The role of tumor epithelial-mesenchymal transition and macrophage crosstalk in cancer progression

Abstract

Purpose of review

The purpose of this review is to summarize the recently published findings regarding the role of epithelial to mesenchymal transition (EMT) in tumor progression, macrophages in the tumor microenvironment, and crosstalk that exists between tumor cells and macrophages.

Recent findings

EMT is a crucial process in tumor progression. In association with EMT changes, macrophage infiltration of tumors occurs frequently. A large body of evidence demonstrates that various mechanisms of crosstalk exist between macrophages and tumor cells that have undergone EMT resulting in a vicious cycle that promotes tumor invasion and metastasis.

Summary

Tumor-associated macrophages and tumor cells undergoing EMT provide reciprocal crosstalk which leads to tumor progression. These interactions provide potential targets to exploit for therapy.

Introduction

Cancer is the second leading cause of death in the US with nearly 2 million new diagnoses and over 600,000 deaths projected to occur in 2022(1). At early stages, most localized cancers are curable, but as tumors begin to invade and metastasize, cure becomes rare. While there have been many advances in screening, early detection, and management of early disease, targeted therapeutics for treatment of advanced disease remain limited. An enhanced understanding of the underlying biology of cancer progression is needed in order to develop novel effective treatments for advanced disease.

Although the initiation of cancer growth is caused by genomic alterations, these alterations do not fully explain tumor progression and metastasis. It is widely appreciated the phenotypic plasticity plays a crucial role in tumor progression(2). One hallmark of tumor plasticity is the epithelial to mesenchymal transition (EMT) in which tumor cells transdifferentiate from an epithelial state to a mesenchymal state where they gain motility and invasive ability that enhances their metastatic capability. EMT has been implicated in nearly every cancer type and plays a fundamental role in tumor progression(3). It has become clear that there is an association between EMT and an immunosuppressive tumor microenvironment and a large body of data demonstrates avenues of crosstalk between EMT tumor cells and immune cells(4).

Macrophages compose a large component of the tumor immune microenvironment. Within tumors, macrophages phenotypes can be altered, often resulting in a phenotype that supports tumor progression. In most tumors, increased infiltration of macrophages is associated with poor outcomes(5). Mounting evidence demonstrates that tumor cells play a role in regulating the recruitment and phenotype of macrophages, and that macrophages alter the tumor cells leading to tumor progression. In this review we will summarize the role of EMT and macrophages in cancer progression, explore mechanisms of crosstalk, and discuss therapeutic strategies for targeting this crosstalk.

The Epithelial to Mesenchymal Transition

EMT is a transdifferentiation process in which cells lose epithelial features and gain mesenchymal traits. This transition was first described in embryogenesis and is known to play a vital role in crucial processes such as development, tissue homeostasis and wound healing(6–8). Cancer cells are also able to take advantage of EMT to increase their ability to invade and metastasize. During EMT, cells lose their tight junctions, apical-basal polarity, and undergo reorganization of the cytoskeleton from a rigid to a more flexible structure – features which switch the cell from immotile to motile(8, 9). In addition, EMT leads to an increased ability for the cell to break down extracellular matrix which allows for tumor cells to invade other tissues (7, 9, 10).

The defining and measurable hallmarks of EMT are loss of epithelial markers such as E-cadherin and gain of mesenchymal markers such as N-cadherin and Vimentin(11). EMT is regulated by several key master-regulator transcription factors including SNAI1/2, TWIST1/2, and ZEB1/2(8). These transcription factors, when highly activated, give cells a fully mesenchymal state, but they can also be partially and transiently activated, giving cells a metastable state of high plasticity(8). Each EMT transcription factor also uniquely activates other traits such as stemness, immune evasion, and changes in metabolism(8). Therefore, different transcription factors may lead to unique EMT states. In cancer, the picture is even more complex in that many cancer cells show EMT-like phenotypes without all the stereotypical EMT features and without high expression of EMT transcription factors(12). This suggests that the cancer cells may undergo aberrant programing which leads to EMT that may not always be dependent on the canonical EMT transcription factors(12).

