Multidimensional outlook on the pathophysiology of cervical cancer invasion and metastasis

Cervical cancer being one of the primary causes of high mortality rates among women is an area of concern, especially with ineffective treatment strategies. Extensive studies are carried out to understand various aspects of cervical cancer initiation, development and progression; however, invasive cervical squamous cell carcinoma has poor outcomes. Moreover, the advanced stages of cervical cancer may involve lymphatic circulation with a high risk of tumor recurrence at distant metastatic sites. Dysregulation of the cervical microbiome by human papillomavirus (HPV) together with immune response modulation and the occurrence of novel mutations that trigger genomic instability causes malignant transformation at the cervix. In this review, we focus on the major risk factors as well as the functionally altered signaling pathways promoting the transformation of cervical intraepithelial neoplasia into invasive squamous cell carcinoma. We further elucidate genetic and epigenetic variations to highlight the complexity of causal factors of cervical cancer as well as the metastatic potential due to the changes in immune response, epigenetic regulation, DNA repair capacity, and cell cycle progression. Our bioinformatics analysis on metastatic and non-metastatic cervical cancer datasets identified various significantly and differentially expressed genes as well as the downregulation of potential tumor suppressor microRNA miR-28-5p. Thus, a comprehensive understanding of the genomic landscape in invasive and metastatic cervical cancer will help in stratifying the patient groups and designing potential therapeutic strategies.


Introduction
Cervical cancer (CC) is the most common tumor of the female reproductive system and the fourth most prevalent cancer type overall [1][2][3].The incidence and mortality rate of CC are the leading causes of death among women and vary among racial and ethnic groups, suggesting that genetic and epigenetic factors may determine the outcome of the disease [4,5].Depending on the histopathology and severity of the cervical lesion, 69% of the CC are classified as squamous cell carcinoma (SCC) and 25% as adenocarcinoma or adenosquamous cancers.The other 6% include small cell carcinoma, rhabdomyosarcoma, primary cervical lymphoma, and cervical sarcoma [5].SCC is most likely to develop from the ectocervix and adenocarcinomas arise from the endocervix [3].Human papillomavirus (HPV) is the key factor as the infections are detected in approximately 99.7% of the CC; however, HPVnegative aggressive CC is also identified.Persistent infections with 14 high-risk HPV genotypes are responsible for nearly all occurrences of cervical squamous cell carcinoma (CSCC) and adenocarcinoma [28].Women infected with HPV develop precancerous phases with low-grade squamous intraepithelial lesions/CIN I (LSIL) and high-grade intraepithelial lesions or CIN II/CIN III (HSIL), and these high-grade lesions eventually lead to invasive CC over a period of time [6][7][8].The role of HPV infection in tumor initiation and progression has been elegantly reviewed by Deligeoroglou et al. [29] and Steinbach and Riemer [30], and reports have been published on the mechanisms of immortalization of the host cells by HPV through signaling pathway dysregulation [11,[31][32][33], genomic instability [34,35], alteration of the microbiome [36][37][38][39], and immune response modulation [40][41][42][43][44]. Therefore, this review only focuses on the cervical microbiome causing inflammation and discusses various events associated with CC metastasis.
Several genetic and epigenetic variations and tumor microenvironment heterogeneity contribute to malignant transformation [12,13].The mutation burden, copy number variations, dysregulated expression of genes, microRNAs (miRNAs), long non-coding RNAs (lncRNAs), circular RNAs (circR-NAs), and DNA methylation status vary among women during different stages of CC [5,[9][10][11].Further, the observed molecular landscape presents the interdependent regulatory mechanisms, which favor survival, development, progression, invasion, and metastasis.Early cancer cells develop into successive generations of cells with accumulated mutations, resulting in vigorous neoplasty.Metastasis begins with the spread of cancer cells from primary tumor sites and invading into the surrounding tissues.Furthermore, these cells can enter hematogenous or lymphogenous circulation as single cells or in clusters, represented by circulating tumor cells (CTCs).Cells then intravasate from the circulation and colonize at the secondary sites [14,15].In terms of biological behavior and genetics, these cells differ significantly from the primary tumor with an abnormal morphology and exponential growth [16,17].Several events and number of molecules are associated with metastasis.These can also be the cervicovaginal microbiome [18], immune cells in the tumor microenvironment such as macrophages and dendritic cells in association with naive T cells [19], pH of the microenvironment [20], and differential expression of oncogenic pathway proteins, such as EGFR1, Wnt/β-catenin, NF-κB, AKT, MMP3, TGF-β that may promote metastasis [21][22][23][24][25][26].Hence, comprehensive multi-omics analysis of CC at different stages can aid in devising effective strategies for CC classification, diagnosis, and treatment.Though the early detection and treatment of CC reduces the associated risk, metastatic and invasive stages with lymphatic dissemination have poor survival statistics.Therefore, new therapeutic approaches that aid in better prognosis of CC is inevitable.In this review, initially, we discuss briefly how HPV-mediated microbiome dysbiosis supports CC progression, followed by the mechanisms underlying the pathological processes during CC cell invasion and metastasis to distant sites through lymph nodes.Furthermore, we have enumerated the somatic deregulation of key genes associated with the metastatic phenotype and their potential in preventing invasion and metastasis in CC.Hence, this study will aid in identifying potential therapeutic determinants to develop better treatment strategies and to prevent CC invasion and metastasis.

Dysregulation of cervical microbiome
It has been established that the cervicovaginal microbiome dysbiosis in concert with HPV infection contributes to CC initiation [18,27].The immune system at the cervical region of the uterus is governed by immunocompetent cells located along the epithelial lining of the cervix [46].A tight regulation of innate and adaptive immunity is established in this region through a balance of the equilibrium between the ability to mount a rapid defensive immune response against invading microbial pathogens and tolerance toward commensal bacteria.During CC progression, these concerted cellular interactions are disrupted leading to a compromised immune response, resulting in low-grade chronic inflammation associated with increased susceptibility to viral and bacterial infection in various CIN stages Fig. 1 [47].
Transition of the cervical epithelium to squamous intraepithelial lesions (SIL) and CC via high risk-HPV (HR-HPV) changes the microbial diversity with an increase in the abundance of Fusobacterium species [48].

