Differential Angiogenic Induction Impacts Nasal Polyp Tissue Growth

In chronic rhinosinusitis with nasal polyps, inflammatory edema drives tissue remodeling favoring anomalous growth of the nasal mucosa, but a proangiogenic contribution of nasal polyp in support of tissue growth is still controversial. The chorioallantoic membrane of chicken embryo model was employed to address the potentiality of nasal tissue fragments to modulate angiogenesis. Fifty-seven fertilized eggs were implanted with polyp or healthy nasal mucosa tissue or were kept as non-implanted controls. The embryos’ size, length, and development stage, and chorioallantoic membrane vasculature morphology were evaluated after 48 h. Quantitative computer vision techniques applied to digital chorioallantoic membrane images automatically calculated the branching index as the ratio between the areas of the convex polygon surrounding the vascular tree and the vessels’ area. Ethics approval and consent to participate: the study was approved by the Human Research Ethics Committee of the Federal University of São Paulo (CAAE number: 80763117.1.0000.5505) and by the Animal Research Ethics Committee of University of São Paulo (nº CEUA 602–2019). Mucosal, but not polyp tissue implants, hampered embryo development and induced underdeveloped chorioallantoic membranes with anastomosed, interrupted, and regressive vessels. Vessels’ areas and branching indexes were higher among the chorioallantoic membranes with polyp implants and controls than among those with healthy mucosa implants. Nasal polyp presents differential angiogenic induction that impacts tissue growth.


Introduction
Chronic sinusitis with nasal polyps (CRSwNP) is a chronic inflammatory disease of the nasal mucosa affecting around 1-4% of the population [1]. One of challenges in understanding the pathophysiology of CRSwNP is the immunological mechanisms underlying the anomalous and recurrent growth of the nasal mucosa, even after therapeutic interventions [2]. The abnormal deposition of extracellular matrix in CRSwNP originates from protruded structures formed by ciliary epithelium, thick basement membrane, and loose stromal tissue with minimal vascularity [3,4]. Moreover, the intense inflammatory infiltrates in the stroma [5,6] and the dysfunction of local fluid control contributes to the polyposis development [7][8][9].
The immunoregulation of chronic inflammation by mesenchymal stem cells (MSC), induction of regulatory T-cells, and increase of IL-10 concentration are promising areas of interest in addressing the abnormal tissue remodelling process in CRSwNP [10,11]. The modulation of genes related to angiogenesis and cell growth differs between nasal polyp and bone marrow-derived MSC. The contribution of nasal Abstract In chronic rhinosinusitis with nasal polyps, inflammatory edema drives tissue remodeling favoring anomalous growth of the nasal mucosa, but a proangiogenic contribution of nasal polyp in support of tissue growth is still controversial. The chorioallantoic membrane of chicken embryo model was employed to address the potentiality of nasal tissue fragments to modulate angiogenesis. Fifty-seven fertilized eggs were implanted with polyp or healthy nasal mucosa tissue or were kept as non-implanted controls. The embryos' size, length, and development stage, and chorioallantoic membrane vasculature morphology were evaluated after 48 h. Quantitative computer vision techniques applied to digital chorioallantoic membrane images automatically calculated the branching index as the ratio between the areas of the convex polygon surrounding the vascular tree and the vessels' area. Ethics approval and consent to participate: the study was approved by the Human Research Ethics Committee of the Federal University of São Paulo (CAAE number: 80763117.1.0000.5505) and by the Animal Research Ethics Committee of University of São Paulo (nº polyp-derived MSC to tissue growth through regulation of fluid balance and angiogenesis has not yet been unequivocally demonstrated [12]. A higher degree of vascular anastomosis was detected in the nasal mucosa compared to bronchial mucosa [13]. In this context, it was demonstrated that mechanical dysfunction impacts the tissue's ability to increase the interstitial hydrostatic pressure during inflammation (transudation/exudation). Stromal expansion subsequently deepens vascular network remodeling, with the formation of new vessels and changes in local hydrostatic relationships, which favors the self-amplified process of capillary fluid output that culminates in edema driven tissue growth [3].
The Gallus domesticus experimental model was convenient for studying the effects of tissue implants upon angiogenesis, allowing simultaneous measurements of edema changes and effects upon tissue development with high reproducibility [14]. This experimental model uses the chorioallantoic membrane (CAM) of Gallus domesticus embryos,which is easy to manipulate and be applied to preclinical in vivo model studies for drug screening and tissue vascular growth [15]. This model evaluates the ability of an implanted tissue to promote angiogenic induction it has been used to study tumors [16,17] and regenerative medicine [18].
This work addresses the ability of nasal polyp tissue to modify the vascular network of the tissue and stimulating angiogenesis in the experimental model of the (CAM) of Gallus domesticus embryos.