EMT was historically considered to be a binary process in which cells switch from fully epithelial to fully mesenchymal, or the reverse through mesenchymal to epithelial transition (MET). However, it has become increasingly clear in the past couple of decades that the EMT process is not a binary switch but instead occurs along a continuum, with metastable intermediary states termed “partial-EMT” or “hybrid-EMT”(13, 14). The intermediary states often exhibit significant plasticity and it has been proposed that the term Epithelial-Mesenchymal Plasticity (EMP) may more accurately describe the EMT process(15). Intriguingly, multiple studies have shown that it is the partial EMT states, rather than fully mesenchymal states, that have the most capacity for metastasis(14, 16).

There are several widely recognized inducers of EMT. Activation of the Wnt/β-catenin pathway leads to EMT through upregulation of SNAI1 and SNAI2(17). TGF-β signaling is perhaps the most widely recognized inducer of EMT, and TGFβ is used widely in vitro and in vivo to study EMT. TGF-β is thought to induce EMT through phosphorylation of Smad proteins(18). The AKT/mTOR pathway, Notch signaling, and receptor tyrosine kinases have also been widely implicated as EMT inducers(19, 20). In addition to known signaling pathways, there are factors in the tumor microenvironment that are known to cause EMT, either through similar pathways or others, including hypoxia, mechanical stress, and release of cytokines from the immune microenvironment(11).

EMT in cancer

Although EMT has long been implicated as an important factor in cancer metastasis, its role has been heavily debated, largely because of the difficulty in detecting EMT in vivo in humans(9). Distinguishing tumor cell signatures from surrounding stromal cells has been a barrier, especially since stromal cells such as fibroblasts inherently express mesenchymal and EMT markers(9). This is especially a problem in bulk level expression data which is the methodology used in most large cancer studies, including the landmark program, The Cancer Genome Atlas (TCGA). Recent bioinformatic techniques to deconvolve bulk data and newer single cell technologies provide solutions to overcome this barrier. Tumors are also often very heterogeneous, meaning detection of EMT is highly dependent on which area of the tumor is sampled. New spatial biology platforms have the potential to help address this problem. However, even with high resolution technologies, the partial, transient, and reversible nature of EMT make it difficult to detect with confidence. Finally, as discussed above, cancers may develop EMT through aberrant signaling that does not depend on the canonical EMT transcription factors and may be context dependent, making it impossible to develop a “one-size fits all” EMT signatures or markers. Nonetheless, evidence of the importance of EMT in cancer progression is vast(21) and it is recognized as a hallmark of cancer(22).

Macrophages

Macrophages are professional phagocytes present within virtually all tissues and active throughout life. They are central components of both innate and adaptive immunity being specialized at detecting and eliminating foreign substances/pathogens, and presenting antigens that activate adaptive immune responses (23). Beyond immunoregulatory functions, macrophages have pivotal roles in tissue development (24), homeostasis, organ-specific functions (25), repair (26) and regeneration (27). The colony stimulating factor 1 (CSF1) receptor (CSF1R) regulates macrophage proliferation and differentiation (28). Consistent with the critical functions of macrophages in numerous physiological processes, Csf1r mutations in animal models and humans are associated with growth retardation (24), abnormal skeletal development (29), brain malformations (30) and even early postnatal lethality (31).

