Footprints of microRNAs in Cancer Biology

MicroRNAs (miRNAs) are short non-coding RNAs involved in post-transcriptional gene regulation. Over the past years, various studies have demonstrated the role of aberrant miRNA expression in the onset of cancer. The mechanisms by which miRNA exerts its cancer-promoting or inhibitory effects are apparent through the various cancer hallmarks, which include selective proliferative advantage, altered stress response, vascularization, invasion and metastasis, metabolic rewiring, the tumor microenvironment and immune modulation; therefore, this review aims to highlight the association between miRNAs and the various cancer hallmarks by dissecting the mechanisms of miRNA regulation in each hallmark separately. It is hoped that the information presented herein will provide further insights regarding the role of cancer and serve as a guideline to evaluate the potential of microRNAs to be utilized as biomarkers and therapeutic targets on a larger scale in cancer research.


Introduction
The hallmarks of cancer were initially proposed by Douglas Hanahan and Robert Weinberg to organize the principles that provide a logical framework for understanding the cellular mechanisms of oncogenesis. These hallmarks were classified into six major categories and had been very influential in cancer research [1]. After about a decade of intense research regarding the fundamentals of cancer biology, a revision regarding the classification was proposed to incorporate the current knowledge of cancer development [2]. The organization of the hallmarks of cancer allowed researchers to understand the core traits of cancer, regardless of the origins of the cancer cells. As our understanding of cancer improved over the past two decades, updated lists of cancer hallmarks have been produced, culminating in the most recent review by Fouad and Aanei in 2017 titled "Revisiting the Hallmarks of Cancer". The distinct hallmarks are selective proliferative advantage, altered stress response, vascularization, invasion and metastasis, metabolic rewiring, abetting microenvironment and immune modulation [3].
Numerous studies have reported the involvement of miRNAs in various human cancers [4][5][6]. This comes as no surprise, as microRNAs (miRNA) have been known to play pivotal roles in many major cellular functions, such as development, differentiation, growth and metabolism [7]. MiRNAs are small, highly conserved non-coding RNAs with an average length of 22 nucleotides that are primarily involved in the regulation of gene expression. These miRNAs may either assume an oncogenic role or tumor-suppressive role in cancer development, depending on the target gene [8,9]. Given the indispensable role of miRNAs in cancer, an in-depth description regarding the involvement of microRNAs (miRNAs) in each cancer hallmark will be presented in this review.

Hallmark 1: Selective Proliferative Advantage
One of the defining characteristics of cancer cells is their ability to grow and proliferate uncontrollably compared to normal cells, which leads us to the first hallmark of cancerselective growth and proliferative advantage. Cell growth and proliferation is a tightly regulated process in normal cells; however, alterations in this process allow the cancer cells to grow and proliferate uncontrollably [10]. This can be achieved via multiple pathways, for example growth ligands, growth receptors, cytosolic signaling and cell cycle regulation [11,12]. By altering these pathways, the cancer cells send signals that promote growth and proliferation while compromising signals that inhibit growth.

Intracellular Signal Pathways Dysregulation in Cancer Cells
Binding of growth factors to receptors on the surface of the cell activates a series of intracellular signal networks namely JAK-STAT, the mitogen-activated protein (MAP) kinase and the phosphatidylinositol 3 (PI3) kinase pathways. Disruption in any one of these pathways may give rise to defective mitogenic signaling. All three pathways are illustrated in Figure 1 and further explained below [13][14][15][16].

JAK-STAT Pathway
The JAK-STAT pathway comprises three main components: tyrosine-kinase-related receptor, Janus tyrosine kinase (JAK) and signal transducer and activator of transcription proteins (STAT). JAK is a family of cytoplasmic proteins that aids in transducing extracellular signals to intracellular downstream cascades. Growth factor molecules bind to their respective receptors on the plasma membrane, which activates intracellular cytoplasmic JAKs that proceed to recruit STATs to the receptor. STATs are phosphorylated by JAKs and other cytosolic serine and threonine kinases, such as ERK1/2. This is followed by its translocation from the cytoplasm into the nucleus. STAT then binds to promoters of the target gene and activates gene transcription.

MAPK Pathway
The mitogen-activated protein kinase (MAPK) pathway, also known as the Ras-Raf-MEK-ERK pathway, mediates cell proliferation. In response to the binding of growth factor to the receptor, inactive RAS (a GTPase) is converted to its active form. RAS activation is followed by RAF activation, which subsequently leads to the phosphorylation of downstream MEK1/2 followed by ERK1/2. Activated ERK1/2 then mediates the transcription of target genes such as c-MYC and others.

Phosphatidylinositol 3 Kinase (PI3K) Pathway
PI3K-Akt-mTOR pathway is also involved in cell proliferation. Binding of growth factors to receptors located on the plasma membrane activates PI3K, which catalyzes the conversion of phosphatidylinositol 4,5-biphosphate (PIP2) to phosphatidylinositol 3,4,5-triphosphate (PIP3). PIP3 then executes its function by recruiting Akt to the membrane to be activated by phosphoinositide-dependent kinase 1 (PDK1). Akt activation then relieves the inhibition on mTORC1, which is mainly involved in protein synthesis. On the other hand, phosphatase and tensin homolog (PTEN) dephosphorylates PIP3 and impairs Akt activation.

Cell Cycle Dysregulation in Cancer Cells
Cell cycle is divided into the interphase (G 1 , S and G 2 phase) and mitotic phase. The proliferative fate of a cell not only depends upon its entry into the cell cycle but also upon its complete progression through each phase. Various regulatory proteins act as gatekeepers of the cell cycle and deregulation often leads to formation of cancer. The regulatory proteins that are actively involved in safeguarding the cell cycle are retinoblastoma (RB) protein and cell cycle checkpoint proteins such as cyclin-dependent kinases (CDK), cyclins and cyclin-dependent kinase inhibitors (CDKi). Figure 2 illustrates the cell cycle regulated by crucial cell cycle checkpoint proteins.

Retinoblastoma Pathway
The RB pathway comprises RB-CDK4/6-cyclin D. Its function is to control the cell cycle progression from the G 1 phase to the S phase. Cyclin D binds to CDK4/6 to form a complex, and together they phosphorylate Rb and render Rb inactive. Hypo-phosphorylated Rb initially binds to E2F and repress its action. Nevertheless, Rb phosphorylation leads to release of E2F from transcriptional repression, which then allows for transcription of S-phase-promoting genes, thereby promoting cell progression from G1 to S phases [10].

Cell Cycle Checkpoint Proteins
Cyclin and cyclin-dependent kinases always work in tandem. The most crucial cyclin-CDK complexes in the cell cycle are cyclin E-CDK2 and cyclin D-CDK4/6. Both complexes are associated with G1 phase to S phase progression. On the other hand, cyclin-dependent kinase inhibitors (CKI) contribute to abrogation of cell cycle progression mediated by growth-inhibitory signals. Examples of CKIs include p27, p21 and p16. Cell cycle progression from the S phase to G2 phase and mitotic phase is also tightly regulated by other cyclin-CDK complexes and various regulatory proteins [17].

