Advertisement

Research Techniques Made Simple: Studying Circular RNA in Skin Diseases

  • Rong Yang
    Affiliations
    Department of Dermatology, The University of Texas Southwestern Medical Center, Dallas, Texas, USA
    Search for articles by this author
  • Richard C. Wang
    Correspondence
    Correspondence: Richard C. Wang, Department of Dermatology, Harold C. Simmons Comprehensive Cancer Center, The University of Texas Southwestern Medical Center, Dallas, Texas, USA.
    Affiliations
    Department of Dermatology, The University of Texas Southwestern Medical Center, Dallas, Texas, USA

    Harold C. Simmons Comprehensive Cancer Center, The University of Texas Southwestern Medical Center, Dallas, Texas, USA
    Search for articles by this author
      Circular RNAs (circRNAs) are a unique class of covalently closed, single-stranded RNAs. High-throughput sequencing has uncovered the abundance and complexity of circRNAs. Changes in levels of circRNAs correlate with diverse disease states, including many skin diseases. CircRNAs can function as microRNA inhibitors, protein interactors, or mRNAs. Although circRNAs do have unique topological features, they share many similarities, including primary sequence, with their linear orthologs, so carefully controlled experiments are required to detect and study them. Here, we summarize some protocols used in the identification, validation, and characterization of circRNAs. We also discuss ways to repress and overexpress specific circRNAs to assess potential unique functions for these molecules. These techniques may be useful in exploring how circRNAs contribute to skin disease.

      Abbreviations:

      BSJ (backsplice junction), circALTO (circular MCV-T/ALTO), circE7 (circular E7), circRNA (circular RNA), HPV (human papillomavirus), IP (immunoprecipitation), m6A (N6-methyladenosine), MCC (Merkel cell carcinoma), MCPyV (Merkel cell polyomavirus), miRNA (microRNA), QRT-PCR (quantitative real-time reverse transcriptase–PCR), RNA-seq (RNA sequencing), RNase (ribonuclease), rRNA (ribosomal RNA), SCC (squamous cell carcinoma), shRNA (short hairpin RNA), siRNA (small interfering RNA)

      Introduction

      Unlike linear RNA, circular RNA (circRNA) is covalently closed and single-stranded RNA (Table 1). CircRNAs are derived from precursor RNAs, typically pre-mRNAs (
      • Kristensen L.S.
      • Andersen M.S.
      • Stagsted L.V.W.
      • Ebbesen K.K.
      • Hansen T.B.
      • Kjems J.
      The biogenesis, biology and characterization of circular RNAs.
      ). Precursor RNAs undergo not only canonical splicing but also backsplicing, in which a downstream 5′ splice site is covalently linked to an upstream 3′ splice site, usually with lower efficiency (
      • Li X.
      • Yang L.
      • Chen L.L.
      The biogenesis, functions, and challenges of circular RNAs.
      ). Canonical splicing generates linear RNAs, whereas backsplicing produces circRNAs and alternatively spliced linear RNAs (Figure 1). Exonic circRNAs, derived from one or more exons, are the most abundant and studied circRNAs and comprise over 80% of identified circRNAs. Additional classes of circRNA molecules derived from precursor RNAs include exon-intron circRNAs, circular intronic RNAs, and circularized transfer RNA introns. CircRNAs were first thought to be rare byproducts of erroneous splicing events that lacked physiological relevance. Improvements in high-throughput sequencing and sequence alignment have revealed tens of thousands of circRNAs. Moreover, detailed evaluation of some circRNAs have revealed that circRNAs can possess unique biological properties and functions.
      Table 1Differences between CircRNA and Linear RNA
      RNABiogenesisTranscript LevelStructure5′ and 3′
      circRNABacksplicingLowCovalently closed circleNo 5′ tail or 3′ polyA
      Linear mRNACanonical splicingHighLinear5′ tail and 3′ polyA
      Abbreviation: circRNA, circular RNA.
      Figure thumbnail gr1
      Figure 1The biogenesis of circRNAs. CircRNAs are primarily derived from pre-mRNAs. Pre-mRNA can undergo canonical splicing to produce linear RNA and an intron lariat, a loop with a 2′–5′ phosphodiester linkage at the branch site. Alternatively, flanking inverted repeat elements or RBP sites may facilitate backsplicing events to generate exonic circRNA or EIciRNA. Although intronic lariats are typically degraded, ciRNAs have been shown to have greater stability. circRNA, circular RNA; ciRNA, circular intronic RNA; EIciRNA, exon-intron circular RNA; m7G, 7-methylguanosine; RBP, RNA-binding protein.
      CircRNAs have been identified in many different kinds of skin diseases (Table 2). In infectious diseases, a number of DNA viruses relevant to skin diseases, including Epstein-Barr virus, Kaposi’s sarcoma herpesvirus, human papillomaviruses (HPVs), and human polyomavirus all generate circRNAs (
      • Tagawa T.
      • Kopardé V.N.
      • Ziegelbauer J.M.
      Identifying and characterizing virus-encoded circular RNAs [e-pub ahead of print].
      ). Many circRNA studies have focused on circRNAs enriched in cancer, and several studies have identified circRNAs of potential relevance in cutaneous squamous cell carcinomas (SCCs) (
      • Sand M.
      • Bechara F.G.
      • Gambichler T.
      • Sand D.
      • Bromba M.
      • Hahn S.A.
      • et al.
      Circular RNA expression in cutaneous squamous cell carcinoma.
      ), melanoma (
      • Tang K.
      • Zhang H.
      • Li Y.
      • Sun Q.
      • Jin H.
      Circular RNA as a potential biomarker for melanoma: a systematic review.
      ), Merkel cell carcinoma (MCC) (
      • Abere B.
      • Zhou H.
      • Li J.
      • Cao S.
      • Toptan T.
      • Grundhoff A.
      • et al.
      Merkel cell polyomavirus encodes circular RNAs (circRNAs) enabling a dynamic circRNA/microRNA/mRNA regulatory network.
      ;
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      ), and HPV-driven SCCs (
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      ). Differentially expressed circRNAs have been identified in several inflammatory skin diseases, including acne (
      • Liang J.
      • Wu X.
      • Sun S.
      • Chen P.
      • Liang X.
      • Wang J.
      • et al.
      Circular RNA expression profile analysis of severe acne by RNA-Seq and bioinformatics.
      ), atopic dermatitis (
      • Moldovan L.I.
      • Tsoi L.C.
      • Ranjitha U.
      • Hager H.
      • Weidinger S.
      • Gudjonsson J.E.
      • et al.
      Characterization of circular RNA transcriptomes in psoriasis and atopic dermatitis reveals disease-specific expression profiles.
      ), psoriasis (
      • Qiao M.
      • Ding J.
      • Yan J.
      • Li R.
      • Jiao J.
      • Sun Q.
      Circular RNA expression profile and analysis of their potential function in psoriasis.
      ), and systemic lupus erythematosus (
      • Wang X.
      • Ma R.
      • Shi W.
      • Wu Z.
      • Shi Y.
      Emerging roles of circular RNAs in systemic lupus erythematosus.
      ). Finally, marked differences in the circRNA expression profile of chronologically aged skin have been noted (
      • Wang L.
      • Si X.
      • Chen S.
      • Wang X.
      • Yang D.
      • Yang H.
      • et al.
      A comprehensive evaluation of skin aging-related circular RNA expression profiles.
      ). Despite the increasing interest in circRNAs relevant to skin disease, the biological functions of most circRNAs have not been explored, and their potential utility as disease markers has not been examined.
      Table 2Skin Disease–Related CircRNAs
      Skin Disease (circRNA Origin)circRNAPutative FunctionsReferences
      Infectious (viral)EBVcircBART, circRPMS1, circLMP2, circBHLF1(
      • Tagawa T.
      • Kopardé V.N.
      • Ziegelbauer J.M.
      Identifying and characterizing virus-encoded circular RNAs [e-pub ahead of print].
      )
      KSHVcircvIRF4, circPANs(
      • Tagawa T.
      • Kopardé V.N.
      • Ziegelbauer J.M.
      Identifying and characterizing virus-encoded circular RNAs [e-pub ahead of print].
      )
      HPVcircE7Translated to E7; transforming oncoprotein(
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      )
      HPyVMCPyV circMCV-T/circALTO,

