Application of the chick embryo chorioallantoic membrane in neurosurgery disease.

The chick embryo chorioallantoic membrane (CAM) is a highly vascularized extraembryonic membrane. Because of its ease of accessibility, extensive vascularization and immunodeficient environment, the CAM has been broadly used in the oncology, biology, pharmacy, and tissue regeneration research. The present review summarizes the application of the CAM in neurosurgery disease research. We focused on the use of the CAM as an assay for the research of glioma, vascular anomalies, Moyamoya Disease, and the blood-brain barrier.

Introduction

The chick embryo chorioallantoic membrane (CAM) is an extraembryonic membrane. Its main function is to exchange gases and nutrients, which is supported by a dense capillary network 1-3. Due to its extensive vascularization and its ease of use, the CAM is a widely utilized research tool. Because of its extensive vascularization and easy accessibility, morphofunctional characteristics of angiogenesis in vivo as well as the efficacy and mechanisms of action of pro- and antiangiogenic molecules have been studied using the CAM 2-4. Due to the lack of a fully developed immune system in the chick embryo 3, 5, the CAM is suitable for tumor engraftment to study various aspects of the angiogenic and metastatic potential of human malignancies3, 4, such as glioma 6-8, colorectal cancer 9, 10, leukemia 11, ovarian cancer 4, prostate cancer 12, 13, and osteosarcoma14. In addition, the CAM has been used to test the biocompatibility of organic and inorganic materials 15-17, the ability of bacterial strains to invade epithelial barriers 18, vascular anomalies and Moyamoya Disease 19-24, the organogenetic capacity of xenograft or homograft 25, 26, the mechanisms of wound healing 27, 28, and the differentiation of the blood-brain barrier 29, 30. The aim of the present review is to summarize the application of the CAM in the field of neurosurgery research. We focused on the use of the CAM as a model for the research of glioma, vascular anomalies, Moyamoya Disease, and the blood-brain barrier.

General structure and function of the CAM

Chick embryo development lasts 21 days. At 3.5 days of incubation, the allantois of the chick embryo appears as an evagination from the ventral wall of the endodermal hindgut. At 4.5 days of incubation, the allantois pushes out of the body of the embryo into the extraembryonic coelom. Its proximal portion lies parallel and caudal to the yolk sac. The narrow proximal portion is the allantoic stalk and is wrapped around the umbilical artery and umbilical vein. The distal portion is full of allantoic fluid, and gradually separates from the embryo and enlarges. This is the allantoic vesicle, which rapidly enlarges from day four to day ten. Fluid accumulation distends the allantois so that its terminal portion resembles a balloon 2, 3, 31. An extensive morphometric investigation has shown rapid extension of the CAM surface area from 6 cm2 at day six to 65 cm2 at day 14 32. In this process, the mesodermal layer of the allantois fuses with the adjacent mesodermal layer of the chorion to form the CAM. The CAM adheres to the acellular inner shell membrane that almost surrounds the entire embryo and other egg contents 2, 3, 31, 33. Histologically, the CAM consists of three layers, the chorionic epithelium, which is the somatic mesoderm attached to the shell membrane, the allantoic epithelium which is the splanchnic mesoderm facing the allantoic cavity, and the intermediate mesodermal layer, which is between the chorionic and allantoic epithelium and enriched in blood vessels and stromal components (Fig. 1)2, 3, 31, 33.

Figure 1

Schematic drawing of the general structure of a 12 day old chick embryo in the egg and the three-tissue layers of the chick chorioallantoic membrane (CAM).

The general vascular system of the CAM consists of the capillary plexus located between the chorionic epithelium and allantoic epithelium, and the larger umbilical artery and vein within the allantoic vesicle (Fig. 2) 34. The umbilical artery emerges from the embryonic abdominal wall, branches into two primary chorioallantoic arteries, and then into six or seven generations. The earlier fifth and sixth generations of blood vessels are parallel to the CAM surface and the later generations change direction abruptly, passing almost vertically in the two-dimensional capillary plexus. The capillary plexus is drained by a single chorioallantoic vein to the embryo 35, 36.

Figure 2

The general vascular system of the chorioallantoic membrane (CAM). Capillary plexus of CAM of a 10 day old chick embryo (a) and the umbilical atery and umbilical vein free-floating within the allantoic vesicle (b).

