Disproportion in Pericyte/Endothelial Cell Proliferation and Mechanisms of Intussusceptive Angiogenesis Participate in Bizarre Vessel Formation in Glioblastoma

Glioblastoma (GBM) is the most malignant tumor in the brain. In addition to the vascular pattern with thin-walled vessels and findings of sprouting angiogenesis, GBM presents a bizarre microvasculature (BM) formed by vascular clusters, vascular garlands, and glomeruloid bodies. The mechanisms in BM morphogenesis are not well known. Our objective was to assess the role of pericyte/endothelial proliferation and intussusceptive angiogenic mechanisms in the formation of the BM. For this purpose, we studied specimens of 66 GBM cases using immunochemistry and confocal microscopy. In the BM, the results showed (a) transitional forms between the BM patterns, mostly with prominent pericytes covering all the abluminal endothelial cell (EC) surface of the vessels, (b) a proliferation index high in the prominent pericytes and low in ECs (47.85 times higher in pericytes than in ECs), (c) intravascular pillars (hallmark of intussusceptive angiogenesis) formed by transcapillary interendothelial bridges, endothelial contacts of opposite vessel walls, and vessel loops, and (d) the persistence of these findings in complex glomeruloid bodies. In conclusion, disproportion in pericyte/EC proliferation and mechanisms of intussusceptive angiogenesis participate in BM formation. The contributions have morphogenic and clinical interest since pericytes and intussusceptive angiogenesis can condition antiangiogenic therapy in GBM.


Introduction
Glioblastoma (GBM) is a highly malignant tumor with bizarre microvasculature (BM) [1][2][3][4]. The microvascular patterns in GBM include (a) thin-walled vessels with findings of microvascular sprouting (the "classic" vascular pattern), which is not the object of this study, (b) vascular clusters, (c) vascular garlands, and (d) glomeruloid bodies [1][2][3][4]. In this work, we focus our attention on cluster/garland-like/glomeruloid structures, also termed plexus, curled cell cords, and glomeruloid microvascular proliferations, which are the patterns that form the BM [1,5]. Glomeruloid bodies, the most complex and demonstrative of these bizarre vascular structures, are made up of closely packed anastomosing capillaries with irregular and tortuous, narrow lumina. Their main cell components are endothelial cells (ECs) and numerous pericytes [5][6][7]. Although glomeruloid bodies can be observed in other conditions in the brain (e.g., cancer metastasis) [8], they are an important finding for GBM diagnosis.
The mechanisms of high neovascularization in GBM have received great attention, including the presence of bizarre vessels and resistance to antiangiogenic therapy, which can be related to different types of angiogenesis [1][2][3][4][5][9][10][11][12][13][14][15][16][17]. It has been suggested that intussusception (intussusceptive microvascular growth) could be a mechanism of compensatory vascular growth occurring in human glioma [18,19], with advantages and inconveniences over classic sprouting angiogenesis [18]. Thus, sprouting and intussusceptive angiogenesis could be complementary mechanisms in human gliomas. In intussusceptive angiogenesis, the preexisting vasculature is split and remodeled by intravascular pillars, which are formed through several procedures, including endothelial bridges, merged adjacent capillaries with modified contacting walls, and vessel loops surrounding interstitial tissue structures (ITSs) [20][21][22][23][24][25][26][27]. Likewise, it has been highlighted that the prevalence and association of some of these intussusceptive mechanisms can determine the predominant pattern of pathological vessels [28,29]. The study of this possibility is of interest to explain the histogenesis of bizarre vessels in GBM, as well as therapy resistance due to a switch from sprouting to intussusceptive angiogenesis. EC proliferation in intussusceptive angiogenesis is low and compensated for by the spreading and thinning of ECs [24,[30][31][32]. Therefore, it is also of interest to explore the overall proliferation index (PI) in BM and specifically in each cell type (EC and pericyte).
Given the above, the objective of this study was to assess the role of EC and pericyte proliferative activity and intussusceptive angiogenic mechanisms in the formation of the BM in GBM. To this end, in the BM, we explore (a) the main types and distribution of cells that comprise or are related to it, (b) the proliferative activity in its global components and in the ECs and pericytes separately, and (c) the findings supporting the participation of vessel intussusception (intussusceptive angiogenesis) in its morphogenesis.

