rates over 50 % refer to the video article of Dohle et al. [14].

Fig. 8
figure 8

CAM cultivation protocols. The e x ovo protocol (a–c) requires (a) a gentle cracking of the egg using a blade and transferring of the contents into a petri dish or weighing boats (b) with plastic a plastic cover (c). The in ovo cultivation (d–f) starts with gentle rotation for 3 days. On embryo development day 3 the eggshell is opened with a sterile tweezers (d) and subsequently covered with a laboratory wrapping film (e). On the day when the experiment starts, the shell above the air pouch of the egg is extended to a diameter of approximately 3 cm (f) enabling further experimental manipulation

Full size image

In ovo cultivation is an important alternative that improves significantly the survival of the embryos. Fertilized eggs are transferred to the incubator equipped with an automatic rotator for 3 days. Turning prevents the embryo from sticking to the shell membranes, as it would if it is left in one position too long. Good results can be obtained by turning the eggs the first thing in the morning, again at noon, and the last thing at night. Alternatively, the use of an automatic rotator enables slow but constant movement of the eggs.

The process of in ovo cultivation is initiated also at day 3 post-fertilization. At this time, a hole of approximately 3 mm in diameter should be created in the eggshell with a sterile tweezers and covered with a laboratory wrapping film to prevent dehydration and possible infections (Fig. 8d). The eggs are then returned to the incubator with a relative air humidity of 65 % and a temperature of 37 °C in a static position until use (Fig. 8e). This initial small incision changes the pressure inside the egg and prevents binding of the CAM with the shell membrane. At day 7 or later, the hole is extended to a diameter of approximately 3 cm in order to provide access to the chorioallantoic membrane (Fig. 8f). This approach still enables experimentation through a window and it offers an almost unchanged physiological environment for the developing embryo.

Depending on the country and regulations that provide oversight to the use of animals, avian embryos are allowed to develop until the hatching stage, usually at day 21 after fertilization. This cultivation protocol has been used broadly e.g. to study vascular development and (pro/anti) angiogenesis, tumor growth and metastatic progression, and to evaluate the effect of chemo-, radiation-, and photodynamic therapies, see Table 2.

Table 2 Broad applicability of the CAM in vascular assays
Full size table

CAM as a screening platform

The CAM has served as an in vivo platform for experimentation for more than 50 years. This system has been used for the study of vascular development and angiogenesis, tumor growth and metastasis [32], respiratory properties [33], ion transport [34, 35], selective vascular occlusion therapies, biocompatibility of engineered materials, drug distribution and toxicology. Next we review this literature and delve into the methodology used to provide readers with choices for future experimentation.

Vascular development and angiogenesis

As previously discussed, because of its rapid vascular growth, the CAM has been the model of choice to evaluate the effect of a wide number of compounds on growing vessels. Many pro- and anti-angiogenic agents have been tested by quantifying the morphological responses of the CAM vasculature. These compounds have included growth factors [6, 36], hormones [37, 38], natural molecules, anti-cancer agents [39], gases [40], organometallic compounds [41–43], pro-angiogenic molecules [37], antibiotics [44], antibodies and synthetic small molecules [7, 36, 45, 46]. The efficacy of such compounds has been evaluated using both qualitative and quantitative methods based on the assessment of the vascular morphology and density. Initially, this was done by scoring the extent of vascularization on a graded scale of 0–4 [47]. Alternatively, vascular growth was also assessed by quantifying the number of vessels that invaded, against gravity, an avascular matrix in which pro-angiogenic compounds were casted (Fig. 9a–c) [9, 10]. Injection of either fluorescent compounds or contrasting agents (India ink) favored visualization and eased the quantification (Fig. 9d).

Fig. 9
figure 9

Angiogenesis assays in the CAM. a Scheme of angiogenic growth into an avascular matrix. This assay requires growth against gravity and invasion into specific polymerized extracellular matrices. A variety of matrices can be used for this purpose; b Nylon mesh squares (pore size 250 μm; Tetko Inc) supporting a polymerized collagen gel containing FGF2 were placed on top of the CAM. The matrix on the left includes, in addition to FGF2, an anti-angiogenic factor (Thrombospondin1) the one on the right has a mixture of FGF2 and VEGF. The CAM was injected with india ink to visualized small vessels; c Higher magnification showing the small vessel inside the previously avascular matrix; d The CAM can also be injected with dextran-FITC to visualize angiogenic vessels

Full size image

More recently, a number of attractive imaging techniques (Table 3) were applied to the CAM to facilitate visualization and offer objective quantification. Even MRI and PET imaging have been used, but the resolution of these techniques might not be sufficient for the identification of vascular structures (Fig. 10).