The immunosuppressive microenvironment initiated by the HPV infection is inhabited by Sneathia and other
Fusobacterium species as it progresses to SIL.Later in the CC stage, Fusobacterium necrophorum colonizes abundantly and supports CC progression by modulating the cytokine profile in the tumor microenvironment [48].Rubinstein et al. [49] reported that F. nucleatum is associated with colorectal carcinogenesis through FadA adhesinmediated phosphorylation of E-cadherin, resulting in the internalization of E-cadherin together with FadA.This in turn stimulates the accumulation of β-catenin in the cytoplasm leading to the transcriptional activation of NF-κB signaling and hence promoting tumor cell proliferation.Currently, there is no conclusive evidence supporting the role of microbiome dysbiosis in CC metastasis.Therefore, further research is needed to obtain adequate knowledge about the spectrum of the cervical microbiome and its molecular determinants to envisage preventive and curative strategies for CC cell migration, invasion, and metastasis.

Metastatic hallmarks
Metastasis is a long evolutionary multistep process that initiates the dissemination of cancer cells.Most of the cells fail to develop into a fully-grown tumor at another site, while some stay alive but fail to produce any detectable clinical manifestations and few cells initiate metastasis by evading Fig. 1 The role of cervical epithelial cells in balancing the equilibrium of the local immune system.Cervical microenvironment and the local microbial diversity alter local and systemic immune response, which play an important role in the progression of CIN toward CC.Lactic-acid-producing bacteria acidify the vaginal milieu pH to 4.6 during eubiosis, with lactic acid as the primary metabolite, thereby providing a non-inflammatory environment.The dysbiotic environment would have a lower redox potential during microbial vaginosis, a phenomenon that encourages the growth of a diverse bacterial species, resulting in the increased pH.Virulence factors produced by the diverse bacterial species undermine epithelial barrier integrity, degrade mucin, and create a pro-inflammatory milieu all the immune checks.It is highly unlikely that the entire process of metastasis is regulated by a single gene.The regulation of metastasis involves activation and deactivation of various genes, and these alterations can be permanent or transient [50].

Lymph angiogenesis: progression of cervical cancer metastasis
Microvascularization is an early regulation, which is inevitable during the gradual progression of CSCC.Numerous signaling cascades, which can be attributed to the dysregulation of various genes, contribute majorly for the development of angiogenesis and lymphangiogenesis, which is indispensable for the progression of CIN toward invasive CSCC.Pathological changes associated with tissue inflammation or tumor progression are involved in the aberrant angiogenic and lymphangiogenic development, which is a prerequisite for the dissemination of the tumor to distinct sites [92,93].Hematogenous and lymphogenous spread of tumor cells are the two major routes well studied for cancer dissemination and are associated with poor prognosis.Although blood and lymph vessels share same origin, they have differences in their structure and function.Lymphatic vessels are derived from the vascular endothelium, whose formation is regulated by Prospero homeobox protein 1 (PROX1), a lymphatic specific transcription factor [93,94].Association of PROX1 with COUP-TFII upregulates the vascular growth factor receptor-3 (VGFR-3) signaling by vascular endothelial growth factor-C (VEGF-C) and allows the migration of cardinal vein lymphatic progenitor cells to form primary lymphatic plexus at the adjacent mesenchyme.This migration of the progenitor cells is balanced by the collagen and calcium-binding EGF domain1 protein.Lymph angiogenesis is then initiated via VEGF-C signaling [93].On the other hand, angiogenesis is regulated by VEGF-A through VEGFR-1 and VEGFR-2 receptor signaling [95].The lymphatic vessels have approximately threefold wider lumen than the blood capillaries, and the tumor cells prefer to disseminate through the lymphatic system due to the high levels of hyaluronic acid, which gives protection from the blood serum toxicity and provides a safe path for the migrating cells.The low flow rate of the lymph causes minimum shear stress to the cells, which makes it an energy-efficient mechanism for the evading cells.The lymph endothelial cells with a leaky arrangement on the surface of the lymphatic capillaries and lacking basement membrane will in turn support the easy migration of the tumor cells to their target sites [94].Metastasized tumor cells will have additional traits such as mutations and genetic heterogeneity to overcome the lymphatic barrier.Analysis of the tumor driver genes and their associated mutations are highly correlated with cell survival, cell fate, and maintenance of genome [96].Genetic heterogeneity due to the unstable genome of the subpopulations of cells can expand and cope in adverse environment and affect tumor evolution.Genetically modified subpopulations can entirely suppress the growth of primary clones imparting resistance to therapy and immune checks [97].In most solid tumors, metastasis through the lymphatic system is observed in the early stage where the infiltrated tumor cells can either migrate to different sites or remain dormant in the lymph vessel.

