Tumor microenvironment: Sanctuary of the devil

Highlights

• Interactions of malignant and non-transformed cells create tumor microenvironment.

• Microenvironment plays a pivotal role in survival and drug resistance of cancer cells.

• Combination therapy targeting both tumor and microenvironment is being conducted.

Abstract

Tumor cells constantly interact with the surrounding microenvironment. Increasing evidence indicates that targeting the tumor microenvironment could complement traditional treatment and improve therapeutic outcomes for these malignancies. In this paper, we review new insights into the tumor microenvironment, and summarize selected examples of the cross-talk between tumor cells and their microenvironment, which have enhanced our understanding of pathophysiology of the microenvironment. We believe that this rapidly moving field promises many more to come, and they will guide the rational design of combinational therapies for success in cancer eradication.

Introduction

In the past two decades, many oncogenes and tumor suppressor genes have been identified. Subsequently, these genes were mapped to the signaling pathways that regulate cell growth or apoptosis, and various anti-tumor agent and therapeutic methods were developed. The survival of cancer patients has been significantly extended; however, the tumor relapse or recurrence is almost always developed eventually with resistance to the initially effective drugs. The cancer therapy has encountered a bottleneck. An emerging concept is that the maintenance and expansion of tumors also strongly depend on external signals from their microenvironment [1], [2], [3].

To fully understand tumor development and progression, a deeper knowledge of the interactions between cancer cells and their microenvironment is needed [3], [4]. Important open questions have been announced:

• What are the specific components of a Tumor Microenvironment (TME)?

• Which signals do cancer cells transmit to and receive from the TME?

• What is the role of infiltrated immune cells to tumor progression?

• How do cancer cells develop chemoresistance in the TME?

Indeed, a series of studies work on the tumor microenvironment were published throughout the past few years. Here we put a brief overview of the TME and highlight a few major recent studies focusing on the cross-talk between tumor cells and their microenvironment. We hope that this mini-review can offer a taste of this rapidly moving field, as well as suggestions regarding further advances.

The TME is the cellular environment in which the tumor exists. Apart from the tumor cells, the TME includes surrounding blood vessels, the extracellular matrix (ECM), other non-malignant cells, and also signaling molecules [4], [5]. By using cell-type-specific markers, researchers have identified different types of normal cells in the TME, including stromal cells, fibroblasts, immune cells (such as T lymphocytes, B lymphocytes, natural killer cells and natural killer T cells, Tumor-associated macrophages, etc.), as well as pericytes and sometimes adipocytes (Fig. 1). The stromal cells and fibroblasts in TME can secrete growth factors, such as hepatocyte growth factor (HGF), fibroblast growth factor (FGFs) and CXCL12 chemokine, which can not only promote growth and survival of malignant cells but also function as a chemoattractant that stimulates the migration of other cells into the TME [6]. Different T cell populations and B cells can be found at the invasive margin of tumors; however, it is still controversial whether the presence of these cells in TME reflects a good or bad prognosis [7]. For many solid tumors, the appearance of natural killer cells and natural killer T cells in TME predicts a good prognosis [8]. Tumor-associated macrophages are abundant in most human and experimental murine cancers, and their activities are usually pro-tumorigenic [9]. Beyond the contributions of specific cell types to TME, the extracellular matrix (ECM) is another major component. ECM is composed of a large collection of biochemically distinct components including proteins, glycoproteins, proteoglycans, and polysaccharides with different physical and biochemical properties [10]. ECM can not only provide a physical scaffold for all cells in the TME but also is abundance of key growth factors. Different cell types in the TME supply distinct ECM proteins. ECM plays a critical role in the development of tumor, which is commonly deregulated and becomes disorganized in later stage of tumor progression. Abnormal ECM can also dysregulate the behavior of stromal cells in the TME and facilitate angiogenesis and inflammation [11]. Interestingly, primary tumors of diverse metastatic potential differ in their composition of ECM components. Indeed, the composition of the extracellular TME has been used a predictor of clinical prognosis [2], [11]. The tumor vasculature is abnormal in almost every aspect of its structure and function [12]. The vasculature of the tumor is always inadequate to meet the demands of the growing mass, leading to hypoxic and acidotic regions of the tumor. When a quiescent blood vessel senses an angiogenic signal from the hypoxic conditions in the TME, angiogenesis is stimulated and heterogeneous new vessels with chaotic branching structures sprout from the existing vasculature [13]. In addition, the tumor blood vessels exhibit an uneven vessel lumen and are usually leaky, which raises interstitial fluid pressure leading to unevenness of blood flow and nutrient, as well as drug distribution, in the TME. This, in turn, increases hypoxia and facilitates tumor development. These unique characteristics of the TME distinguished it from the corresponding normal tissues, and there is emerging evidence that extrinsic stimulations mediated by the microenvironment play a pivotal role in survival and drug resistance of the tumor cells [14].