      TSPyV circALTO
      Transcriptional enhancer or miRNA inhibitor(
      • Abere B.
      • Zhou H.
      • Li J.
      • Cao S.
      • Toptan T.
      • Grundhoff A.
      • et al.
      Merkel cell polyomavirus encodes circular RNAs (circRNAs) enabling a dynamic circRNA/microRNA/mRNA regulatory network.
      ;
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      )
      Neoplastic (viral/human)SCCcircE7; multipleTransforming oncoprotein(
      • Sand M.
      • Bechara F.G.
      • Gambichler T.
      • Sand D.
      • Bromba M.
      • Hahn S.A.
      • et al.
      Circular RNA expression in cutaneous squamous cell carcinoma.
      ;
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      )
      MCCcircMCV-T/circALTOTranscriptional enhancer or miRNA inhibitor(
      • Abere B.
      • Zhou H.
      • Li J.
      • Cao S.
      • Toptan T.
      • Grundhoff A.
      • et al.
      Merkel cell polyomavirus encodes circular RNAs (circRNAs) enabling a dynamic circRNA/microRNA/mRNA regulatory network.
      ;
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      )
      Neoplastic (human)MelanomaciRS-7 (CDR1AS),

      Multiple
      Inhibits melanoma progression, potential biomarker(
      • Hanniford D.
      • Ulloa-Morales A.
      • Karz A.
      • Berzoti-Coelho M.G.
      • Moubarak R.S.
      • Sánchez-Sendra B.
      • et al.
      Epigenetic silencing of CDR1as drives IGF2BP3-mediated melanoma invasion and metastasis.
      ;
      • Tang K.
      • Zhang H.
      • Li Y.
      • Sun Q.
      • Jin H.
      Circular RNA as a potential biomarker for melanoma: a systematic review.
      )
      Inflammatory (human)PsoriasisciRS-7, circZRANB1,

      multiple
      Potential biomarker(
      • Moldovan L.I.
      • Hansen T.B.
      • Venø M.T.
      • Okholm T.L.H.
      • Andersen T.L.
      • Hager H.
      • et al.
      High-throughput RNA sequencing from paired lesional- and non-lesional skin reveals major alterations in the psoriasis circRNAome.
      ;
      • Qiao M.
      • Ding J.
      • Yan J.
      • Li R.
      • Jiao J.
      • Sun Q.
      Circular RNA expression profile and analysis of their potential function in psoriasis.
      )
      ADMultiplePotential biomarker(
      • Moldovan L.I.
      • Tsoi L.C.
      • Ranjitha U.
      • Hager H.
      • Weidinger S.
      • Gudjonsson J.E.
      • et al.
      Characterization of circular RNA transcriptomes in psoriasis and atopic dermatitis reveals disease-specific expression profiles.
      )
      LupusHsa_circ_0044235,