Molecular determinants of lymphatic dissemination in CSCC
The role of lymph node in CC is recognized and is included in the examination of lymph node metastasis (LNM) for staging the CC by the International Federation of Gynaecology and Obstetrics (FIGO) classification in 2018.Occurrence of LNM in the pelvic region reflect stage IIIC1 while that in para-aortic region reflect stage IIIC2 [98,99].CC is mainly known to metastasize utilizing three main routesdirect invasion into the neighboring tissue, hematogenous dissemination, and lymphatic dissemination [5].Pelvic and para-aortic lymph nodes are the first sites for tumor draining and nodal metastasis in CC [100,101].Further, high lymphatic vessel density (LVD) with elevated levels of lymphangiogenic factors such as VEGF-C and VEGF-D are identified in pre-invasive neoplasia (CIN3) [5,[102][103][104].VEGF, an angiogenic factor with well-established prognostic value in gynecologic cancers [105], is overexpressed in early stage of CSCC and is positively correlated with microvessel density (MD) at the early stage of CSCC.This suggests the role of VEGF as a potential marker for developing benign tumor to an invasive state.VEGF is also linked to the incidence of LNM of CSCC [106,107].Studies have also shown that hypoxia-induced transcription factor (HIF-2α) with its increased expression in association with VEGF reduces the patient survival rate in CSCC.This correlation between HIF-2α and VEGF was clearly observed in tissue samples with high FIGO stages of LNM [108].VEGF activation can also be mediated by the guanine nucleotide exchange factor (GEF) NET1 in CSCC without influencing MMP2 and MMP9 expression.NET1-mediated activation of RHOA, FAK, JNK, NF-κB and Wnt signaling was observed during the initiation and progression of various cancers.Among the risk factors, LNM indicates poor prognosis at early or late stages of CSCC [109] and higher the risk of recurrence.
Similarly, along with VEGF expression, metabolic parameters such as total lesion glycolysis (TLG) are significantly higher during lymphatic metastasis of CSCC.Therefore, combined evaluation of the TLG and VEGF may help in predicting the LNM [110].
The receptor for activated C kinase 1 (RACK1), a scaffold protein overexpression in CC, showed increased ability for lymph node metastasis.HPV16/18 proteins E6 and E7 have shown significant association with the upregulation of RACK1 in CC with an increased expression of N-cadherin and SNAIL and downregulation of epithelial markers such as E-cadherin and ZO-1 [70].E6 stabilizes RACK1 through O-GlcNAcylation at Ser122 and promotes tumor invasion and metastasis by inhibiting miR1275, which inhibits LGALS1 gene that encodes galac-tin1.Therefore, inhibition of miR-1275 increases galac-tin1 expression, which in turn activates MEK/ERK, FAK and AKT signaling involving integrin-β1 and promotes LNM in CSCC [70,71].Previous evidence has shown that the galectin1 activates NRP1 signaling to phosphorylate VGFR2 during endothelial vascular cell migration and angiogenesis [111].Additionally, the activation of miR-221-2p is also positively correlated with LNM through the downregulation of THBS2 protein [112].The enriched expression of miR-221-3p found in CSCC-secreted exosomes inhibits the expression of VASH1 by lymphatic endothelial cells and induces lymph angiogenesis and LNM.The increased density of peritumoral lymphatic vessel density (PLVD) than MVD indicates that CSCC prefers lymphatic mode of metastasis than hematogenous dissemination.miR-221-3p also increased the phosphorylation of AKT and ERK1/2 proteins independent of VEGF-C, demonstrating that miR-221-3p-VASH1 axis share common pathway with VEGF-C in lymphatic vessel sprouting and metastasis [113].This indicates that miR-221-3p is involved in context-dependent activation of microvessel development.The molecular changes in the tissue biopsy show an upregulation in EMT-inducing transcription factors TWIST1 and SNAIL in association with metastasis and lymphovascular space invasion (LVSI) Fig. 2 [114].Further, studies have correlated the expression of lncRNA with cellular invasion in CC, which explains their potential role as a biomarker for cancer metastasis.Mining the lncRNA profile of human cancer revealed 234 lncRNAs associated with pelvic lymph node metastasis (PLNM).Among these lncRNAs, MIR100HG and AC024560.2are highly associated with the deregulation of gap junction proteins.The aberrant expression of gap junction proteins assists LNM in ductal breast cancer, oral SCC and ovarian adenocarcinoma.Validating the expression of these lncRNAs in CC tissues may provide promising evidence for their diagnostic potential at an early stage [121].Another study by Shang et al. [127] demonstrated the role of lncRNA regulation in fatty acid metabolism and tumor metastasis in cancer stem cells (CSCs).Reprogramming of fatty acid metabolism by lncRNAs mediated through FABP5 in CC activates VEGF-C and promotes LNM.

Tumor microenvironment in lymphatic dissemination of CSCC
Tumor-associated macrophages (TAMs), matrix metalloproteinases, interleukin-2 (IL-2) and TGF-β present in the microenvironment are the key factors involved in lymphatic dissemination [115][116][117][118][119]. Oany et al. [120] reported the upregulation of immune and inflammatory response genes in CC.High percentage of protein tyrosine phosphatase receptor-type C + (PTPRC + ) cells in cervical tumor are associated with enhanced tumor-infiltration by T-BET + cells and FOXP3 + cells [120].In addition, the tumor microenvironment participates in tumor dissemination by increasing the permeability of the endothelial barrier of the lymphatic vessels [106].Recently, research has advanced in elucidating the role of cancer associated fibroblasts (CAFs) in tumor progression.Reconstitution of CAFs enables the non-metastatic tumor to undergo proximal LNM in CC cells cotransplanted in athymic mice [122].Wei et al. [124] demonstrated that the metastasis of CSCC to the lymph node is highly influenced by the subpopulations of CAFs in the tumor microenvironment.They have identified a specific subset of CAFs with elevated periostin expression, which activates integrin αvβ3 and αvβ5 signaling and triggers the phosphorylation of FAK (Tyr397) and SRC (Tyr416) that results in the direct degradation of VE-cadherin.These may breach the endothelial barrier and establish LNM in CSCC.Therefore, identifying protumorigenic CAFs and targeting the FAK-SRC axis could form an efficient target for the prevention of tumor metastasis [124].Another important factor associated with lymph metastasis is the extracellular matrix (ECM) remodeling.ECM component includes network of collagen, fibronectin, laminin, elastin and proteoglycans, which provide structural and mechanical stability to the tissue and regulate the availability of cytokines and growth factors in the microenvironment.In CSCC, HPV E6/ E7 modulates carcinogenic process by regulating the ECM protein CTHRC1 and upregulating its expression in CC tissue.E6/E7 expression activates POU2F1 by inhibiting p53, which in turn activates CTHRC1-mediated Wnt/PCP signaling pathway and promotes cancer cell migration and invasion [125].Studies have also showed that overexpression of MMP-2 (gelatinase A) and MT1-MMP (collagenase) by HPV16 E6/E7 facilitates ECM degradation and cancer invasiveness [126].

