LGR4 (Leucine-rich repeat (LRR)-containing G-protein-coupled receptor 4) or GPR48 (G-protein- coupled receptor 48), together with LGR5 and 6, is a member of the receptor GPCR superfamily, LGR subfamily B (II), highly homologous to glycoprotein hormone receptors, e.g., LHR (Luteinizing Hormone Receptor), TSHR (Thyroid Stimulating Hormone Receptor) (1–3). LGR4, 5, and 6, are critical regulators of embryonic development, and also have been shown to contribute to several cancers (4–6). They are responsible for adult stem cell maintenance in vivo and stem cell survival ex vivo (7, 8), via direct interaction with ZNRF3 (Zinc And Ring Finger 3), LRP5/6 (LDL Receptor Related Protein 5) and Frizzled receptors, the components of the Wnt signaling pathway (9).

Structurally, all three LGRs (4–6) have highly conserved seven-transmembrane (7TM), rhodopsin-like regions and either 18 (LGR4 and 5) or 13 (LGR6) horseshoe-like leucine-rich repeats in the N-terminal (NT) extracellular domain responsible for ligand-binding, flanked with cysteine-rich sequences known as LRRNT and LRRCT (10). The intracellular region of LGRs ends with the C-terminal (CT) sequence, and is responsible for ligand-mediated signal transduction ( Figure 1 ).

All three members of the human GPCR subfamily B share some genetic similarities. LGR4 and 5 have 46% homology, whereas LGR4 and 6 share 44% sequence similarity. LGR5 and 6 are molecularly closer to one another and share 54% identity (13). All three receptors are also evolutionarily conserved across vertebrate species, e.g., human LGR4/LGR5/LGR6 share 90%, 82%, and 84% similarity with their mouse orthologs: Lgr4/Lgr5/Lgr6, respectively (13). Two paralogs (lgr4 and lgr6) are also found in Teleosts (13), the Actinopterygii intraclass representing 96% of all the existing fish species. Besides vertebrates, the orthologs of Lgrs are also found in invertebrates, e.g., Dlgr2 in a fruit fly, Drosophila melanogaster, is involved in exoskeleton morphogenesis and hardening (14). As indicated by the NCBI gene database, in humans, LGR4 is located on chromosome 11, in the domestic mouse (Mus musculus) on chromosome 2 and in fruit flies on the sex chromosome X.

UCSC Genome browser data (published in 2019) indicates LGR4 is commonly expressed in different human tissues, e.g. kidney, heart, gastroesophagus, pancreas, achieving the highest expression level in the ovary. A comprehensive atlas of LGR4 protein expression in different mouse and human tissues was generated by Yi et al. (15). Using custom-made antibodies against LGR4 and employing immunostaining technique, these authors showed that LGR4 is present in the epidermis and hair follicle of the skin, kidney, pancreatic β-cells, and epithelial cells of both male and female reproductive organs. Besides assessing tissues from healthy subjects, Yi et al. also examined colon tumor samples extracted from colon cancer-suffering patients, confirming high expression of LGR4 in the biopsies of this type of cancer. The role of LGR4 expression in development and progression of different types of cancer has also been recently reviewed by Ordaz-Ramos et al. (4).

Traditionally, members of the LGR subfamily are described as Wnt (Wingless-related integration site) signaling facilitators, since they enable signal transduction by the WNT ligands. In this context, the most critical is interaction of the N-terminus of LGRs (the extracellular domain) with R-spondins (Roof plate specific spondins, RSPOs), first defined as LGR4 and 5 ligands in 2011 (16, 17). The RSPO family comprises of four members: RSPO1 to 4, sharing between 40-60% homology with one another (18, 19). Structurally, RSPOs interact with LGRs via their Furin-like domains. All four RSPOs can induce canonical Wnt signaling via all three members of the GPCR subfamily B (19). RSPO3 is capable of inducing the non-canonical planar cell polarity Wnt pathway during gastrulation, head cartilage formation, and craniofacial development (20). Notably, RSPOs typically do not possess signaling activity on their own and require WNT ligands to execute their functions (21).