Human Tissue Samples
The archives of Histology and Anatomical Pathology of the Departments of Basic Medical Sciences of La Laguna University, University Hospital, and Eurofins ® Megalab-Hospiten Hospitals of the Canary Islands were searched for cases of glioblastoma multiforme. Specimens (paraffin blocks) were obtained from 66 cases. Patients were Caucasian (38 males and 27 females, aged 17-88 years). All samples were studied by conventional histologic techniques. From them, 40 cases with more evident zones showing BM (including glomeruloid bodies) were used for immunochemistry procedures and immunofluorescence in confocal microscopy. Ethical approval for this study was obtained from the Ethics Committee of La Laguna University (Comité de Ética de la Investigación y de Bienestar Animal, CEIBA 2021-3069), including the dissociation of the samples from any information that could identify the patient. The authors, therefore, had no access to identifiable patient information.

Light Microscopy
Specimens for conventional light microscopy were fixed in a buffered neutral 4% formaldehyde solution, embedded in paraffin, and cut into 3 µm thick sections. Sections were stained with hematoxylin and eosin (H&E).

Evaluation of the PI in the BM of GBM
Using the monoclonal antibody Ki-67, which reacts with a protein expressed in the cell-cycle phases G1, G2, S, and M, the PI (growth fraction) was evaluated in the BM of GBM using a high-power objective lens (40×, magnification 400) and immersion objective lens (100×, magnification 1000) and by two independent observers. When the entire nucleus or a part of it was observed to express Ki-67, the cell was considered positive, and the index was established depending on the proportion of positive cells in relation to the total cells evaluated. In addition to Ki-67, we used the PHH3 as a very specific way to detect proliferative activity, thereby complementing and confirming the results obtained with Ki-67. Mitoses were also observed, and the weak staining of the pericytes by muscle-specific actin (MSA) was very useful since, while it revealed the cell type, it did not hide the mitosis. To specifically calculate the PI in the ECs or pericytes of the BM, double immunochemistry for Ki-67 and CD34 (EC marker) or αSMA (pericyte marker) was performed, respectively. For the quantitative study of cells with or without nuclear Ki-67 expression, 200 nuclei of cells expressing their cytoplasmic marker were examined in the BM of each selected case. In addition, samples of two adjacent sections of the same block tumor with superimposed images were used to confirm the previously obtained data. Pearson's chi-squared test was used for statistical analysis.

Microvascular Patterns in GBM
Except for vasculogenic mimicry, the major vascular patterns in GBM were a thinwalled microvasculature ( Figure 1A) with findings of microvascular sprouting ("classic" vascular pattern), vascular clusters ( Figure 1B-D), vascular garlands ( Figure 1E), and glomeruloid bodies ( Figure 1F,G), with the last three forming the BM. Transitional structures were observed between clusters and garland-like and glomeruloid structures (glomeruloid bodies and precursor structures) ( Figure 1H,I). These structures were frequently joined by vessels of varying size ( Figure 1J,K).

Cellular Components of Bizarre Vessels in GBM
ECs and pericytes were the two main types of cells in bizarre vessels in GBM. CD34 and αSMA were expressed in ECs and pericytes, respectively. H-caldesmon, desmin, and calponin only were expressed by pericytes in a few vessels. In most of the cluster/garlandlike/glomeruli (Type I BM) (about 85% of BM), prominent pericytes showed a cell body increased in size, with a plump aspect. These prominent, plump αSMA + pericytes covered all the abluminal EC surface of the vessels in the glomeruloid bodies and precursor structures (Figures 1E-K and 2A,B). With relative frequency, the slit-like lumen of these vessels was virtual or narrow, and the confluent αSMA + pericytes around the folded vessels appeared as aggregates (the flattened unstained EC cytoplasm was not detected), except when double immunochemistry for CD34 and αSMA was undertaken (Figure 2A   like/glomeruli (Type I BM) (about 85% of BM), prominent pericytes showed a cell body increased in size, with a plump aspect. These prominent, plump αSMA + pericytes covered all the abluminal EC surface of the vessels in the glomeruloid bodies and precursor structures ( Figures 1E-K and 2A,B). With relative frequency, the slit-like lumen of these vessels was virtual or narrow, and the confluent αSMA + pericytes around the folded vessels appeared as aggregates (the flattened unstained EC cytoplasm was not detected), except when double immunochemistry for CD34 and αSMA was undertaken (Figure 2A,B).  In approximately 50% of the clusters and occasional vascular garlands (Type II BM) (about 15% of BM), pericytes partially covered the abluminal surface of ECs and presented an elongated cell body with long slender processes and an elaborate system of longitudinal and circumferential branches encircling the ECs ( Figure 1C,D).