The ontogeny of macrophages is still incompletely understood, although the current widely accepted notion is that a huge proportion of macrophages in the adult host are replenished by circulating monocytes derived from bone marrow hematopoietic stem cells (HSC). Some tissue-resident macrophages arise from erythro-myeloid progenitors (EMPs) that develop in the yolk sac (32–34). These embryonic-derived macrophages are genetically distinct from HSC progeny and can persist into adulthood with the ability to self-renew (32, 35). Different tissues harbor varied proportions of embryonic- or monocyte-derived macrophages (36) which can shift with age (37) and disease (34, 38). While the mechanisms are yet to be defined, these findings support that the developmental programs within macrophages influence how they respond to disease insults. It should be noted that all lineage tracing studies that examined macrophage ontogeny were performed in murine models and confirming whether human macrophages share similar origins will be difficult, if not impossible, to test rigorously.

Macrophages are highly plastic, capable of rapidly changing their gene expression in response to different stimuli. Historically, macrophages have been categorized into two phenotypes: classically activated “M1” and alternatively activated “M2”. The M1/M2 concept was proposed in 2000 by Mills and colleagues based on their observation that macrophages from C57Bl/6 and BALB/c mice had opposite responses to IFN-γ or LPS stimulation (39). Today, this binary classification is used to simplistically define macrophage function as pro-inflammatory (M1) or anti-inflammatory (M2). Based solely on in vitro work, subgroups of M2 macrophages have also been advocated: M2a, M2b, M2c and M2d – a classification dependent upon the nature of the stimuli and resulting transcriptional changes (40). Over the years, the dichotomous M1/M2 model has become increasingly confusing with a plethora of M1 and M2 markers discovered through in vitro experiments rarely translating to in vivo situation (41). The concept has also been consistently challenged (41–49) and even Mills et al. acknowledged in their original proposition that this classification “certainly could be an oversimplification” (39). In essence, the M1/M2 concept was an in vitro construction where macrophages were extracted from their native environment and thus, would have undergone dramatic phenotypical changes and may not at all resemble their in vivo counterparts. Moreover, as many have argued (42, 44, 47, 49, 50), the M1/M2 nomenclature is not well supported by large-scale transcriptomic data which instead favors a broad spectrum of activation states. Hence, for the remainder of this review, “M1-like” and “M2-like” were used to refer to these conventional activation states.

Macrophages in cancer

Mounting evidence in the past decade has demonstrated that macrophages are key regulators of various pathologies including cancer. They are an abundant immune cell type within the tumor microenvironment, termed tumor-associated macrophages (TAMs), where they can constitute more than 50% of the tumor’s hematopoietic tissue (51, 52). Murine models of cancer revealed that TAMs can arise from both circulating monocytes and yolk sac-derived macrophages (38, 53, 54), and different origins can contribute differently to tumor development(38). Regarding activation, as discussed above for macrophages, TAMs were originally classified as tumoricidal (M1-like) versus protumoral (M2-like). However, this notion warrants reconsideration due to latest discoveries that TAMs have overlapping M1-like and M2-like markers (52, 54, 55) and thus, do not comport with the conventional macrophage polarization model. Moreover, while M2-like TAMs have been largely clinically linked with poor prognosis (56–60), more recent studies have also associated M1-like TAMs with aggressive cancer biology (61–65).

Irrespective of the M1/M2 classification, the density of TAM infiltrate in many types of human cancers is strongly correlated with unfavorable prognosis (56, 66–68). Furthermore, high expression of the macrophage growth factors CSF1 (69–73) and interleukin (IL)-34 (73–75) that bind CSF1R are also associated with tumor progression and poor patient survival. When macrophages are appropriately activated, they can have tumoricidal capabilities (76–78) particularly given their high phagocytic capacities. However, in malignancies, substantial experimental evidence have shown that cancer cells hijack TAM functions to support their own growth and survival (78, 79) mediated via direct cellular interaction (80, 81), efferocytosis (82, 83), secreted molecules (84) or extracellular vesicles (85, 86). This creates a pathological crosstalk whereby cancer cues manipulate TAMs, which in turn support virtually all aspects of tumor development and progression including proliferation, angiogenesis and EMT as well as provide a protective immunosuppressed niche that allows tumor survival (78, 79, 87). Targeting TAMs and their protumoral functions are, therefore, attractive and highly relevant areas of research in the pursuit of effective anti-tumor and combinatorial cancer therapies.

