Keywords

1 Introduction

Thyroid cancer is the most common endocrine malignancy, and it has been characterized in recent decades by the fastest increase in incidence among all neoplasms [1, 2]. Among differentiated thyroid cancer (DTC), papillary thyroid cancer (PTC) is the most common and accounts for 90% of all histotypes. Usually, PTCs are slow-growing and indolent tumors associated with lymph node metastases in 30–90% of cases [3, 4].

The cornerstone of the PTC treatment is the surgical removal of the tumor with or without radioiodine therapy. Despite an overall good prognosis, a portion of cases progresses to aggressive and refractory disease, showing neck recurrence or distant metastases. The angiogenic switch is the critical process that leads a tumor to local invasion and distant metastases through the growth of new vessels [5]. A deep understanding of the molecular mechanisms that transform the thyroid cancer microenvironment into an angiogenic microenvironment is at the center of the research into new prognostic markers and therapeutic targets [6]. Examining these processes could improve many steps of thyroid cancer management: identifying cytologically indeterminate nodules needing surgical removal, thus reducing diagnostic thyroidectomies; differentiating aggressive cancers to modulate surgical removal extent and approach and the follow-up strategies; improving the targeted therapy of metastatic patients with refractory disease.

Therapeutic options for thyroid cancers, indeed, encompass the unilateral or total surgical removal of the thyroid, followed or not by thyroid hormone suppression and ablative therapy with radioactive iodine (RAI) treatment. In addition, a lateral or central neck dissection is also performed in the case of intra- or preoperative evidence of lateral or central neck lymph node involvement. Distant metastatic patients need further therapies since their tumors are often refractory to T4-mediated thyroid-stimulating hormone (TSH) suppression and RAI therapy [7, 8]. Research in the field of new molecules for treating such patients is focused on the angiogenic processes.

The network of angiogenic cytokines and other mediators prompts the interaction between thyroid cancer cells, thyroid normal and stromal cells. Among these factors, there is evidence that genomic, transcriptomic and proteomic variables can be linked to prognosis and therapeutic purposes at different levels. Systems biology and multi-omic approaches are the new research strategies that trace the link between all these variables.

2 The Thyroid Cancer Microenvironment and Angiogenesis

The combination of stromal cells (SCs) and extracellular matrix (ECM) components is the core of the thyroid cancer microenvironment [9]. The interaction between SCs and ECM components gives support and nutrition to cancer cells through a series of mediators, such as growth factors and cytokines [10].

Cancer-associated fibroblasts (CAFs) produce mediators and drive inflammation, immune response, metabolism, and drug resistance [11]. Indeed, poor survival and lymphatic spread characterize a subtype of papillary thyroid cancer with peculiar CAF-related proteins associated with the mutation V600E of the BRAF gene (BRAFV600E) and the loss of PTEN [12,13,14].

Tumor-associated macrophages (TAMs) are another component of the thyroid cancer microenvironment that show high production of IL-10 and low production of IL-12 [15,16,17]. TAMs allow angiogenesis through the secretion of vascular endothelial growth factor (VEGF), platelet-derived growth factor (PDGF), and basic fibroblast growth factor (bFGF) and cause remodeling of the extracellular matrix through the synthesis of matrix metalloproteases (MMPs) [18, 19].

Tryptase-positive mast cells are another cellular marker of thyroid cancer invasiveness and extrathyroidal extension, showing a higher presence than adenoma [20]. Mast cells produce IL-6, TNF-α, CXCL8/IL-8, CCL25/TECK, CXCL10/IP-10, CXCL1/GRO-α, and VEGF, all involved in the transition from epithelium to mesenchyme, angiogenesis, and lymphangiogenesis [21].

Granulocyte colony-stimulating factor and CXCL8/IL-8, produced by thyroid cancer cells, recruit neutrophils that release oncostatin-M and VEGF-A, as well as granules containing elastase, promoting angiogenesis and cancer proliferation and invasiveness [22].

A differential gene expression characterizes the stroma of thyroid cancer compared to normal thyroid. Different genes, indeed, implied in the control of apoptosis, metabolism, and cell response to hypoxia and proliferation, trigger these cellular and proteomic differences in the SCs and ECM components [23].

Neovascularization in the cancer microenvironment includes three different mechanisms [24]:

  1. 1.

    angiogenesis, the sprouting of new vessels from pre-existing ones;

  2. 2.

    vasculogenesis, the formation of new vessels starting from endothelial precursor cells (hemangioblasts);

  3. 3.

    vasculogenic mimicry, the ability of tumor cells or other non-endothelial cells to complete or form new vessels without vascular endothelial cells.

The first mechanism is the most important and accounts for the longitudinal splitting of existing vasculature into two functional vessels (intussusceptive angiogenesis) and the loop-shaped expansion of the vessel (looping angiogenesis) [25].

Multinodular goiter, Graves’ disease, and cancer show increased vascularity. On the other hand, microvascular density (MVD) has been shown to correlate with disease-free survival in thyroid cancers, particularly in PTC and follicular thyroid carcinoma (FTC) [26]. The metastatic spread pattern differs among tumor types, probably because of the influence of proangiogenic and antiangiogenic factors on the phenotype, as well as the expression of receptors, extracellular matrix components and, finally the differential gene expression profile. Consequently, adenomas, microcarcinomas, PTCs, FTCs, undifferentiated, and medullary thyroid carcinomas show considerable differences in metastatic spread.