Table 3 Imaging techniques used for the CAM and/or chicken embryo visualization
Full size table
Fig. 10
figure 10

Selected imaging techniques used to visualize the CAM vasculature of chicken embryo. a Confocal microscopy after blue fluorescent polymer injection, by Dr. Grace Lee and Dr. Steven, with permission from Mentzer’s lab [231]; b light microscopy of Smallpox virus pocks on the CAM (from Public Health Image Library); c Scanning electron microscopy (unpublished data, courtesy of Dr. Valentin Djonov, Bern, Switzerland); d fluorescence microscopy (with green fluorescent plasma marker and orange fluorescent flow tracers), by Dr. Kenji Chamoto and Dr. Steven Mentzer, with permission from Mentzer’s lab [231]; e non-invasive imaging of in vivo blood flow velocity using optical Doppler tomography, with permission from Chen et al. [48]; f Cells escaping CAM primary tumors were visualized in the embryos with the vasculature highlighted with red fluorescent-tagged LCA, with permission from Springer [73]; g Fluorescence (red) distribution in the vessels by means of laser scanning microscopy (phase contrast image (green), with permission from Elsevier [232]; h co-registered PET and CT images of a chicken embryo in the shell after 18F-Na injection, with permission. This research was originally published in JNM. Warnock et al. [55]. © by the Society of Nuclear Medicine and Molecular Imaging, Inc. (Color figure online)

Full size image

Several visual vessel-counting methods [61] or automated approaches [62–64] have been also developed (Table 4). Quantification, has frequently relied on descriptors such as vascular density [62, 65], vessel branching points/mm2 [64, 66], vascular length [67], or endpoint density/mm2 [68, 69], see Table 4. An example of automatic in vivo and in ovo quantification technique developed by Nowak-Sliwinska et al. is presented in Fig. 11.

Table 4 Summary quantification techniques to assess capillary networks
Full size table
Fig. 11
figure 11

Schematic representation of an image processing procedure used to characterize the CAM vascular network; a The original 16-bit image; b The same image after the “first phase” thresholding to extract main vessels; c The image after a posteriori shading correction filter; d The second parallel phase after the Laplacian filter; e The result after the conversion to an 8-bit image; f The result after the skeletonization process; g The corresponding histogram of non-vascularized meshes/mm2 (surface distribution of non-vascularized regions). Adapted from [64]

Full size image

Gene expression profiling associated with the physiological CAM development [70], as well as with the angiogenic switch during experimental tumor progression [71] has been reported, but the lack of initial sequence information for the chick has limited the use of this more molecular approach. Instead these types of studies have accelerated more in the retina (77).

Lymphangiogenesis

Lymphangiogenesis in the CAM emerges during mid-development (days 5–9 post-fertilization) following the initial growth of the blood vascular system and via expansion of blind-ended sprouts [79]. Isolated lymphatic endothelial progenitor cells are recruited to the tips of growing vessels and grafting quail-chicken chimera experiments showed unequivocally that the allantoic mesoderm has lymphangiogenic potential [12]. Concurrent to haemangiogenesis expansion, lymphatic vessels enlarge and split offspring vessels, accompanied by transient capillary expression of alpha smooth muscle actin (SMA) and recruitment of polarized mural progenitor cells. VEGF-A and hepatocyte growth factor (HGF) are well known angiogenesis inducers in normal and pathologic conditions. Cimpean et al. [80] showed that VEGF-A/HGF combination was able to promote a strong angiogenic response and expression of prospero homeobox protein 1 (Prox1) in the lymphatic endothelial cells of the CAM.

The lymphatics of the chorioallantoic membrane (CAM) are drained by lymphatic trunks of the umbilicus that are connected to the posterior lymph hearts. Intra-embryonic lymphatics are drained via paired thoracic ducts into the jugulo-subclavian junction. Similarly to mammalian lymphatics, the lymphatics of the CAM are characterized by an extremely thin endothelial lining, pores, and the absence of a basal lamina [81]. Molecularly, and in analogy to the mammalian system, lymphatic endothelial cells in the CAM are characterized by the expression of VEGFR-2 and -3 and are also Prox1 positive. Application of VEGF-C, the ligand of these two receptors, on the differentiated CAM, induces proliferation of lymphatic endothelial cells and formation of enlarged lymphatic sinuses. Under normal development, the sources of VEGF-C are the allantoic epithelium and the wall of the larger blood vessels. The fact that the lymphatics of the CAM travel along the side of the larger blood vessels (Fig. 6) indicate that the high level of VEGF-C from the wall of these vessels contributes to pattern the normal distribution of lymphatics in the CAM. The presence of a lymphatic system in the CAM opens the possibility of studying its function and response to a large number of compounds. Lymphatics have been shown to remodel in response to pathological conditions and play a critical role in tumor metastasis [82]. Papoutis et al. [83] studied the effects of two rat tumor cell lines (C6 glioma and 10AS pancreatic carcinoma) that differ in their VEGF-C expression profiles. Whereas the 10AS cell line expresses high levels of VEGF-C, the C6 glioma cells only weakly express VEGF-C. They showed that only the 10AS cells induce lymphangiogenesis.

