Circulating Long Non-Coding RNAs as Novel Potential Biomarkers for Osteogenic Sarcoma

Simple Summary Long non-coding RNAs (lncRNAs) can be detected in a liquid biopsy. We herein discussed the origin, methods of detection, measurement and potential functions of lncRNAs in blood. Furthermore, we used a systematic literature search to identify thirteen circulating lncRNAs whose expression was associated with bone tumor and we examined their impacts on clinical decision-making in the management of osteosarcoma. Abstract Circulating cell-free nucleic acids recently became attractive targets to develop non-invasive diagnostic tools for cancer detection. Along with DNA and mRNAs, transcripts lacking coding potential (non-coding RNAs, ncRNAs) directly involved in the process of tumor pathogenesis have been recently detected in liquid biopsies. Interestingly, circulating ncRNAs exhibit specific expression patterns associated with cancer and suggest their role as novel biomarkers. However, the potential of circulating long ncRNAs (c-lncRNAs) to be markers in osteosarcoma (OS) is still elusive. In this study we performed a systematic review to identify thirteen c-lncRNAs whose altered expression in blood associate with OS. We herein discuss the potential impact that these c-lncRNAs may have on clinical decision-making in the management of OS. Overall, we aimed to provide novel insights that can contribute to the development of future precision medicine in oncology.


Introduction
Osteosarcoma (OS) is a highly aggressive malignant bone tumor, frequently occurring in children and adolescents with an annual incidence of over three per million worldwide [1][2][3]. OS represents different pathological entities based on clinical, radiological, and histopathological features. For instance, based on histopathological features, osteosarcoma can be classified into distinct subtypes with the osteoblastic, chondroblastic, and fibroblastic OS, respectively, being the most common [4].
Nowadays, various clinical practices for OS have been notably implemented, including chemotherapy, radiotherapy, surgery, and targeted therapy; yet, the prognosis for OS still remains poor [5,6]. In fact, approximately 20% of patients showed clinical metastasis at presentation, with a 5-year survival rate less than 30% [7]. For this reason, OS strongly demands reliable, non-invasive, and clinically useful biomarkers.
The precise mechanism of lncRNAs release into the extracellular environment is not completely understood. Hypotheses have arisen that tumor cells, cancer-adjacent normal cells, immune cells, and other blood cells may all release lncRNAs [56,57], as shown in Figure 1. A few studies reported that lncRNAs can be encapsulated into membrane vesicles, such as exosomes or microvesicles (EV), prior to being secreted extracellularly. In such a conformation, the circulating lncRNAs have shown a higher degree stability, probably due to EVs offering protection against the nuclease-mediated degradation that may occur in the extracellular space and in body fluids [20,58,59] (Figure 1). On the other hand, other studies have suggested that the secretion pathway of lncRNAs may also occur in a similar manner to that for miRNAs. As such, lncRNAs might also be released into body fluids in an EV-independent fashion as complexes with high-density lipoproteins (HDLs) or protein Argonaute 2 (AGO2) [60] (Figure 1).

Origin of Circulating lncRNAs
The precise mechanism of lncRNAs release into the extracellular environment is not completely understood. Hypotheses have arisen that tumor cells, cancer-adjacent normal cells, immune cells, and other blood cells may all release lncRNAs [56,57], as shown in Figure 1. A few studies reported that lncRNAs can be encapsulated into membrane vesicles, such as exosomes or microvesicles (EV), prior to being secreted extracellularly. In such a conformation, the circulating lncRNAs have shown a higher degree stability, probably due to EVs offering protection against the nuclease-mediated degradation that may occur in the extracellular space and in body fluids [20,58,59] (Figure 1). On the other hand, other studies have suggested that the secretion pathway of lncRNAs may also occur in a similar manner to that for miRNAs. As such, lncRNAs might also be released into body fluids in an EV-independent fashion as complexes with high-density lipoproteins (HDLs) or protein Argonaute 2 (AGO2) [60] (Figure 1).
The hypothesis of an EV-independent mechanism for lncRNAs secretion might seem less likely given the high abundance of ribonucleases in serum, plasma, and other bodily fluids that can dramatically affect the stability of lncRNAs in the extracellular environment. However, one can speculate that circulating lncRNAs can be capable of resisting the RNase activity through modifications such as methylation, adenylation, and uridylation [61] or via the formation of higher order structures [62]. The origin of circulating lncRNAs. Two major sources of circulating lncRNAs have been postulated so far. LncRNAs can be encapsulated in extracellular vesicles (EV), predominantly exosomes. On the other hand, lncRNAs can also be released from live cells in an EV-independent fashion, thus, similar to circulating miRNAs, circulating lncRNAs might be detected in complexes with protein or high-density lipoproteins (HDL). The latter mechanism is likely to offer = circulating lncRNAs less protection against ribonucleases that are normally present in the extracellular space and body fluids. Created by Servier Medical Art (http://smart.servier.com/ (accessed on 15 May 2021)), licensed under Creative Commons Attribution 3.0 Unported License (https://creativecommons.org/licenses/by/3.0/ (accessed on 15 May 2021)).