Methods
The study was approved by the Human Research Ethics Committee of the Federal University of São Paulo (CAAE number: 80763117.1.0000.5505) and by the Animal Research Ethics Committee of University of São Paulo (nº CEUA 602-2019). Twenty-two participants aged between eighteen and sixty-five years old were recruited. Twelve of these patients were eosinophilic nasal polyp tissue donors (> 5 eosinophils per high power field) undergoing surgical polypectomy and ten patients were healthy middle meatus mucosa donors undergoing partial middle concha bullosa resection.

Angiogenesis and Embryo Development
The effects of tissue samples upon angiogenesis were assayed in the chorioallantoic membrane of chicken embryo model [19]. Fifty-seven fertilized eggs (conventional, free from food contaminating human pathogens, Lohmann breed) (Yamaguishi farm, Sao Paulo, Brazil) were used. Three experimental groups were established: (1) "Polyp"; (2) "Healthy Mucosa"; (3) "Control". The control arm consisted of a sham manipulation of the embryo without an implant.Eighteen embryonated eggs were implanted with polyp fragments, twenty-seven eggs were implanted with healthy mucosal fragments and eleven eggs were kept as non-implanted controls. The number of fertilized eggs varied according to the day of implantation, so a tissue sample may have been implanted in more than one egg when possible. In the cases where more than one egg was implanted by the same sample, the mean of results was considered for analysis purposes.
The bioassay procedure was carried as previously described by Pereira-Lopes et al. 2010 [20]. Each egg was cleaned with ethanol 70%, individually identified, and incubated in a Zagas® hatcher at 37.6 °C and 85% humidity. After 96 h (E4, fourth day), 4 mL of albumin was removed from each egg, pooled, and kept under sterile conditions at 4 °C. Then, in order of the exposed embryos to external environment, small windows were opened in the eggshells and were covered by pieces of the surgical field Ioban™ (3 M) before the eggs were returned to the hatcher. Fortyeight hours later (E6, sixth day), the eggshell windows were reopened, and the surgical samples were collected (polyp and healthy mucosa groups) and warmed to 37 °C. Tissue fragments were divided into 4 mm, recovered by warmed sterile egg albumin by brief immersion, and positioned on the surface of the developing CAMs. The results were evaluated after 48 h incubation with the implants (E8, eighth day). The eggs were removed from the hatcher and put into the refrigerator (4 °C) for 3 h. Eggshells were opened, and the embryos and the CAMs were isolated, cleaned, processed, and macroscopically evaluated. The membranes were preserved with a small amount of formaldehyde, distended on plastic Petri dishes, and scanned in a Hewlett-Packard G2400 Scanjet (2400dpi, real size). The embryos were measured, weighed, fixed with formaldehyde, and photographed with a Nikon SMZ1500 magnifier in order to document the developmental stage of the embryos. Eye, legs, wings, and beak were considered for determining the Hamburger-Hamilton stage [21]. The digital images of the individual membranes were converted into threshold binary images andanalyzed. The results on embryo size, length, and development stage were correlated with the results on the CAM vasculature morphology.