Assessment of vascular permeability, tumor growth, and metastasis

Vessel permeability

The permeability of the vasculature can be assessed in real-time using the traffic of fluorescent dextrans of various molecular weights i.e. macromolecular drug carriers (70 kDa) and antibodies (150 kDa) into the tumor interstitial space. Whereas large dextrans of 2,000 kDa are sequestered within the lumen of the tumor vasculature [84, 85]. These alternative sizes of dextrans offer real-time monitoring power to study vascular leak in the CAM vasculature during the sequential development phases of endothelial proliferation (angiogenesis), cytodifferentiation and senescence [86–88]. It also provides the ability to determine the pharmacodynamics and pharmacokinetics of drugs diffusing from the CAM vasculature [89] or in response to engrafted tumor explants [85]. Regional and temporal differences in vessel permeability within the tumor microenvironment may also be captured at high resolution using an intravital imaging approach.

An enhancement of the vascular permeability in the CAM has been studied using VEGF [90] or techniques such as low-dose photodynamic therapy [20] or a low-frequency contrast material-enhanced ultrasound on the vascular endothelium [91]. In the latter studies diameter of affected vessels, number of extravasation sites, extravasation rate, area, and location, as well as changes in endothelial cells and basement membrane were evaluated.

Tumor growth, and metastasis

During past three decades several studies have reported successful use of the CAM in cancer biology and tumor angiogenesis (Table 5). The first report of an angiogenic response in the CAM was induced by Jensen Rat Sarcoma and was described in 1913 by Murphy [92]. To date many tumor cells placed on the CAM have been shown to engraft and follow all steps of tumor progression: grow, angiogenesis, invasion, extravasation and metastasis. The spontaneous metastatic events occurring after cell inoculation or intravasation into the vasculature has provided valuable information to understand tumor progression and metastatic potential of specific cell lines/primary tumor fragments.

Table 5 Tumor cell lines and tumor tissue successfully used in the CAM to study tumor growth
Full size table

Table 5 summarizes the cell types successfully grown on the CAM. Moreover, the vascularization of neoplastic and normal tissue fragments, either fresh or cryopreserved, grafted on the CAM has been evaluated in detail [71].

The highly vascularized network of the CAM sustains tumor viability and within a few days after cell grafting (Fig. 7) it can offer the means to study intravascular invasion of tumor cells in vivo [23, 73, 107]. In fact, trafficking of tumors cells in the chicken embryo has been shown by real-time imaging [73]. Furthermore, since the CAM is an isolated system, the half-life of many molecules such as small peptides tends to be much longer in comparison to mammalian models, allowing study of potential anti-metastatic compounds that are only available in small quantities [73, 93]. Table 6 summarizes tumor types shown to metastasize in the CAM.

Table 6 Tumor cell lines shown to exhibit spontaneous cell intravasation, vasculotropism and metastasis in the CAM
Full size table

Another mechanism of cell dissemination that takes place along the blood vasculature is called vasculotropism. In this process tumor cells appear to grow along the existent blood vessels, rather then randomly distributed within the CAM mesoderm. This was shown for human HT-1080 fibrosarcoma [110], human A2780 ovarian carcinoma (Nowak-Sliwinska P, unpublished data), and human melanoma [113].

Selective anti-vascular therapies and diagnosis

Photodynamic therapy

The well-vascularized surface of the CAM has several advantages as a model for studying the vascular effects of photodynamic therapy (PDT). PDT is based on the interaction of a systemically or topically applied photosensitizing agent (photosensitizer, PS) with molecular oxygen after irradiation with light of an appropriate wavelength [114]. The initial reaction, leading to generation of reactive singlet oxygen, activates a cascade of chemical and physiological responses, which results in a temporary or permanent occlusion of the irradiated vasculature [115]. The angio-occlusive effect of PDT in the CAM has been has been monitored and described in detail by Debefve et al. [51].

Various light sources [29] and different types of photosensitizers have been tested on the CAM model [116–119] enabling the comparison between different treatment conditions. The photosensitizers can be administrated in the CAM intraperitonealy [120], topically [121] or, most often, intravascularly [18, 20, 42, 116, 117, 122]. Since the photosensitizers are usually hydrophobic, to facilitate its administration, they are distributed in the form of liposomes [123] [117], biocompatible solvents like N-methyl-pyrrolidone, as well as 0.9 % sodium chloride [124]. In addition, polymeric nanoparticles were also used for this purpose (e.g. poly(lactide-co-glycolide, PLGA or PLA) [118]. An overview of delivery systems tested on the CAM is listed in Table 7.

Table 7 Drug delivery systems tested on the CAM
Full size table

Photodynamic therapy (PDT) is known to provoke an angiogenic response [115]. Therefore its combination with a variety of topically [19, 42, 140–142] or intravenously administered [32] anti-angiogenic compounds have been studied using the CAM for their ability to suppress vascular growth.

The CAM was also exploited to determine the relationship between the progression between tissue oxygen concentration during PDT and the resulting vascular damage. Measuring the local pO2 is also of interest for other types of treatments, such as radiotherapy or photodynamic therapy, as it is well established that tissue oxygenation plays an important role on the outcome of such treatments. The measurements of the pO2 in the CAM was performed, using a micro-O2-electrode [143], the delayed fluorescence [121] or luminescence quenching of a ruthenium derivative [144] monitored with a dedicated optical, fiber-based, time-resolved spectrometer or a fiber-optic oxygen sensor [32].