Immune evasion of CSCC in LNM
Primary lesions disseminate into the lymphatic drainage through sentinel lymph node (SLN) [128].The migratory antigen-presenting macrophages or dendritic cells activate CD8 + T cells in lymph node tumor-draining site, which provides first line of defense [128][129][130].However, developing tumor cells target tumor-draining lymph nodes (TDLNs) to create a pre-metastatic niche with suppressed immune response, harboring dysfunctional anti-tumor T cells [131].
The increase in the level CD1a + dendritic cells at the TDLN provides evidence that dendritic cell maturation is inhibited at the lymph node by tumor-derived growth factors, such as prostaglandin-E2 (PGE2), TGF-β, IL-6, and VEGF [128,[132][133][134][135][136].Further, blood vessel remodeling, lymphangiogenesis, and increased chemokine and cytokine secretion alter the immune cell composition and give rise to a 'tumorsupportive' microenvironment [131].Moreover, high Treg counts along with increased cytokine release and a consistent immune suppressive microenvironment were observed with high levels of TNF-α, IL-6, IL-10 and reduced IFN-γ expression [128].Metastasis was preceded by low CD8 + T Fig. 2 Illustration showing the molecular signaling mechanisms by which CSCC undergoes epithelial-mesenchymal transition, lymph angiogenesis, and metastasis via the lymphatic system cell/FoxP3 + Treg ratios, creating pre-metastatic niches in the tumor-draining lymphatic basin [137].The recruitment [138] and polarization of CD14 + monocytes into the suppressive PD-L1 + M2-like macrophages (co-expressing CD14 and/or CD163) by primary tumor cells (possibly via the secretion of CCL2) prevent the antigen presentation and immune response in the lymphatic niche [139].These M2-macrophage-like cells are not only incapable of generating proper CD8 + T cell responses but favor Treg expansion, and facilitate the production of pro-angiogenic and pro-tumor-invasive factors [128,140,141].Elevated rates of Tregs and CD4/CD8 T cell ratios were observed prior to metastatic involvement in cervical TDLN [142].As the metastasis progresses in TDLN, myeloid regulatory cells and memory T cells accumulate resulting in the release of exosomes carrying immune modulatory elements and soluble mediators, leading to "tumor-supportive" microenvironment [131,142,143].The accumulation of Tregs inhibits lymph node-resident conventional dendritic cell (LNR-cDC) activation causing the conversion of Ag-specific naive T cells into Tregs in TDLN [143,144].Transcriptomic analysis of CC data from The Cancer Genome Atlas (TCGA) by Yang et al. [145] suggested the association of LNM with upregulation of immune biomarkers.In high-risk group, LNM was associated with FABP4 and CXCL2 upregulation, while TEKT2 and LPIN2 were downregulated [145].Upregulation of FABP4 promotes EMT by activating AKT pathway [146], whereas CXCL2 promotes tumor growth and angiogenesis via NF-κB signaling [147].Expression of CXCL2 is positively correlated to the neutrophil activation and poor prognosis.Additionally, the expression of PD-L1 was upregulated in high-risk group and promotes CC metastasis to the lymph node [69,148].PD-L1 is characterized for its role in masking the tumor cells by binding to the PD-1 receptor expressed on the T cells and inhibiting immune response, thereby creating an immunosuppressive microenvironment [149].PD-1 phosphorylates and activates PD-L1-integrinβ4 (ITGB4) signaling pathway followed by the activation of AKT-GSK3 pathway.This results in the upregulation of SNAI1 and MMP1 and decrease in the cytokine production [150].The PD-L1-integrin axis also increases the glucose metabolism and tumor cell proliferation through SNAI1-mediated downregulation of SIRT3 in CC [148], suggesting that the reprograming of Warburg effect is essential for the tumor cell dissemination.

Migration and colonization of CSCC to distinct organs
According to the SEER database analysis from 2010 to 2015, there are four major sites for CC metastasis.Lungs are the usual site of metastasis and colonization with a rate of 56% followed by liver with 16%, bone with 23%, and brain with 2% [151].Cox multivariate analysis of the patient data obtained from SEER database confirm that increasing age, non-squamous type, advanced stage, metastasis through pelvic lymph node and poor differentiation are risk factors for lung metastasis [152].It was reported that the pulmonary metastasis occurs heterogeneously and in most of the cases the patients are asymptomatic [152,153].Studies on the metastatic niches of CC in large population showed lungs as the most usual site for colonization.Major challenge is to differentiate lung SCC from metastasized CSCC.It is reported that HPV modulates p16 upregulation and suppresses pRB-E2F signaling by inhibiting CDK4 cyclindependent kinase.In CSCC, p16 expression is relatively higher than pulmonary SCC as it is upregulated by HPV infection [153,154].Therefore, along with p16 expression, the presence of HPV DNA is used for differentiating the metastasized SCC from the lung SCC [154].When the malignant tumor cells enter metastasis, the surviving cells may infiltrate to the distant organs.These infiltrated cells relapse at their new niche, eventually proliferate to form tumor, and distort the function of the host organ.Colonization of the host organ is the slowest step of tumor dissemination, and there is no adequate reference for the mechanism by which CSCC infiltrate to its distant organs.But recent advances in the technologies to access the detailed insight of the CTCs help to conceptualize the triggers for organ infiltration [155].CTCs modify the chemokine gradient for tumor invasion and colonization.This invasive front is comprised of myeloid progenitor cells, TAMs, CAFs and newly formed blood vessels [156] along with Notch, Wnt, TGF-β and cytokines signaling [157].