Quantitative Morphology
Quantitative morphological information about the CAM was obtained by applying computer vision techniques to the imaged membrane. The general approach of CAM images processing consisted of: (i) manually selecting five square regions of interest (ROI) with dimensions of 1024 × 1024 pixels from each CAM image; (ii) automatically segment regions corresponding to the blood vessels in each ROI of the whole set of images; and (iii) extract quantification metrics from binary and grayscale images. Metrics like the plain vessel area were directly measured and the branching index was calculated by the ratio between the areas of the convex polygon surrounding the vascular tree known as the vascular region and the area of the vascular tree known as the vessels' area. The Branching Index (BI) was calculated using the equation described below: (BI) = 1 -convex hull area ratio. Convex hull area ratio = vessels' area/vascular region's area. (Fig. 1) The vessels' area in each ROI provides information about the growth and elongation of blood vessels, whereas branching index informed about the blood distribution and CAM area coverage. This provided complementary information about the angiogenesis processes taking place in the CAM membrane [19,22,23].

Statisticsanalysis
The Kolmogorov-Smirnov test was applied to verify the normality of each quantitative distribution of the data [24] Kruskal-Wallis test with Tukey's multiple comparisons procedures and Spearman's correlation coefficient were used for nonparametric data analysis [24]. ANOVA and Student's tests were employed for parametric data analysis. All tests were performed with a significance level of 95% for p < 0.05. SPSS® for Windows 19.0.0 was used for calculations.

Embryonic Development
The surgical tissue fragments from "Polyp" and "Healthy Mucosa" groups induced divergent effects upon both chicken embryo and the extraembryonic CAM tissue after being implanted in the developing eggs. Coherent effects could be observed upon the embryo growth and maturation stage, which implies cell or tissue signaling ( Table 1).
The healthy mucosa implants were associated with delayed development, lower weight, and shorter dimensions of the embryos compared to embryos implanted with nasal polyp tissue implants (Table 1). Mucosal implants reduced or halted the embryo development, while the polyp tissue implants allowed the embryo development when compared after the standard hallmarks of normal chicken embryo development [21].

Vascular Growth and CAM Morphology
Qualitative aspects of the CAM morphology in the E8 (eighth day) stage indicated that vasculature expansion and organization were modified by the polyp and normal mucosal tissues generating compatible effects upon the embryo and extraembryonic tissues (Figs. 2 and 3). The implanted polyp fragments presented bright colored and adherent to the CAM (Fig. 2 A) with some integration to vasculature. The general aspect of vasculature was similar to normal controls (Fig. 2B).The implanted healthy mucosa fragments presented pale and not adherent to the CAM, with no integration to vasculature (Fig. 2 C).
Qualitative differences were observed in the CAM tissue compared to the normal controls (Fig. 3A). Noticeably, the morphology suggestive of neovascular formation was frequently observed in the implant region of polyp fragments ( Fig. 3B and C) and the polyp fragments were compatible with effective perfusion (Fig. 3B and C). Bleeding and interrupted growth as was observed in 44% of the eggs in the healthy mucosa group, resulting in premature death (Fig. 3D). Minor hemorrhagic foci and delayed development in milder outcomes were seen in all of eggs implanted with healthy mucosa (Fig. 3E). A rapid CAM expansion is expected to occur within the chosen experimental period  (E4-E8), being small CAMs associated with growth interruption events (Fig. 3D, F). Curved neovessels displayed radially from the polyp implants, producing atypical vascular trees (Fig. 3B and H), while mucosa implants induced regressive modifications of the CAM's vasculature, even when the membrane succeeded to develop and grow (Fig. 3I).
The mucosa tissue implants reduced both the growth and the branching of the CAM vasculature of the implanted eggs, respective to normal controls in the E6-E8 period, as referred by the values of the vessel's areas and branching indexes (Fig. 4). Mean vascular areas were lower in the mucosa group than in the control group (Fig. 4A).The vessel areas were slightly more dispersed in the mucosa group than in the other two groups, with the distribution asymmetry being lower reflecting the regressive effects shown in Fig. 3F, I.
The CAM vasculature of the polyp group eggs succeeded to produce the expected branching indexes compared to normal controls (Fig. 4B).This parameter was lower in the mucosal group than in the other two groups, as shown in (Fig. 4B). The result was the same whether the average value or the upper-end values were considered in the analysis.
In summary, implanted polyp and healthy mucosa tissue fragments produced divergent effects upon the CAM vasculature and the development of the embryo after a 48 h period (E6-E8).

