Arterial and microvascular supply of cerebral hemispheres in the nude mouse revealed using corrosion casting and scanning electron microscopy

Abstract

Morphological analyses of cerebral vascularization are not only important for the characterization of the anatomical and physiological relationships between vascular and nervous tissue, but also required to understand structural modifications that occur in many pathological conditions affecting the brain. The aim of this study was to generate a three‐dimensional vascular map of the cerebral hemispheres in the nude mouse brain, a widely used animal model for studying tumour biology. We used the corrosion casting (CC) technique to isolate blood vessels from 30 nude mouse brains. All casts were analysed using scanning electron microscopy (SEM), which generated quantitative data regarding vessel length and diameter as well as inter‐vascular and inter‐branching distances. We identified three different topographical regions: (i) the cortical region, characterized by a superficial wide sheet of vessels giving rise to terminal perforant vessels that penetrate the grey matter; (ii) the inner part of the grey matter, in which dense capillary nets form many flake‐like structures extending towards the grey–white matter boundary, where perforant vessels finally change direction and form a well‐defined vascular sheet; and (iii) the white matter layer, characterized by a more disorganized vascular architecture. In this study, we demonstrate the accuracy of the CC‐SEM method in revealing the 3D‐topographical organization of the vascular network of the normal nude mouse brain. These baseline data will serve as a reference for future anatomical investigations of pathological alterations, such as tumour infiltrations, using the nude mouse model.

Introduction

Ink‐gelatin injections and the observation of specimens by light or confocal microscopy have been, for a long time, the most common methods to study brain vessel morphology. However, these approaches did not provide any definitive information on the topographical disposition of the vessels, and a detailed vascular map could only be imagined by combining distinct serial sections.

The corrosion casting (CC) method, combined with visualizations using scanning electron microscopy (SEM), crucially contributed to overcoming these limitations, allowing the observation of the overall three‐dimensional (3D) anatomy of the blood vessels of an organ without interference from any other non‐vascular tissues (Murakami, 1971; Northover et al. 1980; Kanzaki et al. 1982; Ohtani et al. 1983; Lametschwandtner et al. 1984, 1990; Schraufnagel, 1987; Hodde et al. 1990; Murai et al. 1992; Walocha et al. 2002, 2012; Sangiorgi et al. 2003; Krucker et al. 2006; Gorczyca et al. 2010).

The CC method consists of a polymerizing resin injection followed by the digestion of all the organic components around the vessels. The obtained specimens, which represent the imprints left on the cast, can then be analysed by SEM to better visualize the structural and ultra‐structural details concerning, for example, the endothelial cell membrane (Reidy & Levesque, 1977; Ohtani & Ohtani, 2000) or the indirect signs of capillary blood flow regulation systems (Lewis, 1957; Nakai et al. 1981; Rodriguez‐Baeza et al. 2003; Verli et al. 2007). As a consequence, SEM not only allows the highlighting of macroscopic differences between normal and pathological vessels, but also provides a powerful method for the detailed quantitative and qualitative analysis of 3D ultrastructural modifications (Miodonski et al. 1980; Miyoshi et al. 1995; Heinzer et al. 2006), which were previously visible only in two dimensions using transmission electron microscopy (Wells & Edelman, 1976; Mironov et al. 1994; Ohkuma et al. 1997; Secomb et al. 2000).

In this context, while the organization of cortical blood vessels in the human brain has been described both in physiological (Dahl, 1973; Levesque et al. 1979; Duvernoy et al. 1981; Reina‐De La Torre et al. 1998; Zagórska‐Swiezy et al. 2008) and in some pathological conditions, such as traumatic injuries, arteriovenous malformations (Rodriguez‐Baeza et al. 2003), and degenerative (Meyer et al. 2008) and ischaemic processes (Ohtake et al. 2004), the microvascular architecture of the nude mouse brain has not been reported so far. Moreover, the nude mouse is one of the most widely used animal models for studying tumour biology, because of its intrinsic immunodepression. As a consequence, the characterization of the normal 3D morphology and topography of medium‐sized and capillary vessels is of special interest as it allows for an improved anatomical perspective when assessing the pathological modifications induced by the neoplastic process.

The aim of this study is to draw a 3D anatomical map of the normal vascular architecture in the nude mouse brain cortex.