      Hsa_circ_0068367
      Potential biomarker(
      • Luo Q.
      • Zhang L.
      • Li X.
      • Fu B.
      • Guo Y.
      • Huang Z.
      • et al.
      Identification of circular RNAs hsa_circ_0044235 and hsa_circ_0068367 as novel biomarkers for systemic lupus erythematosus.
      )
      Severe acnecircRNA_0084927,

      multiple
      Potential biomarker(
      • Liang J.
      • Wu X.
      • Sun S.
      • Chen P.
      • Liang X.
      • Wang J.
      • et al.
      Circular RNA expression profile analysis of severe acne by RNA-Seq and bioinformatics.
      )
      Aging (human)Chronological aginghsa_circ_0137613,

      multiple
      Potential biomarker(
      • Wang L.
      • Si X.
      • Chen S.
      • Wang X.
      • Yang D.
      • Yang H.
      • et al.
      A comprehensive evaluation of skin aging-related circular RNA expression profiles.
      )
      Abbreviations: AD, atopic dermatitis; circALTO, circular ALTO; circBART, circular BART; circBHLF1, circular BHLF1; circE7, circular E7; circLMP2, circular LMP2; circMCV-T, circular MCV-T; circPANs, circular PAN; circRNA, circular RNA; cricRPMS1, circular RPMS1; circvIRF4, circular vIRF4; EBV, Epstein-Barr virus; HPV, human papillomavirus; HPyV, human polyomavirus; KSHV, Kaposi’s sarcoma herpesvirus; MCC, Merkel cell carcinoma; MCPyV, Merkel cell polyomavirus; miRNA, microRNA; SCC, squamous cell carcinoma; TSPyV, Trichodysplasia Spinulosa Polyomavirus.
      We describe methods that may be used to identify, validate, and characterize circRNAs, including those involved in skin disease (Figure 2). First, we describe bioinformatic approaches to identify circRNAs from RNA sequencing (RNA-seq) with or without ribosomal RNA (rRNA) depletion to enrich for circRNAs. Second, we describe experiments necessary to validate potential circRNA identified in silico. Third, we describe protocols that may be useful in characterizing the properties of novel circRNAs. Finally, strategies used to assess the biological functions of circRNA through overexpression or repression of specific circRNAs are discussed (Figure 2 and Table 3). These techniques should be broadly applicable to the increasing number of circRNAs discovered in both normal skin and skin diseases.

      Summary Points

      • Circular RNAs (circRNAs) are covalently closed single-stranded RNAs, which are produced by noncanonical backsplicing and have different properties from linear RNAs.
      • CircRNAs can be identified through the identification of backsplice junctions in RNA sequencing data and validated through northern blotting and RT-PCR.
      • Assessment of the N6-methyladenosine modification status; half-life; association with polyribosomes, also known as polysomes; and subcellular localization for specific circRNAs are critical for the characterization of circRNAs and may yield insights into function.
      • Rigorously controlled loss-of-function (small interfering RNA/short hairpin RNA, CRISPR/Cas) and overexpression experiments (plasmid based) are required to determine the biological functions of circRNAs.

      Limitations

      • Because most circRNAs are typically expressed at much lower copy numbers than their cognate linear RNA in cells, results should be interpreted cautiously to avoid interference from cognate linear RNAs.
      • When researching the function, it is hard to tell the function is from circRNA or its linear isoform.
      Figure thumbnail gr2
      Figure 2Schematic representation of approaches to studying circRNAs. CircRNA studies may include the identification, validation, characterization, and functional analysis of circRNAs. After identification through bioinformatic prediction, circRNAs can be validated using techniques such as (a) RT-PCR with primers specific to linear or circRNAs and (b) northern blot with specific probes. (c) Cell fractionation assays, (d) FISH, (e) assessment of circRNA half-life, (f) m6A IP, and (g) polysome (polyribosome) profiling are some of the techniques that are used to characterize specific circRNAs and help to yield possible insights on function. Illustrations of biochemical fractionation and FISH methodologies show examples of circRNAs, which are frequently cytoplasmic, although nuclear circRNAs have also been described. (h) CircRNA minigene constructs allow for the expression and mutagenesis of abundant amounts of specific circRNAs to facilitate functional analyses. (i) By targeting BSJ sites, circRNAs can be specifically knocked down by RNA interference (siRNA/shRNA). (j) If specific flanking intronic regions are important for the biogenesis of circRNA, CRISPR/Cas9–mediated engineering can be used to generate a KO of a specific circRNA. Careful controls to exclude off-target effects of RNA interference and KO approaches should be included. BSJ, backsplice junction; circRNA, circular RNA; gRNA, guide RNA; IP, immunoprecipitation; KO, knockout; m6A N6-methyladenosine; QRT-PCR, quantitative real-time reverse transcriptase–PCR; RNase, ribonuclease; rRNA, ribosomal RNA; shRNA, short hairpin RNA; siRNA, small interfering RNA.
      Table 3Key Considerations in Performing circRNA Studies
      Study PhaseTechniqueAdvantagesDisadvantages
      circRNA enrichmentRNase R treatmentEnrich for circRNAsMay degrade low abundance circRNA; does not degrade intron lariats or structured RNAs
      IdentificationRNA-seqIdentify novel circRNALinear RNA also can be sequenced combined with circRNA
      ValidationNorthern blotNo reverse transcription and amplification stepTime consuming, decrease sensitivity
      RT-PCR/QRT-PCRHigh sensitivityMay result in RT or amplification artifacts
      CharacterizationFractionationLocalization may yield clues on functionInadequate isolation, hard to isolate nucleus from cytoplasm
      FISHVisualization of circRNA localizationInterference from cognate linear RNA
      Half-life analysesInformation on differences between linear and circRNA stabilityNo information on circRNA function
      m6A IPDetermine the modification statusNonspecific RNA can be precipitated
      Polysome analysesProvide evidence for translation abilityProvide indirect evidence and require additional experimentation
      FunctionOverexpression with plasmidsFunctions more easily identified after overexpressionTrans-spliced and rolling circle artifacts may be generated
      siRNA/shRNAFunctions may be identified with specific knockdownLimited sequences to target; off-target effects to linear RNA possible
      CRISPR/Cas9Knockout the expression of circRNALinear splicing may be impacted; circRNA-specific deletion may not be possible
      Abbreviations: circRNA, circular RNA; m6A IP, N6-methyladenosine immunoprecipitation; QRT-PCR, quantitative real-time reverse transcriptase–PCR; RNase, ribonuclease; RNA-seq, RNA sequencing; shRNA, short hairpin RNA; siRNA, small interfering RNA.