The Role of miRNAs in Selective Proliferative Advantage
MiRNAs are involved in the regulation of proliferation-related signal pathways; miRNAs can exert either direct or indirect effects on proliferative signal pathways. Direct effects are mediated by directly targeting the 3 untranslated region (3 UTR) of mRNA of genes that directly activate or inhibit cell proliferation. Indirect effects occur by targeting upstream or downstream regulators of genes that directly regulate cell proliferation.
MiRNAs are known to affect and dysregulate intracellular signal pathways. It was found that miR-101 directly targets the 3 UTR of MAPK kinase 1 (MEK1) mRNA, which is involved in the MAPK/ERK pathway. This results in attenuation of diffuse large B cell lymphoma (DLBL) cell proliferation [18]. Another miRNA that is implicated in the regulation of the PI3K/Akt pathway is miR-20a, which modulates its oncogenic actions in multiple myeloma by negatively regulating PTEN, a tumor suppressor that exerts its inhibitory effects on Akt activation and ultimately cell proliferation [19]. MiRNA may also disrupt the functions of cell cycle checkpoint kinases directly. Retinoblastoma (Rb), one of the crucial players in cell cycle progression, is a direct target of miR-590 in T cell acute lymphoblastic leukemia reported by Miao et al. [20]. Downregulation of Rb by miR-590 leads to an increase in cell proliferation. Table 1 summarizes other miRNAs involved in this hallmark with regards to their target, mechanisms of action and cancer types.

Hallmark 2: Altered Stress Response
The next hallmark is an altered stress response favoring overall survival. Cells have a variety of adaptations and responses to stress that damage the cells such as DNA repair, apoptosis, autophagy and senescence. Cells might be eliminated to maintain the overall health of the tissue under stressed conditions; however, cancer cells in their quest for continual growth and survival will alter their stress response [37]. Four types of stress response, namely DNA repair, apoptosis, autophagy and senescence, are highlighted below.

DNA Repair Pathways
The cells in our body are constantly faced with various genotoxic insults on a daily basis that can lead to DNA lesions. DNA lesions can either occur endogenously or exogenously (ultraviolet rays, ionizing radiation, genotoxic chemicals). These DNA lesions must be repaired promptly, as they may cause genomic instability and contribute to the overall survival of cells. DNA lesion repair is curated by the DNA damage response in our body. The typical DNA damage repair (DDR) includes: (1) sensing the damage; (2) checkpoint activation; (3) DNA damage repair.
Initiation of DDR starts with the formation of the DNA damage sensor MRN complex comprised of MRE11, Rad50 and NBS1 at the site of the lesions. Each of the players in this complex has its own distinct role. MRE11 (meiotic recombination 11 homolog) exhibits DNA exo-and endonuclease activity. Nijmegen breakage syndrome protein 1 (NBS1) recruits other DNA damage regulatory proteins to the lesion site, whereas Rad50 binds DNA ends together [38]. Binding of the MRN complex to the site of the lesion then activates and recruits ataxia-telangiectasia mutated (ATM) or ataxia-telangiectasia and rad3-related (ATR). ATM is normally recruited when there is a double-strand break (DSB), whereas ATR is recruited when there is a single-strand break (SSB) or replication stress [39]. BRCA1 is a key player in DNA repair machinery that is activated by both ATM and ATR [40]. H2AX, acting as DNA damage sensor, is recruited to the lesion site and is phosphorylated by ATM into γH2AX. Then, γH2AX recruits other DNA damage repair factors to initiate the repair process [41]. Poly(ADP-ribose) polymerase 1 (PARP1) is one of the first critical responders to DNA damage (DNA strand breaks), which acts by mounting the DNA repair mechanism accordingly [42].
Checkpoint activation is crucial for DNA repair processes. Activated ATR phosphorylates checkpoint kinase 1 (Chk1), which then goes on to bind to CDC25C, marking it for degradation by the ubiquitin pathway. This action results in inhibition of cell cycle progression. Chk1 also exerts its inhibitory action on cell cycle progression by activating p53. Chk2 is activated by ATM and shares the same function as Chk1 [43].
DNA damage repair is based on the decision to undergo cell cycle arrest, DNA repair or apoptosis mediated by p53. P21 and WEE1 kinase are the downstream effectors of p53. P21 induces cell cycle arrest, whereas WEE1 prevents entry into the mitotic phase. As for DNA repair, p53 modulates the repair mechanism by regulating genes involved in DNA repair pathways, such as base excision repair, mismatch repair, nucleotide excision repair, translesion DNA synthesis (TLS), non-homologous end joining (NHEJ) and homologous recombination (HR) [44,45]. The DNA repair mechanism is illustrated in Figure 3. The miRNAs are able to regulate the DNA repair mechanism by targeting the players involved in the repair machinery via direct or indirect action. DNA damage repair involves vast networks that are interrelated, while miRNA regulation of any players in the network may affect the repair process, either positively or negatively. PARP1 is the direct target of miR-7-5p, leading to abrogation of DNA damage repair in small cell lung cancer as reported by Lai et al. [46]. The same finding was reported in cervical cancer by Yang et al. [47]. Furthermore, miR-203a-3p negatively regulates ATM in ovarian cancer cells, leading to cell cycle arrest [48]. Other examples of miRNAs that participate in DNA repair are listed in Table 2.

Autophagy
Autophagy is a cellular process defined as the breakdown of damaged proteins or organelles and recycling of the macromolecular components for the benefit of other metabolic processes in the cells. Dysregulation of autophagic pathways may either promote or inhibit tumor progression. Cancer cells may hijack the autophagy mechanism and use it to their advantage in ensuring overall survival [59].
The autophagy process is typically made up of five stages, namely initiation, vesicle nucleation, vesicle elongation, vesicle fusion and degradation, which take place in the cell cytoplasm. Autophagy may be initiated in response to nutrient starvation or any other signals. Under normal conditions, the mTORC1 complex inhibits the ULK1 complex, which sets off the autophagy cascade by inducing the formation of phagophore; however, in cases of nutrient starvation, mTORC1 complex inhibition of the ULK1 complex is released, initiating autophagy. The mTORC1 complex is made up of three components, namely mTOR, mLST8 and RAPTOR. The ULK1 complex is comprised of ULK1, ULK2, ATG13, ATG101 and FIP200.
The second stage includes the formation of Beclin1 complex induced by the ULK1 complex, which is made up of various autophagic proteins, such as BCL2 interacting protein (Beclin1), activating molecule in Beclin-1-regulated autophagy (AMBRA1), phosphatidylinositol 3-kinase catalytic subunit type 3 and regulatory subunit 4 (VPS34 and VPS15), UV radiation resistance-associated gene protein (UVRAG) and ATG14. Beclin1 complex formation can be negatively regulated by BCL-2 and BCL-X L to inhibit autophagy.
Vesicle elongation is modulated by two ubiquitin-like conjugation systems that aid in the formation of autophagosomes. For the first conjugation system, ATG5 is conjugated to ATG12 with the help of enzymes such as ATG7 and ATG10. The ATG5-ATG12 conjugate binds to ATG16L1 to act as a facilitator to conjugate microtubule-associated protein 1A/1B LC3 (LC3-I) to phosphatidylethanolamine (PE). The second system involves conjugation of LC3-I with PE to form LC3-II mediated by ATG4B, ATG7 and ATG3. This stage ends with incorporation of LC3-II into the autophagosome membrane.
The fourth stage revolves around the fusion of autophagosome with lysosome mediated by SNARE proteins, including syntaxin-17 (STX17), synaptosome-associated protein 29 (SNAP29) and vesicle-associated membrane protein 8 (VAMP8), together with Rab7. Degradation of autophagosome contents is then carried out by the pH-sensitive enzymes in the lysosome [59,60]. The autophagy mechanism is illustrated in Figure 4. The Role of miRNAs in Altering Autophagy Mechanisms Autophagy can be regulated by miRNAs at various stages of autophagy signal pathways by exerting either direct or indirect actions. Direct autophagy-promoting action is seen when overexpression of miR-423-5p occurs in hepatocellular carcinoma following treatment with sorafenib, which increases the levels of ATG7 and LC3-II [61]. Indirect miRNA regulation can be seen in miR-423-5p, which induces autophagy by directly targeting Bcl-2-like protein 11 (Bim), a negative regulator of Beclin1, which is a crucial autophagy regulator in gastric cancer [62,63]. Regarding the autophagy-inhibiting miRNA, miR-409-5p targets and negatively regulates FIP200, which is crucial for assembly of the ULK1 initiation complex in ovarian cancer as reported by Cheng et al. [64]. Other examples of miRNA involved in autophagy are depicted in Table 3.