EMT and macrophages co-exist in cancer

The association of EMT or mesenchymal states in tumor cells and high levels of macrophage infiltration has been noted in various cancers including glioblastoma, lung, breast, gastric, oral, liver, melanoma (88–94). A pan-cancer analysis of nearly 8,000 specimens from the TCGA representing 22 distinct cancer types evaluated EMT state (EMT low, intermediate, or high) and corresponding immune cell infiltration scores(95). 16 of the 22 cancer types showed enrichment of macrophages in EMT-high compared to EMT-low samples and enrichment of macrophages had positive correlation with EMT score in all cancer types. In another pan-cancer analysis, Wang et al. performed multi-omics analyses of 17 cancer types and created an EMT-CYT index to quantify the EMT-immunity axis. They found that M2-like macrophages were associated with increased EMT-CYT index(96).

Beyond correlation in human specimens, a vast body of literature supports the existence of crosstalk in tumor cells undergoing EMT and macrophages through in vitro and in vivo work. We will highlight several recent and relevant studies below.

The role of macrophages in inducing EMT

The correlation of EMT and macrophages in human tumors has led to a wide array of in vitro and in vivo studies to explore the mechanisms of possible macrophage-mediated EMT induction.

Chemokines are small, secreted molecules that regulate cell positioning and recruitment into tissues. Multiple chemokine/chemokine receptor pairs have proven important in macrophage-induced EMT in cancer. The chemokine ligand, CCL2, and its receptor, CCR2, are known mediators of monocyte and macrophage recruitment and have been implicated in progression of various cancers(97). In breast cancer, the CCL2/CCR2 axis has been shown to induce EMT in tumor cells through the PI3K/AKT pathway(89). Chen et al. showed that culture of breast cancer cells in conditioned media from M2-like macrophages induced EMT in tumor cells through increased β-catenin signaling(89). The authors showed that M2-like macrophages expressed high levels of CCL2 and inhibition of CCR2 attenuated macrophage-induced EMT and β-catenin signaling (89). Furthermore, they demonstrated that culture of cancer cells in M2-like macrophage conditioned media led to increased phosphorylated Akt, which was attenuated by PI3K/AKT or CCR2 inhibitors. Together, this suggests that TAMs may induce EMT through a CCL2/AKT/β-catenin pathway which may be targetable with CCR2 or PI3K/AKT inhibitors.

Another chemokine involved in monocyte and macrophage signaling is CCL5, which may be more abundantly secreted by macrophages than CCL2(98). CCL5 released by TAMs induces EMT and stem cell-like features in prostate cancer cells through a β-catenin/STAT3 pathway(98). The CCL20-CCR6 chemokine-receptor pair has also been implicated in macrophage-induced EMT. Kadomoto et al. demonstrated that co-culture of renal cell carcinoma (RCC) cells with human monocyte cell line-derived macrophages led to increased EMT and migration of RCC cells through Akt activation via CCL20-CCR6 signaling. This was abrogated by CCL20 neutralizing antibody or Akt inhibitor, AZD5363(99).

As discussed above, TGFβ is a known inducer of EMT and is secreted by M2-like macrophages in an immunosuppressive environment. In lung cancer, M2-like macrophages are the primary source of TGFβ, which drives nuclear translocation of Zeb1 in tumor cells, initiating EMT(94). In head and neck squamous cell carcinoma, co-culture of tumor cells with macrophages led to increased TGFβ and EGF secretion from macrophages which induced EMT in tumor cells (100).

Other factors secreted by macrophages that have been shown to induce EMT in tumor cells include IL-6 and oncostatin M (OSM). In lung cancer, Che et al. demonstrated that IL-6 released by macrophages induced the COX-2/PGE2 signaling pathway which led to β-catenin activation and subsequent EMT in tumor cells (101). In an in vivo glioblastoma model, Hara et al. demonstrated that depletion of macrophages by clodronate liposome reduced the presence of mesenchymal-like tumor cells. Their work also revealed that macrophages release OSM which activates STAT3 in tumor cells thereby inducing EMT (102).