Materials and methods

Thirty athymic male BALB/C nude mice (Charles River, Calco, Italy) with a mean weight of 32.5 g were used. This study was carried out in accordance with the recommendations of the Guide for Care and Use of Animal Experimentation. The protocol was approved by the ethical committee of the University of Insubria, Varese, Italy.

The animals were killed with an intraperitoneal injection of 3 mL of ketamine and 0.5 mL of carbocaine. To improve blood fluidity, a subcutaneous injection of 0.1 mL of low‐weight heparin was administered 20 min before the procedure. Soon after the animals’ death, a latero‐cervical right incision was performed, the fascial and muscular layers were opened by a sharp scalpel, and the common carotid artery was exposed under a dissecting stereomicroscope (Leica WILD M3C). A 24G cannula was then inserted into the common carotid artery, driven up into the internal carotid artery and fixed to the vessel by two silk ligatures. The abdominal cavity was then opened by laparotomy, all viscera were displaced, and the inferior cava vein was exposed and punctured to permit blood outflow.

The pre‐casting treatment started with an infusion of 20 mL of heparinized solution (19 mL of 0.9% NaCl solution and 1 mL of heparin) to prevent blood clotting and 10 mL of phosphate‐buffered saline solution to clear the remaining blood from the vascular bed and minimize the amount of blood in the reservoirs.

We proceeded with the fixation of vessels by injecting 10 mL of Karnovsky solution (0.25% glutaraldehyde and 0.25% paraformaldehyde in 0.1 m Na‐cacodylate buffer at pH 7.2) to prevent leakage of the resin and reduce the modifications to the endothelial cells during the injection of the medium.

Approximately 10 mL of resin (MERCOX – SPI supplies) was mixed with 0.2 mL of catalyser (benzoyl peroxide) diluted (20%) with methyl methacrylate (MERCOX ‐ SPI supplies) and then injected by a peristaltic infusion pump very slowly (1 mL min⁻¹) through a cannula, until an increase in pressure was observed. The injection pressure was monitored by a manometer. When reflux from the venous vessel became evident, the injection pressure was reduced to almost zero. The cannula was then extracted from the common carotid artery, which was ligated to prevent any reflux of the resin.

After partial polymerization of the acrylic compound (approximately 10 min), the mouse was immersed in a warm water bath (60 °C) for 1 h to complete the hardening process. Afterwards, the entire head was explanted and immersed in a 15% KOH solution to digest all the tissue around the vessels. This solution was changed daily for approximately 1 week, with thorough washing in distilled water supplemented with 5 mL Mycostatin to prevent fungal growth.

The resulting casts were then dissected, while still immersed in distilled water, under a stereomicroscope with microscissors and surgical, neurosurgical and ophthalmic microscopic tools. The obtained specimens were treated for SEM observations as follows: dehydrated in graded alcohol, critical‐point dried in an Emitech K850 CPD apparatus, mounted on aluminium stubs on adhesive film, and coated with 10 nm of gold in an Emitech K250 sputter‐coater. Due to their size, some specimens required metallic bridges to maintain good conduction throughout the stub (Lametschwandtner et al. 1980) These bridges consisted of thin copper filaments attached to the aluminium stub and to the specimen with a silver paste. The specimens were then observed in a Philips XL‐30 FEG SEM at 10 kV.

3D micrometric analysis

We analysed 150 samples of brain vessel casts (five for each brain: two from each hemisphere and one from the deep white matter). All the samples were analysed by SEM 3D micrometric analysis software, and the data were recorded. SEM stereo images were input to proprietary surface reconstruction software, of which the implementation details are thoroughly discussed elsewhere (Reina‐De La Torre et al. 1998; Binagli et al. 2004). The software reads a stereo‐pair of SEM micrographs, selects a point on the first picture, identifies the same point on the second picture, computes the height of all the key points identified in both micrographs, and connects the thus‐obtained points with a Delaunay triangulation to reconstruct the spatial shape of the original specimen.

All VRML files were studied with a second proprietary software (MicroMetric v1.1.3, © 2004, M. Raspanti), which yields mathematically correct 3D measurements (Raspanti et al. 2005). The collected quantitative data were evaluated using SPSS statistical software.

Results

To simplify the topographical description of the 3D microvascular architecture of the nude mouse brain, we considered three major specialized regions, from the surface to the depth: (i) the cortical layer, corresponding to the grey matter; (ii) the subcortical layer, corresponding to the boundary between the grey and the white matter; and (iii) the white matter.