Apoptosis
Apoptosis is a type of cell death mechanism that is considered a normal biological process. It governs and maintains the balance between cell survival and cell death. A deranged apoptotic mechanism may prompt the cells to undergo malignant transformation, or in other words to exhibit cancer-promoting effects.
Apoptosis occurs via two distinct pathways, namely intrinsic and extrinsic pathways. The extrinsic (death receptor) pathway occurs when the death receptors on the surfaces of the cells are stimulated by extracellular ligands such as TNF (tumor necrosis factor), Fas-L (Fas ligand) and TRAIL (TNF-related apoptosis-inducing ligand). The binding of the ligand to the respective receptors induces the assembly of death-inducing signalling complex or DISC, which is comprised of three components: Fas-associated death domain (FADD) and protein caspases 8 and 10. Downstream caspases such as caspase 3 and caspase 7 are then activated, thereby triggering apoptosis.
Regarding the intrinsic or mitochondrial-mediated apoptotic pathway, activation is triggered by various intra-or extracellular stress signals, such as oxidative stress, irradiation, toxic agents and others. Incoming stress signalling activates pro-apoptotic proteins such as BAX and BAK, which induces mitochondrial outer membrane permeabilization (MOMP). This action is counteracted by antiapoptotic BCL-2 family proteins (BCL-2 or BCL-X L ) or MCL-1. Cytochrome c and SMAC are then released from the mitochondrial intermembrane space into the cytosol. Cytochrome c binds with apoptotic protease-activating factor 1 (APAF1) to form apoptosome, which functions by activating caspase 9 and goes on to further activate caspase 3 and caspase 7, ultimately leading to apoptosis. SMAC aids in the apoptotic pathway by inhibiting the caspase inhibitor X-linked inhibitor of apoptosis protein (XIAP).
Both extrinsic and intrinsic pathways merge via the action of caspase 8, which cleaves and activates BH3-interacting death domain agonist (BID). BID exerts its role by activating BAX and BAK [75][76][77]. The apoptotic process is illustrated in Figure 5. The miRNAs can regulate apoptosis through promotion or inhibition, depending on the target. MiR-224 is overexpressed in breast cancer and directly targets CASP9 (caspase 9), thereby inhibiting apoptosis [78]. A recent finding reported that XIAP, an antiapoptotic gene, is negatively regulated by a novel miRNA, miR-CHA1, leading to apoptosis induction in non-small cell lung cancer [79]. APAF1, an activator of caspase 9 in the intrinsic pathway, is inhibited by miR-484 in non-small cell lung cancer, thereby abrogating apoptosis [80]. Other examples of miRNAs involved in apoptosis are depicted in Table 4.

Senescence
Senescence refers to irreversible cell cycle arrest. Senescence can be divided into three different groups: replicative, stress-induced and oncogene-induced senescence. Replicative senescence occurs due to telomere shortening, which is largely responsible for cell replication. Stress-induced senescence occurs in response to various stress stimuli, such as radiation, oxidative stress, cytotoxic agents and other genotoxic stress. Oncogeneinduced senescence takes place following overexpression of oncogenes such as RAS G12V or BRAF V600E [91]. Senescence can act as a potent barrier against carcinogenesis by halting cell proliferation and forcing the cell to undergo permanent cell cycle arrest. Dysregulation of the senescence pathway may either promote or inhibit malignant transformation [92].
The two main signalling pathways involved in senescence are the p16/pRB pathway and p53/p21 pathway. Upon receiving stress signalling from upstream regulators, p16 INK4A (cyclin-dependent kinase inhibitor) is activated. Bmi-1 (B-cell-specific Moloney murine leukemia virus integration site 1) acts as a negative regulator of p16. Activation of p16 results in its binding to CDK4/6, thereby inhibiting the phosphorylation of pRB (retinoblastoma protein). Inactive Rb blocks cell cycle progression by binding to and inactivating transcription factor E2F, which promotes the entry of cells from G1 to S phases. With this process, cell cycle progression is blocked and senescence is induced. Upstream signalling upregulates the expression of p14 ARF , which functions by inhibiting the activity of MDM2, a p53 inhibitor. Following this, p53 is activated, which in turn induces the expression of its downstream effector p21, a cell-dependent kinase inhibitor (CKI) that blocks cell cycle progression by inhibiting the formation of the cyclin-CDK complex. This in turn leads to activation of senescence [93]. The process of senescence is illustrated in Figure 6. The senescence pathway is highly regulated by miRNA and can cause positive or negative effects on senescence, depending on the target genes affected. For example, Bmi-1, a negative regulator of p53, is downregulated by miR-128 overexpression, thereby directly inducing senescence in glioma cells [94]. Furthermore, miR-30 evades senescence by indirectly targeting two key senescent effectors, p16 and p53, via the downregulation of CHD7 and TNRC6A, respectively. CHD7 is a cotranscriptional activator of p16, whereas TNRC6A is involved in p53 activation [95]. Other examples of miRNAs involved in senescence are listed in the Table 5.

Hallmark 3: Vascularization
Vascularization is another hallmark of cancer, whereby cancer cells promote the formation of blood vessels to deliver nutrients for fast-growing solid tumors. The most well-known process of vascularization is angiogenesis. In normal cells and tissues, the angiogenesis is a controlled process that is turned on or off depending on the needs of the cells; however, in cancerous cells and tumors, the angiogenesis process is continuous and there is a dysregulation of pro-and antiangiogenesis factors [104]. This continuous activation of angiogenesis allows the cancer cells to form blood vessels to obtain sufficient nutrients for continuous growth and proliferation. There are other ways tumors can achieve vascularization, such as vascular co-option, intussusceptive microvascular growth and vasculogenic mimicry [105].