Hematopoietic stem cells (HSCs), which were localized in bone marrow (BM), are one of the best characterized adult stem cell types. The relevance of the BM microenvironment in regulating HSC behavior has only recently been established [15], [16], [17]. Complex bidirectional interactions between the BM niche and HSCs were essential for the maintenance of HSC quiescence and normal hematopoiesis. Moreover, there are growing evidence supporting the idea that the BM microenvironment also plays a pivotal role in the initiation and propagation of leukemia [1], [18].

Recent data indicated that leukemic cells hijack the homeostatic mechanisms of normal HSCs and take refuge in the BM niche [19], [20] (Fig. 1). By using dynamic in vivo imaging system in a mouse model, Colmone et al. showed that leukemic cell growth disrupts normal HSC BM niches, and the niches were converted into an environment that favors cell proliferation and growth in parallel with the leukemogenic events [20]. The normal oxygen saturation is relatively low in BM compared to other organs. It is advantageous for stem cells to localize at the sporadic hypoxic niches, since the exposure to high levels of ROS induces senescence and dysfunction in stem cells. It has been proved in both animal models and clinical samples that the BM is highly hypoxic in the context of hematologic malignancies [21]. Researchers believe that the progression of leukemia is associated with expansion of hypoxic niches and stabilization of the oncogenic hypoxia inducible factor-1α (HIF-1α). In a very recent study, Duan et al. recorded that normal BM niche of immune-deficient mice was damaged by the dissemination of human acute lymphoblastic leukemia (ALL) [22]. The authors characterized the initial HSC niche in a mouse model, which was primarily composed of nestin-expressing mesenchymal stem or stromal cells (MSCs). With the progression of leukemia, these nestin positive cells in the niche were mostly lost and replaced by α-smooth muscle actin (αSMA) expressing stromal cells. Furthermore, similar results were observed in mice bearing xenografts from primary human ALL samples. Moreover, Alexandre et al. also confirmed that B-CLL clone has markedly impact on MSCs and disrupts BM niches in vivo by quantifying the colony forming unit-fibroblasts (CFU-Fs) [23].

Osteoblasts are first defined as specialized, terminally differentiated products of MSCs. Subsequently, osteoblasts have been implicated in normal hematopoietic processes, as they can synthesize dense, cross-linked collagen, and several additional specialized proteins, which comprise part of the HSC niche. Previous studies in mice have associated mutations in osteoblasts with impaired renewal and expansion of HSCs [24], [25]. More recently, a new study showed that mice with an activating β-catenin mutation that is specifically expressed in osteoblasts develop acute myeloid leukemia (AML) [26]. Kode and colleagues generated a mouse model with continues expression of constitutively active β-catenin that confined to osteoblasts. Mice with this mutation exhibited cellular anomalies that were characteristics of AML in humans. By carrying out three different transplantation experiments, the authors confirmed that the mutation in osteoblasts is sufficient to drive the development of leukemia in mice.

Furthermore, it has been proposed that the reprogramming of tumor stroma is a potent enabler of malignancy. Scherz-Shouval et al. described that regulator heat shock factor 1 (HSF1) is frequently activated in cancer-associated fibroblasts (CAFs), which indirectly promotes malignancy in adjacent cancer cells by enabling proliferation, invasion and metastasis [27].

Interactions between tumor infiltrating immune cells and tumor cells have been of great interest because of the dual roles of immune cells and the produced factors. It is a commonly held belief that immune responses prevent and inhibit tumor development; however, recent evidence has also suggested that immune cells in the TME interact intimately with the transformed cells to promote oncogenesis [28], [29]. Successive changes occurring at the tumor site during tumor progression resemble chronic inflammation, and the process has been described with the metaphor that “tumors are wounds that do not heal”, which seems to be largely orchestrated by the tumor and seems to promote tumor survival. Studies have shown that tissues with chronic inflammation generally exhibit high cancer incidence [7]. In addition, the established tumors often exist with an immunosuppressive microenvironment that can block productive antitumor immunity.