Neoplastic Glial cells and BM in GBM
Neoplastic glial cells expressing GFAP were observed around the BM in GBM (Figur 2C-F). This finding was highlighted by double immunochemistry (Figure 2C,D) and dou ble immunofluorescence ( Figure 2E,F) for GFAP and αSMA. Rows of neoplastic glial cell were seen separating neighboring glomeruloid bodies from each other or from their pre cursor structures ( Figure 2E) and occasionally within them ( Figure 3A-D).

Microglia and BM in GBM
Isolated Iba1 + microglial cells were observed in areas adjacent to the BM in GBM using double immunofluorescence for Iba1 and αSMA ( Figure 3E,F). Microgliocytes showed an increased body, with an amoeboid aspect.

Pericyte and EC Proliferation in the BM
Different results in terms of the PI in pericytes and ECs were obtained for Type I BM, as well as between pericytes of both types of BM (Table 1). In the BM, including all the cells in glomeruloid bodies and their precursor structures, the expression of Ki-67 and PHH3 was increased ( Figure 4A-C). The percentage of nuclei expressing Ki-67, with respect to total nuclei (PI), was 43.57% ± 9.5% in Type I BM and 3.41% ± 0.76% in Type II BM (Table 1). In Type I BM, by double immunochemistry for CD34 and Ki-67 ( Figure 4D), the PI in ECs was 0.88% ± 0.29%. Through double immunochemistry for αSMA and Ki-67 ( Figure 4E,F), the PI in pericytes was 42.11% ± 5.76%. Therefore, the relationship between the PI of pericytes and ECs in Type I BM showed that, in pericytes, it was 47.85 times higher than in ECs. Samples showing the relative positions of ECs and pericytes by superimposed images of two adjacent sections of the same block ( Figure 5A,B) confirmed the data. This proliferative activity in pericytes was also confirmed using immunofluorescence for PHH3 and αSMA in confocal microscopy ( Figure 5C). Although the double immunochemistry procedure to explore nuclear Ki-67 expression was performed separately for each cell type (cytoplasmic expression of CD34 for ECs and αSMA for pericytes), the Ki-67 nuclear expression was also seen in the other cell type with unmarked cytoplasm, identifiable by its location (Figure 4D-F). Mitotic figures were only detected in pericytes of Type I BM ( Figure 5D,E). Through the same procedures (double immunochemistry for Ki-67 and CD34 or αSMA) ( Figure 5F), 50% of clusters and occasional garland-like structures (Type II BM) presented a PI of 1.12% ± 0.21% in pericytes and of 2.21% ± 0.88% in ECs. Therefore, the PI of pericytes in Type I BM was 37.59 times higher than in Type II BM. In some clusters, vessels with characteristics of Types I and II BM (transitional forms) were observed. These clusters were not used to calculate the PI.

Intravascular Pillars
By double staining for CD34 and SMA, two main types of intravascular pillars were observed depending on tissue components: (a) with extensions of CD34 + endothelium only, and (b) with a cover formed by CD34 + endothelium and a core formed by SMA + pericytes.
In the first type of pillar, intraluminal bridges formed by CD34 + ECs and their processes were seen extending between opposite walls of small vessels that form the glomeruloid bodies and their precursor structures ( Figure 6A-C). The pillars were thin, and their appearance and disappearance were detected in a series of individual views in confocal microscopy ( Figure 6D1-4). Multiple pillars were present in some vessels ( Figure 6E).
Contacts between ECs of opposite vessel walls were frequently observed ( Figure 6F-H). These contacts were apical between prominent ECs ( Figure 6F) or planar between flattened ECs ( Figure 6G,H). The planar contacts could show prominence of the pericytes toward the vessel lumen (apparent push) ( Figure 6G) or not ( Figure 6H).