Cytokine-chemokine gradient driving tumor invasion to host niche
Disseminating tumor cells (DTCs) migrate as single cells or as a group.The success in the establishment of tumor metastasis depends on the ability of CTCs to overcome various hindrance in the hematogenous or lymphogenous pathways.Emerging studies have shown that collective migration of the metastatic cells has higher invading ability and successfully establishes tumor migration to distinct organs.They move in clusters possessing strong adhesion along with the tumor microenvironment, which may provide resistance toward clinical interventions [155,158,159].Recently, Pein et al. [159] demonstrated the role of metastasis-associated fibroblasts (MAFs) in migration and colonization of breast cancer cells to the lungs by modulating the expression of chemokines.For the evolution of a supportive niche in the lungs, migrating breast cancer cells secrete interleukins IL-1α/β.Additionally, they also observed a subset of lung fibroblasts that expresses CXCR3, which binds to CXCL9/10.The secreted interleukins induce chemokines CXCL9/10 in the MAFs via NF-κB signaling.CXCL9/10 fuel the proliferation and metastatic outgrowth in lungs by activating the MAFs and support the colonization of CTCs in the lungs [159].Chemotactic signals modulate the interaction of the cancer cells within the cluster and its migration into host organ.Another study on the role of CXCL17 in the lung metastasis of the breast tumor cells provides additional evidence to substantiate the chemokine facilitation of tumor dissemination.CXCL17 induces accumulation of the CD11b + Gr-1 + myeloid-derived suppressor cells (MDSCs) in the lungs and promote tumor angiogenesis and extravasation of cancer cells to the lungs through the overexpression of PDGF-BB [160].These explain the engagement of chemotactic signals in the colonization of primary tumors and could be the possible mechanism supporting CC metastasis and colonization in pulmonary tissue.Another study demonstrated that the overexpression of NRAS in the CTCs promote lung colonization by IL-8 mediated the expression of CXCL5 and pro-platelet basic protein (PPBP).These factors may target the CXCR1 receptor of pulmonary blood vessel and facilitate the adhesion of the tumor cells to the lung vasculature as well as the recruitment and homing of CXCR2 receptor-expressing myeloid cells to the microenvironment.Thus, chemokine secretion-mediated activation of CXCR1/2 receptor at the metastatic site helps the CTCs to form pre-metastatic niche in the lungs [161].
The incidence of bone metastasis in CC is relatively rare and increases with the advanced stage.Previously, Matsuyama et al. [162] reported the association of bone metastasis of CC with its clinical stages among stage I (4%), stage II (6.6%), stage III (8%) and stage IV (22.9%) [162] of bone metastasis with high frequency to the lumbar spine and to the pelvic bones [162,163].A case report of 70-year-old women with FIGO stage IIIA also showed metastasis within fibula, calcaneum and right tibia [164].The ECM composition of the bone favors the cancer cells to colonize and proliferate in the bone as it tightly controls the bone remodeling.The major organic component of bone ECM is collagen1, which is involved in providing bone strength and stiffness [165] and also supports survival, proliferation and differentiation [166].The accumulation of collagen type I at the pre-metastatic niche of the host organ distorts the alignment of ECM and supports the colonization of CTCs [165,167].Like the lung colonization of breast cancer cells, chemotactic paracrine signaling of CXCL5 induces bone metastasis via activation of CXCR2 receptor on the bone marrow [168].Primarily, chemokine gradient produced by the CTCs shapes the tumor microenvironment at the host pre-metastatic niche, promoting angiogenesis and recruiting antitumoral leukocytes, TAMs and MDSCs to support the metastatic invasion of host site.The studies have shown that the secretion of endothelin-1 by the circulating breast cancer cells modulates TGF-β production in osteoblast leading to bone metastasis [169].The metastasis of cervical cancer directly to the brain is a rare occurrence [170,172]; instead, spreading of the tumor cells to the lungs as it is the most common site and subsequently to the brain is observed [170].Another possible pathway for dissemination is from the pelvic veins to the vertebral venous plexus and then to the brain parenchyma via the venous sinus of the brain [171].
Liver metastasis was found to be significantly correlated to brain, lung and bone metastasis, as liver has rich blood supply from portal as well as arterial venous system making cancer cells to spread easily [173].The incidence of liver metastasis is equally low, accounting for only 1.2-2.2percent of all cases and only 5% of the case developed hepatic metastasis alone without any extrahepatic comorbidity [174].The liver metastasis occurs later in the progression of primary CC; thus, early surgical excision of primary lesions can successfully prevent hepatic metastasis [175].Current understanding on the mechanism of distinct organ metastasis by CC cells is very limited.Further studies are warranted to establish the characteristic features associated with CTCs of CSCC and its colonization and will help to unravel the mechanisms promoting CC cell migration and colonization.

Genetic and epigenetic features in cervical cancer metastasis
Initiation and development of CC is a complex process [8].Although HPV infection is the critical determinant for the risk of developing CC, not all women infected with HPV develop CC.Host genetic variants and environmental factors also add to the risk towards the susceptibility to CC [176].Further, the genomic and epigenomic profiling of CC at different stages along with the gene expression analysis supports the role of these events in CC initiation, progression, invasion and metastasis [9,10].Hence, comprehensive genomic and epigenomic land scape of CC can help in stratification of affected individuals and may facilitate effective treatment strategies [10].
Although mutations are prevalent in CC, only a few can act as driver mutations, which are expected to initiate and promote growth, and these are common mutations identified in cancer invasion and metastasis [9].Additionally, HPV infection can result in increased mutational spectrum [41].Furthermore, when metastasized to unrelated lineage microenvironment, the tumor cells can acquire new mutations [9], which are genotypically favorable for malignant cells [180].Genomic profiling of CC patients reported frequent mutation in EP300, MUC4, MUC16, SYNE1, KMT2C, PIK3CA, FLG, KMT2D, DST, MAPK1, and TTN [10].Also, the mutation signatures depict DNA mismatch repair deficiency (COS-MIC Signature 6), APOBEC cytidine deaminase (COSMIC Signature 2), and spontaneous deamination of 5-methyl cytosine (COSMIC Signature 1) pattern [10].Integrated molecular profiling of primary and recurrent/metastatic CC from same individual reported elevated mutation burden and copy number alteration.Mutations in epigenetic regulators such as NSD1, ARID1A, CTCF, ARID1B, KMT2C, SETBP1, PBRM1, and KMT2D were specific to recurrent/metastatic CC [9].APOBEC-related mutation signature along with reduced expression of APOBEC3A was also observed for recurrent/metastatic CC.Also, in some cases of CC, nuclearencoded sigma factor 6 (SIG6) mutation is associated with defect in mismatch DNA repair [9].
Our bioinformatic analysis of TCGA data of CC revealed reduced expression of miR-28-5p in CC patient (unpublished data).miR-28-5p is located at 3q27.3, which is frequently gained in CC.The significant downregulation in CC could be due to circRNA ArfGAP with FG repeats 1 (cir-cAGFG1), which binds to miR-28-5p and targets HIF-1α, thus promoting proliferation, invasion, and migration as well as suppression of apoptosis by escalating glycolysis through circAGFG1/miR-28-5p/HIF-1α axis as observed in non-small-cell lung cancer (NSCLC) [199].Furthermore, lncRNA CDKN2B antisense RNA 1 (CDKN2B AS1) which is known to be upregulated in colorectal cancer has been reported to bind miR-28-5p in order to regulate proliferation as well as apoptosis inhibition [200].Similar mechanisms of regulation may be evident in CC hence downregulation of miR-28-5p could promote CC progression and invasion.This evidence also suggests the intricate and interrelated regulatory pathways that exist for the suppression or the progression of CC [38,45,181,201,202].