Low-dose PDT enhances vessel permeability, therefore might be used as a drug delivery method [145]. The perturbation of endothelial cell membranes by PDT results in the loss of tight junctions responsible for the sealing properties between adjacent cells, and consequently leaving enlarged gaps between cells that increase the vascular permeability to macromolecules following PDT. This principle is currently being applied for targeted delivery of chemotherapy, as the injection of a reduced amount of a chemotherapeutic drug could be effective due to enhanced local delivery rendered possible by combining Visudyne®-PDT with aspirin [86].

Photodynamic diagnosis

Photodynamic diagnosis (PDD) of cancer using photosensitizers is one of the major advances as a new technique for early cancer detection. The clinical application of hypericin (HY) or aminolevulinic acid (ALA) and its derivative Hexvix® has been demonstrated to improve visualization of urological cancers as compared to white light cystoscopy [146]. This light-induced fluorescence after treatment is an experimental model used in the early diagnosis of cancerous disorders in urology. Malik et al. [147] conducted an in vivo study evaluating the ALA-based fluorescence diagnosis of endometric implants on the CAM. Saw et al. [148] reported on the hypericin-N-methyl pyrrolidone (HY-NMP) formulations in the CAM inoculated human bladder cancer cells as a potential fluorescence diagnostic agent. The NMP formulations investigated were able to produce significantly higher contrast for tumor tissues and at earlier time points than was possible with HY in albumin. MPEG-hexPLA micelles were successfully used as hypericin carriers for ovarian cancer diagnostics [149].

Laser photoangiolysis

Selective vascular ablation (photoangiolysis) using pulsed lasers that target hemoglobin is an effective therapeutic strategy for many lesions residing in the thin tissue that forms the vocal folds. However, vessel rupture with extravasation of blood reduces selectivity for vessels, which is frequently observed with the 0.45-ms, 585 nm pulsed dye laser. The CAM has been useful predicting the effects of photoangiolytic lasers on the vocal fold microvasculature. Specifically, the CAM has been explored to test the effects of distinct laser settings, modes of delivery, active cooling, and wavelengths. Such information has been essential for optimizing the effectiveness of lasers in treating laryngeal pathology while preserving vocal function. One example is the 980 nm Gold laser that achieved selective vessel coagulation [150]. In other study, the pulsed 532 nm potassium-titanyl-phosphate laser was used to achieve selective vessel destruction without rupture using the CAM under conditions similar to flexible laryngoscopic delivery of the laser in clinical practice. It was shown that this treatment was effective for ablating microcirculation while minimizing vessel wall rupture and hemorrhage [151].

Radiotherapy

A few reports have used the CAM in radiotherapy-related studies. Interestingly, chicken embryos are able to tolerate high dose of X-rays under single or multifractionated irradiation, an obvious advantage for these studies. Not surprisingly X rays have been shown to decrease the number of blood vessels, most probably due to vascular targeting within the first hours after irradiation [152]. However, it appears that X-rays promote multiple effects in the CAM: (1) induce early apoptosis of CAM cells, a property that can be exploited to screen radioprotective compounds [153], (2) modulate the synthesis and deposition of extracellular matrix proteins involved in regulating angiogenesis and (iii) affect angiogenesis induced by tumor cells implanted onto the CAM [154]. More recently, the CAM/embryos has been used to test the activity of radiosensitizers. For example, etanidazole, a well-known hypoxic cell radiosensitizer, was evaluated using tumor-bearing chick embryos [21].

Distinct modalities of X-ray beams have also been evaluated in the CAM. In particular, microplanar beams (microbeam radiation, MR) have been compared to broad beams (seamless radiation) [155]. MR promotes capillary damage, with tissue injury caused by insufficient blood supply. Vascular toxicity and physiological effects of MR depend on the stage of capillary maturation and appear within the first hour after irradiation. The authors noticed a dose dependent vacuolization of endothelial cells as well as growth retardation, resulting in a remarkable reduction in the vascular endpoint density. Finally, several preclinical studies in the CAM have used combination therapy and demonstrated an increased effectiveness when radiotherapy is combined with an anti-angiogenic treatment [156, 157].

Combined effects of paclitaxel and single or fractionated doses of ionizing radiation during angiogenesis were also evaluated [22]. X-rays suppression of angiogenesis includes anti-proliferative action and induction of nitric oxide (NO) which also inhibits vascular growth [158]. These results suggested that NO is involved in the anti-angiogenic mechanism of X-rays and that pharmacological manipulation of NO may offer an alternative therapeutic approach for treating pathological conditions involving angiogenesis [159].

Vascular disrupting agents

Vascular disrupting agents (VDA) that destroy irregular tumor blood vessels often exhibit an immediate impact on the vasculature of tumors provoking their collapse after a few applications. They are usually small molecule compounds that bind to the colchicine binding site of tubulin and cause extensive cytoskeletal rearrangements of the microtubules and the actin bundles in endothelial cells [160, 161]. Disruption of small blood vessels, as well as hemorrhages as a result of leaking or broken vessels were observed shortly after administration of VDAs, e.g. verubulin [162], or the microtubule-depolymerizing agent C9 [163]. Moreover, the growth of solid tumors on the surface of the CAM was successfully inhibited by VDAs, e.g. phenyl-3-(2-chloroethyl) urea [164]. New compounds with both, anti-angiogenic and vascular-disrupting properties have been also frequently tested in the CAM model (i.e. deoxypodophyllotoxin [165], vinblastine [166], combretastatin A-4 or its gold derivative [167] ).