Vascularization Mechanisms in Cancer Cells
Vascularization, also known as angiogenesis, is the formation of new blood vessels surrounding a solid tumor into other ducts within the body. Vascularization generally starts when a solid tumor grows to a certain size, as this creates the need for extra nutrients and oxygen to be supplied to the tumor microenvironment for propagation of the primary tumor. This is triggered when there is low oxygen within the tumor microenvironment (Hypoxia). Hypoxia induces HIF1-α (hypoxia-inducible factor-1 alpha) expression, leading to the activation of downstream factors that are crucial for vascularization [106][107][108].
VEGF (vascular endothelial growth factor) is an HIF1-α induction-dependent factor and a potent inducer of tumor vascularization. It was found that anthracycline treatment in prostate cancer-xenografted mice, which blocks the HIF1-α DNA binding potential, attenuates vascular formation via downregulation of the VEGF activity. The results also showed that the reduction of VEGF leads to impaired growth of prostate cancer [109].
In addition, cellular protease was also found to be a contributor in tissue vascularization. An example is matrix metalloproteinase (MMP), a protease that is transcriptionally activated by HIF1-α [110,111]. It was found that fibroblasts surrounding the tumor could also affect angiogenesis; fibroblasts secrete factors crucial for MMP production in neighboring tumor cells [112]. Furthermore, downregulation of MMP attenuates angiogenesis, further supporting the suggestion that vascularization is MMP-dependent [113].
The changes in the genes mentioned earlier affect angiogenesis by modulating the tumor microenvironment, thereby affecting crucial proteins found most predominantly in tight junctions, as well as other cell-to-cell junctions, such as adherens junctions and desmosomes. Additionally, exosomal secretion into the extracellular matrix (ECM) could also affect cell-to-cell junctions, which contribute to angiogenesis [114,115].
The regulation of vascularization via miRNA can be either direct or indirect. Direct regulation can be observed when the miRNA targets both activator and suppressor genes involved in tissue vascularization via 3 -UTR binding on their mRNAs. Similar miRNA-mRNA hybrids occur through indirect regulation; however, these miRNAs target specific factors (transcription cofactors) that influence genes involved directly in vascularization. Control can occur at different levels (exosomal, proteomic, genomic and transcript) of the central dogma of molecular biology, thereby leading to angiogenesis. The microRNA regulation associated with cancer angiogenesis is illustrated in Figure 7.

The Role of miRNAs in the Vascularization of Cancer Cells
Cancer tissue vascularization requires specific signalling from various factors for its formation. These factors are regulated by miRNAs. Two high-risk miRNAs, namely miR-148a and miR-30, which regulate HIF1-α via binding directly to its inhibitor FIH1 (factors inhibiting HIF1-α) in the glioblastoma was reported by Wong et al. [116]. Inhibition of these miRNAs results in the downregulation of the HIF1-α protein, which corresponds to the reduction of VEGF expression and attenuation of vascularization. This is an example of the effects of cofactor targeting via miRNA binding, which influences the activity of transcription factors that directly activate gene expression.
Another interesting miRNA control process occurs when the cancer itself secretes miRNA via exosomes, thereby affecting neighbouring cells. In this case, these would be endothelial cells, which allow for high vascular permeability. This was observed in colorectal cancer cells (CC), whereby exosomal secretion from the CCs containing the miR-25-3p significantly affected the vascular integrity [117]. Another study also found that hepatocellular carcinoma cells (HCCs) overexpressed miR-210, which was found in high abundance in HCC secretion (HSS). Further experimentation revealed exosome-rich miRNA, whereby treatment of HSS on HepG2 resulted in the induction of tubal formation by downregulating SMAD4 and STAT6. Furthermore, direct targeting of the miRNA processing via DROSHA downregulation attenuates angiogenesis [118]. Other examples of miRNAs involved in vascularization are shown in Table 6.

Hallmark 4: Invasion and Metastasis
The process of vascularization also leads to the next hallmark of cancer, which is invasion and metastasis. This hallmark is a defining feature of malignant tumors characterized by their ability to spread and invade neighbouring tissues. Metastasis is also the main cause of cancer mortality, as the cancer cells travel from the tissue of origin and colonize other distant organs or tissues. This leads to cancerous growth on multiple sites in the body, which compromises the bodily function and ultimately leads to death [142].

Mechanisms of Invasion and Metastasis in Cancer Cells
Proliferation of cancer cells can be continuous post-vascularization, as there are now means for providing nutrients to the primary tumor. Metastasis usually follows vascularization. Metastasis is a process of invasion where the primary tumor obtains the means to propagate to other parts of the affected individuals from the primary tumor site. A brief overview regarding metastasis is illustrated in Figure 8. Figure 8. A brief overview of cancer metastasis. Cancer cells loses their adhering factors, leading to intravasation, a process whereby the cancer cells detach from the primary tumor site and invade the circulatory system. Once they are in the circulatory system, cancer cells will be disseminated throughout the body, followed by extravasation, where the cancer escapes the circulatory system and forms a secondary tumor.
Proteins (adhering factors-AF) found on these sites are crucial in intercellular crosslinking (desmosomes), signal exchange (tight junction or gap junction) and cytoskeletal connection (Adherent Junction). Aberrant changes of the AFs can affect cell-to-cell adhesion, allowing the tumor to metastasize [150,151]. Some of the examples of adhering factors found on cell-to-cell junctions and their implications in metastasis are shown in Table 7. Table 7. Adhering factors found on cell junctions and their implications in metastasis.

Cell Junctions Adhering Factors Implication in Cancer Metastases Reference
Gap Junction Connexin 43 Brain cancer cells secrete cGAMP to astrocytes via connexion 43 channels, leading to STAT1-NF-κB-mediated metastasis [152] Zonula Occluden (ZO-1) ZO-1 was downregulated following overexpression of upstream regulator ZIP4 expression. This induced tumor migration [153] Tight Junction Claudin Crucial for cell anchorage. Breast cancer cells with high Claudin-2 have higher liver metastatic potential [154] Occludin Occludin upregulation suppresses metastatic potential of squamous cell carcinoma [155] Adherens Junction Cadherins Deletion of e-cadherin results in the development of both local and distant metastasis [156] Catenins B-catenin nuclear localization is crucial for ZEB1 transcriptional activation, which negatively regulates metastasis [157] Desmosomes Desmoglein 2 Loss of desmoglein enhances tumor invasiveness and migration [158] Armadillo repeat units containing proteins (ARM)/plakoglobin It was found that inhibition of proteins with the ARM structure and plakoglobin enhances the metastatic ability of bladder cancer and lung cancer, respectively [159,160] Desmocollin Desmocollin 3 downregulation leads to Akt pathway activation and decreases e-cadherin abundance in colorectal cancer, thereby enhancing metastatic potential. [161]

The Role of miRNAs in Metastases and Invasion of Cancer Cells
Dysregulation of gene expression usually leads to cancer metastasis. Some cancer cells acquire the characteristics of other cell types in order for them to metastasize. An example can be seen in breast cancer bone invasion, whereby miR-301a-d regulates DKK-1, RUNX-2 and ITGA5 genes, which are involved in osteogenesis, leading to breast cancer osteomimicry. Low expression of miR-30 was observed in tumor samples, and induction of miR-30 expression was followed by attenuation of bone metastases [162].
Furthermore, mir-331 and miR-195 were also reported to have metastatic implications. It was found that these circulating miRNAs could be prognostic markers in luminal A breast cancer. Furthermore, mir-331 and miR-195 target a cohort of genes that is crucial for Akt signalling and epithelial mesenchymal transition (EMT), in which both are crucial for metastasis [163]. Various other examples of miRNAs implicated in cancer invasion and metastasis are shown in Table 8.  6. Metabolic

Drivers of Metabolic Reprogramming
Oncogene RAS: Oncogenic RAS is frequently upregulated in cancer and its aberrant signalling contributes to altered metabolism. RAS signalling promotes glucose uptake by upregulating the expression of glucose transporter GLUT1. RAS signalling also stimulates the glycolytic pathway, which is the master regulator of aerobic glycolysis, also known as the Warburg effect [178].
Oncogene MYC: MYC functions as a transcription factor and is involved in various oncogenic processes. Aberrant MYC signalling is common in cancer, leading to altered metabolism. Active MYC signalling is associated with upregulation of metabolic enzymes, such as lactate dehydrogenase A (LDHA) and pyruvate kinase (PKM2) of the glycolytic pathway [179,180]. MYC also induces the utilization of glutamine as an alternative energy source [181].
Tumor suppressor 53 (TP53): Loss of p53 can contribute to alterations in metabolic pathways. Furthermore, p53 impairs glucose metabolism by inhibiting the transcription of glucose transporters GLUT1 and GLUT4 [182]. The glycolysis pathway that is preferentially utilized by cancer cells is impaired by p53 through direct downregulation of hexokinase 2, an enzyme involved in glycolysis and indirectly through inducing the expression of PARK2, a negative regulator of HIF-1a [183,184].

