Emerging Perspectives on Leukemia Inhibitory Factor and its Receptor in Cancer

Tumorigenesis and metastasis have deep connections to inflammation and inflammatory cytokines, but the mechanisms underlying these relationships are poorly understood. Leukemia Inhibitory Factor (LIF) and its receptor (LIFR), part of the interleukin-6 (IL-6) cytokine family, make up one such ill-defined piece of the puzzle connecting inflammation to cancer. Although other members of the IL-6 family have been shown to be involved in the metastasis of multiple types of cancer, the role of LIF and LIFR has been challenging to determine. Described by others in the past as enigmatic and paradoxical, LIF and LIFR are expressed in a diverse array of cells in the body, and the narrative surrounding them in cancer-related processes has been vague, and at times even contradictory. Despite this, recent insights into their functional roles in cancer have highlighted interesting patterns that may allude to a broader understanding of LIF and LIFR within tumor growth and metastasis. This review will discuss in depth the signaling pathways activated by LIF and LIFR specifically in the context of cancer–the purpose being to summarize recent literature concerning the downstream effects of LIF/LIFR signaling in a variety of cancer-related circumstances in an effort to begin teasing out the intricate web of contradictions that have made this pair so challenging to define.


INTRODUCTION
The interleukin-6 (IL-6) family cytokine LIF was originally discovered as an inducer of differentiation and inhibitor of proliferation in a murine myeloid leukemia cell line, where it originally received its name (1). However, LIF has since been demonstrated to be expressed by a variety of different cell lines with diverse downstream effects. The most well-known function of LIF is its role in maintaining murine embryonic stem cells (mESC) in culture by maintaining their totipotency and enhancing their selfrenewal (2), an effect that is not seen in human ESCs. LIF has essential activities outside of ESC self-renewal and has been demonstrated to play an important role in mediating interactions between the embryo and the maternal environment. During development, LIF signaling is necessary for human blastocyst implantation (3) through mediating the invasiveness of trophoblastic cells (4). As such, LIF may represent a target for non-hormonal contraception (5), and has been suggested as a potential biomarker for the success of in vitro fertilization (6). Additionally, LIF expression is important in suppressing the maternal immune response during embryological implantation (3).
Discovered shortly after the ligand for which it is named, LIFRb is a subunit of both the LIFR and the ciliary neurotrophic factor receptor (CNTFR) (Figure 1). The LIFR is a heterodimer consisting of LIFRb and glycoprotein 130 (gp130), while CNTFR is a trimer of LIFRb and gp130 with an additional CNTF-a receptor subunit. LIF is only one of a whole host of cytokines known to bind to LIFRb (Figure 1). These include ligands that are part of the interleukin-6 (IL-6) cytokine family: namely, oncostatin M (OSM), ciliary neurotrophic factor (CNTF), cardiotrophin-1 (CT-1), and cardiotrophin-like cytokine (CLC) (7). Recently, interleukin-like EMT inducer (ILEI) was determined to be a ligand of LIFRb, though further studies will be necessary to determine the precise receptor complex that ILEI utilizes (8).
It is apparent that LIFRb plays a significant role in post early embryological development stages, as indicated by the rare congenital disease Stüve-Wiedemann Syndrome (SWS), which is caused by a truncated LIFRb subunit. SWS is characterized by skeletal deformities, cardiac and respiratory distress, temperature dysregulation, and mild cognitive impairment (9,10). To what degree these symptoms are due to impaired LIF signaling is challenging to determine, though, as LIFRb has other ligands as previously mentioned. To further illustrate this point, LIFRb knockouts in mice, while not embryonically lethal, result in premature death shortly after birth-likely due to significant neural, metabolic, bone, and placental defects (11). LIF knockouts on the other hand are not lethal, implying the potential for functional redundancy among ligands for LIFRb with regards to development. It is this very concept of functional redundancy, in addition to the fact the LIF and LIFRb exhibit such clear polyfunctionality, that make the pair so obscure within the context of our understanding of cancer. The following will discuss LIF-LIFR signaling from a general perspective, and then transition to a more precise conversation regarding these signaling pathways within cancer. In addition, the review hopes to also touch on how our perspectives of LIF-LIFR signaling have grown more nuanced-with the addition of signaling pathways such as the Hippo pathway, the possible overlap with other LIFR ligands, the mechanisms through which LIF and LIFR have been classified as either pro-tumor growth/metastasis, or tumor growth/ metastasis suppressive.