The CAM blood vessels appear at day four as a network of immature, irregularly spaced tubules with large luminal diameters, scattered in the mesoderm between two layers of epithelium. The CAM vessels grow rapidly until day eight. Then gradually differentiate into the capillary plexus, arterioles, and venules 31, 35. These immature blood vessels lack a complete basal lamina and smooth muscle cells (SMC), so there is little ultrastructural distinction between capillaries, arterioles, and venules. At nine to thirteen days, arterial endothelial junctions become more extensive, and presumptive SMC and connective tissue become apparent 35, 37. Capillaries show no changes in the periendothelial space and veins accumulate SMC and connective tissue to a lesser extent than arteries. Arteries have developed a distinct muscular layer and a more complete basement membrane by day 14 to day 18 35, 37, 38.

The CAM has a very simple vascularized structure, but serves multiple functions during embryo development. The capillary plexus and the CAM confer a respiratory function 2, 39-41. In addition to the respiratory interchange of oxygen and carbon dioxide, it is involved in the calcium transport from the eggshell 42-46, acid-base homeostasis in the embryo 31, 47, 48, and ion and water reabsorption from the allantoic fluid 49, 50. The allantois also serves as reservoir for the waste products excreted by the embryo 2, 34.

Application of the CAM in glioma research

The CAM is especially well suited for xenograft experiments due to its immune-deficient environment and dense capillary network 51. Hurst, et. al. 52 first demonstrated that a human tumor of the central nervous system (meningioma) could be successfully transplanted to the CAM. Since then, Vogel, et. al. 53 directly transplanted numerous intracranial tumors, including cerebellar astrocytoma, glioblastoma multiforme, meningioma, and craniopharyngioma to the CAM. Seventy-one percent of the eggs initially survived the procedure, and 41% of the tumors were viable after one week. Microscopically, the first generation of the inoculated tumor resembled the parent tumors. Upon subsequent transplantation, specimens were necrotic and did not survive past the second or third generation 53.

Tereseviciute, et. al. 54 and Balciuniene, et. al. 51 evaluated the histology and survival of human glioblastoma tumors transplanted to the chicken CAM. The studies showed that glioblastoma cells not only survived in the CAM, but also appeared histologically similar to the original glioblastoma. Surprisingly, the xenograft's immune cells (e.g. macrophages) also survived. Tereseviciute observed invasion of avian vessels in the xenograft 54. However, Balciuniene observed that the transplanted glioblastoma remained an insulated unit that survived from nourishment of the chorioallantoic membrane by diffusion. They also observed minimal invasion of the tumor cells into chicken tissue 51.

Strojnik, et. al. 55 compared the CAM glioma and rat glioma models. The full ranges of cytological features of human glioblastoma were observed in both models. In the rat glioma model, the xenograft tumors were sharply demarcated from the surrounding brain tissue. In the CAM glioma model, smaller tumor nodules were observed growing within the connective tissue and vessel walls. Therefore, Strojnik, et.al. proposed that the CAM glioma model may be more suitable for tumor invasiveness studies than the rat glioma model 55. Furthermore, the CAM glioma model is a popular choice to investigate glioma angiogenesis, growth, invasiveness, and the screening of antitumor drugs because it is less expensive, has a shorter experimental duration, has a simple protocol, and has minor ethical concerns compared to the rat glioma model.

1. The CAM as an effective model to study glioma angiogenesis and antiangiogenesis

Angiogenesis is essential in establishing a network of blood vessels needed to supply oxygen and nutrients required for promoting the survival and invasiveness of glioma cells. Glioma cells release specific growth factors to activate endothelial cell growth of the host tissue in order to become highly vascularized 56. Increased vessel density around the xenograft has been repeatedly confirmed in the CAM glioma model 1, 57-59. Therefore, the CAM glioma model has been widely used to assess the effects of drugs or cell factors on angiogenesis. Table 1 summarizes the studies described in this section.

Table 1

The chick CAM as a model to study angiogenesis and antiangiogenesis of glioma.