Traditionally, the WNT-mediated signal potentiation by LGRs is possible due to RSPOs interaction with E3 ubiquitin homologous ligases RNF43 (Ring finger 43) or ZNRF3, which inhibit activation of Frizzled- the WNT protein receptor (22). Interaction of RSPOs with the extracellular domains of RNF43 and ZNRF3 leads to their clearance from the cell membrane and thus enables Wnt signaling to proceed (22, 23). Interestingly a recent paper by Park et al. (24), indicates LGR4 and LGR5 have distinct roles in Wnt signaling potentiation. LGR4 does not require RSPOs to interact with ZNRF3 and RNF43 ligases and LGR5 does not interact with either of the ligases and is thus a weaker activator of this pathway than LGR4. Besides RSPOs, Wnt signaling potentiation via LGR4 can also be mediated by NORRIN (Norrie Disease Protein), a ligand structurally distinct from RSPOs which does not interact with either LGR5 or 6 (25).

Besides direct effects of LGR4 on the Wnt pathway, many reports indicate the involvement of this receptor in the crosstalk between Wnt and other pathways, e.g., NFκB (Nuclear Factor-kappa-light-chain-enhancer of activated B cells). Using intestinal cells, Lai et al. (10) recently observed that the intracellular domain of LGR4 and 5 can also trigger N-terminus-(ligand-binding)-independent NFκB signaling on their own. Li et al. (37), showed that LGR4 plays a protective role against liver injury. These authors demonstrated that functional LGR4 is expressed in mature hepatocytes and that RSPO1 protects hepatocytes from Tumor Necrosis Factor-α-induced cell death. When Lgr4 is specifically deleted in hepatocytes, the liver becomes more susceptible to acute injury (see section 7). The authors concluded that LGR4 protects hepatocytes from injury by inhibiting NFκB signaling. As shown by Han et al. (38), Wnt signaling potentiated by RSPO2 and LGR4 also interacts with Transforming Growth Factor beta (TGFβ) signaling via Follistatin (FS) (TGFβ antagonist) to drive early myogenesis. These authors showed that LGR4 is critical for a proper myogenic differentiation mediated by RSPO2. Using a mouse model, in vitro and in vivo, Han et al. showed that there exists a positive feedback loop between RSPO2, LGR4 and FST in the context of myogenesis, where FST enables Wnt signaling via RSPO2 and LGR4 at the same time being a primary target gene of their interaction. Recently, Wang et al. (39) revealed an alternative function of LGR4-RSPO binding related to early mammalian hematopoiesis regulation, both in vitro and in vivo, which is unrelated to Wnt. These authors have shown that LGR4 targets TGFβ signaling to modulate human pluripotent stem cell (hPSC) hematopoietic differentiation. Deletion of Lgr4, but not Lgr5, is detrimental for mesoderm development and therefore disables hematopoietic differentiation both in vitro and in vivo. Wang et al. demonstrated that RSPO1-3, but not RSPO4 interact with LGR4 to promote hematopoietic differentiation. Interestingly, they also showed some differences in receptor binding capability, with RSPO2 acting solely via ZNRF3 independently of LGR4, as opposed to RSPO1 and 3. Capability to manipulate LGR4 and RSPOs in the context of hematopoiesis may facilitate the large-scale generation of functional Hematopoietic Stem Cells (HSCs) for potential clinical applications. The study by Luo et al. (11) performed on osteoclasts showed that at the intracellular level, LGR4 can activate Gαq pathway to inhibit GSK3β (Glycogen Synthase Kinase 3 Beta) phosphorylation independently of Wnt signaling, thus preventing NFATC (Nuclear Factor of Activated T-Cells)-mediated osteoclastogenesis (see section 5). In invertebrates (e.g., D. melanogaster), DLGR2, the equivalent of LGR4, interacts with bursicon proteins: BURS/PBURS, activating cAMP pathway (14, 40). These reports altogether suggest that LGR4 participates in multiple signaling pathways, dependent on tissue and cell type.