Detection Methods of Circulating lncRNAs
Difference sources of liquid biopsy (i.e., whole blood, plasma, serum, urine, and gastric juice) can be used to quantify circulating lncRNAs. However, due to the possibility of blood cell RNA contamination, whole blood is the less recommended option so far [63]. The origin of circulating lncRNAs. Two major sources of circulating lncRNAs have been postulated so far. LncRNAs can be encapsulated in extracellular vesicles (EV), predominantly exosomes. On the other hand, lncRNAs can also be released from live cells in an EV-independent fashion, thus, similar to circulating miRNAs, circulating lncRNAs might be detected in complexes with protein or high-density lipoproteins (HDL). The latter mechanism is likely to offer = circulating lncRNAs less protection against ribonucleases that are normally present in the extracellular space and The hypothesis of an EV-independent mechanism for lncRNAs secretion might seem less likely given the high abundance of ribonucleases in serum, plasma, and other bodily fluids that can dramatically affect the stability of lncRNAs in the extracellular environment. However, one can speculate that circulating lncRNAs can be capable of resisting the RNase activity through modifications such as methylation, adenylation, and uridylation [61] or via the formation of higher order structures [62].

Detection Methods of Circulating lncRNAs
Difference sources of liquid biopsy (i.e., whole blood, plasma, serum, urine, and gastric juice) can be used to quantify circulating lncRNAs. However, due to the possibility of blood cell RNA contamination, whole blood is the less recommended option so far [63]. In addition, EDTA-anticoagulant collecting tubes have been suggested to be more suitable for the analysis of circulating lncRNAs [57]. Of note, some studies have found that lncRNAs remained stable in plasma even under multiple cycles of freeze-thaw, incubation at 45 • C, or storage at room temperature for as long as 24 h [56].
Overall, the methods to extract circulating lncRNAs can be divided into two major groups: guanidine/phenol/chloroform-based and column-based protocols. The columnbased method is currently considered more reliable, since organic and phenolic contaminants in TRIzol-based methods might invalidate results [64].
Regarding the measuring and normalization methods, some studies have suggested that the use of an equal volume of input RNA sample may be more accurate than an equal amounts of RNA measured using a NanoDrop spectrophotometer since many diseases, including cancer, may indeed release an higher degree of RNAs into body fluids than healthy control groups, leading to a significantly higher level of circulating RNA in cancer patients that causes misleading results [64].
To date, quantitative real-time PCR (qRT-PCR) is still considered the gold standard for quantitative expression analysis of lncRNAs, including circulating lncRNAs [65]. Microarrays and whole transcriptome analysis (RNA-seq) still have limited uses in this field. In fact, the high throughput potential of microarrays relies on a reference database of targets, which in the case of circulating lncRNAs, is still very limited [21]; the RNA-seq requires huge amounts of starting RNA samples. Additionally, RNA-seq is currently expensive and needs special equipment and/or expert bioinformaticians [64], whereas, the targetedapproach of qRT-PCR is still more accessible, and saves money and time. Accordingly, qRT-PCR can be divided into relative and absolute analyses. In relative quantification methods, the choice of endogenous controls is critical to properly normalize the expression levels. In this regard, it must be noted that no systematic evaluation of reference genes for serum lncRNA has yet been reported, posing some limitations for the relative qRT-PCR method in the analysis of lncRNAs from a liquid biopsy.