Drug delivery systems, drug pharmacokinetics and biodistribution

Drug delivery systems (DDS) aim to distribute therapeutic agents to specific sites of injury or disease. Research on DDS includes: (1) the direct modification of the therapeutic agents to form drug conjugates, (2) the encapsulation of the therapeutic agents into micro or nanosized structures, and (3) the encapsulation or conjugation of the therapeutic agents into scaffolds. Most commonly, these different DDS are studied in vitro with cells in culture and then tested in pre-clinical animal models. However, the leap from in vitro cultured cells to in vivo pre-clinical animal models often leads to unsatisfactory performance which is frustrating due to the high cost of the pre-clinical studies. The CAM has been an excellent platform to test multiple DDS formulations concurrently, thus it has been used as an intermediate step between in vitro analysis and in vivo pre-clinical evaluation in mammals [168].

Pro-angiogenic drugs

In tissue repair, the controlled induction of angiogenesis directed at afflicted sites (ischemic tissue) is of great importance. To test the pro-angiogenic activity of different DDS, investigators frequently apply them directly onto the CAM surface and proceed to monitor the angiogenic response. These pro-angiogenic agents can be delivered as drug conjugates, as micro or nanoparticles, or as part of a tissue scaffold. For example, the delivery of growth factors such as VEGF121, VEGF165, and FGF-2 to induce an angiogenic activity has been extensively studied [6, 36, 128, 136]. Using the CAM assay, researchers have been able to screen different formulations to find that the individual delivery of each of these factors entrapped within a matrix results in the formation of highly leaky vessels [169], whereas the delivery of combinations of these factors or the immobilization of VEGF (either 121 or 165) yields vessels with normal morphology and adequate barrier performance (non-leaky vessels).

Anti-angiogenic drugs

The efficiency of formulations containing anti-angiogenic drugs with the CAM model may be easily evaluated by determining degree of vascular occlusion [30, 56, 125]. Efficacy of paclitaxel, delivered via various types of formulations, was studied on the CAM and elegantly reviewed by Vargas et al. [170]. Recently, surface-engineered dendrimers were used for tumor specific delivery of doxorubicin hydrochloride [138]. The later system enabled an initial rapid release followed by sustained release of doxorubicin hydrochloride with significant anti-angiogenic activity. Tetrac (tetraiodothyroacetic acid) covalently linked to poly (lactide-co-glycolide) as a nanoparticle arrested tumor-related angiogenesis and human renal cell carcinoma growth in the CAM model [131].

Biodistribution

Drug pharmacokinetics can be assessed using the blood sampling or selected organ extraction of fluorescence detection (in case of fluorescent drugs). The later method is widely used with direct fluorescence measurements in the CAM vasculature in vivo [20] or in ex vivo via tissue fluorescence analysis [171]. Due to the small organ size and low blood volume, both development-stage dependent [172] sensitive analysis methods should be used. Consequently, non-invasive methods such as biosensors (e.g. glucose [139]) for the analysis of drug concentration can be advantageous in this model.

Gas/ion transportation

As previously alluded to, the CAM functions as “the lung” of the avian embryo, and thus, it is responsible for gas exchange, but it also provides calcium transport from the eggshell into the embryo, it offers acid–base balance, and it mediates regulation of water, and electrolyte resorption from the allantoic cavity where urinary waste products are discharged.

The CAM is involved in the transport of sodium and chloride from the allantoic sac and calcium from the eggshell to the vasculature. Through dilation of the associated blood vessels (known as chorioallantoic vessels), the embryo is able to thermo-regulate and avoid overheating for long periods of time. Several studies related to gas and water exchange have been reported. For example, the functional role of hypoxia in vascular flow was nicely explored in the studies published by Lindgren and colleagues (128). Initially, the CAM performs gas-exchange functions via the area vasculosa, the main respiratory organ before day 6 [173], but this is gradually substituted by the CAM vessels. By day 14, the CAM capillary network is completely developed and by day 19 the embryonic lung ventilation starts as the CAM degenerates and the embryo hatches. Lindgren et al., showed that a one-time change in oxygenation of the external microenvironment between day 15 and day 19, could affect blood flow. Local hypoxia yield direct vasoconstriction, increasing the flow [174], demonstrating that ventilation is associated with perfusion.

In terms of calcium, the chorionic epithelial cells transport Ca2+ to the embryonic circulation at an impressive rate: 100 nmoles of calcium per hour for 1 cm2 of the CAM surface. The mechanism of this transport seems to rely on the sequestration of calcium within endosome-like vesicles upon the initial uptake phase [94]. Interestingly, in vitro CAM ectodermal cells exhibit faster calcium influx rates when compared to human erythrocytes suggesting alternative calcium entry mechanisms [175]. This unidirectional process was shown to be highly specific for Ca2+ and regulated by vitamins D and K [176]. A detailed review of transepithelial ion transport through the chorionic epithelium has been described elsewhere [35].

