Transposable elements in health and disease: Molecular basis and clinical implications

Abstract

Transposable elements (TEs), once considered genomic “junk”, are now recognized as critical regulators of genome function and human disease. These mobile genetic elements—including retrotransposons (long interspersed nuclear elements [LINE-1], Alu, short interspersed nuclear element-variable numbers of tandem repeats-Alu [SVA], and human endogenous retrovirus [HERV]) and DNA transposons—are tightly regulated by multilayered mechanisms that operate from transcription through to genomic integration. Although typically silenced in somatic cells, TEs are transiently activated during key developmental stages—such as zygotic genome activation and cell fate determination—where they influence chromatin architecture, transcriptional networks, RNA processing, and innate immune responses. Dysregulation of TEs, however, can lead to genomic instability, chronic inflammation, and various pathologies, including cancer, neurodegeneration, and aging. Paradoxically, their reactivation also presents new opportunities for clinical applications, particularly as diagnostic biomarkers and therapeutic targets. Understanding the dual role of TEs—and balancing their contributions to normal development and disease—is essential for advancing novel therapies and precision medicine.

Introduction and Classification of Transposable Elements

Transposable elements (TEs), also known as “jumping genes”, are dynamic DNA sequences capable of relocating within the genome. First discovered in maize by Barbara McClintock in the 1940s, TEs have revolutionized our understanding of genomes by revealing their mobile and plastic nature. Ubiquitous across all life forms, TEs exhibit diverse structures and mobilization mechanisms and act as major drivers of genetic diversity and genome evolution.^[1,2,3]

In humans, TEs constitute approximately 50% of the genome, far exceeding the 1.5% comprised of protein-coding genes [Figure 1A]. Based on their transposition mechanisms, TEs are classified into two major categories: Class I (retrotransposons) and Class II (DNA transposons) [Figure 1A and B]. Retrotransposons move through a “copy-and-paste” mechanism that involves an RNA intermediate, which is then reverse-transcribed into DNA. In contrast, DNA transposons use a “cut-and-paste” mechanism and do not involve an RNA intermediate [Figure 1B–E].^[1,2,3] Retrotransposons are further divided into long terminal repeats (LTRs) and non-LTR retrotransposons. The latter includes long interspersed nuclear elements (LINEs) and short interspersed nuclear elements (SINEs) [Figure 1A and B].

Figure 1.

Structure, abundance and life cycle of TEs in the human genome. (A) Pie chart showing the distribution of various TEs in the human genome. (B) Structural representation of different types of TEs. The frequencies of new insertion events for L1, Alu and SVA in newborn babies are indicated, while DNA transposons and ERVs are considered molecular fossils that are incapable of mobilization in modern humans. (C) Lifecycle of DNA transposon, which encodes the DNA transposase that cuts itself and integrate into another genomic location. (D) Lifecycle of the LTR retrotransposon or ERV. ERV regulators and therapies targeting ERVs are marked as indicated. (E) Schematic of L1 retrotransposition. L1 regulators and therapies targeting L1s are marked as indicated. (F) Phylogenetic tree showing the evolution of L1 subfamilies (L1PA6 to L1Hs) in primates, including rhesus, orangutan, gorilla, chimpanzee, and human. L1Hs, L1 homo sapiens or L1PA1, is the youngest L1 subfamily in the human genome. A: A boxes; AR: A-rich linker; ASO: Antisense oligo; B: B boxes; C: Cysteine-rich domains; EN: Endonuclease; Env: Envelope protein; ERVs: Endogenous retrovirus; Gag: Group-specific antigen protein; ITR: Inverted terminal repeat; LTR: Long terminal repeat; L1: LINE-1; LINE-1: Long interspersed nuclear elements; mya; Million years; ORF: Open reading frame; Pol: Pol protein; RT: Reverse transcription; SINE: Short interspersed nuclear element; SVA: Short interspersed nuclear element-variable numbers of tandem repeats-Alu; TEs: Transposable elements; UTR: Untranslated region; VNTR: Variable numbers of tandem repeats.