In 2016 Luo et al. (11) described LGR4 as a novel receptor for RANKL in the context of osteoclastogenesis. RANK (Receptor-Activator of NFκβ), the known classic receptor for RANKL, induces osteoclast activation upon RANKL interaction. LGR4, on the other hand, inhibits osteoclastogenesis upon RANKL binding, thus tightly regulating bone resorption, which when excessive leads to osteoporosis. In the study by Luo et al., a conditional knockout of Lgr4 (Lgr4Fl/Fl-LysMCre mice) specifically targeting osteoclast precursors, caused increased bone loss and increased activity of osteoclasts as measured by their specific marker TRAP (Tartrate-Resistant Acid Phosphatase). At the molecular level, Luo et al. demonstrated that upon binding RANKL, LGR4 activates Gαq-mediated signaling to stabilize GSK3β and prevent its phosphorylation at Ser9 (i.e. inactivation). Stabilized (unphosphorylated) GSK3β, a known inhibitor of osteoclast differentiation (41), leads to NFATC phosphorylation and its arrest in the cytoplasm, resulting in the deactivation of the osteoclast differentiation program ( Figure 2 ). By using the ligand binding extracellular domain of LGR4 (LGR4-ECD, aa 28-528), truncated in different regions, Luo et al. deciphered that the fragment containing aa 80-396 specifically interacts with RANKL. They tested the therapeutic potential of LGR4-ECD against osteoporosis, through binding RANKL and thus, preventing its binding to RANK on osteoclasts (see section 6). Very recently, Jang et al. (42) and Ko et al. (43) demonstrated the effects of RANKL point-mutations (RANKL-MT3, having 2 histidine residues in the extracellular domain aa. 220-230 replaced by phenylalanine and tyrosine), which rendered RANKL unable to interact with RANK, but did not interfere with its interaction to LGR4 caused its high affinity to LGR4. As expected, this RANKL mutation had an inhibitory effect on osteoclastogenesis. Both authors showed that when applied in vitro or in vivo together with the wild-type RANKL, RANKL-MT3 counteracts the effects of the wild-type RANKL and inhibits osteoclast activation. RANKL-MT3 preferably binds LGR4 and reduces NFATC nuclear translocation, a process responsible for osteoclast maturation. Interestingly, Jang et al. (42) demonstrated that RANKL-MT3 administration in vivo generates natural anti-wild type RANKL antibodies in mice, therefore acts as a vaccine preventing excessive bone resorption. They also showed that RANKL-MT3 can reverse ovariectomy-induced osteoporosis.

LGR4, however, does not always act as a sponge neutralizing RANKL/RANK interaction. A recent publication by the group of Yue et al. (44) indicated RANKL-RSPO2-LGR4 interaction

in breast cancer cells is crucial for DKK1 production via activation of Gαq and β-catenin, enabling further breast cancer metastasis into the bones. Inhibition of this pathway can have a therapeutic potential against tumor spread.

LGR4 interacts with its ligands- RSPOs, Norrin and RANKL via its extracellular domain- LGR4-ECD [also alternatively called LGR4-ED (45)]. LGR4-ECD is a part of the membrane bound form of LGR4, but interestingly, it can also be encountered in nature as a circulating (free) form of LGR4. As such, LGR4-ECD was first described by Hsu et al., who identified a natural Lgr4 splice variant encoding specifically only the ectodomain of LGR4. Hsu et al. (45) showed LGR4-ECD to play an important role as a tight regulator of Wnt signaling during mammalian gonadal development. When administered exogenously in vitro, LGR4-ECD inhibited the Wnt signaling in HEK 293 cells in a dose-dependent manner, whereas in vivo, LGR4-ECD treatment decreased ovarian development and steroidogenesis in rats. In the paper already discussed above, Luo et al. (11) tested the potential of LGR4-ECD as an anti-osteoporotic drug. They used three mouse models characterized by high levels of bone resorption and osteoporosis, namely ovariectomy, RANKL injection (46), and Tnfrsf11b-deficiecy (which leads to a lack of osteoprotective Osteoprotegerin-OPG) (47). In all three models, the authors observed improved bone mass and decreased osteoclast activity in response to LGR4-ECD administration. The fact that LGR4-ECD has an anti-osteoporotic effect indicates its direct interference with RANKL-RANK pathway, which triggers osteoclastogenesis and when excessive, leads to osteoporosis.