Metastasis-regulated differentially expressed genes
GSE26511 dataset from TCGA including 39 samples (20 CC-lymph node-negative samples and 19 CC-lymph nodepositive samples) was analyzed by Ge et al. [204].Approximately 1,263 genes were differentially expressed, and these genes were associated with signaling pathways, cell cycle processes, immune response, regulation of immune system processes, inflammatory response, and cell activation [204].
Dataset analysis of 116 non-metastatic and 10 metastatic samples from "TCGA" indicated differential expression of transcription factor NR5A2 in metastatic tissue compared to non-metastatic tissue of CC [3].Earlier, NR5A2 had been identified to contribute in developing CC during GWAS [10,205].NR5A2 plays an important role in maintaining pluripotency in embryonic stem cells (ESCs) [206] and reprogramming of somatic cells into induced pluripotent stem cells (iPSCs) [207].The genetic heterogeneity among the CC, recurrence and radio/chemotherapy resistance, tumor invasion and metastasis are attributed to the presence of CSCs [208,209].CSCs have self-renewal and multi-lineage differentiation abilities [210].Significant and positive correlation was observed between NR5A2 and vimentin [3].Vimentin is the biomarker of EMT, which is the first step for invasiveness and metastasis [211].EMT is characterized by the reduced expression of E-cadherin (epithelial marker) and increased expression of N-cadherin and vimentin (mesenchymal marker) [3,212].Thus, NR5A2 positively regulates vimentin and activates EMT signaling pathway in metastatic CSCC [3].Transcription factors associated with EMT, such as SNAI (SNAI1 and SNAI2), ZEB (ZEB1 and ZEB2), and TWIST (TWIST1 and TWIST2), can suppress expression of E-cadherin and regulate the EMT through different pathways [213].Matrix metalloproteinase-1 (MMP1) degrades ECM during both physiologically normal and disease processes.In CC, it is upregulated and aids in cancer cell invasion, migration and proliferation via EMT.MMP1 is closely linked with LNM [203,214].Differential gene expression analysis of primary and recurrent/metastatic CC from the same individual suggested upregulation of activated antitumor immunity gene in primary CC and genes involved in EMT and angiogenesis in recurrent/metastatic CC.While CXCL9, SPEG, MUC21, and APOBEC3A were downregulated, POSTN was upregulated in metastatic CC compared to primary tumor [9,203].
We performed in silico analysis with the limma R package [216] for identification of differentially expressed genes (DEGs) in five datasets of metastatic cervical cancers deposited in the Gene Expression Omnibus database [215].Genes with common expression patterns in all the datasets were selected for subsequent analyses.The filtering threshold of p-value < 0.05 and logFoldChange > 1.5 for upregulation, logFoldChange < -1.5 for downregulation was set as the standard filter.The genes that fall under the standard filter criteria were selected as DEGs.We have identified 70 upregulated and 46 downregulated genes in at least two datasets.The overlapping DEGs identified in more than two of the selected datasets were extracted and represented in the UpSetR plot Fig. 3a with the UpSetR package [217].Further, protein-protein interaction (PPI) network was constructed for the overexpressed Fig. 3b and downregulated genes Fig. 3c using Search Tool for the Retrieval of Interacting Genes (STRING) [218].Additionally, we have used the Maximal Clique Centrality (MCC) algorithm of CytoHubba [219] to identify the hub nodes in the co-expressed network.The densely interacting genes with a score ≥ 0.4 were filtered for the hub genes, and top 10 genes were identified based on the MMC score.Among the hub genes identified with a high MMC score, KRT6B, TGM3, ALOX12B, and CRCT1 were found to be upregulated Fig. 3d and DDX58, IFI44, OAS2, IFI44L, and IL15 Fig. 3e were downregulated in metastatic CC.
Further, the gene ontology study identified that the upregulated genes were primarily involved in protein binding and serine-type peptidase activity and the downregulated genes were involved in catalytic activity, localization, and immune response Fig. 4a, b.The functional enrichment analysis of the DEGs was performed for molecular function, biological process, and cellular component using g.Profiler tool [220] with a p-value cut-off score of 0.05.
When pathway analysis was performed on the DEGs from WikiPathways using EnrichR tool [221] with p-value > 0.05 set as the cut-off, suggesting that the upregulated genes were significantly involved in Wnt signaling, chromosomal and microsatellite instability in colorectal cancer, apoptosis modification and signaling Fig. 5a.The downregulated genes showed association with the TGF-β signaling pathway, host-pathogen interaction of human coronavirusinterferon-induced pathways regulating Hippo signaling, MAPK and NF-κB signaling pathways Fig. 5e.Further, using NetworkAnalyst [222], we predicted host-microbiome protein-protein interactions based on the domain-domain binding from the MicrobioLink database [223].Host-microbiome interaction analysis uncovered four upregulated genes among the DEGs, such as CNKSR3, CSNK1A1, EPHB3, and IL1R2, interacting with multiple microbial protein domains Fig. 5b and four downregulated genes, SPTBN1, CFI, DST, and SYT1 interacting with two microbial domains Fig. 5f.Mutations in CC dataset from TCGA were obtained for the DEGs and visualized using MAFTools [224].The mutation frequency for the up-and downregulated genes is shown in Fig. 5c, d and Fig. 5g, h, respectively.The analysis revealed mutations in IDS, MAP4K4, COL4A2 and KHDRBS1 among the upregulated and, in SPTBN1 and DST among the downregulated genes Fig. 5i-n.
Protein expression analysis was performed for DEGs using Human Protein Atlas [225], revealing significant difference in the expression of IDS, MAP4K4, COL4A2, KHDRBS1 in CC and normal cervical epithelial tissues Fig. 6.IDS encoding iduronate 2-sulfatase (ID2S) was overexpressed in CC epithelial tissue compared to the normal tissue with an unknown mutation status.ID2S is primarily reported in mucopolysaccharidosis, and its association with the regulation of breast cancer metastasis has been reported [226,227].MAP4K4 is a threonine/serine kinase [228] that plays a significant role in a variety of physiological processes, including embryonic development, immunological response, inflammation, insulin sensitivity, and metabolic diseases [229][230][231][232]. Overexpression of MAP4K4 is associated with the activation of NF-κB and JNK signaling pathway and EMT induction in hepatocellular carcinoma (HCC) [233].Studies have also shown that SOX6 is a downstream target of MAPK4 that triggers autophagy in CC and it functions through the inhibition of the PI3K-AKT-mTOR pathway and activation of the MAPK/ERK pathway [234].MAP4K4 is reported as one of the potential metastatic genes in colorectal cancer metastasis [235] and CC [236].Downregulation of MAP4K4 in cancer cell lines inhibits cell proliferation and cell growth [237,238], apoptosis induction [239,240], and migration and invasion [241].
COL4A2 encodes type IV collagen and impacts cancer progression and pathogenesis.In HCC, COL4A1 overexpression promotes metastasis [242].COL4A1 overexpression is positively correlated with dendritic cells and macrophage expression in CC [242].Suppression of COL4A2 inhibits migration and proliferation in triple-negative breast cancer [243].Elevated expression of KHDRBS1/SAM68 detected in CC tissues with pelvic node metastasis and was not found in normal tissues [244].The RNA-binding protein (RBP), SRC associated in mitosis of 68 kDa (SAM68), is a member of signal transduction and activation of RNA (STAR) family.It contains a functional RNA-binding KH domain and binds to single-stranded nucleic acids.During mitosis, SAM68 is tyrosine-phosphorylated and associated with Src.The tyrosine phosphorylation mediates the interactions with SH2-and SH3-containing proteins.It participates in cell growth and differentiation [245].SAM68 is reported to play a dual role in cancer, i.e., tumor suppressor as well as an oncogene.Under the influence of extracellular stimuli, the level of SAM68 expression and phosphorylation influences tumor progression.For example, SAM68 phosphorylation inhibited cell proliferation induced by breast tumor kinase (BRK) [246].SAM68 expression is elevated in CC, and its cytoplasmic location is linked to LNM.EMT induction is mediated via AKT/ GSK-3/SNAIL when SAM68 is downregulated, indicating its oncogenic function in cervical cancer [244].Further, SAM68 expression levels were elevated in non-melanoma skin cancer tissue samples, and the increased expression was associated with enhanced growth and proliferation of A431 cells, while knocking down of SAM68 in malignant keratinocytes increases the sensitivity to DNA damage and decreases tumor burden [247].