Hypoxia-inducible factor 1 (HIF-1):
HIF1 is a transcription factor that is stimulated in response to hypoxia. Since the hypoxic environment is commonly found in most cancers, there is no doubt that HIF1 is also frequently upregulated in cancer cells. Additionally, HIF1 signalling can be activated by other factors such as oncogenes. One of the main roles of HIF is as a master regulator of aerobic glycolysis or the Warburg effect in cancer cells. HIF1a activates the glycolytic pathway by upregulating glucose metabolism enzymes, such as LDHA, PKM2, HK1 and HK2, and by increasing glucose transporters (GLUTs) [187][188][189]. HIFs are negatively regulated by von Hippel-Lindau protein (pVHL) and factor-inhibiting HIF1 (FIH-1) [190,191].

Alteration of Metabolic Pathways in Cancer Cells
Carbohydrate metabolism: One of the key players associated with cancer metabolism is glucose. Cancer cells require a greater abundance of glucose than normal differentiated cells to meet their higher energy demand. Instead of relying on more efficient oxidative phosphorylation for glucose production, cancer cells choose to opt for less efficient glycolysis, even under nor-moxic conditions. This phenomenon is referred to as the 'Warburg effect' or as aerobic glycolysis (Figure 9). Glycolysis supplies the cancer cells with the building blocks that are needed for macromolecule synthesis [192,193]. Cancer cells reprogram glucose metabolism to work in their favor through various mechanisms, with a few examples being increasing the number of glucose transporters (GLUTs) and upregulating glycolytic enzymes such as hexokinase, pyruvate kinase and lactate dehydrogenase [194]. The end product of glycolysis, pyruvate, then feeds into the tricarboxylic acid cycle (TCA) to produce citrate as an alternative energy source.
Lipid metabolism: Lipid metabolism is often deregulated in cancer and mostly contributes positively to cancer development. Lipids are of great importance for cells, as they are needed to form the lipid bilayer of plasma membrane, which aids in cell proliferation. De novo lipid biosynthesis by cancer cells exhibits greater resistance against oxidative stress. Furthermore, lipid metabolism contributes to carcinogenesis by providing alternative energy sources and in the synthesis of signalling molecules, such as hormones. Examples of crucial genes involved in lipid metabolism are sterol regulatory element-binding protein (SREBP), fatty-acid-binding protein (FABP) and adipose-differentiation-related protein (ADRP), while the crucial enzymes are fatty acid synthase acetyl-coA synthetase (ACSL), acetyl-CoA carboxylase (ACC) and others [195,196].
Amino acid metabolism: Amino acids are needed as the building blocks for protein synthesis. Glutamine, a critical amino acid, is utilized by cancer cells to produce glutamate for use in the TCA cycle or for glutathione synthesis in the antioxidant system. The main enzyme that participates in this pathway is glutaminase (GLS) [197,198].

The Role of miRNAs in Metabolic Rewiring
The miRNAs participate in the regulation of metabolic rewiring by either inducing or inhibiting the expression of metabolic-related genes. For example, GLUT1 and HKII are negatively regulated by miR-124, which subsequently leads to impaired glycolysis in nonsmall cell lung cancer [199]. A recent report demonstrated that miR-31-5p can promote the Warburg effect by downregulating HIF-1a inhibitor (FIH) activity, which ultimately results in increased glycolysis and ATP production, further sustaining lung cancer cells [200]. Other examples of miRNA involved in metabolic rewiring are depicted in Table 9.

Hallmark: Tumor Microenvironment
The tumor microenvironment (TME) plays a vital role in the development, progression and eventual metastasis of cancer. It has been identified that the fundamental mechanisms governing interactions between various components of the TME and tumor cells encompass vastly dynamic factors, including hypoxia, as well as multiple cell types, such as cancerassociated fibroblasts (CAFs) and macrophages [211]. While cancer cells have been known to secrete a multitude of microRNAs to neighbouring and distant cells via exosomes to augment their functions, mounting evidence has also implicated the role of the TME in contributing to further supplementing malignant cells with factors that favor their survival and progression [211]. A recent review in this area extensively covered the influence of microRNAs in hypoxia, angiogenesis and the interplay of various cell types [211,212]. The most recent research in this area further unraveled the novel microRNAs involved in the TME, which will be covered in more detail in this section.