Intravascular Pillars
By double staining for CD34 and SMA, two main types of intravascular pillars were observed depending on tissue components: (a) with extensions of CD34 + endothelium only, and (b) with a cover formed by CD34 + endothelium and a core formed by SMA + pericytes.
In the first type of pillar, intraluminal bridges formed by CD34 + ECs and their processes were seen extending between opposite walls of small vessels that form the glomeruloid bodies and their precursor structures ( Figure 6A-C). The pillars were thin, and their appearance and disappearance were detected in a series of individual views in confocal microscopy ( Figure 6D1,2,3,4). Multiple pillars were present in some vessels ( Figure 6E).
Contacts between ECs of opposite vessel walls were frequently observed ( Figure 6F-H). These contacts were apical between prominent ECs ( Figure 6F) or planar between flattened ECs ( Figure 6G,H). The planar contacts could show prominence of the pericytes toward the vessel lumen (apparent push) ( Figure 6G) or not ( Figure 6H).  The characteristics, arrangement, and morphological development of the second type of pillar, which showed an endothelial cover and a pericytic core, depended on (a) vessel size, (b) formation of vessel loops and ITSs, as well as opening or not of the loop lumen surrounding the ITSs, and (c) pericyte hyperplasia and hypertrophy. Next, we describe the variants of this type of pillar depending on the conditions outlined above.
In dilated vessels (including some in glomeruloid bodies), pillars could be observed with a different extension and alignment in the vessel lumen ( Figure 7A-J). When the pillars were longitudinal or cross-sectioned, the ECs formed the pillar cover and the pericytes formed the core (ITS) (Figure 7D-H). The cells in the pillar core ranged between thin processes of αSMA + pericytes perceptible in the ends of the pillar ( Figure 7I) to rows of pericytes along it ( Figure 7D). On occasion, pillars were seen originating simultaneously from two opposite points of the vessel wall ( Figure 7J). The characteristics, arrangement, and morphological development of the second type of pillar, which showed an endothelial cover and a pericytic core, depended on (a) vessel size, (b) formation of vessel loops and ITSs, as well as opening or not of the loop lumen surrounding the ITSs, and (c) pericyte hyperplasia and hypertrophy. Next, we describe the variants of this type of pillar depending on the conditions outlined above.
In dilated vessels (including some in glomeruloid bodies), pillars could be observed with a different extension and alignment in the vessel lumen ( Figure 7A-J). When the pillars were longitudinal or cross-sectioned, the ECs formed the pillar cover and the pericytes formed the core (ITS) (Figure 7D-H). The cells in the pillar core ranged between thin processes of αSMA + pericytes perceptible in the ends of the pillar ( Figure 7I) to rows of pericytes along it ( Figure 7D). On occasion, pillars were seen originating simultaneously from two opposite points of the vessel wall ( Figure 7J).   The relationship of pillars with vessel loops and ITSs was a common finding. Vessel loops were observed originating from mother vessels or other loops in the glomeruloid bodies and precursor structures ( Figure 8A,B). Each fully formed vessel loop presented a virtual or open lumen, and its internal EC layer surrounded an ITS made up by αSMA + pericytes ( Figure 8B). In the glomeruloid bodies and precursor structures, the vessel loops and ITSs formed part of a conglomerate of similar structures when the vessel loop lumen was virtual ( Figure 8C). When some of the vessel loop lumen was open, the ITS and its surrounding ECs formed the core and cover of a pillar, respectively ( Figure 8D,E). The relationship of pillars with vessel loops and ITSs was a common finding. Vessel loops were observed originating from mother vessels or other loops in the glomeruloid bodies and precursor structures ( Figure 8A,B). Each fully formed vessel loop presented a virtual or open lumen, and its internal EC layer surrounded an ITS made up by αSMA + pericytes ( Figure 8B). In the glomeruloid bodies and precursor structures, the vessel loops and ITSs formed part of a conglomerate of similar structures when the vessel loop lumen was virtual ( Figure 8C). When some of the vessel loop lumen was open, the ITS and its surrounding ECs formed the core and cover of a pillar, respectively ( Figure 8D,E).  Expression of Ki-67 and PHH3 was observed in the αSMA pericytes surrounded by the loops without open lumen and in those present in the pillar cores ( Figure 8F,G), which were widened, contributing to the formation of true pericyte aggregates in glomeruloid bodies and precursor structures ( Figure 8I).