Allergenicity and toxicity

The worldwide objection to the use of animals in testing of irritants and cosmetics makes the development of a system that does not use mammals highly desirable. Below are some important examples.

Skin irritation

A number of in vitro methods to assess skin and eye irritation/corrosion have been developed as alternatives to the in vivo rabbit tests [177]. Human skin models, like EpiSkin™ and SkinEthic™, that closely mimic the biochemical and physiological properties of the upper parts of the human skin, have undergone extensive formal validation and acceptance procedures in order to be broadly applicable. In Europe, the CAM and the cornea pocket assay have also been accepted for this purpose. In the CAM, potential for irritation can be assessed with specific read-outs on the membrane and/or vessels (hemorrhage, lysis, coagulation, changes small vessel diameter [178]), which is then compared to the effect of 5 % sodium magnesium lauryl-myristyl-6-ethoxysulphate [179].

The chimeric skin-CAM system serves as a promising alternative to current animal-based sensitization approaches, particularly when using human grafts. The first report of a successful human skin graft onto the CAM was described by Goodpasture [58]. An important advantage of the CAM on skin engrafts as compared to in vitro models of organotypic skin explants is that the skin is maintained in a more “physiological status”, meaning dependent on vessels for its oxygenation. By having a continuous blood supply, the grafted skin remains viable for the same length of time as chick embryo survival. Furthermore, epidermal Langerhans cells (LC) are retained in the epidermis for several days after grafting into the CAM, unlike in organotypic explants [179]. Slodownik et al. used allergens to evaluate the LC as a measure of their allergic potency. All agents with known allergic potential induced statistically significant migration of LC [180].

Eye irritation

Ocular tissues, such as the cornea and conjunctiva, are susceptible to injuries and adverse effects, either from administered drugs or additives used in pharmaceutical products [181, 182]. Assessment of the toxicity of ophthalmic formulations and the potential for ocular irritation represents an essential step in the development of new ocular delivery systems.

The CAM has replaced the highly controversial “Draize eye irritancy test” used to determine the irritation potential of cosmetics by placing liquid, flake and powdered substances into the eyes of rabbits. The hen’s egg test (HET–CAM) provides a platform for conjunctival irritation testing, as it responds to irritant substances with an inflammatory reaction similar to that seen in conjunctival tissue [183–185]. Various studies included classification of several classes of irritants, ranging from non-irritant to severe, with high correlations between the systems. This development was encouraging since it suggested that in vitro models could provide valid alternatives for prediction of ocular toxicity [186]. After the application of the test substance (e.g. cholesterol or naltrexone), blood vessels are examined for irritant effects at different time points post-application. The evaluation includes assessment of hyperaemia, hemorrhage and clotting [187]. A time-dependent numerical score is then attributed to each test substance or formulation, to give an irritation score that ranges from slight to severe [187, 188]. A modification of the HET–CAM assay was introduced by Hagino et al. [189] who used trypan blue to determine degree of cell death. Trypan blue has been used in many biological systems to distinguish between living and dead cells. As the amount of trypan blue adsorbed by the CAM can be measured spectrophotometrically, the quantification is simple and reproducible. When compared by simple linear regression, the degree of trypan blue adsorbed by the CAM and that obtained from Draize eye irritation test in the cornea presented high correlation coefficients, indicating that the CAM can be an excellent choice for predicting corneal injury without compromising quality.

The CAM has also been used as a model in surgical retinal research [190]. In fact, several of the new surgical tools, techniques for coagulating retina, microinjection, cannulation techniques, and endoscopic surgery were first tested in the CAM and subsequently evaluated in animal models. Microrobots that are used for intravitreal delivery of as well as other surgical applications were first evaluated on the CAM [191] prior to further testing into a rabbit eye [192].

Organ transplantation and tissue engineering

Due to its immunodeficiency prior to day 15 post-fertilization [13], the CAM has been extensively used to evaluate allogeneic and xenogeneic transplantations. In fact, the CAM is considered as a suitable model to study the implantation/biocompatibility process and to explore differentiation of tissue implants. It has also been used to evaluate tissue remodeling within or in response to the implanted material, regeneration events, and neovascularization. These processes depend on the mechanical and physico-chemical properties of the implanted material [178]. Already in 1916, for the first time, multiple chicken tissues were grafted in the chicken embryos [193]. More recently, allogeneic and xenogeneic tissues were successfully grafted on the CAM, such as liver [194], proliferative endometrium [195], adrenal gland or cerebellum [196], or intact human skin [58, 177], see Table 8.