As mentioned before, upregulation of Wnt signaling accompanied by the increased levels of LGR4 and RSPOs marks a variety of tumors. Thus, neutralizing LGR4 activity with LGR4-ECD may have clinical utility in cancer therapy, particularly in the context of new strategies based on checkpoints inhibitors. Such an approach was tested in the study by Tan et al. (48), who demonstrated that RSPO1-LGR4 interaction promotes the immunosuppressive M2 phenotype in tumor-associated macrophages and is associated with a decreased recruitment of CD8+ T cells capable of neutralizing cancer cells. Using either LGR4-ECD or anti-RSPO1 antibody, Tan et al. were able to revert immunosuppression and inhibit the growth of LLC (Lewis lung carcinoma) tumors and B16F10 melanomas. Also, in the paper by Yue et al. (44), LGR4-ECD was shown to hold therapeutic potential against breast cancer metastasis into bones. These studies demonstrate a significant and a tissue context-dependent therapeutic potential of LGR4-ECD, which is summarized in Figure 3 .

LGR4-ECD mediates its therapeutic effects differentially in different tissues, through binding RSPOs in cancer cells, and binding RANKL in osteoclasts. Therefore, identifying the biding sites for each ligand on LGR4-ECD could enable specificity of its therapeutic effects. In 2013 Wang et al. (12) and then in 2015 Xu et al. (49) described LRRs 3-9 and 4-9, respectively, as the essential domains of LGR4-ECD responsible for binding the furin-like domains of RSPO1). In 2016, by testing 4 differentially truncated forms of LGR4-ECD, Luo et al. (11) concluded that at the primary structure level the region between LRR-NT and LRR8 of LGR4-ECD, corresponding to aa. 28-249, binds RSPO1. Regarding RANKL binding, in silico modeling used by Luo et al. predicted that it binds to aa. 108-346 of LGR4-ECD. Data obtained from experiments with 4 differentially truncated forms of LGR4-ECD indicated that RANKL binds to LGR4-ECD in the region within the aa. 80-369 (LRR1-12). Current knowledge on the structural and three-dimensional interaction between LGR4-ECD and NORRIN is very limited. Based on the paper by Deng et al. (25), NORRIN seems to interact with LGR4 via its β-sheets 3 and 4 (aa. 95-123) since mutations within these particular fragments of NORRIN impair its interaction with LGR4. However, to the best of our knowledge, there is no information available on specific sites within LGR4 protein structure, responsible for NORRIN binding. Specific LGR4-ECD domains responsible for RSPO1 and RANKL binding are depicted in Figure 1B and Figure 1C , in the linear and 3D form.