DNA transposon

A typical DNA transposon consists of a coding sequence that encodes DNA transposase, flanked by two inverted terminal repeat (ITR) sequences. The transposase binds and cleaves the ITRs, excises the transposon, and facilitates its reintegration into a new genomic location^[1,2,3] [Figure 1B and C]. Although DNA transposons are relatively rare in humans compared to other eukaryotes and are now largely inactive (“molecular fossils”), they have played a significant role in genome evolution^[1,2,3]. Notably, they contributed to the emergence of essential genes, such as RAG1 and RAG2, which mediate V(D)J recombination^[4] and gave rise to host-TE fusion proteins that function as key epigenetic regulators.^[5]

LINE-1 (or L1)

As the only active autonomous TE in humans, L1 encodes all necessary components for its own mobilization. RNA polymerase II (Pol II) transcribes full-length L1 RNA, which is then translated into two key proteins: Open reading frame (ORF) 1 (RNA-binding chaperone) and ORF2 (a bifunctional enzyme with endonuclease activity at AT-rich DNA and reverse transcriptase activity). Retrotransposition proceeds via target-primed reverse transcription (TPRT) initiated at genomic nicks, followed by complementary DNA (cDNA) synthesis and integration [Figure 1E].^[1,2,3] Recent structural studies have resolved the biochemical architecture of ORF2, identifying its template and target sites.^[6,7]

L1s are thought to have originated from Class II introns and have persisted in vertebrate genomes for over 150 million years, diversifying through mutations and amplifications.^[8] Human L1 elements include the L1Hs (L1 Homo sapiens, the youngest and most active subfamily), L1PA (primate-specific and older than L1Hs, e.g., L1PA2 and L1PA3), and L1M (mammalian-specific and ancient, predating primate divergence) [Figure 1F]. Although L1s makes up approximately 17% of the human genome with approximately 500,000 L1 copies, most are truncated or mutated, with only about 100 copies remaining capable of retrotransposition.^[9] Nonetheless, hundreds to thousands of L1 elements per cell remain transcriptionally active and can serve as DNA regulatory elements, influencing gene expression.^{[1,2,3,10,11]}

Alu

Alu is a subfamily of SINEs, approximately 300 bp long, derived from the 7SL RNA gene.^[12] Alu elements are characterized by two internal repeats separated by a short spacer region. Among the Alu subfamilies, AluY is the most recent and active, followed by AluS and AluJ. AluY plays a significant role in shaping human genome evolution and gene regulation.^[12]

Alu does not encode the proteins required for its own mobilization. Instead, Alu relies on L1-encoded ORF2 for reverse transcription (RT) and genomic integration [Figure 1E]. Alu RNAs (transcribed by Pol III) hijack L1’s enzymatic machinery in trans, enabling their proliferation throughout the genome. Despite their dependence on L1, Alu mobilizes at frequencies comparable to L1, with estimated insertion rates of 1:200–1:20 (L1) and 1:40–1:20 (Alu) per birth [Figure 1B].^[13]

SINE-variable numbers of tandem repeats (VNTR)-Alu (SVA)

SVA is specific to hominids and is the youngest transposon family in the human genome. It consists of Alu-like and SINE-R sequences separated by VNTRs [Figure 1B]. SVAs constitute approximately 0.2% of the human genome and are grouped into subfamilies SVA-A through SVA-F [Figure 1A].^[14] Similar as Alu, SVA belongs to the nonautonomous retrotransposon. SVA RNAs (transcribed by Pol II) hijack L1-encoded ORF2 for RT and genomic integration, with estimated insertion rates of 1:500–1:63 per birth [Figure 1B and 1E].^[13]

Long terminal repeats (LTR)