LIF-LIFR SIGNALING
The LIFR complex is a heterodimer consisting of gp130 and LIFRb. Intracellularly, the LIFRb/gp130 receptor complex famously signals through the JAK/STAT pathway and is constitutively associated with a janus associated kinase (JAK) family member-JAK1, JAK2, and TYK3 (12). The most demonstrably important is JAK1, as various knockout models for JAK1 exhibit significantly dampened responses to LIF as well as other IL-6 cytokines (13). Unlike other IL-6 family members, LIF has a high affinity for both gp130 and LIFRb, and it is hypothesized that an ordered binding process is unlikely (14). Once bound to either subunit, LIF induces receptor heterodimerization, leading to the activation of a JAK1. Once activated, JAK1 phosphorylates tyrosine residues on both LIFRb and gp130, which provide docking sites for various signal cascade components including signal transducer and activator of transcription 3 (STAT3) and the cytokine signaling inhibitor phosphatase SHP2. The activation of SHP2 by JAK1 is generally thought to be the mechanism through which the MAPK and PI3K pathways are activated, as S H P 2 a c t i v at i o n i s r e q u i r e d f o r t h e d o w n s t r e a m phosphorylation of ERK1/2 (15). Although LIFR-mediated activation of PI3K/AKT pathway is less understood than others discussed in this review, it is generally accepted that SHP2, and perhaps GAB1, bind to the p85 subunit of PI3K in ESCs (16). This ultimately leads to the activation of the downstream transcriptional regulator mTOR. Of these three discussed pathways thus far (JAK/STATs, MAPK, and PI3K/ AKT), JAK/STAT3 appears to be dominant, as STAT3 has 4 binding sites on both the LIFRb and gp130, whereas SHP2 has one. As such, study of LIF and its receptor have been primarily focused on the JAK/STAT pathway. More information on the biochemical nature of this process can be found in an excellent review published by Nicola and Babon ( 7 ). Once phosphorylated, STAT3 forms a homodimer with another STAT3, and enters the nucleus where it acts as a transcription factor for various genes associated with increased proliferation and enhancing stem cell self-renewal, most notably Myc and Nanog ( Figure 2) (17).

Downstream LIFR Signaling and Crosstalk
Activation of STAT3 is followed by the rapid upregulation of the inhibitory protein, suppressor of cytokine signaling 3 (SOCS3). As the name implies, SOCS3 acts to inhibit the JAK/STAT pathway by binding to and leading to the ubiquitination of JAK1 and gp130, as well as competing with SHP2 for binding sites on the LIFRb-gp130 heterodimer, inhibiting MAPK signaling (18,19). Regulation of LIFR signaling, though, does not seem to be solely dependent on transcriptional activity.
Research into early developing mouse embryos suggest that the PI3K and JAK/STAT pathways exist in a tentative balance with MAPK, with the prior necessary for ESC self-renewal and survival, and the latter with differentiation (7,20), (Figure 2). LIF-induced pluripotency is highly dependent on the activation of STAT3 (21,22) and cells expressing a non-functional STAT3 and grown in the presence of LIF are induced to differentiate (22).
In contrast to mESCs where the JAK/STAT is acting as a suppressor of differentiation, in developing murine mammary tissue JAK/STAT acts as a pro-apoptotic signal, and MAPK as a pro-survival signal. During post lactational regression, elevated levels of LIF were shown to induce cellular apoptosis in a STAT3dependent manner via lysosomal mediated cell death (23)(24)(25), and LIF-induced STAT3 also leads to an upregulation of OSM and the OSM receptor (26). During ductal elongation, LIF was demonstrated to signal primarily through ERK1/2 as a survival signal. This is relevant in that it shows outcomes of LIF signaling are markedly different across tissue types-a concept that will be developed further as we begin to discuss LIF in cancer.