miRNAs Involved in Cancer-Associated Fibroblasts (CAFs)
The tumor microenvironment consists of various cell types. Among those that play the most vital roles in the progression of the disease are cancer-associated fibroblasts (CAFs). CAFs have been known to interact with tumors via a multitude of mechanisms, including exosomes, which convey biological instructions through the transport of metabolites, long non-coding RNA (lncRNAs), proteins and microRNAs [213]. In a reciprocal manner, exosomes from tumors are able to transform the function of CAFs, which often leads to an enhanced microenvironment favoring the survival and development of tumors, while exosomes from CAFs can be internalized by tumor cells, and in most cases can partake in the progression and metastatic formation of cancers [213].
Recent research on CAF-secreted microRNAs affecting cancers of the oral cavity revealed that they regulate tumor-inhibitory and tumor-promoting pathways. The study is done by using oral squamous cell carcinoma (OSCC) patient samples identified that miR-382-5p was overexpressed in CAFs as compared to normal fibroblasts. In vitro assays showed that CAFs overexpressing miR-382-5p promoted the migration capabilities and invasiveness of OSCC cells [214]. In contrast, another research demonstrated that miR-34a-5p was able to repress OSCC cell proliferation and metastasis. This microRNA, which directly targets AXL, was also found to be capable of inhibiting tumorigenesis in xenograft models. Collectively, this study indicated that activation of the miR-34a-5p/AXL axis was able to confer aggressiveness to OSCC via the AKT/GSK-3β/β-catenin/Snail signalling cascade [215].
A recent study revealed that exosomal miR-139 from CAFs was able to repress the progression of gastric cancer by inhibiting matrix metalloproteinase 11 (MMP11) [216]. Initial experiments indicated that there is a significant downregulation of miR-139 in CAFs of the gastric cancer microenvironment. Further analysis indicated that MMP11 was the direct target of miR-139. The researchers attempted to shuttle miR-139 into CAFs to increase their bioavailability to gastric cancer cell lines in vitro and to stomach tumors in vivo, and found that both experiments resulted in drastic decreases in MMP11 expression. Further analysis showed significant reductions in invasiveness in vitro and repression of tumor progression and metastasis in vivo. The data from this study indicated that miR-139 produced in gastric CAFs may repress the progression and development of gastric cancer metastasis by modulating the level of MMP11 in the surrounding tumor microenvironment. Another study showed that the suppression of miR-214 in CAFs leads to increased migration and invasion abilities of stomach cancer cells [217]. This study further revealed that these characteristics were induced by the microRNAs' target FGF9, which is linked to the further development of EMT. Experimentation regarding the overexpression of this miRNA suppressed the migration and invasion of gastric cancer cells in vitro. Moreover, it was found that the use of mimetics led to elevation of E-cadherin and suppression of Vimentin, N-cadherin and Snail, denoting repression of EMT of GC cells; thus, this study indicates that miR-214 is able to repress the tumor-promoting capabilities of CAFs via targeting of FGF9 to regulate the EMT process of gastric cancer cells.
Interestingly, some microRNAs have been found to suppress the conversion of normal fibroblasts into cancer-associated fibroblasts. It was demonstrated that miR-124 produced by human ovarian surface epithelial cells (hOSECs) was able to suppress the conversion of normal fibroblasts into cancer-associated fibroblasts (CAFs) in ovarian cancer [218].
Analysis of exosomal miRNA of ovarian cancer samples showed a marked decrease in the expression of miR-124, whereas the opposing finding was found in normal hOSECs. Further analyses indicated that normal fibroblasts with suppressed miR-124 exhibited characteristics of CAFs, including upregulation of α-SMA and FAP, which led to enhanced migratory and invasive capabilities. Experiments to reverse the condition via ectopic expression of miR-124 in CAFs led to the attenuation of α-SMA and FAP expression and counteracted motility and invasion traits. Further, the direct target for miR-124 is the sphingosine kinase 1 (SPHK1) gene transcript involved in the regulation of cell proliferation, adhesion, chemotaxis, migration and tumor growth, among others. This study provided evidence that ovarian cancers, via downregulation of miR-124, mediate the CAF transition to mold the tumor microenvironment for optimal oncogenesis. Further to this, it was reported that miR-141-3p was able to attenuate gastric-cancer-mediated transformation of normal fibroblasts and BMSC into CAFs. Normal fibroblasts with suppressed miR-141-3p exhibited features of CAFs, including enhanced migratory and invasive capabilities. Additionally, miR-141-3p was able to hinder the migration and invasion of gastric cancer cells and repressed the transformation of normal fibroblasts and BMSC into CAFs. Additionally, it was identified that the direct target for miR-141-3p is the STAT4 gene transcript. These data collectively demonstrated that miR-141-3p exerts its actions by regulating the STAT4/wnt/β-catenin pathway [219].

miRNAs in Hypoxia
Hypoxia occurs naturally in the tumor microenvironment (TME) as a result of oxygen deprivation due to cancer growth, and is able to alter cell-to-cell interactions, as well as molecular signalling. The roles of microRNAs in the hypoxic TME have previously been shown elsewhere [211,212].Since then, several research groups have identified additional miRNAs that regulate key pathways during hypoxia, which will be covered in the following section.
Numerous research publications have linked the onset of hypoxia to the acquisition of increased cellular proliferation, migration and invasiveness of cancer cells. It is noteworthy that although hypoxic tumors tend to activate these same capabilities but often activate different mechanisms, different pathways and microRNAs are employed to achieve this end. A recent study demonstrated that miR-590-5p is induced under low-oxygen conditions in colorectal cancers and is able to aid in disease progression by regulating the activity of matrix metalloproteinases. RECK, the direct target of this miRNA, enhances the invasive and migratory ability of cancer cells when suppressed by activating matrix metalloproteinases (MMPs) and filopodia protrusions; thus, this study showed that downregulation of miR-590-5p leads to inhibition of tumor proliferation and metastasis in mouse models of CRC [220].
Further, recent research on hepatocellular carcinoma (HCC) identified that miR-196-5p is inducible under hypoxic conditions and contributes to tumor progression and liver cancer metastasis. Clinical samples from HCC patients showed significantly low levels of miR-196-5p, with further in vitro and in vivo experiments on the ectopic overexpression of this miRNA, demonstrating substantial impairment of HCC growth and metastasis. Additionally, it was found that miR-196-5p exerted its function through regulation of the high-mobility group AT-hook 2 (HMGA2) gene transcript [221].
In another related study, it was discovered that miR-210 mediated the epithelialmesenchymal transition (EMT) in pancreatic cancers under hypoxic conditions [222]. It was reported that as the level of miR-210 increased under hypoxia in PANC-1 cell lines, the expression of HIF-1α and NFκB also elevated in tandem whilst HOXA9 decreased. HOXA9 was proven to be the direct target of miR-210. Ectopic expression of miR-210 under normoxic conditions led to decreasing levels of EMT epithelial markers, which included E-cadherin and β-catenin, while increasing the expression of mesenchymal markers, including vimentin and N-cadherin. This led to the net result of increasing cell migration and invasiveness. Additionally, it was disclosed that NFκB levels also increased, further enhancing the migration and invasiveness of the cells. In contrast, experiments using miR-210 antagonists on hypoxic PANC-1 cells showed reversal of the EMT, HOXA9 and NFκB gene expression, which led to decreased cell migration and invasiveness. The data from this study collectively showed that under hypoxic conditions, miR-210 suppressed levels of HOXA9 to activate the NFκB pathway, which drives EMT in pancreatic cancer.
Studies on specific microRNAs found in different types of cancer cells often show that they target different mRNA transcripts to exert their function, thereby affecting dissimilar pathways. This has been observed with miR-210, which was found to target HOXA9 in pancreatic cancers, although for prostate cancers that target was determined to be the neural cell adhesion molecule (NCAM). A recent study identified that miR-210 expression in prostate cancer is induced by hypoxia and is involved in regulating neural cell adhesion [223]. Taken together, these studies show that miR-210 plays a vital role in regulating cellular responses to hypoxia and provides evidence that the regulation of various adhesion molecules is crucial for the progression of multiple cancers.
In addition to cell migration and adhesion molecules, it has been observed that hypoxia of the TME also induces the dysregulation of microRNAs affecting cellular metabolism. In a recent study on liver cancer, it was discovered that miR-885-5p directly targets Hexokinase 2 (HK2) to regulate the Warburg effect [210]. This study revealed marked suppression of miR-885-5p in HCC tissues and cell lines. Ectopic expression of miR-885-5p in hypoxic models of HCC led to substantial suppression of growth and migration in in vitro and in vivo models. Further, the overexpression of miR-885-5p in vitro led to marked reductions in glucose uptake and lactate production via suppression of several glycolytic enzymes, thereby providing evidence that it is involved in the regulation of the cancer cells' glycometabolic activity. Additional analysis also revealed that miR-885-5p binds to the 3 UTR transcript of hexokinase 2. The data from this study collectively suggested that the miR-885-5p/HK2 axis has additional potential to be explored as a therapeutic target and prognostic biomarker of liver cancer. A further summary of the microRNAs that control the tumor microenvironment is shown in Table 10.