Table 8 The CAM model in transplantations and tissue engineering studies
Full size table

Transplanted tissues can survive and develop in the CAM by either peripheral anastomoses between implanted graft and original CAM vasculature or by new angiogenic growth from the CAM into the transplant. It has been demonstrated that the formation of peripheral anastomoses between pre-existing donor and host vessels is the most common, mechanism involved in the revascularization of embryonic grafts [71]. The assembly of graft and CAM blood vessels occurs between relatively large arteries or veins, resulting in chimeric vessels of varying morphology [203].

The CAM has been also used in stem cell research in which various extracellular matrices or synthetic polymer-derived scaffolds have been evaluated in an effort to construct desired tissue-like structures [173, 176, 198]. The combination of biocompatible materials and stem cells appears to be a future trend that might prove fruitful in comparison to other systems, in large part, due to the neovascular properties of the CAM.

To date tissue engineering has been successful in producing simple avascular tissues, such as skin and cartilage, which are thin enough for oxygen and other nutrients to diffuse passively. To sustain a tissue with large dimensions (greater than 400 μm in thickness), a vascular supply that effectively permeate the implant is essential [204]. However, caveats are also inherent to the system. In particular, inflammation and ulceration might occur in response to implants of non-biological materials. The next generation of tissue engineering products has been largely designed to promote timed release of growth factors and to achieve biocompatibility without promoting inflammation. Moreover, several technologies have been used to evaluate the biocompatibility and/or inflammatory response to biomaterials implanted [34]. More recently the use of mathematical models for tissue-engineered angiogenesis that takes into account the interactions between several of the participating cell types into a porous scaffold has provided important insights to more effectively test and evaluate models of interaction [198].

CAM as a tool to study vascular development

Physiological angiogenesis

The ease and accessibility of the CAM is undisputed. Thus, studies that aimed at evaluating the effect of environmental factors in vascular development have taken advantage of this system in a manner that could not be achieved in mammalian systems. Exposure of CAMs to different degrees of CO2 [205] and O2 [68, 206], for example, have demonstrated the impact of gases on vascular development. While hypoxia and hyperoxia will be discussed more in detail below, this constitutes a classical example of how the system can be exploited to clarify environmental effects on vascular growth. Similarly, hyperglycemia has been evaluated through a simple process: intravitellus glucose injection [61]. Interestingly, the investigators reported that hyperglycemia decreased angiogenesis as early as day 5 and found a significant increase in apoptosis and pericyte association as early as 2 days post-treatment [61]. These results are consistent with recent findings indicating the negative role of hyperglycemia in angiogenic responses through increase of reactive oxygen species and limitation of VEGFR2 at the cell surface [207].

The CAM has also been used to study localization and levels of molecules in relation to their angiogenic/quiescent status [40, 208–212]. Thus, it has served as a valuable model system in which to validate expression of relevant vascular molecular pathways. In addition, quail-chicken chimeras have been extensively used to answer questions related to vascular development. For example, the relative plasticity of arterial and venular endothelium has been tested using quail grafts into chicken coelom that acted as hosts. In these experiments, the investigators found that endothelial cells from the grafts of distinct vessels remain plastic even after the acquisition of molecular markers for arteries and veins [213].

Unfortunately, lack of avian-specific reagents (as well as the initially limited genomic information) provided hindrance to a more robust use the CAM in developmental studies of physiological angiogenesis. In this regard, the recent expansion of retroviral, lentiviral and adenoviral vectors for genomic modification of CAM cells has offered significant advantages to this system [214]. Thus, dominant negative constructs and other tools have been developed to explore the contribution of specific growth factors and signaling cascades in the CAM, in addition to the standard local administration of the growth factor. For example, using manipulation of FGF1 levels, Forough and colleagues showed that neovascularization stimulated by FGF1 could be reduced by selectively targeting AKT signaling [215].

Hypoxia and hyperoxia

Hypoxia during embryogenesis may induce changes in the development of some physiological regulatory systems, thereby causing permanent phenotypic changes in the embryo. Various levels of hypoxia at different time points during embryogenesis were found to affect both anatomical and physiological morphogenesis. Druyan et al. demonstrated that after the 12 h of hypoxic exposure on the CAM, expression of HIF1α, MMP2, VEGF, and VEGFR2 was significantly higher in hypoxic (17 % O2) embryos than in controls. In day 6 expression of HIF1α, VEGF, and FGF2 was significantly higher in hypoxic embryos than in controls [216].

Strick et al. conducted the morphometric measurements of chorioallantoic membrane vascularity in chicken eggs that were incubated in various oxygen atmospheres (12, 16, 21, 45, or 70 %) beginning on day 7, and vascularity was measured at day 14 [217]. They concluded that hypoxia stimulates and hyperoxia inhibits angiogenesis in the CAM in a dose-related manner. Burton and Palmer showed that hypoxia (17 % O2) reduced embryo and CAM growth, but accelerated maturation of the capillary plexus. Hyperoxia (40 % O2) had a less marked effect but appeared to retard the final invagination of the plexus, resulting in a thicker air-blood barrier [218].