Figure 1A provides a graphical representation of the vessels within each investigated area: (A) cortical vessels; (B) perforant vessels; (C) vascular clews belonging to the cortical layer; (D) subcortical vessels belonging to the subcortical layer; and capillaries in the form of (E) disorganized vessels; or (F) vascular clews in the white matter layer.

Figure 1.

(a) Microvascular anatomy and topographic disposition of vessels in nude mouse brain, including (A) cortical vessels, (B) perforant vessels, (C) cortical units, (D) subcortical knee, (E) disorganized white matter capillaries, and (F) vascular clews. (b) Schematic representation of the vascular cortical architecture, including (a) major vessels, (b) medium‐sized vessels, and (c, d) terminal branches.

Grey matter: cortical vessels

Observing the vascular casts from the surface, a wide 3D network of vessels is clearly visible at a low magnification (Figs 1B and 2). We distinguished large‐sized major vessels (Fig. 1B–a), medium‐sized vessels arising from or draining into large‐sized major vessels (Fig. 1B–b), and another two series of terminal vessels (Fig. 1B–c,d). The SEM analysis of the vascular casts showed that these vessels form a vascular tree that extends mainly in two dimensions, forming a sheet along the surface of the brain (Fig. 2).

Figure 2.

Cortical vessels. The cortical plexus is clearly depicted as made up of a wide net of arteries (*) and veins (#) giving off many collaterals and perforant vessels all over the brain surface.

The casts of the cortical large‐sized major arteries are characterized by a longitudinal orientation of ovoid‐shaped endothelial cell nuclei imprints. The major axis of the cells is parallel to the major axis of the vessel. The mean diameter of these arteries is 117.8 ± 10.4 μm, which remains constant till terminal bifurcation. The mean inter‐branching distance (between the collateral origin sites) is 846 ± 5.7 μm. The casts of cortical veins are characterized by randomly placed endothelial cell nuclei imprints of a roundish shape. When these casts are broken, sections show an ovoid shape, sometimes flattened and stretched. The mean major diameter of these sections is 125 ± 7.4 μm.

The casts of medium‐sized arteries, appearing roundish when sectioned, have a mean diameter of 80.8 ± 7.5 μm and give rise, every 703 ± 191.4 μm, to collateral vessels with a mean branching angle of 48 °. The medium‐sized collecting veins have a mean diameter of 92 ± 6.6 μm.

The terminal vessels constitute the last subpial collateral segments of the vessels before entering the cortex. Their mean diameters are 49.5 ± 3.8 μm (Fig. 1B–c) and 20.7 ± 2.7 μm (Fig. 1B–d). The inter‐branching distances gradually decrease from proximal to distal, with a mean value of 47 ± 13.5 μm (Fig. 3).

Figure 3.

Cortical vein and cortical artery. (A) Cortical vein visible as a slightly collapsed structure with a variable diameter. (B) Cortical artery with a tortuous shape but with a quite constant diameter.

In some specimens, it was also possible to observe on the casts many longitudinal grooves caused by muscular and pericytial perivascular sphincters as well as several other circular grooves, always located at the origin of the collaterals (Fig. 4). Several other grooves left by intercellular junctions appear as polygonal impressions on the surface of the casts (Fig. 5).

Figure 4.

Sphincter constrictions, including (A) longitudinal and (B) circular constrictions.

Figure 5.

Ultrastructure of a cerebral medium‐sized artery. Ovoid impressions of endothelial cell nuclei (*) placed along the major axis of the vessels. It is also possible to observe the imprints of intercellular junctions (#) that disclose the real shape of endothelial cells facing the lumen.

Two types of perforating vessels arise from the last segment of the cortical arterial tree (Fig. 1B–d). The first type originates directly from the artery and forms a squared angle with the superficial vascular network. The second type is the terminal branch of the artery, which gradually enters the cortex (Fig. 6). At the level of the grey matter cortical layer, the perforating branches run parallel to each other deep into the tissue until the boundary between the grey and white matter. They enter the cortex for approximately 180/200 μm without yielding collaterals: their mean diameter is 15 μm and intervascular distance is 50 μm. The perforating branches, after giving rise to the capillary network of the cortex, continue until reaching the white matter layer of the subcortical region. The capillary architecture of the cortex is composed of thin vessels (8.3 μm in mean diameter) grouped in flake‐like structures, measuring approximately 200 μm in mean diameter (almost half of the mean distance between two perforant vessels; Fig. 7). The capillaries constituting these structures branch out several times and decrease in diameter before reaching the extremities of the territory of another vascular unit. Sometimes, at this level, it is also possible to observe termino‐terminal anastomosis between the capillaries of two vascular units forming a capillary cortical net.