These studies, along with others, support the idea that macrophages secrete a vast array of factors that can induce EMT in tumor cells. These factors likely differ between tumor types and even within tumor types. It is also possible that multiple factors released from macrophages can act in parallel. In order to target macrophage-induced EMT, it will be important to evaluate factors implicated by in vitro and in vivo studies in large numbers of human specimens across different cancer types to look for shared potential targets.

Role of tumor cells in recruiting and polarizing macrophages

In addition to macrophages inducing EMT in tumor cells, tumor cells themselves can incite macrophage recruitment and polarization to tumor-promoting phenotypes. Chemokines are important regulators of this process. For example, in lung cancer, CCL7 has been shown to induce macrophage infiltration and M2-like polarization in vitro (103). Furthermore, in xenograft models formed by CCL7-low cells, there was reduced macrophage infiltration compared to CCL7-high tumors. Upregulation of CCL7 by LINC01094 mediated SPI1 overexpression was the proposed mechanism (103). The CCR6/CCL20 axis has been shown to mediate macrophage recruitment in colon cancer. Nandi et al. showed that monocytes and macrophages expressed CCR6 and migrated towards CCL20 at least as strongly as to CCL2 both in vitro and in vivo (104). The recruited macrophages then secrete inflammatory factors leading to tumor progression. This process is inhibited by CCR6 deficiency suggesting the CCL20/CCR6 axis may be a potential target to prevent macrophage recruitment.

Other factors have been identified in tumors that may regulate macrophage recruitment and polarization. In colon cancer, Wang et al. showed that PCSK9, a protein associated with poor prognosis, promotes tumor cell EMT and inducesM2-like polarization of macrophages likely via regulating MIF levels. Accordingly, co-culture of PCSK9 knockdown cells with macrophages prevented M2-like polarization and promoted an M1-like phenotype (105). In breast cancer, CECR2 is overexpressed in metastatic compared to primary breast cancer lesions and is also associated with M2-like macrophage infiltration (106). CECR2 inhibition suppressed macrophage recruitment and M2-like polarization in-vitro and reduced tumor metastasis in vivo. It was shown that CECR2 works through the NF-kB pathway to upregulate CSF1 and CXCL1, factors associated with macrophage chemotaxis (106).

It is clear from the multitude of studies demonstrating tumor cell-initiated macrophage recruitment and polarization, as well as those revealing macrophage-induced tumor cell EMT, that these two events are related and interdependent. What initiates the cascade is unclear but likely involves tumor genomic alterations and cues from the tumor microenvironment.

The reciprocal feedback between macrophages and tumor cells

A handful of studies have explored the bi-directional crosstalk between macrophages and tumor cells. Alterations in the tumor cells themselves, perhaps starting the EMT process, may be the inciting event that leads to macrophage recruitment. Conversely, it may be macrophage infiltration that occurs first and initiates the alterations in tumor cells that perpetuate the cycle of crosstalk. Further work is needed to understand the underlying drivers, but the studies described here demonstrate what is known about the mechanisms of this pathological crosstalk.

Nguyen et al. investigated macrophage/EMT crosstalk in the development of RCC, which arises from VHL gene mutation in approximately 70% of cases (107, 108). Through in vitro studies and a Vhl conditional knockout mouse model the authors demonstrated that VHL deficient kidney cells secreted IL-6 which attracted macrophages and polarized them to an M2-like phenotype. The polarized macrophages released TGFβ1 and CCL18 which induced EMT in the VHL deficient cells and led to tumor growth and metastasis(108). Thus, it is likely that both EMT and macrophage signaling play important roles in all VHL deficient kidney cancer, which fits with the findings that kidney cancers have extremely high macrophage infiltration and high expression of EMT-related genes(95).