LIF AND LIFRb EXPRESSION IN CANCER
As summarized in Table 1, LIF and LIFRb expression are linked to a variety of human cancers, many of which are associated with both negative and positive prognostic outcomes. As a whole, it appears that LIF activation of both the JAK/STAT and PI3K/ AKT pathways are associated with the promotion of tumor growth and metastasis. On the other hand, LIFRb expression seems to be connected to the tumor suppressor pathway Hippo, and thereby is correlated with decreased tumor growth and metastasis.
Despite conflicting evidence as to the precise role of LIF across cancer types, some interesting patterns have emerged, including the role of LIF in maintaining cancer stem cells (CSCs) in glioma, chordoma, melanoma, osteosarcoma, and glioblastoma ( Table 1). Generally speaking, cancer stem cell maintenance by LIF and LIFR seem to follow a similar trend as that seen in mESCs: LIF signaling leads to the activation of STAT3, which increases the stem cell like properties in solid tumors through transcriptional regulators. LIF is not always the sole cause, though, as in ovarian cancer IL-6 and LIF work together to stimulate STAT3 phosphorylation and stemness, while the loss of either LIF or IL-6 highly abrogates this process (50). To add to this point, glioma initiating cells (which exhibit stem cell like qualities in glioma) are stimulated to produce LIF following signaling via TGF-b, leading to an increase in STAT3 phosphorylation (39).
Another pattern seen is the propensity for LIF signaling to result in migration and metastasis, something seen in its close relative OSM and IL-6 in multiple cancers, most notably, breast cancer (74,75). In both instances, metastasis is highly dependent on the activity of STAT3, though other pathways such as MAPK, PI3K/AKT, and the Hippo pathway have also been linked to LIF/ LIFRb dependent effects on the oncogenic process. Aberrant JAK/STAT signaling has been linked to a variety of pathological states, including but not limited to various immune disorders such as rheumatoid arthritis, as well as cancers such as prostate and breast cancer (76,77). Specifically, STAT3 overactivity has been associated with the invasion and proliferation of a significant variety of cancer cells both in vitro and in vivo, and as such has been recognized to be a strong oncogene.
But LIF does not seem to be solely dependent on STAT3 activation in order to be pro-oncogenic, and some have even pointed to tumor cell dormancy induction via a LIF : LIFR : STAT3 axis in breast cancer to bone marrow metastasis (62). For example, PI3K overactivation is commonly associated with the increased survival and proliferation of cancer cells. Activation of this pathway via LIF stimulation is correlated with apoptotic resistance in cholangiocarcinoma, but not with increased growth or metastasis (78). Furthermore, in the breast cancer cell lines MDA-MB-231 and T47D, treatment with, as well as transient overexpression of LIF led to increased mTOR activity and the phosphorylation of 4EBP1 and p7056K, which are downstream targets of this pathway and play roles in apoptotic resistance as well as protein synthesis (31). Overall, in these cell lines investigators found that LIF activity was correlated with increased growth in vivo and metastatic qualities in vitro (31). LIFR influences the PI3K/AKT pathway in a variety of cancers including prostate (79,80), gastric (66), hepatocellular (70), nasopharyngeal (81) and rhabdomyosarcomas (57). In some instances, LIFRb expression has been associated with decreased PI3K/AKT activity. Interestingly, while LIF has been demonstrated to exert effects on lipid metabolism in both the brain and in adipocytes via the PI3K/AKT pathway, little research has been done to evaluate how LIF signaling influences cancer metabolism in both glioblastoma and glioma, or in other cancer related pathological states.

LIF-Induced Immunosuppression
The relationship between tumor cells and the immune system is a highly complex process, and extensive evidence suggests that many tumors actively suppress the host immune response as a way to prevent immune-mediated tumor destruction. LIFinduced immunosuppression has been recently demonstrated in prostate cancer cells (55) and glioblastoma (82).
LIF as an immunomodulator/suppressor in cancer represents an important potential target for treatment. In a study of glioblastoma, the presence of high levels of LIF in the tumor microenvironment (TME) was associated with an increased number of tumor-associated macrophages (TAMs). Higher levels of TAMs in the TME was shown to prevent the invasion of CD8+ T cells via the repression of CXCL9 secretion, thus hindering an effective immune response to cancerous tissue (82). To examine this phenomenon in vivo, glioblastoma patient xenograft models in immunocompromised mice were treated with a neutralizing monoclonal antibody (nAb) for LIF. Treatments with LIF nAb greatly reduced TAMs, as well as led to an increased accumulation of CD8+ T cells. Overall, the study found that high levels of LIF were associated with decreased response to anti-PD1 therapies, and that LIF nAb when used in conjunction with anti-PD1 therapy could be a potential therapeutic option for patients with solid tumors exhibiting high LIF expression (82). In 2019, a humanized LIF nAb called MSC-1 entered phase 1A clinical trials and has been recommended to enter phase 2 dose trials for patients with relapsing or non-responsive solid-state tumors (83). Clearly, LIF and the LIFRb have relevant connections to cancer growth and metastasis that warrant additional research and definition.