A Systematic Literature Search Identifies Thirteen Circulating lncRNAs with High Diagnostic Sensitivity in OS
To identify all the circulating lncRNAs whose expression has been reportedly associated with OS, either onset or progression, we ruled out a study in accordance with the preferred guidelines for reporting items for systematic reviews and meta-analyses (PRISMA) [66]. The study protocol of this systematic review was prospectively registered at the international prospective register of systematic review, PROSPERO (CRD42021250424). Briefly, a computerized literature search was performed in PubMed, Embase, and Scopus (last search: May 2021) using the terms "long non coding RNA" or "long untranslated RNA" or "lncRNA" and osteosarcoma and "liquid biopsy" or "serum" or "blood" or "plasma" AND "diagnostic" or "prognosis" or "prognostic" or "survival" or "metastasis". We further applied inclusion and exclusion criteria as described in the registered protocol.
To investigate the diagnostic and/or prognostic values of circulating lncRNAs, all the studies enrolled a number of patients and controls in a 1:1 ratio (62 ± 28 participants, mean ± SD), with a total of 873 OS patients who underwent liquid biopsy analysis for the detection of specific circulating lncRNAs, from 2015 to 2021. Altogether, at the time of OS diagnosis, an increase in ATB, EPEL, FAL1, FGD5-AS1, HNF1A-AS1, LINC01278, LINC01354, MALAT1, TUG1, UCA1 and a decline in HAND2-AS1 and NEF lncRNA expression levels were recorded in patient blood specimens with respect to controls. of circulating lncRNAs change along with disease status of OS patients (pre-operative and post-operative) (Cai 2017 [73], Ma 2015 [80], Wang 2017 [72]). Correlation between the abundance of circulating lncRNAs and survival rate was also measured in five studies    Table 1)). Of note, a positive correlation between the transcript levels measured in the bloodstream and OS tissue was found.   Briefly, all studies except that of Jiang (2020) [70] reported the diagnostic value of circulating lncRNAs, while only three studies investigated whether the expression level of circulating lncRNAs change along with disease status of OS patients (pre-operative and post-operative) ( Table 1)). Of note, a positive correlation between the transcript levels measured in the bloodstream and OS tissue was found.
Overall, we noticed that serum, either fresh or frozen, was the most common liquid biopsy to study circulating lncRNAs. Primarily, total RNA was extracted using TRIzol reagent and was analyzed using quantitative real-time polymerase chain reaction (RT-qPCR). The abundance of circulating lncRNAs was normalized with respect to either GAPDH (six studies) or β-actin (five studies) housekeeping transcript levels ( Table 1).
Different statistical methodologies were applied for assessing the relationship between the clinicopathological parameters of patients and the abundance of certain circulating lncRNAs. Overall, receiver-operating characteristic (ROC) curves were used to evaluate the performance of each lncRNA to discriminate OS patients from controls (reported as area under the curve (AUC), in Table 1). All 13 lncRNAs identified in liquid biopsy showed high diagnostic potential with the long intergenic non-coding RNA LINC01278 being the best performer (AUC = 0.945; 95% CI = 0.908-0.982, p value <0.001) (Zhang 2021 [69]). Furthermore, three studies reported the diagnostic power of circulating HNF1A-AS1 (Cai 2017 [73]), FAL1 (Wang 2017 [72]) and MALAT1 (Huo 2017 [77]) lncRNAs to be more effective than alkaline phosphatase (ALP) in distinguishing osteosarcoma from healthy individuals. Notably, Huo (2017 [77]) showed that combined detection of MALAT1 and alkaline phosphatase (ALP) significantly increased diagnostic sensitivity (Table 1).  The expression levels of HNF1A-AS1, TUG1, and FAL1 were the only ones to be monitored during the course of disease so far. Notably, both studies showed an augmentation of lncRNA expression to be associated with relapse.  [74]), and FGD5-AS1 (Song 2020 [79]) or a decline in LINC01278 (Zhang 2021 [69]) levels, respectively were found to be associated with a poor prognosis in OS.