Discussion
In GBM, we studied the morphogenesis of the BM (clusters, garland-like structures, and glomeruloid bodies) and report as main findings (a) disproportion in the pericyte/EC proliferation (high proliferative activity of pericytes in relation to ECs), and (b) several mechanisms of intussusceptive angiogenesis. Next, we take into consideration (a) the vascular patterns in GBM, (b) the high proliferation of pericytes in relation to ECs, leading to pericyte hyperplasia and hypertrophy, and extension and thinning of ECs, and (c) the mechanisms of pillar formation (mechanisms of intussusceptive angiogenesis) and their participation in bizarre vessel morphogenesis.
In the classification of the main microvascular patterns in GBM, we followed the previous systematizations of other authors [1][2][3][4][5]. Thin-walled vessels are considered an exponent of classic microvascular sprouting ("classic vascular pattern"), while the other patterns (cluster/garland-like/glomeruloid bodies) are considered bizarre (malformed/aberrant) microvasculature, which was the object of our study. Given the observation of multiple transitional structures in the BM, we encompassed them as glomeruloid bodies and precursor structures. In the classification of GBM microvasculature, some authors grouped the patterns into two types depending on the prognostic significance. Thus, microvascular sprouting and vascular clusters were considered as Type 1 and vascular garlands, glomeruloid bodies, and vascular mimicry were considered as Type 2, with a more negative influence on progression-free survival in the latter [3]. The patterns in this classification partially coincide with those in the two types of our study in BM, based on the characteristics of the pericytes (patterns with vessels of thin walls and signs of sprouting, and vascular mimicry were not included in our study). Based on pericytic characteristics, our classification of BM facilitated PI comparison between different BM structures, including the PI between pericytes and ECs, demonstrating important imbalances. The classification, therefore, allowed us to respond to one of the approaches to BM formation. Given that our objective was focused on and limited to the mechanisms of BM formation, future studies with more extensive data are required for the clinical meaning and therapeutic implications of these mechanisms, above all in the response to antiangiogenic treatment.
Our observations revealed that the high PI in the BM corresponds mostly to prominent αSMA + pericytes, to the extent that there is a manifest disproportion between the pericyte and EC PI (47.85 times higher in pericytes than in ECs). This high PI in pericytes concurs with the activation of endogenous pericytes in response to orthotopic mouse gliomas [33]. Abundant pericytes/vascular smooth muscle cells, a heterogeneous population [34,35] (e.g., the expression or not of MSA), have been previously observed in the microvasculature of glioblastoma [5,36,37]. This behavior of pericytes also contributes to the heterogeneity of angiogenesis and blood vessel maturation in human tumors [38,39]. The marked hyperplasia and hypertrophy of pericytes, which occupied the entire EC abluminal surface of the frequently folded bizarre vessels, suggest that their contractile force can be an important factor in vessel folding and contraction (e.g., capillaries with tortuous and narrow lumen in glomeruloid bodies). Likewise, the great proliferation of pericytes in the BM has morphogenic and clinical interest since pericytes facilitate (a) GBM growth through immunosuppression (reduction in T-cell activation) and anti-inflammatory response, and (b) GBM cell resistance to temozolomide or chemotherapeutic efficacy by targeting glioma stem cell-derived pericytes, which disrupt the blood-tumor barrier [40][41][42][43][44][45]. Likewise, the inhibition of NG2 proteoglycan, which promotes tumor vascularization [46] and is expressed by pericytes, induces a decrease in GBM microvessels [47,48]. The presence of structures of BM with different pericyte PI (PI of pericytes in Type I BM was 37.59 times higher than in Type II BM) and of transitional forms also suggests successive changes in preexisting pericytes, including the emergence of high proliferative activity. Further studies are required on the factors that participate in the different proliferative response of pericytes and ECs in the BM of GBM.
The low PI in ECs of the BM also suggests that they form vessel loops mainly by migration and extension in this aberrant type of microvascularization. A similar EC extension has been described in intussusceptive angiogenesis by vessel loop formation in the ovarian pedicle after ovariectomy and human tumor xenograft [25,26]. Although the PI in ECs is low with respect to that of pericytes, it confirms proliferation and survival of ECs in BM [49].