Tumor and Metastasis Suppression
In contrast to LIF typically being associated with the increased invasion and metastasis of cancer, LIFRb expression has been shown to be correlated with the opposite. While LIFRb is typically downregulated in a variety of cancers, it's co-receptor gp130 is ubiquitously expressed in the human body, even detectable in serum, and it's expression pattern across types of cancer is highly variable (84)(85)(86). In a variety of cancers, LIFRb expression has been associated with higher patient survivability, and increased metastasis-free survival ( Table 1), and that depletion of this receptor is somehow linked to decreased cellular adhesion and more aggressive cancer phenotypes through the inactivation of the Hippo pathway. We will discuss the Hippo pathway in further depth later on in this review.
Although LIFRb signaling and its downstream targets have been well studied, how this receptor is regulated in cancer is poorly understood. Some have postulated that decreased LIFRb expression occurs via an epigenetic mechanism such as LIFRb promoter methylation, which has been observed in breast (87), clear renal cell carcinoma (64), hepatocellular carcinoma (68,69), and colorectal cancer (88). Unfortunately, there is little research overall in regard to the mechanisms by which LIFRb is regulated in cancer. Expression of LIFRb is downregulated by miR-125a (58), miR-125b (89), and miR-9 (59) in a variety of human cancer cell lines. There is also some evidence pointing to hypoxia downregulating LIFRb, as hypoxic conditions decreased LIFRb expression in breast cancer cells and multiple hypoxic responsive elements have been identified in the LIFRb promotor (62).That same group also identified histone deacetylase (HDAC) as a potential mechanism for LIFRb downregulation, which is supported by evidence indicating that LIFRb is upregulated when breast cancer cells were treated with HDAC inhibitors (28) (62). Notably, LIFRb was upregulated in gastric cancer cells in vitro following transfection with the long noncoding RNA LNC-LOWEG, and this was correlated with decreased capacity for migration (65). In myeloid and placental cell lines, the LIFRb gene was shown to be regulated by the transcription factor RUNX1, which has been shown to be important in leukemia, as well as breast cancer (90,91). While LIFRb expression seems to be negatively correlated with breast cancer growth and metastasis (59,60,87), high expression of LIF is positively correlated (27,(30)(31)(32). The triple negative breast cancer cell line MDA-MB-231 was shown to highly express LIF, and treatment with LIF neutralizing antibodies impeded proliferation (27). On the other hand, overexpression of the LIFRb in this same cell line resulted in decreased metastasis in vivo, with no effect on proliferation (59). However, it should be noted that others have found that MDA-MB-231 cells are unresponsive to LIF, and that this cell line had undetectable levels of LIFRb expression (59,62). This is an excellent example of the challenge in discerning the role of LIF and LIFRb in cancer, as even in a single cell line their effects are debated. Conversely, in pancreatic cancer high LIF expression is correlated with lower metastasis free survival (54), whereas induction of LIFRb expression in pancreatic cancer cell lines in vitro and in vivo decreased proliferation and migration, increased E-cadherin expression, and was associated with more favorable patient outcomes (73).
This begs the question: Why is decreased LIFRb expression associated with worse outcomes, especially metastasis, when the majority of its downstream signaling pathways are classically described as oncogenic in nature? High expression of LIF could potentially lead to decreased expression of LIFRb via internalization and degradation, as LIF binds to LIFRb with a high affinity and an over 24-hour half-life until ligand/receptor disassociation, as demonstrated in kinetic studies (92), but this is purely speculation. Although the precise reasons may differ among cancers, few have made significant forays into the underlying molecular mechanisms by which LIFRb plays a role as a metastasis and tumorigenic suppressor. The most relevant underlying molecular mechanisms demonstrated have defined connections to the tumor suppressor Hippo pathway, and links between LIFRb and the Hippo pathway have been demonstrated in breast (59), clear renal cell carcinoma (64), and gastric cancer (93). Although LIF-LIFR signaling activates a variety of pathways associated with cancer progression such as JAK/STAT and MAPK, more potent and significant activators of these pathways already exist and are potential targets for treatment. This is not to say that LIF and LIFRb are not relevant; however, but rather that the focus of the conversation surrounding them in cancer should be shifted towards how LIF and LIFR can be understood through the lens of tumor suppression and promotion via the less understood Hippo pathway. The potential therapeutic and physiological significance of the relationship between LIF/LIFRb and the Hippo pathway thus necessitates speaking of their interaction in more depth.