Circulating lncRNAs Associating with OS Show High Degree of Heterogeneity
Genome browsers for research in comparative genomics, evolution, sequence variation, and transcriptional regulation were used to further identify additional information with respect to the thirteen retrieved lncRNAs. Results were extracted and are listed in Table 2.
Here, we found that the majority of the circulating lncRNAs that have been studied for their potential as biomarkers for OS, so far, are very large (above 2000 nucleotides (nt)) with a complex transcriptional organization that produces several different splicing variants (SVs). The longest lncRNA identified was MALAT1, of which the primary sequence is 8779 nt and produces three SVs, while HAND2-AS1 shows the most varied transcriptional regulation, being the gene with the largest number of exons in the group, producing eleven SVs (Table 2). In contrast, FAL1 is the shortest lncRNA (566 nt) consisting of only one SV. Notably, whether only specific SVs are sorted in the secretory pathway to reach the bloodstream is still unknown.
Although the molecular mechanism to produce circulating lncRNAs is still poorly characterized and their biological significance remains elusive, the secretion of circulating cytoplasmic lncRNAs (cc-lncRNAs) is thought to be similar to that underlying miRNA export, which is based on active secretion mediated by membrane-bound vesicles or through a vesicle-free RNA-binding protein dependent pathway [60,85,86]. Instead, circulating nuclear lncRNAs (cn-lncRNAs) might primarily originate from the passive leakage of dead cells. In this scenario, ATB, HNF1A-AS1, LINC01278, LINC01354, LINK-A and UCA1 could play active roles in cell-to-cell communication that might be relevant to disease progression and be worth future investigation.
The presence of a poly-adenylation (A) tail has been documented only for six of the above-mentioned lncRNAs. In fact, evidence shows that EPEL, HAND2-AS1, TUG1, UCA1 carry sequence motifs recognized by the RNA cleavage complex, while ATB and MALAT1 do not present any typical signal [82,[87][88][89].
Finally, three major classes of lncRNAs have been identified in OS patient blood, including four antisense (EPEL, FGD5-AS1, HAND2-AS1, HNF1A-AS1), three competitive endogenous RNAs (ceRNAs) (ATB, LINC01278, LINK-A) and three scaffolds (FAL1, LINC01354, MALAT1). Furthermore, studies have shown that NEF and TUG1 can serve both as a scaffold and ceRNAs [90][91][92][93][94] (Table 2). However, whether similar modes of action are retained by circulating lncRNAs upon contact with any recipient cells is yet unknown. In fact, studies have primarily focused on the potential role of circulating lncRNAs as biomarkers of human diseases, irrespective of their contribution to the pathology. Investigations on this matter would extend our understanding of the biological significance of circulating lncRNAs in OS and other human diseases.

The lncRNAs-miRNAs Crosstalk Is Critical for the Biological Activities of Osteosarcoma Cells
Regardless the lncRNA's presence in the bloodstream, which is relevant to address their potential as markers of OS, evidence of how the thirteen retrieved lncRNAs are involved in OS were also researched and is listed in Table 2. This information might be helpful to further address hypotheses regarding the roles of circulating lncRNAs in the oncogenesis of bone. Despite all the lncRNAs identified herein being previously associated with several other cancer types (Table 2), the mode of how they behave in OS to drive or contribute to pathogenesis remains poorly characterized.
When the thirteen lncRNAs were pooled together and qualitatively analyzed, a lncRNA-miRNAs regulatory axis emerged as the most prominent network associating with the OS pathogenic mechanism, with ATB, FGD5, HNF1A-AS1, LINC01278, MALAT1, NEF, TUG1 and UCA1 lncRNAs being reported for their ability to sponge miRNAs that affect a plethora of cellular targets critical in malignancy (Table 2). Interestingly, according to the available data, the regulation of gene expression by a competitive endogenous RNA (ceRNA) mechanism is indeed emerging as a leading lncRNA function in OS, as well as in a remarkable number of other types of cancer (Table 2) (reviewed in [140,141]). For instance, MALAT1 regulates osteosarcoma progression and facilitates lung metastasis by targeting several miRNA families [142][143][144], promotes thyroid cancer progression by regulating miR-204 [145], and leads to chemoresistance in hepatocellular carcinoma by sponging miR-140-5p [146]. A total of 17 miRNAs have been identified to be downregulated by MALAT1 in a fashion that is critical to drive OS, such as miR-202 [144], miR-206 [142], and miR-26a-5p [147] to mention a few.
Moreover, the ability of EPEL, FAL1, HAND2-AS1, HNF1A-AS1, LINC01354, LINK-A, MALAT1 and UCA1 lncRNAs to modulate the abundance of transcriptional factors, structural proteins, or to affect key cellular pathways, such as Wnt/β-catenin and PTEN/AKT, have been also described [68,[70][71][72]78,115,148,149]. In this regard, Wnt signaling emerged as most regulated by the retrieved circulating lncRNAs. However, whether the regulation of such targets is similarly driven by lncRNAs-miRNAs crosstalk or involves a more complex network is still unknown. Nonetheless, due to the large flexibility to form a variety of different complexes, either with proteins or nucleic acids, it is not surprising that a lncRNA can play different roles at once. For instance, in vitro and in vivo experiments highlighted MALAT1 and UCA1's capabilities to either regulate miRNA expression or scaffold transcriptional regulators in bone oncogenesis [78,[150][151][152].