The presence of intravascular pillars, the hallmark of intussusceptive angiogenesis, in BM was confirmed by their appearance and disappearance in successive serial sections in confocal microscopy. The precursor findings involved in pillar morphogenesis in the BM of GBM coincide with one of the main mechanisms proposed in classical studies on vessel intussusception [20][21][22][23][24][25][26][27]50]. Thus, in vessels with a small lumen, intraluminal endothelial bridges appeared between the opposite vessel walls formed by thin EC filopodial ridges. Our observation of numerous apical and planar contacts of ECs also concurs with the participation of these structures in the initial phase of pillar formation.
A core formed by pericytes, stromal cells, and extracellular matrix, and a cover formed by ECs have been described in well-developed pillars. The only cells present in the pillar core in the bizarre vessels of GBM, mainly in those with a wider lumen, were αSMA + pericytes. The observation of thin processes of αSMA + pericytes between ECs in the pillar ends coincides with evolutive stages II and III of pillar formation described as a bilayer EC arrangement and pillar perforation [49].
Vessel loops have been demonstrated to contribute to pillar formation in the microvasculature [25,26,29,51,52]. Our observations in the BM of GBM reveal that vessel loops form part of an important mechanism of morphogenesis of these vessels by intussusceptive angiogenesis. Indeed, each loop surrounds the pericytes located between it and its mother vessel. These incarcerated pericytes make up an ITS. When the loop lumen is open, the EC sheets that form the internal endothelial layer of the loop and the ITS (encircled pericytes) are partially incorporated into the loop lumen, which is connected from two points (the two connecting segments of the loop) with that of the mother vessel. Thus, the surrounding endothelium and the ITS contribute to forming the pillar cover and core, respectively. While the pillar was intraluminal, the ends appeared connected to the wall of the mother vessel. When the loop lumen remains closed, ECs and pericytes form a solid set. Importantly, pericytes in the core of pillars also showed high proliferative activity, which increased pillar core size.
The contribution of mechanisms of intussusceptive angiogenesis to the BM of GBM is also of clinical interest. GBM is a highly malignant neoplasia, in which GBM angiogenesis and chemotherapy may be conditioned by multiple factors, including tumor location and natural substances [53][54][55]. In addition, GBM is difficult to treat effectively for lengthy survival, with many facets of therapy resistance and tumor recurrence [56][57][58]. Among these difficulties is the resistance to antiangiogenic therapy, in which the switch from sprouting to intussusceptive angiogenesis has been described as an escape mechanism [59].
In conclusion, the facts leading to the formation of the BM in GBM can be summarized as follows (Figure 9): (1) disproportion in pericyte/EC proliferation, highly in favor of pericytes, (2) pericyte hyperplasia and hypertrophy in most of MB, (3) phenomena of intussusceptive angiogenesis, including (a) intraluminal endothelial bridges (intussusceptive microvascular growth), (b) development of intussusceptive pillars with a cover (ECs) and a core (pericytes), and (c) formation of numerous vessel loops, which also participate in pillar development when the loop lumen is open and their internal ECs surround an ITS (pericytes), and (4) persistence of these facts in the most evolved structures of the BM, such as complex glomeruloid bodies. Therefore, a disproportion in the pericyte/EC PI (with high pericyte proliferation) and mechanisms of intussusceptive angiogenesis participate in BM formation in GBM. In addition to the improved knowledge of the mechanisms that participate in the formation of the BM in GBM, these contributions are of clinical interest, since pericytes and intussusceptive angiogenesis can condition antiangiogenic therapy. rized as follows ( Figure 9): (1) disproportion in pericyte/EC proliferation, highly in favor of pericytes, (2) pericyte hyperplasia and hypertrophy in most of MB, (3) phenomena of intussusceptive angiogenesis, including (a) intraluminal endothelial bridges (intussusceptive microvascular growth), (b) development of intussusceptive pillars with a cover (ECs) and a core (pericytes), and (c) formation of numerous vessel loops, which also participate in pillar development when the loop lumen is open and their internal ECs surround an ITS (pericytes), and (4) persistence of these facts in the most evolved structures of the BM, such as complex glomeruloid bodies. Therefore, a disproportion in the pericyte/EC PI (with high pericyte proliferation) and mechanisms of intussusceptive angiogenesis participate in BM formation in GBM. In addition to the improved knowledge of the mechanisms that participate in the formation of the BM in GBM, these contributions are of clinical interest, since pericytes and intussusceptive angiogenesis can condition antiangiogenic therapy.  Informed Consent Statement: Patient consent was waived, due to the study using a historical histological collection of tissue samples that were dissociated and contained no personal or identification data of the patients. The authors, therefore, had no access to identifiable patient information.
Data Availability Statement: All the data are reported in the present paper.