LIF-LIFR ACTIVATION OF THE HIPPO PATHWAY The Hippo Pathway
The Hippo pathway, first discovered in Drosophila for its role in organ development, is a signaling cascade of particular interest to researchers due to its frequent dysregulation in human cancers (94). The primary effectors of this pathway are the transcription cofactors yes-associated protein (YAP) and transcriptional coactivator with PDZ binding motif (TAZ) (Figure 3). YAP and TAZ bind to a diverse array of transcription factors, the most important of which are members of the TEA Domain (TEAD) family. The Hippo pathway is activated by a variety of upstream cellular inputs, including various growth factors, cellular adhesion, and metabolic status resulting in the activation of salvador (SAV1). The core Hippo kinase cascade is as follows: SAV1 interacts with and activates MST1/2 (the mammalian homologue of the Hippo protein in drosophila) via an autophosphorylation event in the activation domain of MST1/2. Once activated, MST1/2 phosphorylates LATS1/2, leading to the recruitment of MOB1 to LATS1/2, whereupon MOB1 is also phosphorylated by MST1/2. The LATS/MOB1 complex is what engages and phosphorylates YAP (Ser127) and TAZ (Ser89). The phosphorylation of these serine residues generates binding sites for cytoplasmic 14-3-3 proteins, which sequester YAP and TAZ to the cytoplasm leading to their degradation. Dysregulation by increased expression or activation of YAP and TAZ have been found to be associated with malignant transformation and oncogenesis in numerous cancers, and thus their regulation (both at the transcriptional and protein levels) has become an area of importance in oncology, especially breast cancer (95)(96)(97)(98). Thus, the Hippo pathway and its core kinases are tumor suppressors, while YAP and TAZ are oncogenes. The relationship between LIFRb and the Hippo pathway is what initially defined LIFRb as a metastasis suppressor in breast cancer (59).

Mechanism of LIFR-Mediated Hippo Activation
The activation of the Hippo pathway via the LIFR was originally discovered by Chen et al. (59) (59). Interestingly, membrane localization of scribble was thought to require the expression of E-Cadherin, though the results of this study indicated otherwise.
Scribble is important in the maintenance of cellular polarity and has been demonstrated to have effects on the MAPK signaling pathway, as well as the Hippo pathway (95,99). In the context of the Hippo pathway, scribble acts as a scaffolding protein for MST1/2 and LATS1/2 and TAZ (95). Upon localization to the cell membrane, this complex is active, and can begin the phosphorylation cascade that ultimately results in the cytoplasmic sequestration of YAP and/or TAZ via binding by 14-3-3 proteins (Figure 3).
In a recent study on clear renal cell carcinoma (CRCC), LIFRb was found to be consistently downregulated in more aggressive cancers, likely due to promoter methylation and copy number variation (64). Silencing LIFRb expression in CRCC cell lines led to an increase in the nuclear localization of YAP and enhanced migration and invasion. Most importantly, the silencing of YAP partially reversed this phenotype, indicating that loss of LIFRb-promoted transformation is likely due to decreased Hippo activity, and therefore, increased transcriptional activity of YAP (64). The study in CRCC did not determine the mechanism by which LIFR activated the Hippo pathway, and in the years since Chen et al. originally established the Hippo-LIFR connection, the mechanism underlying the localization of scribble to the cell membrane via LIFR activation has not been determined. This is an important gap in the literature. Loss of cellular polarity is a hallmark of EMT, and if polarity-associated proteins such as Scribble are required for LIFR to activate the Hippo pathway then this gap could partially explain how LIF-LIFR signaling can have such a stark difference in downstream effect across disparate types of tissues and especially within cancer. But this is not the only association between LIF/LIFRb and other effectors in the Hippo pathway. Interestingly, LIF activity has also been associated with the activation of YAP via the Src family kinase YES.