Concluding Remarks
The role of lncRNAs in OS tumor development has only recently been investigated, yet several studies have shown that the deregulation of a number of lncRNAs influence the occurrence and progression of osteosarcoma, as reviewed in [153].
Many of these lncRNAs proved to have a detectable expression levels in either serum or plasma samples, making them promising biomarker candidates for non-invasive diagnostics. However, the clinical application of the so-called circulating lncRNAs in OS remain elusive. We herein systematically searched, summarized, and discussed all the studies to show the relationship between circulating lncRNA expression levels and OS that can be helpful to address future intervention of circulating lncRNAs in OS management.
The expression levels of thirteen circulating lncRNAs consistently correlated with those measured in OS tissues and have a high potential diagnostic value. In particular, dysregulation of seven lncRNAs (EPEL, FGD5-AS1, FAL1, HNF1A-AS1, LINC01278, MALAT1 and TUG1) that can be detected in OS blood also coincide with the clinical stage of the disease, metastatic progression or survival, and, above all, with therapeutic response.
Nonetheless, there is still a long way to go to adopt circulating lncRNA in OS clinical practice. In fact, all circulating lncRNAs in OS have only been reported in a single study with the exception of TUG1, therefore there is a need for systematic validation studies that investigate multiple lncRNAs with well-characterized and diverse patient samples. In other terms, the research cohort size should be bigger and selection bias should be reduced as much as possible.
This study also highlights a few other critical points that future investigations should consider to support the exploitation of circulating lncRNAs in the management of OS. For instance, blood preparation and endogenous controls in qRT-PCR analysis of circulating lncRNAs still require a standardization methodology. The choice of anticoagulant, the volume required for sample collection, and the temperature for storing the samples need to be more uniform to keep the analysis among different groups consistent. Controversial results might arise through the use of different quantitative standards in qRT-PCR. In this regard, a recent study identified the lncRNA RP11-204K16.1, XLOC_012542, and U6 small nuclear RNA as the most stable reference genes for circulating lncRNA analysis in serum for cervical cancer patients [154]. Similarly, a set of these reference genes should be further identified for the quantitative analysis of circulating lncRNAs in OS.
In summary, this study acknowledges both the pros and cons of the use of circulating lncRNAs as biomarkers for OS. The low degree of invasiveness, affordability and time saving procedures make a few circulating lncRNAs, whose expression coincide with the clinical stage of the disease, promising novel biomarkers to add to current clinical practices for the management of OS. However, the lack of standard methodologies and current small sample size still pose a high risk of bias and strongly limit their use.
Several studies have discovered that lncRNAs play critical regulatory roles in the formation of micrometastases through modulating specific signaling pathways in cancer cells [155,156]. Furthermore, the early detection of abnormal expression levels of several serum lncRNAs was linked to the late onset of metastases [157]. As a result, repeated serum lncRNA samples may aid in the detection of micrometastasis, which is only partially detectable using traditional diagnostic approaches. Therefore, along with their application as OS biomarkers, circulating lncRNAs might also be novel candidate targets. In fact, the presence of lncRNAs originating from tumor tissues in the bloodstream strongly suggests a role in cell-to-cell communication that might be relevant to oncogenesis. However, limited studies have been done in this field so far, which mainly point out a role in angiogenesis promotion [158] or in the modulation of how the surrounding cells respond to circulating miRNAs [159].
Overall  results might arise through the use of different quantitative standards in qRT-PCR. In this regard, a recent study identified the lncRNA RP11-204K16.1, XLOC_012542, and U6 small nuclear RNA as the most stable reference genes for circulating lncRNA analysis in serum for cervical cancer patients [154]. Similarly, a set of these reference genes should be further identified for the quantitative analysis of circulating lncRNAs in OS.
In summary, this study acknowledges both the pros and cons of the use of circulating lncRNAs as biomarkers for OS. The low degree of invasiveness, affordability and time saving procedures make a few circulating lncRNAs, whose expression coincide with the clinical stage of the disease, promising novel biomarkers to add to current clinical practices for the management of OS. However, the lack of standard methodologies and current small sample size still pose a high risk of bias and strongly limit their use.
Several studies have discovered that lncRNAs play critical regulatory roles in the formation of micrometastases through modulating specific signaling pathways in cancer cells [155,156]. Furthermore, the early detection of abnormal expression levels of several serum lncRNAs was linked to the late onset of metastases [157]. As a result, repeated serum lncRNA samples may aid in the detection of micrometastasis, which is only partially detectable using traditional diagnostic approaches. Therefore, along with their application as OS biomarkers, circulating lncRNAs might also be novel candidate targets. In fact, the presence of lncRNAs originating from tumor tissues in the bloodstream strongly suggests a role in cell-to-cell communication that might be relevant to oncogenesis. However, limited studies have been done in this field so far, which mainly point out a role in angiogenesis promotion [158] or in the modulation of how the surrounding cells respond to circulating miRNAs [159].
Overall, the functions of circulating lncRNAs are still unknown. Understanding the mechanisms to regulate the expression levels of circulating lncRNAs might provide new clues on the oncogenesis of OS and new tools in translational medicine