Su et al. demonstrated a GM-CSF/CCL18 axis in breast cancer cells which functioned through activation of the NF-kB pathway (109). Co-culture of mesenchymal breast cancer cells induced M2-like polarization in freshly isolated human monocytes through GM-CSF secretion by the tumor cells. GM-CSF-activated macrophages in turn secrete CCL18 which induces EMT in tumor cells. Interestingly, EMT was reversed when tumor cells were replated in plain media. However, when replated with activated macrophages, the mesenchymal state and high CCL18 expression were maintained. This suggests a protumoral positive feedback loop that was further demonstrated in vivo using a humanized cancer mouse model where neutralizing antibodies against GM-CSF or CCL18 led to prolonged metastasis-free survival (109).

Crosstalk has also been investigated in colon cancer in which M2-like macrophages localize at the leading edge of tumors associated with decreased expression of E-cadherin and increased expression of vimentin (110). Wei et al. found that CD163+ TAMs at the invasive front were associated with an increased ratio of mesenchymal vimentin-expressing circulating tumor cells and with aggressive tumor features. Co-culture of macrophages with colon cancer cells demonstrated the ability of macrophages to initiate EMT in the tumor cells through increased IL-6 secretion which then upregulates FoxQ1 in tumor cells through the JAK2/STAT3 pathway. In addition, co-culture of tumor cells and macrophages led to increased levels of CCL2 which could further recruit macrophages (110).

As more studies explore the role of macrophage/EMT crosstalk in tumors, it is expected that we will discover a multitude of genes and proteins involved that will likely vary between tumors or may interact in an additive manner. It will be important to find common regulators of this crosstalk in order to develop therapeutic ways to prevent or target it. Several potential strategies will be discussed in the next section.

Therapeutic implication of macrophage/EMT crosstalk in cancer

Impact on immune checkpoint inhibitors

There are multiple mechanisms of macrophage and EMT cancer cell crosstalk that may serve as therapeutic targets (Fig. 1). For example, the development of immune checkpoint inhibitors has transformed cancer therapy and provides a new avenue to harness the immune system in fighting tumors (111).The underlying theory is that tumors hide themselves from T-cell killing through upregulation of immune checkpoints, and thus T-cells have been the main focus in immunotherapy development. However, there is evidence that macrophages may function upstream of T-cells and may be as important or more important players in tumor progression and powerful potential targets of immunotherapy. In fact, it has been proposed that PD-1/PD-L1 inhibitors may function through inhibition of macrophage induced immunosuppression (112).

Figure 1. Mechanisms of crosstalk between macrophages and tumor cells undergoing EMT.

The macrophage is on the left and tumor cell on the right. Factors released from macrophages that induce EMT in tumor cells through various mechanisms are depicted on the upper half of the cells. Factors released from tumor cells that lead to macrophage recruitment and polarization are shown in the lower halves of the diagram. Drug categories, and examples of each, that may target this crosstalk are depicted in red. Figure created with BioRender.com.

There is a reason to believe that crosstalk between tumor cells and macrophages may involve T-cells, and further multi-dimensional studies are required. For example, Hara et al. found that OSM released from macrophages led to activation of STAT3 in glioblastoma tumor cells and a mesenchymal like state. Through co-culture with CD8 T-cells, the authors demonstrated that the macrophage-induced mesenchymal tumor cells were associated with increased numbers and cytotoxicity of T-cells. In addition, the mesenchymal tumor cells were more prone to T-cell killing (102). It is possible that EMT or a mesenchymal state in other cancers may promote cytotoxic T-cell function as well, possibly making these tumors more susceptible to immune checkpoint inhibitors. Future studies will be necessary to shed light on the role of macrophage and T-cell crosstalk.