Figure 6.

Perforant branches. (A) The first type of perforant artery that forms a sharp square angle with the origin vessel. (B) The second type of perforant artery that originates terminal to the cortical vessels.

Figure 7.

Cortical capillary unit. The specimen was dissected under the stereomicroscope, and disconnected from the other units to be better visualized. (A) Low magnification picturing the flake‐like structure with a dense spherical capillary net. (B) Note the afferent and efferent vessels going to and originating from the cortical unit itself.

Subcortical vascular layer: the boundary between grey and white matter

After providing vascular supply to the grey matter through cortical capillaries, the perforating vessels run deep into the grey matter and reach its inferior limit, in close contact with the white matter. At this level, they form a subcortical vascular sheet where they turn almost parallel to the cortical surface (Fig. 8). All these terminal vessels, measuring 13 μm in mean diameter, have a quite tortuous shape and constitute a well‐defined vascular sheet extending parallel to the cortical superficial vascular network. These vessels do not anastomose to each other, but they run for approximately 400 μm before ending in thin capillaries principally oriented towards the grey matter.

Figure 8.

Subcortical vessel. The vessel was carefully dissected from the other capillaries under the stereomicroscope to better visualize its shape. A cortical artery gives rise to a perforant vessel that when reaching the boundary between grey and white matter runs parallel to the cortical surface forming a vascular knee (*).

White matter: capillary vessels

The vascularization of the white matter originates from the capillary plexuses of the cortex or directly from some perforant vessels. This vascular network is characterized by a general architectural disorganization subserved by tortuous capillaries, which are freely distributed in space (Fig. 9). Moreover, these vessels do not follow any order and do not seem to be related to any specific cellular structure (Fig. 9B). The only visible examples of specialization consist of several isolated vascular clews 200 μm in diameter originating from a perforant cortical vessel and draining into two or more veins, possibly related to deep nuclei (Fig. 9A).

Figure 9.

White matter capillaries. (A) The vascular net of capillaries does not follow any precise organization. (B) A more organized vascular structure is characterized by an afferent and an efferent vessel sited in the white matter.

Discussion

Due to its immunodepressed status, the nude mouse is one of the most widely used animal models in experimental studies concerning tumour biology. For this reason, to better understand the angiogenic effects caused by implanted tumours as well as the modifications imposed on normal cerebral tissues and vessels by other pathological conditions, the accurate knowledge of the normal structure and 3D organization of vessels is of primary relevance.

Currently, the CC technique combined with SEM analysis constitutes the best method to investigate vascular architecture in an organ by focusing on its 3D structural and ultrastructural features without any interference from other organic tissues (Dahl, 1973; Duvernoy et al. 1981; Lee, 1995).

In this study, we used the CC‐SEM method to provide, for the first time to our knowledge, a detailed map of the normal micro‐vascularization in the nude mouse brain. We found that, in particular, medium‐sized and capillary vessels have specific characteristics that are strictly dependent on their topography and on the functional role of the cells they supply. In particular, from the cortex to the subcortical level, we observed the following structures:

1. cortical vessels (Fig. 2);

2. perforant vessels entering the grey matter (Fig. 6);

3. cortical capillaries organized in flake‐like vascular structures (Fig. 7);

4. subcortical vessels forming a vascular sheet in the boundary between the grey and white matter (Fig. 8); and

5. white matter capillaries or several more organized vascular clews (Fig. 9).

The cortical vascular sheet is composed of a wide network of arteries and veins extending throughout the cortical surface. These vessels are characterized by a high degree of inherent elasticity due to their intrinsic architecture. Moreover, a quantitative and qualitative distinction can be made between two types of cortical vessels. The first type consists of large arteries and veins, which lack collaterals and are mainly associated with blood transport into the cortex. The second group includes medium‐sized vessels and their collaterals, which may have a distributive role in the cerebral parenchyma. Moreover, as described by previous authors, the vascular architecture of the cortex seems to be organized very similarly to the cellular organization of cortical neurons (Rowbotham & Little, 1965; Reina‐De La Torre et al. 1998).