The following processes (and genes) are involved in the activity of the stromal cells surrounding thyroid cancer: cell survival (RIPK5), proliferation (PTGS2, DUSP5), apoptosis (ZFP36L1, IER3), metabolism (SLCA2A3), organization (RAB7B), response to hypoxia (HIF1A, TUFT1, BHLHB2), and protein degradation (SKP1, KLK-4) [23].

As already explained, a network of processes develops from these differentially expressed genes, allowing the components of stroma and matrix to communicate with cancer cells, inducing angiogenesis and aggressiveness. The release of exosomes by the thyroid cancer cells is another crucial moment of this communication [27].

From a transcriptomic point of view, dysregulation of miRNA influences the features of different types of thyroid cancers through the effects on proliferative signals, resistance to apoptosis, and epithelial-mesenchymal transition [28].

Thyroid hormones, T3 and T4, ligate their receptor site on the ανβ3 integrin and modulate, together with HIF-1α, the activity of VEGF and bFGF on endothelial and other vascular cells [26]. TSH also shows a proangiogenic activity because it stimulates VEGF production through a protein C kinase pathway [24, 29].

In addition, thyroid microvascular activation depends on cAMP- and mTOR-mediated pathways, linked to the iodine deprivation and reactive oxygen species (ROS) production that causes VEGF release [30].

VEGF is not a single factor, but it encompasses a family of proteins, VEFG-A, -B, -C, -D, -E and PDFG, each with a receptor differentially expressed in various cells [31]. Angiogenesis is mediated by VEGF-A, -E and VEFGR-2-neuropilin (NRP)-1, -2, while lymphangiogenesis by VEGF-C, -D and VEGFR-2, -3 [31]. Moreover, cells different from vascular ones, can undergo the influence of VEGF through autocrine and paracrine phenomena, including the following pathways: ROS production, lysophosphatidic acid signaling, c-Jun N-terminal kinase, NF-kB, PI3K/Akt signaling, AP-1 and SP-1 [32]. VEFG-C, angiopoietin-2, VEGFR-2 and -3 over-expression correlates with increased tumor size, aggressiveness and metastasis formation of thyroid cancer [33]. bFGF and FGFR, independently from VEFG, promote thyroid cancer angiogenesis [34].

MMPs, zinc-endopeptidases, contribute to angiogenesis because their degradation of ECM components stimulates the release of angiogenic factors stored with heparan sulphate [35]. MMP-9, in particular, is regulated by epidermal growth factor (EFG) through focal adhesion kinase (FAK) phosphorylation. The balance between MMPs and tissue inhibitors of metalloproteinases (TIMPs), produced by the tumor microenvironment, is at the basis of a switch towards an invasive phenotype, mainly due to increased MMP production [36].

3 Angiogenic Mediators and their Prognostic Value

The introduction of gene classifiers, primarily based on BRAF mutations, on the fine-needle aspiration specimen improved the selection of patients with cytologically indeterminate nodules affected by malignant disease [37]. However, even if every genomic variant of thyroid cancer were described, this information would not be fully useful in clinical practice: many mutations can even be found in normal cells, and many thyroid cancers do not show any mutation [38].

Aggressive DTC shows BRAF mutated, galectin-3, HBME-1, CK19 and estrogen receptor beta. In particular, BRAFV600E is typical of the tall cell variant of PTC, and the same is true for the following hub genes: COL5A1, COL1A1, COL10A1, COL11A1, CCL20, and CXCL5 [39]. BRAF is also altered in the radioiodine refractory DTC and cases with central neck nodal metastases, mainly associated with miR-146b-3p, miR-146b-5p, and miR-222 [40]. Moreover, BRAFV600E identifies cases with a high risk of recurrence. In the case of anaplastic thyroid cancer, BRAFV600E reduces TSP-1 expression, inducing the stabilization of pericytes, which contributes to the secretion of VEFG, PDGF and other factors by vascular and stromal cells, with migration and invasion of neoplastic cells into the new vessels [41].

VEGF-A, -C, PDGF-BB and angiopoietin-2, although higher in patients with the neoplastic and benign disease compared with healthy patients, cannot identify aggressive disease, metastatic disease or large cancers [36]. Nevertheless, VEGF-C, angiopoietin-2, KDR, Flt-4, and TEK, are highly expressed during the angiogenic shift of thyroid cancer and correlate with tumor size, nodal and distant metastases. Both mechanisms, reduced TSP-1 expression induced by BRAFV600E and VEGF overexpression, are associated with an increased microvascular count [24].

Angiogenic stimulation accompanies cell cycle activation in thyroid cancer, which is proven by the association of VEGF with FAL1 and cyclin D1 [42]. However, a regulatory effect of VEGF production pertains to the effects of TSH, opening perspectives for therapies based on recombinant human TSH.

Among circulating markers, miRNAs play a predominant role: in addition to the already cited miR-146b-3p, miR-146b-5p, and miR-222, the exosomes containing miR-21-5p produced by thyroid cancer cells show a significant proangiogenic effect [27, 43].

Medullary thyroid carcinoma (MTC) with metastases is characterized by overexpression of VEGFR-2 and EGFR. Its adverse prognosis correlates with HIF-1α, induced by hypoxia and other signals that coordinately induce VEGF expression [44].

Prostate-specific membrane antigen (PSMA) is a favorable prognostic marker of MTC when found in its neo-vessels, and it could be used as a target for radio-guided imaging and therapy [45]. Moreover, the MMP-2/TIMP-2 ratio is another prognostic factor for MTC [46].

PTEN loss and BRAFV600E seem to drive collagen deposition from an increased number of fibroblasts and its cross-linking, conferring an aggressive behavior to thyroid cancer [14, 47]. Meanwhile, the role of other factors, like macrophages and T lymphocytes, is still uncertain [9].