Background in ESCs
The connection with Src family kinases is a little researched facet of LIF/LIFR signaling. One such member of this family, YES, is a tyrosine kinase that activates YAP. The activation of YES by the LIFR has been shown to be relevant to LIF-induced stem cell maintenance, and unlike other pathways discussed thus far, seems to have little to no crosstalk with other LIF signaling pathways such as MAPK and JAK/STAT, at least in the context of mESCs (100). Research has shown that LIF-induced ESC selfrenewal in mice is highly dependent upon LIF-mediated YAP-TEAD4 activation via YES, and these researchers determined that YES activity downstream of LIF had more profound effects on self-renewal than LIF-STAT3 (101). Although the precise mechanism is understudied, the proposed model is as follows. YES binds to the gp130 receptor subunit of the LIFR via an SH2 domain and is activated by JAK1. The active YES then then goes on to phosphorylate and activate YAP. YAP binds to and stimulates transcription with TEAD2, leading to the expression of the pluripotency factor OCT3/4 (101).

Increased YAP Activity in Cancer
Recent studies on LIF in cancer have further demonstrated LIF-LIFR mediated YES activation. In a human in vitro pancreatic cancer model, LIF expression was highly correlated with increased YAP activity (54). In this instance, researchers were trying to understand the relationship between STAT3 and human KRAS driven pancreatic adenocarcinoma (PDAC), and this team hypothesized that LIF functioned in an autocrine manner, stimulating the growth of pancreatic cancer cells as well as their formation of 3D spheres in culture. They found that increased KRAS activity resulted in an increased expression of LIF. This effect was lost when downstream signaling proteins in the MAPK pathway were inhibited, suggesting that LIF upregulation in PDAC is dependent on MAPK activation. In general, LIF was found to be overexpressed in human pancreatic carcinomas relative to normal tissue, and that in a pan-cancer analysis LIF was significantly more upregulated in cancers with a mutation in KRAS. The silencing of LIF, though either genetic means or neutralizing antibodies resulted in an increased phosphorylation of YAP at ser127, and the activation of upstream Hippo pathway kinases (54). Furthermore, LIF nAbs used with gemcitabine significantly reduced the growth of patient-derived xenograft (PDX) tumors in vivo. These results are in stark contrast to earlier findings in breast cancer demonstrating that LIF-LIFR signaling activated the Hippo pathway, thereby inhibiting YAP and TAZ. Additionally, a study of gastric cancer found that higher levels of LIF and LIFR were associated with increased proliferation, invasion, and metastasis (93). Interestingly they determined that LIF-LIFR signaling actually inhibited Scribble localization to cell membranes, thereby preventing the inactivation of YAP through the Hippo pathway. When YAP was inhibited via shRNA, the effect of LIF-LIFR signaling on cancer growth and migration was lost (93). In a dose dependent manner, LIF decreased the phosphorylation of MST and LATS, implying that LIF-LIFR signaling is somehow inhibiting the Hippo pathway and allowing YAP to remain active, rather than directly activating YES to activate YAP (93).
To further the complexity, in a model of nasopharyngeal carcinoma (NPC), cells constitutively expressing a cytoplasmic variant of LIF had a markedly lower expression of YAP as well as phosphorylated YAP at ser127, suggesting that although YAP expression was decreased, a higher proportion of YAP remained active relative to controls (46). Depletion of LIFRb resulted in an increased expression of YAP, and a higher level of pYAP was also demonstrated-though, this could simply be due to the fact that more YAP was physically present in the cell. Clearly, though, this is showing another link between LIF/LIFRb expression and YAP. There are further links between LIF/LIFRb in YAP expression, as LIFRb expression has been negatively correlated with YAP expression in clear renal cell carcinoma (64). In nasopharyngeal carcinoma (NPC) high levels of LIF are associated with higher degrees and radio resistance, tumor progression, and decreased DNA repair (81). Overall, the findings in both PDAC and NPC suggest that the relationship between LIF-LIFR signaling, YES-YAP activity and the Hippo pathway are significantly more nuanced than originally described in breast cancer models. Combined with the findings of LIF signaling leading to upstream inhibition of the Hippo pathway in gastric cancer, it is clear that LIF-LIFR-Hippo pathway interactions are highly tumor dependent. This should be unsurprising at this point, considering how this has been a recurring motif for LIF not only in cancer but physiologically as well. Looking across all cancers that LIF has been associated with, a closer examination of the Hippo pathway's involvement in that cancer, if one has not been found, should be necessitated.