Targeting EMT in tumor cells

Therapeutics targeted at EMT inducers, regulators, or effectors is one strategy to overcome tumor progression. Multiple small molecule TGFβ inhibitors or antibodies against TGFβ have been developed and are currently in phase 1 and phase 2 clinical trials for solid tumors including Galunisertib, Vactosertib, AVID200, and Fresolimumab (113). Therapies targeting upstream signaling pathways associated with EMT such as WNT/β-catenin inhibitors and AKT inhibitors are in very early stages of phase 1 clinical trials (113). Overall, due to the plasticity and heterogeneity of EMT in cancer and the difficulty of measuring it, therapies targeting EMT have been slow in development. However, advances in single cell and spatial technologies will undoubtedly enhance our understanding of EMT in cancer leading to development of new therapies.

Targeting macrophages in cancer

Another therapeutic method is to inhibit macrophage function in tumors, thereby reducing their ability to enhance EMT. One strategy is to block the CCL2/CCR2 axis. A CCL2 blocking agent, carlumab, showed promise in inhibiting tumor growth in vivo but unfortunately has not been as effective in clinical trials for metastatic castrate resistant prostate cancer and other solid tumors (114, 115). CCR2-directed therapy has been shown to be safe and tolerable in a phase 1b clinical trial of pancreatic cancer in combination with FOLFIRINOX chemotherapy and showed improved response rates compared to chemotherapy alone (116). Findings from ongoing trials will shed light on the efficacy of these therapies.

CSF1 or CSF1R blockade is another strategy under investigation, which has been shown to alter the macrophage composition towards a more antitumor population (117). CSF1R blockade has shown efficacy in clinical trial of the rare tumor, diffuse-type tenosynovial giant cells, where 71% of patients had significant clinical benefit (118). Multiple ongoing trials will help determine the efficacy of these therapies in other tumors (119).

A novel method under investigation is the use of chimeric antigen receptor (CAR) T cells to target specific macrophage populations. Rodriguez-Garcia et al. created T cells with a CAR directed to FRβ which has been found to be highly expressed on M2-like macrophages (120). This therapy was effective in reducing FRβ+ macrophages in murine models and also induced CD8+ T cell recruitment and reduced tumor growth (120).

Another novel future strategy may be through use of drug carrying nanomaterials. Nanomedicine has several advantages over traditional drug therapies including the ability to carry multiple drugs targeted at different factors and the ability to be engineered to physically penetrate the targeted cell (121). For example, Zhang and Palmer developed a hemoglobin modified liposome which undergoes specific uptake by macrophages through CD163 receptor mediated endocytosis. These liposomes could be loaded with any agent to destroy or reprogram macrophages (122). A multitude of other nanomaterial based strategies have been developed that may prove useful in macrophage-targeted therapies (121).

Future directions

Due to the heterogeneity of molecular factors involved in the EMT/macrophage crosstalk present in tumors, it will be important going forward to target common pathways that may be involved with a variety of specific genes or proteins. For example, STAT3 plays a role in multiple tumor cell/macrophage interactions and STAT3 inhibitors exist that could be effective in EMT-high cancers. The Wnt/β-catenin, PI3K/AKT, and NF-kβ pathways are also common players in macrophage/EMT crosstalk and may be important pathways that could target both EMT and macrophage infiltration.

Conclusion

The cyclical crosstalk between macrophages and mesenchymal tumor cells or those undergoing EMT plays a crucial role in tumor progression. This crosstalk presents a powerful potential target in advanced cancers where current effective therapies are limited. Further studies are needed to understand how the different associated genes, proteins, and pathways in this crosstalk interact and vary between cancer types and between individual tumors within any cancer type. It will be crucial to find similarities that can be broadly targeted in order to effectively perform drug development and clinical trials. Additional work is also needed to understand the complexities of macrophage/tumor cell crosstalk and other factors in the microenvironment such as T-cells. Such understanding may lead to better directed use of current therapies such as immune checkpoint inhibitors, and lead to more effective immunotherapies.