Hallmark 7: Immune Modulation
Cancer cells are constantly being monitored and eliminated by the immune system; therefore, cancer cells must have a way to avoid detection by the immune system, which is achieved via immune modulation. This achieved is through cancer immunoediting, which involves the selection of cancer cells that are able to evade the immune system. Cancer immunoediting involves three phases, namely elimination, followed by equilibrium and finally escape [233]. The role of miRNAs in immunomodulation of the cancer microenvironment is well recognized. Recent reviews in this area have extensively covered the influence of miRNAs in immune-modulatory molecules and immune cells as well as immune checkpoints such as PD-1/PD-L1 [234][235][236][237]. The latest research in this area has further unraveled the novel miRNAs that contribute to the survival of tumors by dysregulating several key elements of the immune system involved in the surveillance of cancers, including tumor-associated macrophages (TAMs), T cells, myeloid-derived suppressor cells (MDSC) and natural killer (NK) cells, which will be covered in more detail in this section.

miRNAs and Tumor-Associated Macrophages
The roles of tumor-associated macrophages (TAMs) in cancers have been well studied. TAMs are known to exhibit functional plasticity, which enables wide-ranging phenotypes from immune-stimulating to immune-suppressive states [238]. The phenomenon of TAM reprogramming and expansion to promote oncogenicity is purportedly influenced by tumor-derived factors [239].
Cancers that carry the p53 mutation have reportedly been able to reprogram macrophages to support tumor growth via exosomes containing miR-1246. Colon cancer cells expressing mutp53 release miR-1246-enriched exosomes to adjacent macrophages, which reprograms them into an anti-inflammatory immunosuppressive state, favoring enhancement of TGF-β activity [240]. This immunomodulatory mechanism alters the microenvironment substantially to induce cancer progression and subsequent metastasis, leading to poor survival of colon cancer patients.
Recent research has also showed that high expression of miR-21 in TAMs is associated with the progression of tumor growth [239]. It was further revealed that miR-21-depleted TAMs promote the antitumor response by undergoing transcriptional network rewiring, which contributes to a proinflammatory angiostatic state. This results in enhancement of cytotoxic T cell activity via expression of cytokines and chemokines, showing that regulation of miR-21 has potential therapeutic implications.
Aside from tumors, transformed TAMs are also able to secrete miRNA-enriched exosomes to further regulate the microenvironment to favor cancer progression. It was identified in epithelial ovarian cancer (EOC) samples that miR-29a-3p and miR-21-5p released by TAMs to neighbouring CD4 + T cells were able to inhibit STAT3 expression. Additionally, these miRNAs also induce imbalance of Treg/Th17 cell regulation, leading to further inhibition of STAT3 [241]. These data imply the possible development of an EOC treatment by targeting exosomal miRNAs.
Recently, research has shown that miR-148b, which controls the expression of colonystimulating factor 1 (CSF1), is downregulated in metastatic hepatocellular carcinoma cells (HCC) [242]. The authors also reported that the dysregulation of miR-148b led to progressive growth and metastasis of HCC via CSF1/CSF1 receptor-mediated TAMs infiltration, concluding that miR-148b plays a suppressor role in HCC.

miRNAs and Myeloid-Derived Suppressor Cells (MDSCs)
Myeloid-derived suppressor cells (MDSCs) are a subset of cells normally present in infection and cancer microenvironments. This heterogeneous group of cells are known to have immunosuppressive activities, although their exact role has not been elucidated.
It was recently discovered that MDSC expansion in gastric cancers was induced by miR-107 secreted by tumor cells [243]. The authors found that miR-107 was not only abundant in gastric cancer cells but also in the secreted exosomes. The exosomes, when taken up by MDSCs, target and suppress expression of DICER1 and PTEN genes, which are responsible for regulating MDSC proliferation and activation of the PI3K pathway, respectively. This study concluded that gastric cancers were able to induce the expansion and activation of MDSC, while the downregulation of miR-107 serves as a novel therapeutic intervention for gastric cancer.
Another study showed that a subset of MDSCs of the stomach express miR-130b during helicobacter-induced spasmolytic polypeptide-expressing metaplasia (SPEM) [244]. It was determined that mir-130b was required for T cell proliferation suppression and that its levels in the blood correlate to the metaplastic changes of the stomach. These data imply that expression of miR-130b in gastric MDSCs could be explored as a marker for metaplastic changes that potentially lead to stomach cancer.
The presence of MDSCs is a known impediment that negatively affects cancer immunotherapy. Huber and colleagues in their study on melanomas, determined that a set of microRNAs, which included miR-99b, miR-100, miR-125a, miR-125b, miR-146a, miR-146b, miR-155 and let-7e, is associated with MDSCs and enables treatment resistance via immune checkpoint inhibitors. These microRNAs, which have been shown to be prevalent in tumor samples, CD14 + monocytes and plasma, were determined to be responsible for converting monocytes into MDSCs and were correlated with myeloid cell infiltration. The abovementioned MDSC-related microRNAs were, therefore, indicated as plausible blood markers for the prediction of immunotherapy outcomes [245].
In a recent study on glioma cells, it was shown that miR-10a and miR-21 found in glioma-derived exosomes (GDE) were responsible for initiating an MDSC-induced immunosuppressive microenvironment [246]. In their study, it was demonstrated that GDE effects on MDSCs were achieved through targeting of RAR-related orphan receptor alpha (RORA) and phosphatase and tensin homolog (PTEN) pathways. This study concluded that glioma cells are able to exert extensive differentiation and activation effects on MDSCs through secreted exosomes under hypoxic conditions.

miRNAs and Natural Killer (NK) Cells
Natural killer (NK) cells have been identified as leading effector lymphocytes of the innate immune system against the formation of tumors. Currently, it is widely accepted that the cytotoxic activity of NK cells is decreased in many forms of cancers. These observations have been postulated to be related to microRNA dysregulation.
In a recent study, it was found that miR-20a elevation in colorectal cancer (CRC) cells was responsible for evasion of immune surveillance by NK cells [247]. Although preliminary investigations on this miRNA indicated that overexpression and knockdown did not affect CRC cell growth in vitro, further cytotoxicity assays showed that miR-20a knockdown increased CRC cell sensitivity to NK cell activity. The direct target of miR-20a was identified to be NKG2D ligand major histocompatibility complex (MHC) class-I-related chain gene A (MICA) transcripts. This study, therefore, postulated that miR-20a targets MICA to regulate the sensitivity of CRC cells to NK cells.
In another recent study, it was demonstrated that miR-130a targets STAT3 to increase the cytotoxic activity of NK cells against non-small cell lung cancer (NSCLC) cells [248]. The findings indicated that miR-130a was markedly reduced and STAT3 was notably elevated in NK cells isolated from NSCLC patients. Further functional studies by overexpressing miR-130a reversed the capability of NK cells to increase cytotoxicity against A549 lung cancer cells. In a related study on lung cancer, miR-218-5p was able to suppress the cytotoxic activity of NK cells towards lung adenocarcinoma (LA) by targeting serine hydroxymethyl transferase 1 (SHMT1). Further experimentation by attenuating miR-218-5p resulted in IFNγ and TNF-α secretion in IL-2-activated NK cells [249]. These studies demonstrated that manipulation of microRNAs is a viable strategy for potentiating NK cell immunotherapy against multiple types of lung cancers.
It was additionally discovered that the microRNA cluster Mirc11 was able to disrupt inflammatory responses of NK cells but not their cytotoxic activity against B16-F10 melanoma. The loss of the Mirc11 cluster, which consists of miRNA-23a, miRNA-24a and miRNA-27a, appears to significantly reduce the expression of proinflammatory factors in in vitro experiments and also hindered interferon-γ-mediated clearance of melanoma in animal models by NK cells [250].
In a recent study on liver cancers, it was demonstrated that HCC metastasis in the lungs was driven by miR-561-5p/CX3CL1 signalling [251]. This study disclosed that three miRNAs, namely miR-137, miR-149-5p and miR-561-5p, were identified to be present in patients with pulmonary metastasis stemming from HCC. Bioinformatics analyses and chemokine expression profiling determined CX 3 CL1 to be the probable target of miR-561-5p. Moreover, it was found that high levels of this miRNA were responsible for attenuating the anticancer activity of CX 3 CG1 + NK cells via CX 3 CL1; therefore, these results demonstrated that downregulation of miR-561-5p in CX 3 CG1 + NK cells could potentially be a strategy for developing cellular anticancer treatment effectors.