ILEI: A NOVEL LIGAND FOR LIFRb
Interleukin-like EMT inducer (ILEI) is a cytokine-like protein of the FAM3C family that is speculated to have a four-helical bundle structure similar to LIF and has been implicated in a number of pathophysiological contexts, including Alzheimer's and cancer metastasis (102). A recent study by Howe and colleagues (8) identified ILEI as a ligand for LIFRb based upon a yeast 2-hybrid screen that was confirmed with crosslinking and immunoprecipitation experiments.
A series of experiments by Howe and colleagues sought to elucidate potential mechanisms by which TGF-b induced metastasis and CSC renewal in breast cancer, in which they found chronic stimulation of normal murine mammary gland (NMuMG) cells with TGF-b led to an increase in both LIFRb and ILEI protein expression. Furthermore, they demonstrated that that ILEI activated STAT3 in a dose-dependent manner that was dependent on LIFRb expression.
Immunocompromised mice injected with NMuMG cells expressing LIFRb and ILEI had significantly higher host tumor burden and metastasis relative to controls, and this effect was partially lost in mice infected with cells with LIFRb and ILEI knocked down. Intriguingly, in mice injected with NmuMG cells originally expressing LIFRb, expression of LIFRb was lost in sites of tumor outgrowth, as well as metastasis. This could be alluding to the role LIFRb seems to play in tumor initiation, and CSC renewal, while simultaneously acting as a metastasis suppressor.
The induction of ILEI and LIFRb expression by TGF-b is particularly interesting result, as TGF-b has been associated with the increased transcription of LIF in a number of cancers including in melanoma (72), thymic epithelium (103), glioblastoma (39), and in tumor associated stromal fibroblasts (104). Furthermore, there are some lines of evidence suggesting that TGF-b works in conjunction with the oncogenic transcriptional regulator c-Myc and OSM to cause the malignant transformation of human mammary epithelial cells (105)(106)(107). The relationship between OSM, TGF-b, and STAT3 implies there may be a crucial connection between the downstream effects of IL-6 cytokines and TGF-b.

LIFR-HIPPO Activation via Alternative Ligands
As to date, published studies have only pursued a link between LIF/LIFR-mediated activation of the Hippo pathway, and one may be inclined to wonder if other LIFRb ligands (especially OSM, considering both LIF and OSM can utilize the LIFR complex) also have the capacity to activate the Hippo pathway. Evidence of a role for other ligands is supported by the fact that transgenic mice who are LIF -/-(thought the LIFR is intact) exhibit only mild physiological deficits, whereas LIFRb -/die shortly after birth. Furthermore, in trophoblastic cell lines, it was shown that there is some degree of functional overlap between OSM and LIF in downstream effect (108). While hereditary LIFRb mutations result in the rare developmental disease Stüve-Wiedemann syndrome, women who have a deficiency in LIF expression face the problem of infertility with little other apparent physiological differences. Therefore, if there is in fact LIFR activation of the Hippo pathway across multiple cell lines, it is highly likely that other ligands have the capacity to result in pathway activation, especially considering the Hippo pathway's significant importance during development.
There is at least some tangential evidence of a relationship between other IL-6 cytokines and the Hippo pathway, especially YAP. In a murine heart failure model, YAP-TEAD activity was demonstrated to result in the upregulation of OSM and the OSMR and was directly associated with the dedifferentiation of cardiomyocytes. Interestingly, there was also a link between OSM activity, and a further upregulation of YAP, indicating there may be a potential positive feedback loop between OSM and YAP (109). In a murine model of breast cancer to bone metastasis, OSM was demonstrated to cause the upregulation and secretion of amphiregulin (AREG), a growth factor that lead to the differentiation of osteoclasts (75). Although the authors of this study did not elucidate the mechanism of AREG upregulation, separate studies have shown that YAP-TEAD activation in a human breast cancer model directly lead to an AREG increase (110), and similarly TAZ-TEAD induced migration and invasion of BC cells is partially abrogated when AREG is knocked down (111). Although this potential mechanism is purely speculative, this certainly begs the question as to whether or not OSM is modulating AREG expression through YAP, as there is already some evidence indicating that LIFRb/gp130 complex has the capacity to activate YAP through the protein YES.