Author	Intervention	Intervention effect
Hagedorn 59	RTKi	Antiangiogenesis
Auf G 65	Silencing IRE1	Antiangiogenesis
Nico B 58	Anti-Epo	Antiangiogenesis
Pen A60	IGFBP7	Antiangiogenesis
Jiang L63	Bmi-1-overexpressing	Angiogenesis
Martinho O57	RTKi	Antiangiogenesis
Miranda-Goncalves V8	CHC	Antiangiogenesis
He YQ56	DMPP	Antiangiogenesis

Abbreviations: RTKi: receptor tyrosine kinase inhibitors; IRE1: inositol-requiring enzyme 1α; Epo: erythropoietin; IGFBP7: insulin-like growth factor binding protein 7; Bmi-1: B-cell-specific Moloney murine leukemia virus integration site 1; CHC: Alpha-cyano-4-hidroxycinnamate; DMPP: dimethyl phenyl piperazine iodide.

Dimethyl phenyl piperazine iodide (DMPP) is antiangiogenic. In a dose-dependent manner, U87 glioma cell xenografts treated with DMPP had less growth and angiogenesis compared with the untreated cells 56. However, DMPP did not inhibit U87 cell proliferation in vitro, even at high concentration. This implies that the tumor regression observed in the xenografts was due to DMPP blocking angiogenesis 56.

Gliomas exhibit high glycolytic rates, and monocarboxylate transporters (MCTs) play a major role in the maintenance of the glycolytic metabolism through the proton-linked transmembrane transport of lactate. In the CAM glioma model, Alpha-cyano-4-hidroxycinnamate (CHC, MCT inhibitor) induced a decrease in the tumor perimeter and the number of blood vessels compared with the untreated group. However, treatment of the untransplanted CAM with CHC does not affect the angiogenesis of the CAM, indicating that CHC impaired tumor glycolytic metabolism 8.

Martinho, et. al. utilized the CAM glioblastoma model to study the effect of receptor tyrosine kinase inhibitors (RTKi) on angiogenesis and tumor proliferation in vivo. They observed that the RTKi reduced the number of blood vessels recruited to tumors and inhibited tumor growth. Combined with additional in vitro experiments, they identified new receptor tyrosine kinase (RTK) targets of RTKi. The CAM glioblastoma model could be used to identify potential predictive biomarkers to anti-RTK therapies and provide a basis for specific targeted therapies in the future 57.

Insulin-like growth factor binding protein 7 (IGFBP7) is a tumor suppressor gene, which has been proposed to participate in senescence, apoptosis and late phase angiogenesis of tumor cells 60. After treatment of the CAM U87 glioma model with IGFBP7, the sizes of formed vessels around the tumor were reduced by 40% compared with untreated tumors. Immunohistochemical examination confirmed that the presence of IGFBP7 promoted the recruitment and differentiation of vessel pericytes, which stabilized the tumor blood vessels and suppressed tumor growth 60.

Erythropoietin (Epo) is a cytokine with pleiotropic effects, which exhibit a wide variety of biological functions in non-hematopoietic tissues, such as angiogenesis 61. It has been observed that the expression of Epo is greatly increased in high-grade gliomas compared to normal brain, but the biological function was unclear 62. To determine whether Epo was stimulating angiogenesis in the gliomas, Nico, et. al. treated the CAM glioma model with an anti-Epo antibody. They observed that angiogenesis was inhibited, suggesting that Epo may play a role in glioma angiogenesis 58.

The CAM glioma model was also used to evaluate gene regulation on tumor angiogenesis. B-cell-specific Moloney murine leukemia virus integration site 1 (Bmi-1), an onco-protein, has been implicated in the progression of gliomas. Jiang, et. al. observed that Bmi-1-overexpressing glioma cells induced more second- and third-order vessels in the CAM model compared to control cells. This was due to the activation of the NF-kB signaling pathway by Bmi-1. NF-kB induces expression of vascular endothelial growth factor C (VEGF-C), which plays a pivotal role in tumor vascularization 63, 64.

Auf, et. al. inoculated inositol-requiring enzyme 1α (IRE1) dominant-negative U87 cells on the CAM. They observed the size of the tumor was significantly smaller and tumor angiogenesis was decreased compared with control cells. However, the IRE1 dominant-negative U87 cells also exhibited a highly invasive aspect arising from the tumor masses and spreading along blood and lymphatic vessels inside the CAM. The overall structure of the chicken membrane appeared disorganized compared to the control cells 65. This may suggest that some antiangiogenic therapy may have limited benefits in the treatment of glioma.