Discussion
Angiogenesis contribution to the pathophysiology of nasal polyposis is still controversial. The pathognomonic tissue remodeling of this clinical condition is frequently attributed to the underlying chronic inflammatory process. However, the combination of the chronic inflammatory tissue remodeling with the associated edema is required to produce the extensive structural impairment that characterizes CRSwNP [4]. The role of endothelial cells in the production and evolution of CRSwNP was explored in this work based on the tissue growth hypothesis in driving edema. The poor fluid compartment control was assumed to be the result of functional effects on the vascular wall, as polyp tissue required three times higher transudation volume to reach the same interstitial hydrostatic pressure of the healthy nasal mucosa during inflammation [3].
Inflammatory mediators that produce tissue remodeling and modulate angiogenesis induction in chronic processes can indeed produce early effects on the differentiation of endothelial cells, inducing an increase in capillary permeability, tip cell migration, and vascular budding. Late functional changes and vessel stabilization are probably myofibroblasts and pericytes dependent, but the effects of early soluble mediators and extracellular matrix-derived molecules upon endothelial cell morphology were previously detected and could be studied even in endothelial cell cultures.The Gallus domesticus allows simultaneous measurements of edema change sand effects upon tissue development with high reproducibility,without causing animal suffering [14,28]. Implants of cells or tissue fragments do not result in rejection by the egg, which provides a nutrient-rich environment capable of maintaining tissue viability for a few hours and, in some situations, for several days [29,30].
Healthy mucosa implants reduced or halted embryo development and CAM vasculature growth. The measured vascular areas were smaller in the eggs from the healthy mucosa group, and the branching indexes were lower than those observed in the eggs of the control and polyp groups. The healthy mucosa effects were attributed to the presence of interrupted vessels, and a smaller and less anastomosed network. Quantitative results were a priori independent on the CAM size, as the measurements considered size standardized ROI areas. Differences would have been described as being even greater if the membrane size was considered and the whole vessel areas were measured. Underlying pathophysiological events were interpreted in terms of the failure of the healthy mucosa fragments to contribute timely to their perfusion through pro-angiogenic biochemical signaling. The qualitative evidence of directional growth of capillary vessels was absent in the periphery of healthy mucosa implants, while it was abundant in the periphery of polyp implants. Also, the mucosa implants presented no signs of effective perfusion at the end of the experiment, while the polyp implants looked like the original tissue that was implanted and adhered to the CAMs' surfaces.
Experimental results support the previous inference about the role of induction of angiogenesis in CRSwNP in the disease relapses after surgical removal of polyps, a behavior known to be dependent on its clinical endotype.
The sustained implant viability in the CAM model relies on the development of a neovascular assembly for the effective tissue perfusion that occurs through coordinated angiogenesis processes [31].This possibility has been previously explored for studies of tumor and artificial tissues implants [16,21]. While the healthy mucosa implants produced only regressive effects throughout the CAM area, the biochemical signaling of polyp fragments synergized the embryo's innate inflammatory mechanisms and the growing vascular network to ease liquid and cell percolation in the implanted regions.
The outcome of the polyp implants' interactions in the model was well-perfused CAM membranes with atypical vessel tree topologies. The interstitial hydrostaticpressure and fluid compartment control by the implants were probably relevant to this peculiar behavior of polyp implants. Indeed, nasal polyp interstitial hydrostatic pressure was reported to be very low, when compared to healthy nasal mucosa [3]. On the other hand, the highly positive interstitial hydrostatic pressure in the stroma of malignant tissues is associated with perfusion impairment, making inadequate delivery of therapeutic agents [32]. Loss of complexity and atypical symmetry could be easily suggested by the  Fig. 3 for both experimental groups. Quantitative protocols for analysis of angiogenesis were recently reviewed and several techniques can be of help in future studies [25,26,33]. Present results show that both the measured vascular area and the branching index measured in the healthy mucosa implanted eggs were lower than expected for the controls. It could be concluded that both types of tissue implants interacted and modified the CAM vasculature as the healthy mucosa produced negative effects and the polyp as modified growth patterns positively.