      Bioinformatics approaches for circRNA detection

      To discover novel circRNAs, sequencing of rRNA-depleted total RNA is commonly used to identify expressed circRNAs (
      • Kristensen L.S.
      • Andersen M.S.
      • Stagsted L.V.W.
      • Ebbesen K.K.
      • Hansen T.B.
      • Kjems J.
      The biogenesis, biology and characterization of circular RNAs.
      ). Although it is possible to identify circRNAs from rRNA-depleted total RNA, treatment of samples with ribonuclease (RNase) R, a 3′-to-5′ exoribonuclease that digests linear RNAs but not structured or circRNAs, enriches for the detection circRNAs. However, pretreatment with RNase R may deplete some lower abundance circRNAs and limits the ability to quantitate circRNA levels compared with cognate linear RNAs. CircRNAs can be identified by using algorithms that align reads that specifically map to backsplice junctions (BSJs), rather than canonical splice junctions. CircRNA prediction algorithms differ significantly in their sensitivity and specificity. Indeed, marked differences between several circRNA identification pipelines have been reported using the same RNA-seq datasets (
      • Hansen T.B.
      • Venø M.T.
      • Damgaard C.K.
      • Kjems J.
      Comparison of circular RNA prediction tools.
      ). CIRI2, CIRCexplorer, and Mapsplice all show high accuracy and good sensitivity in the identification of bona fide circRNAs confirmed by RNase R treatment. Although most algorithms have been developed for detecting circRNA from genomic sequences, pipelines designed for the specific identification of viral circRNAs have also been developed (
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      ). Bioinformatic pipelines are still likely to yield many false positives, so several criteria can be utilized to help eliminate potential false positives. Bona fide circRNAs are frequently abundant, detected in multiple independent datasets, and conserved across multiple species. For example, bioinformatics analyses revealed that the HPV16-derived circular E7 (circE7) was highly abundant, with levels comparable to linear splicing events; identified in multiple RNA-seq datasets; and conserved across multiple high-risk HPV types (
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      ). Similarly, circular MCV-T/ALTO (circALTO), a Merkel cell polyomavirus (MCPyV)–derived circRNA, could be identified as conserved in multiple polyomavirus species (
      • Abere B.
      • Zhou H.
      • Li J.
      • Cao S.
      • Toptan T.
      • Grundhoff A.
      • et al.
      Merkel cell polyomavirus encodes circular RNAs (circRNAs) enabling a dynamic circRNA/microRNA/mRNA regulatory network.
      ;
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      ).

      Experimental validation of predicated circRNAs

      After the identification of putative circRNAs through bioinformatics approaches, validation of specific circRNAs via wet bench techniques is essential, because false BSJs are identified by most bioinformatic pipelines. The most commonly used molecular methods include RT-PCR with divergent primers and northern blotting (Figure 2a and b). One critical difference from standard protocols is the pretreatment of total RNA with RNase R, typically in the presence of an RNase A/B/C inhibitor. True circRNAs should be relatively resistant to digestion by RNase R, whereas the abundance of linear RNAs should be markedly decreased by RNase R treatment.
      Because circRNAs lack polyadenylated (polyA) 3′ tails, reverse transcription must be completed using random hexamer priming with or without rRNA depletion. To detect circRNAs, primers flanking the BSJ of the putative circRNA should be designed (Figure 2a). These divergent primers should not yield a product on the cognate linear splicing products of the parental gene (Figure 2a). Conventional, inward-facing primers for the predicted cognate linear RNAs should be designed as a control both for circRNA quantification and RNase R treatment. Ultimately, the product from the divergent PCR reaction should be sequenced to confirm the expected BSJ.
      Northern blots are not subject to artifacts generated by reverse transcription or PCR amplification and may be an orthogonal, and possibly more accurate, method to validate the presence and abundance of circRNAs. Probes specifically designed to bind only the BSJ site may be used to confirm the presence of the circRNA but will not allow for the assessment of the relative abundance of the circRNA relative to total RNA generated from the gene of interest. Probes that detect sequences shared by both circular and linear RNAs may be used to detect circRNAs if used in conjunction with RNase R treatment (Figure 2b). Limitations of northern blotting include a requirement for large amounts of starting total RNA and relative low sensitivity compared with amplification-based methods. For circALTO and circE7, both RT-PCR and northern blotting approaches confirmed the presence of putative circRNAs, allowing for robust validation of the bioinformatic predictions (
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      ;
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      ).