2. The CAM as a model to study glioma invasion

The CAM glioma model has been utilized to study the invasion potential of glioma (Table 2). The research has focused on the invasion of the chorionic epithelium and the blood vessels by the tumor cells. The transplanted glioma cells or specimens invaded the chorionic epithelium and grew into the mesenchymal connective tissue. Invasion into neighboring vessels and metastasis to distant organs has also been observed 1, 6, 54, 59, 66. Hagedorn, et. al. established a glioma progression model using the CAM. Glioma cell invasion into the CAM was clearly visible by immunostained sections. The CAM blood vessels served as guiding structures for migration of the glioma cells. Moreover, several cancer-progression genes (e.g. CYR61, LUM) were up regulated after 48 h post-implantation 59.

Table 2

The chick CAM as a model to study glioma invasion.

Author	Intervention	Intervention effect
Westhoff 6	NF-kB inhibition	Anti-invasion
Auf G 65	Silencing IRE1	Anti-invasion
Teruszkin 66	perillyl alcohol	Anti-metastatic

Abbreviations: IRE1: inositol-requiring enzyme 1α

After the establishment of the invasion model, Westhoff, et. al. used the CAM glioma model to demonstrate the importance of transcription factor NF-kB in glioma invasion. Inhibition of NF-kB, either by transfection with an IkBa superrepressor (SR) construct or pharmacologically by treatment with Disulfiram, greatly inhibited the invasion of the glioma cells into the CAM 6.

Inhibition of invasion has been studied using both genetic and pharmacological inhibitors. As mentioned previously, silencing of IRE1 in U87 glioma cells exhibited invasion and spreading along blood and lymphatic vessels inside the CAM. The overall structure of the chicken membrane was disorganized. Abnormal hyperplasic vessels were closely associated with the tumor nodules. 65. Teruszkin, et. al. demonstrated the anti-metastatic potential of perillyl alcohol on glioma C6 cells using the CAM model. C6 cells intrinsically metastasize when inoculated in the CAM model. C6 cells treated with concentrations of 0.3 and 0.03% perillyl alcohol did not invade the lower CAM, whereas the untreated cells did 66.

3. Other applications of the CAM in glioma research

The CAM glioma model has also been applied to developing new drugs and treatments for individual therapy (Table 3). Warnock, et. al. first imaged tumor glucose metabolism and protein synthesis in the CAM U87 glioblastoma model by positron emission tomography (PET) via catheterization of CAM blood vessels with fludeoxyglucose (18F-FDG). Warnock, et. al. proposed the use of this model to screen tracers and optimize imaging parameters 7.

Table 3

Other applications of CAM in the research of glioma.

Author	Applications
Shoin69	Screening sensitive chemotherapy drugs for the clinical individual treatment
Grodzik68	Testing new anti-glioma drug
Durupt67	Test anticancer effect of oncolytic adenoviruses
Warnock7	Screening of new tracers and optimization of imaging parameters

Durupt, et. al. used the CAM glioma model to test the anticancer effect of oncolytic adenoviruses. Histological and immunohistochemical analyses of tumor sections confirmed large areas of necrosis, adenovirus replication and syncytia formation in the glioma tumors. This suggests that the CAM glioma model is also a promising tool to be used in preclinical evaluation of oncolytic viruses 67. Carbon nanoparticles have the potential to become a new anti-glioma drug. Grodzik, et. al. showed that carbon nanoparticles inhibited angiogenesis, invasion, fibroblast growth factor-2 (FGF-2) and vascular endothelial growth factor (VEGF) expression 68.

Shoin, et. al. used the CAM glioma model to assess the efficacy of anticancer drugs with clinical effects in 21 patients with previously untreated malignant glioma. They observed an over-all predictive accuracy of 90% and concluded that the use of the chick CAM model was a good predictive test for tumor sensitivity to chemotherapeutics. 69.