SPTBN1 is a member of the β spectrin family and is important for the stability of lateral membrane.It also functions as a TGF-β adaptor protein and modulates Smad complex [248].SPTBN1 has been associated with several malignancies, and its expression varies according to the stage and type of cancer.SPTBN1 expression decreased significantly with the progression of epithelial ovarian cancer.The decreased expression resulted in cell proliferation, migration, and increased vimentin expression [249].It inhibited the progression of ovarian cancer via the JAK/STAT3 pathway.Its expression was much lower in more invasive ovarian cancer cell lines, suggesting that it may play a role in tumor aggressiveness [249].Further, the BAP31/SPTBN1 complex was found to regulate CC progression via the TGF-β/ SMAD signaling pathway under the regulation of miR-362 [250].In primary breast cancer tissues, the SPTBN1 expression was shown to be considerably downregulated as compared to normal, and it was identified to be a potent inhibitor of EMT and breast cancer progression [251].Similarly, SPTBN1 is downregulated in HCC and loss of its expression stimulates Wnt signaling, thereby promoting the acquisition of stem cell-like characteristics, and eventually, contributing to the growth of malignant tumors [252].
Further, analysis from two datasets (GSE6791 and GSE9750), various isoforms of kallikreins (KLKs) -KLK8, KLK7, KLK10) were found upregulated in metastasized CC.Kallikrein (KLK) proteins are a subgroup of serine proteases.KLKs play an important role in carcinogenesis by altering cell proliferation, EMT, compromising oxygen balance, and degrading ECM.They also allow tumor cell detachment by infiltrating through ECM barriers and promoting metastatic dissemination [253].Kallikrein-related peptidase 8 (KLK8) is one such protein that has been identified to be highly expressed in cervical, ovarian and endometrial cancers, indicating that it could be a viable target for therapy [254,255].KLK8 induced colorectal cancer development by triggering EMT [256].The wound healing and Transwell migration assays revealed that KLK8 promoted cell proliferation, invasion and metastasis [256].The upregulation was also associated with metastasis of head and neck SCC to cervical lymph nodes [257].Further, it was identified Fig. 3 Differential gene expression and pathway analysis a An UpSetR plot of DEGs between metastatic and non-metastatic CC tissues retrieved from GEO database (DEGs; with log-fold change > =|1.5| and p-value < = 0.05 cut-off) across five CC studies, b Protein-protein interaction network of the upregulated genes, c protein-protein interaction network of the downregulated genes, d top 10 hub genes upregulated in metastatic CC, e top 10 hub genes downregulated in metastatic CC.The hub genes are identified by the highest number of connections in the network.The color scale ranges from yellow (fewer interacting) to red (higher interacting) indicating the relative importance of the hub genes ◂ Fig. 4 Functional enrichment analysis between metastatic and nonmetastatic genes in CC a GO terms identification on the upregulated genes in molecular function, biological process and cellular compo-nent, b GO terms identification on the downregulated genes in molecular function, biological process and cellular component to be significantly downregulated in breast cancer tissues [258].Though the mechanisms of metastasis progression are not fully understood, it can degrade collagen type IV and fibronectin, both of which are required for cancer invasion and metastasis [259].
In CC tissues and serum, kallikrein 7 (KLK7) levels were reported to be higher, both of which are associated with tumor growth, invasion and metastasis [260,261].KLK7 was found to be regulated by histone modifications and promoter methylation.Trichostatin A (TSA) enhanced KLK7 expression in cervical and pancreatic cancer cell lines, and the transcription factor specificity protein 1 (SP1) supported the effect of TSA on KLK7 by transcriptional activation [262].KLK7 facilitates various processes involved in cancer, such as cell growth, proliferation, migration, angiogenesis via hydrolyzing cytokines, ECM, and membrane proteins [263].Overexpression of KLK7 in pancreatic cancer was associated with urokinase-type plasminogen activator receptor (uPAR), which inhibited cellular adherence to vitronectin [264].Further, in melanoma, KLK7 overexpression inhibited cell adhesion by lowering integrin expression and inducing spheroid formation by increasing MCAM/CD146 expression [190].The expression in oral SCC was similarly shown to increase as the malignancy progressed from stage I to stage IV [265].The brain cancer cells also showed high KLK7 expression, which significantly induced the invasion and metastasis [266].
KLK10 is another member of serine protease family closely related to glandular kallikreins [267].The gene was found to be downregulated in breast and prostate cancer cell lines, and tumor growth was inhibited in vivo upon its overexpression, indicating its tumor-suppressor function [268].Its expression is regulated via CpG hypermethylation in ovarian, breast, prostate cancer as well as in acute lymphoblastic leukemia [269,270].Further, overexpression of KLK10 reduced the proliferation and invasive potential of tongue cancer cells [271].Metastasis-associated protein 2 (MTA2) has been linked to metastasis and tumor growth.Downregulation of MTA2 prevented migration and invasion while increasing KLK10 expression via SP1 transcription factor in CC cell lines [272].KLK10, on the other hand, was found to be upregulated in colorectal cancer, the loss of Fig. 5 Pathway and network analysis a Top ten significant pathways enriched for pathway analysis through WikiPathways in upregulated genes, b host-Microbiome protein-domain interaction on upregulated genes, c mutation frequency of upregulated genes, d mutated genes which are upregulated in metastatic CC, e top ten significant pathways enriched for pathway analysis through WikiPathways in downregulated genes, f host-microbiome protein-domain interaction on downregulated genes, g Mutation frequency of downregulated genes, h mutated genes which are downregulated in metastatic CC i-n Needle plot for the mutations in IDS, MAP4K4, SPTBN1, COL4A2, KHDRBS1 and DST expression of which induced apoptosis, caspase 3 activity, and inhibited proliferation via the PI3K/AKT/mTOR signaling pathway [123].
Protocadherins (PCDH) belong to the non-classical subfamily of cadherins.These are highly expressed in brain and mediate cell to cell interactions [273].Several studies have reported that PCDH is abnormally expressed in numerous malignancies and show either a carcinogenic or anti-tumor effect [274], suggesting distinct family members with effects that are tumor dependent.PCDH7 was found to be overexpressed in NSCLC and reported to show oncogenic activity by hyperactivating MAPK pathway [275].PCDH7 expression was also shown to be upregulated in prostate cancer cells [276].Colony formation, migration, and invasion were all impeded by the loss of its expression.It may play a role in the activation of the AKT and ERK pathways during the advancement of prostate cancer [276].The expression was reduced in gastric cancer, particularly in cases of LNM.Cell migration, invasion, and E-cadherin expression were all considerably reduced because of PCDH7 downregulation [277].Furthermore, in CC, decreased expression of PCDH7 was linked to cancer cell metastasis, migration and invasion [278].The role of MAP4K4, SPTBN1, PCDH7, KLK7, KLK8, KLK10 and SAM68, which were identified to be differentially expressed among metastatic compared to non-metastatic CC in our bioinformatic analysis and their effect on the metastatic progression of various other type of cancer is summarized in Table 2.