      Characterization of circRNAs by subcellular localization, half-life analysis, N6-methyladenosine immunoprecipitation, and polyribosome profiling

      The subcellular localization of an RNA is an important component of understanding its potential functions. Biochemical fractionation of the cell is one method commonly used to determine whether RNAs are cytoplasmic or nuclear (Figure 2c). Cytoplasmic and nuclear fractions can be separated through sequential solubilization steps with RNA prepared separately from each fraction. The efficiency of fractionation can be assessed using cytoplasmic and nuclear RNA controls (e.g., 18S and MALAT1, respectively). The levels of RNA can be assessed by quantitative real-time reverse transcriptase–PCR (QRT-PCR) and may be expressed by normalizing to the enriched fraction. RNA FISH may also be used to determine the subcellular localization patterns (Figure 2d). FISH is performed using probes spanning the circRNA BSJ. Specimens may be first treated with RNase R to decrease competing signals from linear RNAs. Because many circRNAs have also been reported to be enriched in extracellular vesicles, specific protocols may be used to isolate and identify circRNAs in these structures of emerging importance (
      • McBride J.D.
      • Rodriguez-Menocal L.
      • Badiavas E.V.
      Extracellular vesicles as biomarkers and therapeutics in dermatology: a focus on exosomes.
      ;
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      ).
      Analyzing the half-life of linear and circRNA is commonly used to investigate the stability of circRNA (Figure 2e). RNA is harvested from cells at different time points after treatment with a transcriptional inhibitor, such as actinomycin D. Normalization to stable cellular RNAs (e.g., 18S rRNA) facilitates the analysis. Most studies have revealed that circRNAs are much more stable than linear RNAs, perhaps owing to the circular structure. For example, whereas the half-life of linear transcripts from the MCPyV early region is 4–8 hours, no degradation of circALTO could be detected within 24 hours (
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      ).
      N6-methyladenosine (m6A) is an abundant epigenetic modification of RNA and has been reported to recruit specific RNA-binding proteins to mediate diverse functions. It has been reported to be enriched on circRNAs and is required for efficient circRNA production and also promotes cap-independent translation (
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      ,
      • Yang Y.
      • Fan X.
      • Mao M.
      • Song X.
      • Wu P.
      • Zhang Y.
      • et al.
      Extensive translation of circular RNAs driven by N6-methyladenosine.
      ;
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      ). Immunoprecipitation (IP) and quantification of m6A RNA is a common protocol used to assess the modification status of specific circRNAs (Figure 2f). Specifically, m6A or IgG control antibodies are used to precipitate total RNA and the relative enrichment of specific RNAs, including positive controls such as SON, a control mRNA with abundant m6A modifications, can be quantitated by QRT-PCR. Alternatively, m6A IP may be coupled with high-throughput sequencing for a more comprehensive analysis of modified sequences. These modifications have been found to be critical determinants of RNA function including circRNA formation and translation (
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      ,
      • Yang Y.
      • Fan X.
      • Mao M.
      • Song X.
      • Wu P.
      • Zhang Y.
      • et al.
      Extensive translation of circular RNAs driven by N6-methyladenosine.
      ;
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      ).
      Polyribosome (also known as polysome) fractionation analysis is a useful method to detect RNAs that encode proteins (Figure 2g) (
      • Collier A.E.
      • Wek R.C.
      • Spandau D.F.
      Human keratinocyte differentiation requires translational control by the eIF2α kinase GCN2.
      ). After gentle cell disruption, lysates are subjected to centrifugation over a sucrose gradient. Fractions containing monosomes and polysomes are separated, and RNA is isolated from the distinct fractions. CircRNAs predicted to be translated should be enriched in the polysome fraction. In a small survey, 83% (10/12) of circRNAs with both coding potential and m6A modifications were found to be associated with polysomes and, thus, likely to be translated (
      • Yang Y.
      • Fan X.
      • Mao M.
      • Song X.
      • Wu P.
      • Zhang Y.
      • et al.
      Extensive translation of circular RNAs driven by N6-methyladenosine.
      ). Although biochemical characterizations, such as associations with polysomes, may reveal potential functions, molecular genetic approaches to assess function are necessary to explore the function of possible circRNAs.