012/
results might arise through the use of different quantitative standards in qRT-PCR. In this regard, a recent study identified the lncRNA RP11-204K16.1, XLOC_012542, and U6 small nuclear RNA as the most stable reference genes for circulating lncRNA analysis in serum for cervical cancer patients [154]. Similarly, a set of these reference genes should be further identified for the quantitative analysis of circulating lncRNAs in OS.
In summary, this study acknowledges both the pros and cons of the use of circulating lncRNAs as biomarkers for OS. The low degree of invasiveness, affordability and time saving procedures make a few circulating lncRNAs, whose expression coincide with the clinical stage of the disease, promising novel biomarkers to add to current clinical practices for the management of OS. However, the lack of standard methodologies and current small sample size still pose a high risk of bias and strongly limit their use.
Several studies have discovered that lncRNAs play critical regulatory roles in the formation of micrometastases through modulating specific signaling pathways in cancer cells [155,156]. Furthermore, the early detection of abnormal expression levels of several serum lncRNAs was linked to the late onset of metastases [157]. As a result, repeated serum lncRNA samples may aid in the detection of micrometastasis, which is only partially detectable using traditional diagnostic approaches. Therefore, along with their application as OS biomarkers, circulating lncRNAs might also be novel candidate targets. In fact, the presence of lncRNAs originating from tumor tissues in the bloodstream strongly suggests a role in cell-to-cell communication that might be relevant to oncogenesis. However, limited studies have been done in this field so far, which mainly point out a role in angiogenesis promotion [158] or in the modulation of how the surrounding cells respond to circulating miRNAs [159].
Overall, the functions of circulating lncRNAs are still unknown. Understanding the mechanisms to regulate the expression levels of circulating lncRNAs might provide new clues on the oncogenesis of OS and new tools in translational medicine   The study was conducted according to the guidelines of the international prospective register of systematic review PROSPERO (registration number: CRD42021250424).

Data Availability Statement:
No new data were created in this study. The data presented in this study are openly available in PubMed as listed in the references.

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