miRNAs and T Cells
T cells are major constituents of the adaptive immune system and are capable of distinguishing altered cancer cells from normal cells; however, wide-ranging immunosuppressive mechanisms found in the tumor microenvironment to evade detection enable the continuing survival of tumors, and in some cases further deteriorate prognosis through transforming the functions of T cells.
In recent research, it was revealed that miR-24-3p hinders T cell activity by targeting FGF11 in nasopharyngeal carcinoma (NPC) [252]. In this study, enrichment of miR-24-3p was observed in exosomes of NPC cell line and patient sera samples. Knockdown experiments reversed the inhibition of T cell proliferation, Th1 and Th17 differentiation and induction of Tregs. It was also discovered that miR-24-3p directly targets FGF11 for its activity, while tumor FGF11 levels were positively correlated to CD4 + and CD8 + T cell counts in vivo, which were predicative of favorable patient disease-free survival.
It is widely accepted that progressing tumors derive mechanisms to hijack the PD-1/PD-L1 immune checkpoint via microRNAs to dysregulate T cell functions [237]. MiR-140 was significantly suppressed in Helicobacter pylori (Hp)-positive gastric cancers [253]. PD-L1 was identified to be the direct target of miR-140 in patient samples. Further experimentation to overexpress miR-140 demonstrated that gastric cancer proliferation could be suppressed through the regulation of PD-L1 levels. Moreover, in vivo research also showed that miR-140 repressed the growth of tumors in mice models of gastric cancer. The increase in cytotoxic CD8 + T cells and reduction of MDSC and Tregs in the immediate tumor microenvironment were determined to be the main factors contributing to the effects of miR-140 treatment. These data collectively indicate that miR-140 was able to target PD-L1 to exert an antigastric cancer response.
In recent research on breast cancer, it was discovered that miR-149-3p plays a vital role in resuscitating CD8 + T cell deletion by downregulating inhibitory receptors and enhancing cytokine secretion [254]. The researchers found that PD-1-overexpressing CD8 + T cells showed significantly lower levels of miR-149-3p, which was predicted to bind to the 3 UTRs of T cell inhibitor receptors PD-1, TIM-3, BTLA and Foxp1 mRNA transcripts. By using mimetics of miR-149-3p to treat CD8 + T cells, the authors managed to enhance their killing capacity on 4TI mouse breast tumor cells, which was largely attributed to reversal of T cell inhibitor receptor expression, reduction of apoptosis and secretion of effector cytokines, including IL-2, TNF-α and IFN-γ. Based on these findings, it was speculated that miR-149-3p could potentially be developed into an effective antitumor immunotherapeutic agent.
In studies on colon cancer, the indoleamine 2,3-dioxygenase 1 (IDO1) transcript was identified as the target for miR-448 to regulate the antitumor function of CD8 + T cells [255].
In vivo experiments indicated that overexpression of IDO1 promoted xenograft tumor growth in immune-competent mice but not in nude mice. Additional studies on the downregulation of IDO1 via ectopic expression of miR-448 mimetics markedly reduced IDO1 protein expression levels, which unequivocally led to the inhibition of apoptosis of CD8 + T cells. The findings in this study suggest that miR-448 is able to suppress IDO1 to enhance CD8 + T cell activity against colon cancers.
In silico approaches to identify the correlations of miRNAs to cancers have also become increasingly important due to the increasing wealth of bioinformatics data. One such study identified that miR-195 is potentially involved in inhibiting lung adenocarcinoma progression by enhancing CD4 + T cell activation [256]. Further analysis identified that CD4 + T cells were the subset of lymphocytes involved in infiltration of lung adenocarcinoma through activation of miR-195-targeted genes. The data from this study collectively indicated that miR-195 is able to act as an inhibitor of lung adenocarcinoma by enhancing CD4 + T cell activity via the CCDC88C/Wnt signalling pathway. A further summary of miRNAs involved in immune modulation is shown in Table 11.

The Role of microRNAs in Cancer Biology beyond the Hallmarks of Cancer
Most recently, it has come to light that the hallmarks of cancer themselves are not sufficient to comprehensively describe the full length and breadth of cancer biology. A plethora of modulations that enable cancer cells to be resistant to therapy and underlying mechanisms of disease relapse have been reported and described extensively, with sufficient evidence making it now apparent that these mechanisms extend beyond the definitions of the hallmarks of cancer.
Mechanisms that have been implicated in therapy resistance and disease relapse include formation of polyploid or multinucleated giant cancer cells, which dedifferentiate somatic cells and endow them with stem-cell-like properties through the giant cell cycle, which encompasses a dormancy phase prior to reactivation and stabilization as well as the phenomenon of anastasis, in which cancer cells are able to recover themselves and become more malignant upon removal of apoptotic stimuli, senescent-like cancer cells and antiproliferative drug-resistant cancer cells [257][258][259][260][261][262]. Compounding matters further is the intratumor heterogeneity that enables different populations of cancer cells to reside within the same tumor [259]. Nevertheless, one unifying factors across all of these mechanisms is epigenetic modulation [257][258][259].
MicroRNAs have been shown to be involved in the epigenetic machineries of various types of cancer cells, which have been reviewed and described extensively elsewhere [263,264]. Some examples of microRNAs involved in epigenetic modulation include miR-200a, miR-148a, miR-19a, miR-96, miR-25 and miR-29b-3p [263,264]; therefore, it could be worthwhile further elucidating the epigenetics mechanisms that are governed by microRNAs in cancer biology, as they could provide us clues to circumvent the various hurdles currently associated with therapy resistance and disease relapse in cancer.

Conclusions
The hallmarks of cancer are the most accurate models currently available to summarize cancer biology in a nutshell. Nevertheless, as this review has shown, the hallmarks of cancer are not sufficient anymore to describe cancer biology in its entirety and could benefit from a revisit. Based on the cumulative evidence gathered herein, it would not be too far-fetched to suggest epigenetic modulation as an additional hallmark of a cancer cell. Furthermore, this review has made it clearly evident that the footprints of microRNAs are all over the various hallmarks of cancer and beyond in cancer biology. The compelling evidence demonstrates that miRNA promotes or inhibits cancer progression in various cancer types by regulating genes that could serve as either tumor suppressors or as oncogenes; thus, the outcome of miRNA regulation in the target genes could either be positive or negative in terms of cancer progression. It has been shown previously that one miRNA can regulate many genes. On that note, from this review, we can see that there is an overlap of miRNA regulation, whereby one miRNA can be involved in regulating different cancer hallmarks, which occurs by targeting many genes (Table 12). Adding to the complexity of matters is the fact that microRNAs themselves are epigenetically regulated, and in turn regulate other genes via various epigenetic modulation, creating an epigenetic feedback loop [263,264], which should be taken into consideration when studying the function and roles of microRNAs in cancer biology.   Tumor microenvironment Hepatocellular carcinoma [210] Author Data Availability Statement: Not applicable.

Conflicts of Interest:
The authors declare no conflict of interest.