      Functional analyses of circRNAs

      The biological relevance of validated circRNAs should ultimately be tested through overexpression or knockdown functional analyses. Overexpression of a specific circRNAs is one common method used to determine circRNA function (Figure 2h). Minigenes containing the proposed circRNA and flanking sequences may be expressed in cell lines of interest. Backsplicing events tend to occur with low efficiency, so the sequence may be constructed with sequences that enhance circRNA formation. Complementary intronic sequences or QKI protein-binding sites flanking upstream and downstream of the circularized exon (
      • Conn S.J.
      • Pillman K.A.
      • Toubia J.
      • Conn V.M.
      • Salmanidis M.
      • Phillips C.A.
      • et al.
      The RNA binding protein quaking regulates formation of circRNAs.
      ) have both been reported to enhance circularization. Changes to the primary sequence of the minigene may be used to confirm the BSJ sequence, localization sequences, RNA-binding protein motifs, potential internal ribosome entry sites, or other primary sequence elements thought to be important for circRNA function (
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      ).
      Overexpression of a circRNA facilitates the analysis of its effect on cell growth through a variety of assays (e.g., proliferation, differentiation, gene expression). Several circRNAs function by binding microRNAs (miRNAs) to prevent them from inhibiting other RNA targets (i.e., miRNA sponges), so screens to identify potential miRNA binding partners de novo or tests for specific targets based on homology have both been successful (
      • Tagawa T.
      • Kopardé V.N.
      • Ziegelbauer J.M.
      Identifying and characterizing virus-encoded circular RNAs [e-pub ahead of print].
      ). Some circRNAs have also been reported to function as mRNAs. Both native and epitope-tagged open reading frames can be used to detect translated circRNAs. One challenge related to overexpression strategies is that expression plasmids also generate abundant linear RNAs, including artifactually trans-spliced RNAs and rolling circle transcription products. Thus, it is critical to assess the presence of such linear transcripts and exclude their contribution to relevant phenotypes through the use of RNA knockdowns that are specific to linear RNAs. If possible, circRNA function should be confirmed with knockdown experiments completed in parallel.
      The specific depletion of circRNAs is another strategy used to explore the biological function of circRNAs. RNA interference experiments using small interfering RNA (siRNA) or short hairpin RNA (shRNA) to knock down circRNAs specifically is a common strategy (Figure 2i). Both siRNAs and shRNAs should be designed to target the BSJ specifically, which constrains the design and may thus limit knockdown efficiency. Additional control siRNAs should be designed, including ones targeting relevant sequences shared by both circular and linear isoforms, and others targeting a linear region excluded by the circRNA. After confirming the specificity of the siRNAs to the targeted ortholog(s), these molecules can then be used to determine the specific effects of circRNA knockdown. In addition, CRISPR/Cas9–mediated genome modification may be used to modify endogenous circRNA loci (Figure 2j). Specifically, CRISPR has been used to delete sequences (e.g., Alu site) required for circRNA formation with minimal impacts on linear RNAs from the same locus to generate circRNA-specific knockouts (
      • Zhang Y.
      • Xue W.
      • Li X.
      • Zhang J.
      • Chen S.
      • Zhang J.L.
      • et al.
      The biogenesis of nascent circular RNAs.
      ). Loss-of-function experiments need to be interpreted with caution because circRNAs usually originate from loci that also generate linear RNAs, which might also be inhibited. Specific rescue of the knockdown by re-expression of an shRNA-resistant circRNA is one approach to control for off-target effects from loss-of-function experiments (
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      ).

      Conclusion

      CircRNAs have emerged as a novel, and increasingly appreciated, addition to the armamentarium of RNA molecules that contribute to phenotypic complexity. When compared with other classes of RNA molecules, circRNAs have unique properties, including high stability and enrichment in extracellular vesicles. Despite their relatively low expression, circRNAs have diverse and important biological functions. Most commonly, they have been reported to function as miRNA sponges, such as ciRS-7/CDR1AS (
      • Hansen T.B.
      • Jensen T.I.
      • Clausen B.H.
      • Bramsen J.B.
      • Finsen B.
      • Damgaard C.K.
      • et al.
      Natural RNA circles function as efficient microRNA sponges.
      ); protein interactors, such as circular PABPN1 (
      • Abdelmohsen K.
      • Panda A.C.
      • Munk R.
      • Grammatikakis I.
      • Dudekula D.B.
      • De S.
      • et al.
      Identification of HuR target circular RNAs uncovers suppression of PABPN1 translation by CircPABPN1.
      ); and translation templates, such as circular ZNF609 (
      • Legnini I.
      • Di Timoteo G.
      • Rossi F.
      • Morlando M.
      • Briganti F.
      • Sthandier O.
      • et al.
      Circ-ZNF609 is a circular RNA that can be translated and functions in myogenesis.
      ;
      • Yang Y.
      • Fan X.
      • Mao M.
      • Song X.
      • Wu P.
      • Zhang Y.
      • et al.
      Extensive translation of circular RNAs driven by N6-methyladenosine.
      ). These well-characterized, functional circRNAs provide evidence that circRNAs are regulated and biologically relevant, rather than accidental byproducts of canonical splicing.
      Several circRNAs with potential relevance to skin disease have been characterized, including ciRS-7, which inhibits melanoma progression and is significantly downregulated in psoriasis (
      • Hanniford D.
      • Ulloa-Morales A.
      • Karz A.
      • Berzoti-Coelho M.G.
      • Moubarak R.S.
      • Sánchez-Sendra B.
      • et al.
      Epigenetic silencing of CDR1as drives IGF2BP3-mediated melanoma invasion and metastasis.
      ;
      • Moldovan L.I.
      • Hansen T.B.
      • Venø M.T.
      • Okholm T.L.H.
      • Andersen T.L.
      • Hager H.
      • et al.
      High-throughput RNA sequencing from paired lesional- and non-lesional skin reveals major alterations in the psoriasis circRNAome.
      ); circE7, which produces the transforming oncoprotein E7 in HPV-induced cancers (
      • Zhao J.
      • Lee E.E.
      • Kim J.
      • Yang R.
      • Chamseddin B.
      • Ni C.
      • et al.
      Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
      ); and circALTO, which is detectable in MCC and encodes for ALTO, a transcriptional regulator (
      • Yang R.
      • Lee E.E.
      • Kim J.
      • Choi J.H.
      • Kolitz E.
      • Chen Y.
      • et al.
      Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
      ). Circular variants of CYP2C18, a cytochrome P450 involved in retinoid metabolism, were among the earliest circRNAs to be identified in eukaryotes and are readily detected in the human epidermis, yet nothing is known about their potential role in skin development or function (
      • Zaphiropoulos P.G.
      Exon skipping and circular RNA formation in transcripts of the human cytochrome P-450 2C18 gene in epidermis and of the rat androgen binding protein gene in testis.
      ). These tantalizing examples offer but a small glimpse into the potential impact of the thousands of circRNAs that have been identified in silico. We hope that the methods described in this article will accelerate our understanding of circRNAs and their potential roles in skin biology and the diagnosis and treatment of skin disease.

      Conflict of Interest

      The authors state no conflict of interest.