LGR4 has been shown to protect different types of cells against inflammation, most notably in the digestive system. As mentioned above (see section 4), LGR4 protects differentiated hepatocytes against injury (37). By knocking down Lgr4, Li et al. demonstrated that LGR4 is essential for hepatocyte survival against TNFα-induced damage in primary cultures and in in vivo- models of hepatic injury. The authors demonstrated the activation and internalization of LGR4 upon RSPO1 binding in the hepatocytes. RSPO1 significantly decreased TNFα-induced hepatocyte death, coupled with reduced levels of NFκB-p65 and Caspase-3. Li et al. also showed that RSPO1 protective effects depend on WNT/β-catenin signaling. An earlier study by Planas-Paz (50) described the role of LGR4 and RSPO1 (and ZNRF3/RNF43) in liver zonation and regeneration. The study of Liu et al. indicated Lgr4 deficiency also leads to more pronounced inflammation (inflammatory bowel disease) of the intestine under Dextran Sodium Sulfate (DSS) administration (51). The authors associated LGR4 deficiency with impaired WNT/β-catenin signaling. Reactivation of WNT/β-catenin signaling with genetic or chemical methods caused a significant reduction of inflammation in the Lgr4 mutant mice. This study suggests diagnosis of LGR4 levels may have a potential clinical significance in the context of colitis therapy in humans. LGR4 has also been demonstrated to regulate kidney development (52). By knocking down Lgr4 in mice, the authors revealed that lack of LGR4 during embryogenesis progression promotes increased cell death in the renal peripheral mesenchyme, accompanied by decreased levels of anti-apoptotic protein PAX2 (Paired Box 2). This results in an overall kidney hypoplasia. Wu et al. (53) demonstrated the role of LGR4 in prevention of venous ulcer formation and suppression of inflammation during skin wound healing. In this study, Lgr4 ko mice exhibited poor skin healing and increased proinflammatory profile within the wounds. They also noted a significant drop in LGR4 mRNA levels with the appearance of venous ulcers in human subjects, accompanied by an increase in microRNAs 34a and c, directly targeting LGR4 (see section 9) and causing its downregulation at both, mRNA and protein levels. A recent paper by Chen et al. (54) revealed the role of LGR4 in protection against myocardial ischemia-reperfusion (I/R) injury. When Lgr4 was knocked down in rat cardiac cell line (H9c2), the authors observed increased apoptosis and mitochondrial dysfunction, accompanied by elevated ROS production, decreased ATP production, and inhibition of ERK pathway activation.

A common phenomenon in the studies described above was either impaired Wnt signaling or increased NFκB activity. This seems to be responsible for the detrimental phenotype of Lgr4 knockout in different tissues. On the other hand, as shown by Ge et al. (55) Lgr4 overexpression in the osteoarthritic rat synoviocytes, helps diminish the secretion of proinflammatory cytokines [Interleukin 1 (IL1), TNFα, and Interleukin 6 (IL6)] and the activation of NFκB, resulting in an overall decrease in joint inflammation. These studies collectively highlight the clinical potential of Lgr4 overexpression.

Besides having anti-inflammatory properties, LGR4 plays an important role in metabolism regulation, acting as a positive or negative regulator of various metabolic cues.

So far little is known about mechanisms controlling expression of LGR4. Published data indicate microRNAs, RANKL concentration, and alternative splicing to be important regulators of this process. LGR4 expression at the mRNA and protein levels is controlled by three main microRNAs (miRNAs): 34a and c and miRNA-193-3p. In this context, the study by Cong et al. (72) demonstrated the importance of miRNA34c expression during osteoclast differentiation, showing that miRNA34c suppresses LGR4, (a receptor blocking osteoclastogenesis) and thus enables osteoclastogenesis to proceed. Mechanistically, miRNA34c mediated LGR4 suppression, contributes to increased phosphorylation, and thus, inactivation of GSK3β, which in turn allows NFATC-mediated osteoclast activation. Expression of miRNA34c in osteoclast precursors is regulated by RANKL as well as M-CSF (Macrophage Colony-Stimulating Factor), both stimulating osteoclastogenesis. It is worth emphasizing that negative regulation of LGR4 has an important role in maintaining a basal bone turnover, required for skeleton homeostasis. On the other hand, total inactivation of LGR4 in osteoclast precursors leads to excessive osteoporosis, as described earlier by Luo et al. Besides the bone, miRNA34c, together with miRNA34a orchestrates LGR4 expression in the skin, and in this context, it may have pathological implications. For example, data obtained by Wu et al. (53) indicate miRNA34a and c participate in venous ulcers (pathological skin wound healing) formation via negatively targeting LGR4 in keratinocytes. This disables anti-inflammatory potential of LGR4 and thus potentiates skin wound healing pathology. At the molecular level, the miR34-LGR4 axis regulates NFκB signaling pathway. These authors (53) showed that when miRNA-34 a/c levels increase during the pathological skin wound healing, it leads to increased phosphorylation of GSK3β which results in a further phosphorylation of a component of NFκB pathway- p65, at serine 536. This phosphorylation exerts a pro-inflammatory effect. Negative regulation of LGR4 is also mediated by miRNA-193-3p, which by suppressing LGR4 and indirectly LGR4-dependent transcription factor-ATF4 (Activating Transcription Factor 4), tightly controls the process of bone formation (osteoblast differentiation) (73).