Our results showed that perforating branches do not emit any collateral vessels and have a constant perpendicular orientation with respect to the cortical surface (Fig. 6). It is likely that this disposition strongly facilitates the formation of well‐organized neuronal columns by diminishing the space occupied by the vascular support. In fact, the columnar orientation of neurons appears to be well‐suited to the longitudinal perforating branches (Rowbotham & Little, 1965). If these vessels had a different orientation, it would be difficult to optimize vascular support to the nervous cells. Soon after entering the cortex, these perforant branches do not give off many collaterals, probably because this area has just dendritic arborization, which does not require dense vascular support.

We observed that the deepest part of the grey matter region is characterized by a progressive increase in vascular density and by a reduction in inter‐vascular distance, leading to one of the most highly vascularized areas.

We observed the presence of ‘V‐shaped’ or circular grooves located at the perforating branches origin sites (Fig. 4). These structures might have the role of vascular sphincters that regulate blood flow to the different brain regions (Duvernoy et al. 1981). On the basis of this anatomical evidence, it could be hypothesized that the vasospastic phenomena encountered during several pathological cerebral processes could be related not only to the reduced size of the larger muscular arteries (SAH) but also to a sphincter mechanism visible in medium‐sized vessels (Murakami, 1971; Duvernoy et al. 1981; Motti et al. 1987).

The perforant vessels running within the subcortical region show a sudden modification in their orientation from vertical to parallel to the brain surface, thus forming an anatomical ‘vascular barrier’ between grey and white matter (Fig. 8). Interestingly, such a morphology has been found in many pathological conditions, such as metastases and abscesses, and this pattern explains the high frequency of haemorrhages after traumatic head injury (Mironov et al. 1994; Ohkuma et al. 1997; Rodriguez‐Baeza et al. 1998). We hypothesize that the knee formed by subcortical vessels at the boundary between the grey and white matter plays a regulatory role by reducing the rate and pressure of blood flow, and thus making it easier for metastatic cells or germs to grow in this area. In addition, due to its spatial bidimensional distribution, this vascular sheet may represent a ‘locus minoris resistentiae’ leading to the micro‐haemorrhagic lesions (i.e. gliding contusions) typically observed after acceleration–deceleration brain injuries (Ohkuma et al. 1997).

Going deeper into the white matter, the high vascular density observed in the cortex is replaced by a more disorganized vascular architecture of tortuous capillaries, with no signs of specialization (Hodde & Nowell, 1980; Fig. 9).

The main limitation of this study consists in the possibility of artefacts due to traumatic damage to the casts, obstruction of the vessels secondary to blood clots, or strong spasms of the arterial wall. These factors may potentially lead to a misunderstanding of the correct vascular architecture in the analysed areas.

Conclusions

In this experimental work, we demonstrated the high potential of the CC technique in investigating vessel anatomy in a normal nude mouse brain. Moreover, the high quality of our pictures showed that the combined application of SEM visualization further enhances the accuracy of the morphological representation by revealing the layer‐by‐layer organization of the vascular network.

We believe that these results may have crucial implications for the comprehension of morphological vascular alterations associated with pathological conditions. In fact, especially in cases of brain tumours, normal vessels, recruited by the tumour itself, play a major role in the genesis of the tumour and in supporting neoplastic infiltration. In these conditions, the accurate knowledge of the 3D disposition and shape of normal vessels enables improved recognition and interpretation of the neoangiogenic mechanisms. On the other hand, the characterization of normal anatomy allows the evaluation of the mechanisms of action and the results of anti‐angiogenic experimental therapies aimed at inducing the regression or normalization of vascular tumour‐related alterations.

Further studies, based on the same CC‐SEM approach, will contribute to highlighting the 3D vascular anatomy of additional structures of the nude mouse brain, such as the meninges, ventricular ependyma, deep grey nuclei and brainstem.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Author contributions

Conception or design of the work: SS, MR, AM, MP, SB; acquisition of data, data analysis and interpretation: SS, ADB, MR, AM, MP; drafting the manuscript: SS, ADB, AM, MP, MR; critical revisions of the manuscript: all authors; final approval of the manuscript: all authors.