LTR retrotransposons are flanked by LTRs and are considered remnants of ancient RNA viruses that infected ancestral genomes—examples include human endogenous retroviruses (HERVs) [Figure 1D]. These retrotransposons are classified based on the type of transfer RNA (tRNA) that binds to their primer-binding site to initiate RT. For instance, HERVK designates a group of proviruses that use lysine (K) tRNA as a primer, whereas HERVH uses histidine (H) tRNA.^[15]

Over evolutionary time, these elements have accumulated mutations, internal deletions and recombinations, often leading to the loss of protein-coding sequences essential for replication. These genetic alterations prevent LTR retrotransposons from forming functional viral-like particles or producing reverse transcriptase, rendering them incapable of reintegration. Consequently, these inactive LTRs become fixed in the genome, where they may still modulate gene regulation or chromatin architecture but have lost their ability to mobilize.^[1,2,3]

Molecular Mechanisms of TE Activity

Multilayered regulation of TEs

TEs present a genomic paradox: While they drive genetic innovation, their unchecked activity can threaten genomic integrity.^[1,2,3] To maintain this balance, cells utilize a sophisticated, multitiered regulatory system [Table 1, Figure 1D and E].

Table 1.

Key regulatory pathways that control human transposable elements (TEs).

Regulatory pathways	Description	Key regulators	TEs affected
DNA methylation	m5C methylation of CpG sites in TE promoters for silencing; 6mA DNA methylation	DNMTs, TETs	L1, Alu, ERV
Histone modification	Repressive (H3K9me3, H3K27me3, H2A/H4R3me2, H4K20me3); Activating (H3K27ac, H3K4me3, H4K16ac)	KAP1, HUSH, HUSH2, SETDB1, KRAB-ZFP, NuRD, HDACs, KMT2D, SUV39H1/2, MSL3, PRMT5	L1, SVA, ERV
Chromatin remodeling	Chromatin remodelers regulate TE accessibility	SWI/SNF, INO80, MORC1/2	L1, Alu, HERVK
Transcription factor	Transcription factors activate or repress TE transcription	YY1, RUNX3, MYC, FOXA1, P53, KRAB-ZFP, CTCF, MeCP2, SOX, DPPA2/4	L1, Alu, ERV
RNA interference	Small RNAs (siRNAs, miRNAs, piRNAs) degrade TE RNAs	PIWI, RIG-I, miR-128, let-7	L1, Alu, ERV
RNA modification	m6A and m5C modification regulates the TE RNA fate, including chromatin epigenetics	METTL3/14, YTHDC1, MBD6	L1, Alu, ERV
RNA binding protein	RNA binding protein regulates TE RNA processing and stability	SAFB, PTBP1, MATR3, hnRNPM, DBR1, SRSF10, DIS3, EXOSC6, NEXT, HELZ2, TUT4/7, MOV10, DDX42, SAMHD1, UPF1	L1, ERV
Translation and protein homeostasis	L1 proteins are controlled at the level of translation, protein assembly and degradation	SAMHD1, Condensin I/II, TDP-43, cGAS, PML, PABPC1, PABPN1	L1
DNA damage & repair	DNA damage and repair factors regulate reverse transcription and DNA integration	APOBEC3, SETX, PCNA, Fanconi anemia factor, BRCA1, Homologous recombination	L1, ERV