      Multiple Choice Questions

      • 1.
        Which of the following is true for circular RNAs (circRNAs)? They are ___
        • A.
          Single-stranded RNAs
        • B.
          Produced by backsplicing event
        • C.
          Covalently closed
        • D.
          Typically less abundant than linear RNAs
        • E.
          All of the above
      • 2.
        True or false? Ribonuclease (RNase) R digests all RNAs except circRNAs.
        • A.
          True
        • B.
          False
      • 3.
        Which of the following techniques can specifically confirm the existence of circRNAs that have been identified in silico?
        • A.
          Bioinformatic pipelines that detect backsplice junctions
        • B.
          Northern blot with and without RNase R
        • C.
          PolyA primed reverse transcription and PCR
        • D.
          Western blot
        • E.
          Polysome enrichment
        • F.
          All of the above
      • 4.
        To knockdown a circRNA, the small interfering RNA should be designed to anneal to ___
        • A.
          Backsplice junction
        • B.
          Any potential open reading frames in the circRNA
        • C.
          A linear region outside the backspliced exons
        • D.
          Predicted RNA-binding protein sites
      • 5.
        Which of the following is a limitation in studying circRNA functional studies?
        • A.
          The stability of circRNAs makes them difficult to knock down.
        • B.
          N6-methyladenosine modifications frequently interfere with RNA interference.
        • C.
          Efficiency of circRNAs depend on their subcellular localization.
        • D.
          circRNAs and their cognate linear RNAs share parental gene sequences, making knockdown and knockout experiments more difficult to control.

      Acknowledgments

      This work was supported by grants from the University of Texas Southwestern Cary Council (Dallas, TX), ACS (RSG-18-058-01), and National Institute of Arthritis and Musculoskeletal and Skin Diseases (Bethesda, MD) (R01AR072655) to RCW.

      Author Contributions

      Conceptualization: RY, RCW; Writing - Review and Editing: RY, RCW

      Supplementary Material

      Detailed Answers

      • 1.
        Which of the following is true for circular RNAs (circRNAs)? They are ___
      • CORRECT ANSWER: E. All of the above.
      • CircRNAs are covalently closed, single-stranded RNAs produced by backsplicing RNAs. They are typically less abundant than linear RNAs.
      • 2.
        True or false? Ribonuclease (RNase) R digests all RNAs except circRNAs.
      • CORRECT ANSWER: B. False.
      • RNase R is an exoribonuclease that digests linear RNA, but not circular, lariat, or structured RNAs. RNase R may be used to enrich for circular RNAs.
      • 3.
        Which of the following techniques can specifically confirm the existence of circRNAs that have been identified in silico?
      • CORRECT ANSWER: B. Northern blot with and without RNase R
      • CircRNAs can be identified by northern blots using backsplice-specific probe junctions or, more commonly, performing the blots in the presence and absence of RNase R.
      • 4.
        To knockdown a circRNA, the small interfering RNA should be designed to anneal to ___
      • CORRECT ANSWER: A. BSJ
      • Both small interfering RNA (siRNA) and short hairpin RNA should be designed to target the BSJ. Not all circRNAs are mRNAs, so targeting any open reading frames (ORFs) may not be relevant. Moreover, ORFs are frequently shared by the circRNA and the cognate linear RNAs. Hairpins or siRNAs targeting a linear region outside the backspliced exon will knock down just the linear RNA.
      • 5.
        Which of the following is a limitation in studying circRNA functional studies?
      • CORRECT ANSWER: D. circRNAs and their cognate linear RNAs share parental gene sequences, making knockdown and knockout experiments more difficult to control.
      • Because circRNAs and their linear counterparts frequently share originating loci and sequences, it may be challenging to manipulate one isoform without affecting the other. In overexpression experiments, artifactual linear spliced products may complicate analyses.