Alternative LIF and LIFRb Transcripts
On a final note, very few studies (both in cancer and in other fields) make a clear distinction between the intracellular and secreted forms of LIF. There have been three transcripts of the LIF gene identified in both mouse and human cells: LIF-T, LIF-M, and LIF-D (112,113); to this point we have been almost exclusively discussing the secreted form LIF-D. Regulation of these transcripts is centered around the alternative transcription of the first exon, which contains the secretory signal sequence. While LIF-M can exist in the cytoplasm or can be secreted, LIF-T completely lacks the first exon containing the secretory sequence and is localized to the cytoplasm. Early research showed that alternative LIF transcripts had both a tissue-dependent expression profile, as well as unique functions, with the intracellular transcripts LIF-T and LIF-M demonstrated to initiate proapoptotic signaling independent of the LIFR (113,114). There has been some recent data on these alternative transcripts, though, including a recent study of the African elephant which identified a LIF-M "like" protein participating in p53-mediated apoptosis. The African Elephant genome contains multiple copies of this LIF-M-like gene and was postulated by the authors to be a partial example of a solution to Peto's paradox (115). Interestingly, high expression of an intracellular LIF mutant was associated with more invasive and aggressive tumors in nasopharyngeal carcinoma (46).
LIFRb has an alternative structure as wellthere is both a membrane bound and secreted form (116). Generally, it is hypothesized that soluble LIFRb is that it is meant to bind up latent LIF in the extracellular matrix. As to what regulates this alternative transcript is unknown, though one could speculate that this is a response to a high LIF environment.

CONCLUSION
The IL-6 family cytokine LIF and its receptor subunit LIFRb have come to represent a challenge to understanding the role of inflammatory cytokines in cancer. Despite significant advances in our knowledge of how inflammation drives cancer progression and metastasis, LIF and LIFRb provide particularly poignant demonstrations of how much there is to learn about the processes involved. There has been a significant focus throughout the years on STAT3 being the causal driver of LIF mediated effects in cancer, and not without cause-our primary understanding of LIF is derived through its effects on mESCs via STAT3. Other cytokines, specifically IL-6 and OSM, clearly have more profound effects in cancer through STAT3-this has left LIF in the proverbial wayside, as more potent activators of STAT3 have been targeted for study. Even the case for STAT3 being a driver of metastasis and tumor growth in breast cancer has been challenged, as there have been studies that have shown both LIF and OSM suppressing tumor growth and metastasis via STAT3 in breast cancer cell lines (62,117,118). In recent years, though, the apparent connection of LIF and specifically LIFRb to the Hippo pathway have opened up a new avenue for our broadening understanding of how this cytokine functions. This has, in many ways, left us with more questions than answers: what could explain the data demonstrating that LIF activates YAP via the YES/gp130 pathway, while other studies show that LIFRb inhibits YAP through the Scribble/Hippo pathway? Furthermore, is it possible for other ligands in the IL-6 family to activate these downstream signaling pathways as well? Indeed, all IL-6 family cytokines can bind to gp130, and many can bind to LIFRb. These are just a small sample of many unanswered questions when it comes to LIF and LIFRb in cancer, many of which are enticing avenues of research. With ILEI being a new ligand for LIFRb and considering the development of a nAb against LIF in solid tumors in a clinical trialthere is a significant need in the field of immuno-oncology to more readily define the relationship to the Hippo pathway. Hopefully, this review will act as an aid to any researcher looking to further develop our emerging perspectives of LIF and its receptor in cancer.