Of the lymphoid lineage cells, many different T cell populations have been observed within the TME, especially at the invasive tumor margin and in draining lymphoid organs [30]. Among these, CD8⁺ memory T cells and CD4⁺ T helper 1 (Th1) T cells function as the major anti-tumor immune effector cells. Among the many factors Th1 and CD8⁺ T cells secreted, their characteristically produced cytokine IFN-γ is the most significant cytokine preventing and suppressing the development of cancers. Studies have shown that high numbers of these immune cells in the TME are correlate with a good prognosis in many types of solid tumor [31]. However, not all T cells are anti-tumor effectors. The roles played by CD4⁺Th2 cells, Th17 cells, Tregs, CD4⁺CD25⁺Foxp3⁺ regulatory T lymphocytes in tumor development and survival remain elusive, as contradictory results have been documented both in animal models and in clinical analysis [31], [32], [33]. The infiltrated B cells in TME can promote disease progression by secreting pro-tumorigenic cytokines. Their importance in supporting tumor growth is evident in B cell–deficient mice, which exhibit resistance to engraftment of certain syngeneic tumors. Innate cytotoxic lymphocytes, natural killer (NK) cells and natural killer T (NKT) cells are also found in TME, and their appearance are linked with a good prognosis in many cancers [2], [34].

Among the myeloid lineage cells that are found in TME, myeloid suppressive cells (MDSC), mast cells, and most of the tumor associated macrophages (TAM) serve to promote tumor development [2]. The MDSC are immunosuppressive precursors of dendritic cells, macrophages and granulocytes. They maintain normal tissue homeostasis in response to various systemic infection and injury. It has been demonstrated in several animal models that these TME infiltrated MDSC can promote tumor vascularization and disrupt major mechanisms of immune-surveillance by interfering antigen presentation by dendritic cells (DC), T cell activation, and NK cell cytotoxicity. Mast cells are also shown being recruited to tumors, where they release factors that enhance proliferation of endothelial cells to promote tumor angiogenesis. The presence of tumor associated macrophages (TAM) in TME, frequently exhibit an M2 phenotype, supports tumor angiogenesis and invasion. M2-type TAM-derived cytokines IL6, TNF, IL-1β and IL-23 are generally recognized as dominate tumor-promoting forces [35], [36]. However, the classically activated M1 macrophages are pro-inflammatory and anti-tumorigenic. Similar to TAMs, neutrophils have been shown to have opposing functions in regulating cancer progression and metastasis, indicating their ultimate role in cancer development may rely heavily on the conditions of TME. The DCs are derived from the bone marrow and play a key role in inducing and maintaining the antitumor immunity. However, their antigen-presenting function may be lost or inefficient in the TME. DCs presenting tumor-specific antigens are being developed as vaccines to induce immune responses to regress tumors and prevent relapse [37].

The impact of tumor-infiltrating immune cells has been debated for decades. To our knowledge, these immune cells play dual roles with potential to either eliminate or promote malignancy depending on the complex signals of TME [8], [29]. The prospect of effective immunotherapies for the treatment of patients with cancer is now becoming a clinical reality. Certain antibody therapies have demonstrated the potential for directing a patient's own immune system against tumors. Researchers are making great efforts to understand the tumor microenvironment and characterize the location and status of immune cells and their interaction with tumor cells. Additionally, a deeper understanding of the signaling cascades active in immune recognition of cancers is crucial, which will provide important targets for cancer immunotherapy [2], [38].

The role of the microenvironment in tumor development was originally proposed by Paget in the “seed and soil” hypothesis. There is increasing evidence supporting the idea that tumor development is strongly dependent on cellular interactions with the TME [6], [39] (Fig. 2). Majority of the tumors respond to initial treatment; however, disease relapse is common. It is believed that the cancer stem cells or part of the cancer cells are protected from current therapies through signals from the niches, which eventually leads to the selection of secondary genetic changes and outgrowth of malignant cells with pharmacologic resistance [40], [41]. This mechanism is pivotal during chemotherapy and contributes to disease relapse.