      References

        • Abdelmohsen K.
        • Panda A.C.
        • Munk R.
        • Grammatikakis I.
        • Dudekula D.B.
        • De S.
        • et al.
        Identification of HuR target circular RNAs uncovers suppression of PABPN1 translation by CircPABPN1.
        RNA Biol. 2017; 14: 361-369
        • Abere B.
        • Zhou H.
        • Li J.
        • Cao S.
        • Toptan T.
        • Grundhoff A.
        • et al.
        Merkel cell polyomavirus encodes circular RNAs (circRNAs) enabling a dynamic circRNA/microRNA/mRNA regulatory network.
        mBio. 2020; 11 (e03059–20)
        • Collier A.E.
        • Wek R.C.
        • Spandau D.F.
        Human keratinocyte differentiation requires translational control by the eIF2α kinase GCN2.
        J Invest Dermatol. 2017; 137: 1924-1934
        • Conn S.J.
        • Pillman K.A.
        • Toubia J.
        • Conn V.M.
        • Salmanidis M.
        • Phillips C.A.
        • et al.
        The RNA binding protein quaking regulates formation of circRNAs.
        Cell. 2015; 160: 1125-1134
        • Hanniford D.
        • Ulloa-Morales A.
        • Karz A.
        • Berzoti-Coelho M.G.
        • Moubarak R.S.
        • Sánchez-Sendra B.
        • et al.
        Epigenetic silencing of CDR1as drives IGF2BP3-mediated melanoma invasion and metastasis.
        Cancer Cell. 2020; 37: 55-70.e15
        • Hansen T.B.
        • Jensen T.I.
        • Clausen B.H.
        • Bramsen J.B.
        • Finsen B.
        • Damgaard C.K.
        • et al.
        Natural RNA circles function as efficient microRNA sponges.
        Nature. 2013; 495: 384-388
        • Hansen T.B.
        • Venø M.T.
        • Damgaard C.K.
        • Kjems J.
        Comparison of circular RNA prediction tools.
        Nucleic Acids Res. 2016; 44: e58
        • Kristensen L.S.
        • Andersen M.S.
        • Stagsted L.V.W.
        • Ebbesen K.K.
        • Hansen T.B.
        • Kjems J.
        The biogenesis, biology and characterization of circular RNAs.
        Nat Rev Genet. 2019; 20: 675-691
        • Legnini I.
        • Di Timoteo G.
        • Rossi F.
        • Morlando M.
        • Briganti F.
        • Sthandier O.
        • et al.
        Circ-ZNF609 is a circular RNA that can be translated and functions in myogenesis.
        Mol Cell. 2017; 66: 22-37.e9
        • Li X.
        • Yang L.
        • Chen L.L.
        The biogenesis, functions, and challenges of circular RNAs.
        Mol Cell. 2018; 71: 428-442
        • Liang J.
        • Wu X.
        • Sun S.
        • Chen P.
        • Liang X.
        • Wang J.
        • et al.
        Circular RNA expression profile analysis of severe acne by RNA-Seq and bioinformatics.
        J Eur Acad Dermatol Venereol. 2018; 32: 1986-1992
        • Luo Q.
        • Zhang L.
        • Li X.
        • Fu B.
        • Guo Y.
        • Huang Z.
        • et al.
        Identification of circular RNAs hsa_circ_0044235 and hsa_circ_0068367 as novel biomarkers for systemic lupus erythematosus.
        Int J Mol Med. 2019; 44: 1462-1472
        • McBride J.D.
        • Rodriguez-Menocal L.
        • Badiavas E.V.
        Extracellular vesicles as biomarkers and therapeutics in dermatology: a focus on exosomes.
        J Invest Dermatol. 2017; 137: 1622-1629
        • Moldovan L.I.
        • Hansen T.B.
        • Venø M.T.
        • Okholm T.L.H.
        • Andersen T.L.
        • Hager H.
        • et al.
        High-throughput RNA sequencing from paired lesional- and non-lesional skin reveals major alterations in the psoriasis circRNAome.
        BMC Med Genomics. 2019; 12: 174
        • Moldovan L.I.
        • Tsoi L.C.
        • Ranjitha U.
        • Hager H.
        • Weidinger S.
        • Gudjonsson J.E.
        • et al.
        Characterization of circular RNA transcriptomes in psoriasis and atopic dermatitis reveals disease-specific expression profiles.
        Exp Dermatol. 2021; 30: 1187-1196
        • Qiao M.
        • Ding J.
        • Yan J.
        • Li R.
        • Jiao J.
        • Sun Q.
        Circular RNA expression profile and analysis of their potential function in psoriasis.
        Cell Physiol Biochem. 2018; 50: 15-27
        • Sand M.
        • Bechara F.G.
        • Gambichler T.
        • Sand D.
        • Bromba M.
        • Hahn S.A.
        • et al.
        Circular RNA expression in cutaneous squamous cell carcinoma.
        J Dermatol Sci. 2016; 83: 210-218
        • Tagawa T.
        • Kopardé V.N.
        • Ziegelbauer J.M.
        Identifying and characterizing virus-encoded circular RNAs [e-pub ahead of print].
        Methods. 2021; (accessed June 30, 2021)https://doi.org/10.1016/j.ymeth.2021.03.004
        • Tang K.
        • Zhang H.
        • Li Y.
        • Sun Q.
        • Jin H.
        Circular RNA as a potential biomarker for melanoma: a systematic review.
        Front Cell Dev Biol. 2021; 9: 638548
        • Wang L.
        • Si X.
        • Chen S.
        • Wang X.
        • Yang D.
        • Yang H.
        • et al.
        A comprehensive evaluation of skin aging-related circular RNA expression profiles.
        J Clin Lab Anal. 2021; 35: e23714
        • Wang X.
        • Ma R.
        • Shi W.
        • Wu Z.
        • Shi Y.
        Emerging roles of circular RNAs in systemic lupus erythematosus.
        Mol Ther Nucleic Acids. 2021; 24: 212-222
        • Yang R.
        • Lee E.E.
        • Kim J.
        • Choi J.H.
        • Kolitz E.
        • Chen Y.
        • et al.
        Characterization of ALTO-encoding circular RNAs expressed by Merkel cell polyomavirus and trichodysplasia spinulosa polyomavirus.
        PLoS Pathog. 2021; 17e1009582
        • Yang Y.
        • Fan X.
        • Mao M.
        • Song X.
        • Wu P.
        • Zhang Y.
        • et al.
        Extensive translation of circular RNAs driven by N6-methyladenosine.
        Cell Res. 2017; 27: 626-641
        • Zaphiropoulos P.G.
        Exon skipping and circular RNA formation in transcripts of the human cytochrome P-450 2C18 gene in epidermis and of the rat androgen binding protein gene in testis.
        Mol Cell Biol. 1997; 17: 2985-2993
        • Zhang Y.
        • Xue W.
        • Li X.
        • Zhang J.
        • Chen S.
        • Zhang J.L.
        • et al.
        The biogenesis of nascent circular RNAs.
        Cell Rep. 2016; 15: 611-624
        • Zhao J.
        • Lee E.E.
        • Kim J.
        • Yang R.
        • Chamseddin B.
        • Ni C.
        • et al.
        Transforming activity of an oncoprotein-encoding circular RNA from human papillomavirus.
        Nat Commun. 2019; 10: 2300