Besides microRNAs, Lgr4 expression can also be modulated by its ligands. For example, Luo et al. (11) demonstrated that expression of Lgr4 in osteoclast precursors is regulated by RANKL concentration and intensity of its interaction with RANK. The authors discovered that when RANKL concentration increases in the extracellular space and it binds RANK to excessively initiate osteoclastogenesis, a negative feedback loop mechanism turns on to stimulate the expression of Lgr4 ( Figure 2 .). This enables a tight control of osteoclast activation. The already discussed paper by Hsu et al. (45), indicated alternative splicing as another Lgr4 expression regulatory mechanism. They discovered the reading frame shift during Lgr4 translation, resulting from the introduction of an early termination codon, preceding translation of seven-transmembrane domains of Lgr4. Alternative splicing of Lgr4 leads to formation of a 500 amino acid protein, containing solely its extracellular domain which acts as a decoy receptor for RSPO2 and NORRIN during gonadal development (see section 6).

LGR4 has emerged as a G protein-coupled receptor that controls multiple pathways and biological processes, traditionally regulating WNT-mediated signaling and crucial during development of multiple organs: eye, reproductive tract, and kidney. LGR4 is required for stem cell maintenance in the small intestine. As recently discovered, it also plays Wnt-independent anti-inflammatory and energy expenditure-orchestrating roles ( Figure 4 ). Thus, LGR4 is important at a systemic level in the maintenance of whole-body homeostasis. At the molecular level, LGR4 participates in different intracellular signaling pathways, including Wnt, NFκB, TGFβ, and cyclic AMP. This indicates LGR4 to be involved in a plethora of mechanisms controlling cellular events in a tissue-specific manner. LGR4 expression level needs to be tightly controlled: too little or too much of LGR4 can lead to impairments, such as osteoporosis, delayed puberty, obesity, and cancer. Besides the significance of the membrane-bound LGR4, its soluble form- LGR4-ECD exhibits promising therapeutic potential in osteoporosis and cancer/metastasis progression treatment. Altogether, this makes LGR4 an interesting therapeutic target in several clinical contexts. However, the role of LGR4 in many cell types and tissues, such as the pancreatic beta cells, remains to be defined and is an emerging area of interest.

It might be particularly interesting to study this receptor and its regulation in the context of autoimmune and metabolic diseases (e.g., Type 1 and Type 2 diabetes) given the common anti-inflammatory action of LGR4 in different tissues as well as its emerging role in metabolism. Such studies might reveal novel signaling functions and physiological roles of LGR4 and provide new avenues of therapeutic interest. Studies assessing the presence of LGR4-ECD in circulation, and its role in physiological and pathophysiological conditions, might have therapeutic relevance as well.

JF prepared the initial and final versions of the manuscript and figures. NGK edited the final version of the manuscript. NLR edited the final version of the manuscript. SD edited the initial and final versions of the manuscript. RCV discussed and edited the initial and final versions of the manuscript and figures. All authors contributed to the article and approved the submitted version.

We also gratefully acknowledge our funding sources: Juvenile Diabetes Research Foundation (JDRF)-sponsored postdoctoral fellowship no. 3-PDF-2020-936-A-N (to JF); work in Vasavada laboratory is supported by grants (to RCV) from the National Institutes of Health (R01DK125856, R01DK102893), City of Hope Start-up funds, and the Wanek Family Project to Cure Type 1 Diabetes at City of Hope; work in Dhawan laboratory is supported by grants (to SD) from the National Institutes of Health (R01DK120523), City of Hope Start-up funds, the Wanek Family Project to Cure Type 1 Diabetes at City of Hope, and Human Islet Research Network (NIH) UC4 DK104162.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.