BM niches have been proposed to have roles in the development and progression of hematopoietic malignancies [42]. Leukemic cells undergo spontaneous apoptosis once they are removed from the BM microenvironment and cultured in vitro without supportive stroma, highlighting the importance of external signals from the microenvironment in maintaining these cells [43]. BM stromal cells play an important role in regulating hematopoietic cell development through the production of cytokines, chemokines, and intracellular signals initiated by cellular adhesion [44]. Moreover, these stromal cells are pivotal for maintaining quiescence of HSCs. During the past few years, researchers found that these molecular mechanisms may also facilitate leukemia stem cell (LSC) survival [1].

It is well known that co-culturing with BM derived mesenchymal stromal cells can protect AML cells from the chemotherapeutic drug–induced apoptosis; however, the mechanisms through which these cells promote malignancy in their neighbors are less well understood. Notably, CXCL12, CCL2, IL-8, and other cytokine/chemokine levels were increased in co-culture supernatants. Chemokine CXCL12 elaborated by osteoblasts, CAR cells or MSCs is one of the key factors mediating the crosstalk between leukemic cells and the TME, which regulates the homing and engraftment of leukemia (stem) cells into the BM niche. It has been shown that culturing of AML cells with CXCL12 promotes their survival, whereas adding neutralizing CXCR4 antibodies, CXCL12 antibodies, or the CXCR4 inhibitor AMD3100 significantly decreases it [43]. Moreover, weekly administration of anti-human CXCR4 antibody to mice previously engrafted with human AML cells leads to a dramatic decrease of human AML cells in BM, blood, and spleen in a dose- and time-dependent manner [45]. Taken together, these data indicate that CXCL12–CXCR4 interactions in the BM microenvironment contribute to the chemoresistance of leukemic cells. In addition to the CXCL12–CXCR4 interactions, changes of several other canonic signaling pathways have been found during the interaction between leukemia cells and BM niche, which may be responsible for the TME-mediated drug resistance [1], [46]. For example, constitutive activation of PI3K/mTOR signaling occurs in the majority of human AML [47], and the nuclear factor kappa B (NFκB) pathway is activated in BM in chronic lymphocytic leukemia (CLL) patients [48]. CCL2 and IL-8 are involved in the IL-17 signaling pathway which is highly involved in the inflammatory process in the tumor microenvironment [49]. Additionally, CD44 is a key regulator of leukemia stem cells homing to BM niches and maintenance of their primitive state [50]. HIF-1α was demonstrated to regulate CXCL12 gene expression in both MSCs and endothelial cells, resulting in selectively high in vivo CXCL12 level in TME, which increased migration and homing of leukemia cells [51].

The vasculature within the tumor microenvironment is also known to have tumor-promoting effects, such as tumor formation and progression [52]. Rafii and colleagues recently showed that tumor endothelial cells participate in an intricate crosstalk with lymphoma cells and promote lymphoma cell invasiveness and resistance to chemotherapy [53]. The results from a series of in vitro and in vivo experiments demonstrated that FGF4 produced by lymphoma cells activates FGFR signaling in endothelial cells, leading to overexpression of Notch ligand Jag1. In turn, endothelial cell derived Jag-1 were shown to activate Notch2 signaling in neighboring lymphoma cells that contribute to the lymphomagenesis, expansion, invasiveness and resistance to chemotherapy. And besides, CCL2 paracrine signaling has been shown to improve the angiogenesis in breast tumors, leading to decreased survival [54].

It is becoming increasingly clear that exosomes, 40–100 nm membrane vesicles of endocytic origin secreted by most cell types, have specialized functions in intercellular signaling. The role of exosomes as mediators between cancer cells and TME has gained increasing attention [55]. Two recent studies have provided evidence that these small vesicles contribute to the tumor progression and resistance to therapy. Melo et al. investigated the microRNAs (miRNAs) contained in exosomes to tumor progression. Six miRNAs, as well as the corresponding pre-miRNA and key components of the RISC-loading complex, were observed in the cancer exosomes derived from breast cancer cell line MDA-MB-231. These results suggest active miRNA biogenesis in exosomes. Notably, incubation of MCF10A cells (a non-tumorigenic human mammary epithelial cell line) with cancer exosomes induced oncogenic changes, whereas exosomes from normal cells and serum of healthy donors did not [56]. Boelens et al. found that stromal-derived exosomes transfer 5′-triphosphate RNA to breast cancer cells, leading to the RIG-I mediated radiation resistance. Moreover, the Notch pathway in cancer cells was subsequently activated through STAT1, which expands radiation-resistant cells [57].