Intussusceptive angiogenesis

While most of the growth of the CAM vasculature occurs through sprouting angiogenesis, towards the end of vascular expansion and associated with remodeling processes, intussusceptive microvascular growth also occurs. The process of intussusception is one by which capillaries are generated via the intravascular division of the lumen. This occurs through the growth of thin intraluminar walls (pillars) that proceed to fuse with the opposite side of the vessel. In this manner, the insertion of transcapillary pillars, provides a means to divide vessels further increasing the complexity and number of individual vessels in the vascular plexus. The pillars increase in thickness to include fibroblasts and extracellular matrix increasing in girth and ultimately contribute to separate the vascular segment into two fully independent vessels.

While pillars are discrete structures within the blood stream, relatively isolated from extravascular soluble tissue factors, that are not only identifiable by intravital 2D imaging, but amenable to detailed morphometric analysis. Recent studies have indicated that the number of pillars is the highest between day 8–10 of CAM development and that they are exquisitely responsive to levels of VEGF [219–221]. The pillars are lined with normal-appearing endothelium suggesting a normal responsiveness to intraluminal flow fields [222]. The effect of mechanical forces in shaping pillar geometry, suggested their potential role in initiating pillar formation and potentially explain why they only occur later in CAM development. Most pillars, particularly at vessel bifurcations, demonstrated only small changes in geometry during culture. Lee et al. [223] in their study on physical forces on the CAM concluded that these spatial dynamics reflect a responsiveness of blood vessels to intravascular flow fields and indicate the importance of mechanical forces in shaping both intussusceptive angiogenesis and vessel structure in the microcirculation.

Vasculogenic mimicry

Vasculogenic mimicry (VM) a process by which tumor cells acquire functional features of endothelium and contribute to the formation of the vascular wall has been successfully studied in the CAM [224]. Ribatti et al. [97] showed for the first time the occurrence on VM in the melanoma grafted CAM model after implanting B16-F10 melanoma cells. Since, VM represents an alternative mechanism of tumor vascularization in which tumor cells actively participate in the formation of blood vessels. Its demonstration in the CAM provides an important validation step that can be further explored.

Physical forces in the CAM

Hemodynamic forces, including vascular circumferential tension and shear stress, have been predicted to influence angiogenesis. Recently, Branum and colleagues tested this hypothesis directly by evaluating the effect of pharmacological induction of acute and chronic bradycardia in chicken embryos. Contrary to predictions, the investigators found that vascular density of smaller vessels was reduced, but larger vessels were not significantly affected [225].

In terms of vascular remodeling, physical forces are also known to contribute to the progression of the uniformed primary plexus into a hierarchic vascular tree. It is also known that arterialization of veins can occur when subjected to high shear stress forces. For example saphenous veins used in coronary bypass grafts to substitute occluded arteries show signs or arterialization. Using the CAM, investigators have explored whether molecular identity of arteries and veins is also governed by flow. In an elegant series of experiments le Noble and colleagues showed that arterial-venous differentiation is in fact a flow driven highly dynamic process that relies on the plasticity of the endothelium [226].

In addition to genetic and biochemical signaling, the expansion and remodeling of a vascular supply is controlled by microenvironmental signals such as mechanical forces conveyed by extracellular matrix (ECM) and by flow. Although capillary development is driven by angiogenic mitogens, cell sensitivity to them might be modulated by a complex set of physical interactions between cells and ECM that alter the cellular cytoskeleton [227]. Changes in capillary cell shape and function can be produced by altering ECM elasticity or topography, as these provide mechanical stresses that impact traction forces [228]. Mechanical forces may also stimulate vascular remodeling and capillary growth in vivo [229, 230]. Kilarski et al. [230] found that tissue tension is generated by activated fibroblasts or myofibroblasts during wound contraction followed by the induction of new vasculature. These mechanical forces pulled vessels from the preexisting vascular bed with functional circulation that expanded as an integral part of the growing granulation tissue through vessel enlargement and elongation.

The mechanical forces associated with blood flow include wall shear stress (the frictional force tangential to the vessel wall) and circumferential strain (a blood pressure related force perpendicular to the direction of flow); wall shear stress being the dominant mechanical force in the smooth and continuous flow of the peripheral CAM microcirculation [222]. Although wall shear stress cannot be directly measured, numerical simulations enable not only the calculation of wall shear stress but the mapping of these mechanical forces to the vessel wall. The spatial relationships revealed by these maps provide important insights into the role of mechanical forces in shaping the structure of vessels and alters their responses during pathological conditions.

Closing remarks

Despite enormous volume of information generated by in vitro tests, in vivo experimentation offers a stronger reliable value when assessing angiogenic responses. In vivo assays usually require special skills and ethical committee approvals, thereby limiting the number of tests that can readily be performed. As presented in this review, the CAM model has been used effectively by several disciplines including biology, medicine and bioengineering to determine vascular responses and to explore questions related to blood vessel and lymphatic morphogenesis and physiology. Numerous and continuous improvements in imaging and quantification continue to offer strengths to the model. Furthermore, while not mammalian, the CAM offers a valuable intermediate step that is both cost and time-efficient. Finally, considering the pressures with minimizing the use of animal models and following the “3R” rule (Replacement, Refinement and Reduction) the CAM provides an excellent intermediary step prior to pre-clinical evaluation.