Conclusion
The pathogenesis of CC is intricate and interconnected to the key determinants of molecular interactions and biological effects.HPV infection along with other risk factors is known to initiate the CC by recreating aberrant local microenvironment.Further, genetic and epigenetic modifications along with altered gene expression may favor the progression and metastasis of the cancer cells.The role of microbiota is significant in the modulation of microenvironment to facilitate the infection by HPV and also to promote metastasis of the tumor.Hence, understanding and elucidating the functional properties of the diverse species in the cervix will aid in deciphering the role of cervical microbiota (viral, fungal and bacterial) in the progression of CIN to CC and metastasis.Sentinel lymph nodes are the first nodes to receive lymphatic drainage from the primary lesion and promote LNM due to immune suppression by the tumor microenvironment.Further, lymphatic system can facilitate metastasis to distant organs, such as bone, lung, liver, and brain.Hence, identification of LNM and modulation of its microenvironment is required to prevent metastasis to distant organs and to provide effective treatment.Number of molecular pathways and their downstream functions are altered because of viral proteins, HPV integration, somatic mutation, copy number variation, epigenetic alteration, differential expression of the genes, miRNAs, lncRNAs and circRNAs, and interestingly, each of these factors is interconnected such that they favor incremental evolution, development and progression of CC.Implication of multi-omics approach to determine the genomic landscape can aid the stratification of CC patients and formulating effective treatment strategies.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made.The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material.If material is not included in the article's Creative Commons licence and your intended use is not Lymph node metastasis and cancer progression [286] permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.To view a copy of this licence, visit http:// creat iveco mmons.org/ licen ses/ by/4.0/.

Fig. 6
Fig. 6 Immunohistochemistry image representation from Human Protein Atlas for IDS, MAP4K4, COL4A2, KHDRBS1, showing significant upregulation in CC and downregulation in normal cervix tis-

Table 1
Major signaling pathways and genes reported in CC metastasis

Table 2
Evidence for the role of shortlisted differentially regulated genes on the metastatic progression of various cancer types