Application of the CAM in vascular disease research

There is a close relationship between the abnormalities during the embryonic period and the formation of some congenital intracranial vascular anomalies, such as arteriovenous malformation (AVM). The dense capillary network and the characteristic of continuous angiogenesis of the CAM provide a model for studying this vascular anomaly. Jedelska, et. al. successfully cultivated two AVM biopsy specimens on the CAM. Lesions adjacent to the adipose tissues were used as a negative control. Both AVMs induced an angiogenic response that was particularly prominent after four to eight days, and the adipose tissue did not elicit any vascular reaction. Histopathologic examination revealed that the AVM tissue derived from the CAM retained similarity to the tissue of origin and the transplanted AVM tissue was reperfused with the CAM blood. This research showed that AVM can grow reliably on the CAM assay and has the propensity to induce angiogenesis and re-vascularization. The CAM AVM model may be a potential platform for studying the generative mechanism and therapy of AVM 19.

Several authors have demonstrated that injecting mouse aortic endothelial cells transfected with a retroviral expression vector harboring a human FGF2 cDNA (pZipFGF2 MAE cells) into the allantoic sac induced neovascularization and the appearance of hemangioma-like lesions on the CAM. Therefore, the CAM/FGF2-overexpressing pZipFGF2 MAE cell system may be a powerful tool to study the genesis of vascular lesions 19-22, 70.

Moyamoya Disease is a progressive intracranial vascular disorder particularly affecting the circle of Willis and its arteries that results most commonly in transient ischemic attacks 71. An indirect anastomosis operation is one of the effective operations to treat ischemic Moyamoya Disease. The blood supply used may be from different human tissues, such as the anastomotic branches of the scalp arteries, temporal muscle, omentum, galea, or dura. The best tissue is dependent on the ability to produce new vessels. In order to evaluate the adequacy of materials utilized for anastomosis in Moyamoya Disease, Imaizumi, et. al. compared the ability of different human tissues to induce angiogenesis on the CAM. Fresh tissue samples were obtained at the operation and inoculated on the CAM. Neovascularization was evaluated 96 hours later. The scalp artery exhibited the highest activity of angiogenesis, followed by the omentum and galea, while the temporal muscle, dura, and fat tissue showed poor activity 24. Another study showed that the outer layer of the epidural exhibited significantly higher angiogenic activity than the inner layer of the dura 23. This research has laid the theoretical foundation for indirect anastomosis in the treatment of ischemic Moyamoya Disease.

Application of the CAM blood brain barrier research

The CAM has also been applied to study the differentiation of the blood brain barrier (BBB). Janzer, et. al. transplanted purified astrocytes in the CAM and observed that the vessels associated with the grafts are impermeable to Evan's blue, a vascular tracer that does not pass through BBB vessels. However, the vessels associated with the meningeal- or skin-derived fibroblast grafts allowed visible leakage of Evan's blue. This is the first time that the astrocytes induced BBB features in the non-neural origin vessels on the CAM 72.

Lobrinus, et. al. observed that the CAM vessels can express neurothelin, a transmembrane glycoprotein of the immunoglobulin superfamily and a marker of chicken BBB-forming endothelial cells 29, 73, when exposed to astrocyte-conditioned medium but not when exposed to glioma- or endothelial cell-conditioned medium 74. This suggested that an unknown soluble factor secreted by astrocytes might play an important role in inducing expression of BBB properties in endothelial cells. However, Holash, et. al. observed that mature astrocytes do not induce expression of neurothelin or another molecular marker, the glucose transporter (GT) in the CAM vessels. However, when a fetal chick neural tube was grafted on the CAM, GT was induced in the endothelium 75, 76. Taken together, astrocyte-induced formation of the BBB remains controversial.

Conclusion

In summary, this review represents the application of CAM in the field of neurosurgery disease. The CAM is a valuable alternative to rodent in vivo models for the study of angiogenesis and invasiveness of glioma, gene and protein expression in tumor formation, screening antiglioma drugs, origin and treatment of vascular anomalies, and differentiation of the BBB. Compared with rodent models, the main advantages of the CAM in vivo assays are low cost, simplicity, short experimental duration, easy dynamic observation, minor ethical concerns and a naturally immunodeficient environment. There are some limitations in the application of the CAM, including nonspecific inflammatory reactions and less specific antibodies available. However, the complete sequencing of the chicken genome and the maturity of transgenic chicken technology will be helpful to develop more specific antibodies and give the CAM experimental model more attention.