Regarding morphogenesis and tissue regeneration, it also should be argued if the slightly vascularized polyp tissue could modify the CAM vasculature structure without affecting the embryo development and growth because polyp and control group embryos reached the equivalent size, weight, and development stage. At least for the interval E6-E8, the alternative possibility is that the polyp implants actively stimulate host cell populations. As mucosal growth abnormalities are common clinical findings in CRSwNP, this implies biochemical signaling for tissue growth and remodeling produced by the polyp tissue cells [34,35].
Inflammatory cells and stem cells present in the implants were considered the main cell populations able to interact with the developing egg tissues to produce the observed effects. Inflammatory cells are expected to either produce regressive changes or inflammation in the egg. Human mesenchymal stem cells were previously reported to remain alive in the egg environment, to be able to migrate to embryo tissues and to interfere with vasculogenesis and tissue growth [29,36]. It was previously demonstrated that the nasal polyp-derived stem cells differentially modulated the angiogenesis-related genes expression, VEGF, and FGF10 [12]. VEGF and FGF10 proteins were reported to participate in the development of embryo limbs, which was an embryo differentiation process taking place in the same stage the assays were conducted [37,38]. Inflammatory cells from CRSwNP, however, were recently reported to produce increased immunostimulation through the Receptor for the Activator of Nuclear factor Kappa-B (RANK) pathway [27,39]. The inflammatory response stimulation by RANK and RANK-ligand coupling, besides the induction of a Th2 type response, cooperates to undermine the barrier function ofthe growing vasculature, producing edemaprone neovessels [27]. It was assumed that the mesenchymal stem cells present in the implanted polyp fragments together with the polyp inflammatory cells, while preserving to some extent, could signal for the extended survival of both the implant cell populations and host tissues, and reduce inflammatory rejection of the tissue implants. Not only are the production of proangiogenic growth factors important, but also the migratory capacity and the possibility of neovasculature stabilization may be determined by the vascularization behavior of structures implanted on the CAM, as reported in the interaction between glioblastoma and epithelial stem cells [40]. The proposition that the results should be attributed to angiogenesis signalling by polyp mesenchymal stem cells in such an inflammatory environment was favoured by setting the forty-eight hours period between the tissue implantation and data harvesting. During this interval, cell migration and proliferation were less likely to occur extensively, and labile inflammatory cytokines co-stimulatory effects would still be observed as well.
Further studies should confirm the polyp and mucosaderived mesenchymal stem cells'immunoregulatory and proangiogenicinteractions. The application of undifferentiated mesenchymal cells derived from the human placenta in cell therapy is defended because the anti-inflammatory and immunomodulatory properties of these cells were demonstrated to stimulate tissue regeneration and repair [41].

Conclusion
This study demonstrated a differential angiogenic induction by the nasal polyp when compared with healthy nasal mucosa.
Funding the authors declare that they received no funding for this study.

Declarations
Competing interest The authors declare that they have no competing interests.

Consent for publication
We confirm that all of the listed authors meet the criteria for authorship and originality and that none of the material in the manuscript is included in another manuscript, have been published previously, or are currently under consideration for publication elsewhere.