APOBEC3: Apolipoprotein B mRNA editing enzyme catalytic subunit 3; BRCA1: Breast cancer type 1 susceptibility protein; cGAS: Cyclic GMP-AMP synthase; CTCF: CCCTC-binding factor; DBR1: Debranching RNA lariats 1; DDX42: DEAD-box helicase 42; DIS3: DIS3 homolog, exosome endoribonuclease and 3′-5′ exoribonuclease; DNMTs: DNA methyltransferases; DPPA2/4: developmental pluripotency associated 2/4; ERV: Endogenous retrovirus; EXOSC6: Exosome component 6; FOXA1: Forkhead box A1; HDACs: Histone deacetylases; HELZ2: Helicase with zinc finger 2; HERVK: Human endogenous retrovirus K; HUSH: human silencing hub; INO80: INO80 complex ATPase subunit; KAP1: Krüppel-associated box domain–associated protein 1; KMT2D: Lysine (K)-specific methyltransferase 2D; KRAB-ZFP: Krüppel-associated box domain zinc finger proteins; L1: Long interspersed nuclear elements 1; let-7: Lethal-7; MATR3: Matrin 3; hnRNPM: Heterogeneous nuclear ribonucleoprotein M; MBD6: Methyl-CpG binding domain protein 6; MECP2: Methyl-CpG binding protein 2; METTL3/14: Methyltransferase 3/14, N6-adenosine-methyltransferase complex catalytic subunit; miRNA: MicroRNA; miR-128: MicroRNA-128; MORC1/2: MORC family CW-type zinc finger 1/2; MOV10: Mov10 RNA helicase; MSL3: MSL complex subunit 3; MYC: MYC proto-oncogene, bHLH transcription factor; NEXT: Nuclear exosome targeting complex; NuRD: Nucleosome remodeling and deacetylase; PABPC1: Poly(A) binding protein cytoplasmic 1; PABPN1: Poly(A) binding protein nuclear 1; piRNA: PIWI-interacting RNA; PIWI: P-element induced wimpy testis; PML: Progressive multifocal leukoencephalopathy; PRMT5: Protein arginine methyltransferase 5; PTBP1: Polypyrimidine tract binding protein 1; RIG-I: Retinoic acid-inducible gene I; RUNX3: Runt-related transcription factor 3; SAFB: Scaffold attachment factor B; SAMHD1: SAM and HD domain containing deoxynucleoside triphosphate triphosphohydrolase 1; SETDB1: SET domain bifurcated histone lysine methyltransferase 1; SETX: Senataxin; siRNA: Small interfering RNA; PCNA: Proliferating cell nuclear antigen; SOX: SRY-related high-mobility group; SRSF10: Serine and arginine rich splicing factor 10; SUV39H1/2: SUV39H1/2 histone lysine methyltransferase; SWI/SNF: Switch/sucrose-nonfermentable; TDP-43: Transactive response DNA-binding protein 43; TE: Transposable element; TETs: Ten-eleven translocation family proteins; TUT4/7: Terminal uridylyl transferase 4/7; UPF1: UPF1 RNA helicase and ATPase; YTHDC1: YTH N6-methyladenosine RNA binding protein C1; YY1: Yin Yang 1.

DNA methylation

The first layer of TE regulation involves epigenetic modifications [Table 1]. At the DNA methylation level, DNA methyltransferases (DNMTs) add methyl groups to CpG islands in TE promoters, thereby inhibiting transcription.^[16] This mechanism is especially important in germline cells to prevent the transgenerational transmission of active TE.^[17] Interestingly, the Ten-eleven translocation protein family can demethylate 5-methylcytosine (m5C) to promote TE transcription and can also repress L1 expression independently of its enzymatic activity.^[18]

Histone modification

Repressive histone marks, such as H3K9me3 and H3K27me3, establish heterochromatin at LINEs, SINEs, and LTRs to silence their transcription.^[19] Several silencing complexes have been identified: (1) The KAP1/TRIM28 complex recruits SETDB1 to deposit H3K9me3, particularly targeting young L1s, SVAs, and endogenous retroviruses (ERVs).^[20] (2) The HUSH complex—MPP8, TASOR, and PPHLN1—recognizes intronless cDNAs and reinforces H3K9me3 deposition on active retrotransposons.^[21,22] Interestingly, the HUSH complex is antagonized by an alternative complex, HUSH2, which incorporates TASOR2 (FAM208B) instead of TASOR. Unlike HUSH, TASOR2 promotes L1 expression, demonstrating how paralog switching can reverse regulatory outcomes.^[23,24] (3) H3K9me3-generating methyltransferase, SUV39H, is involved in suppressing retrotransposons in mouse embryonic stem cells.^[25] (4) The NuRD complex partners with retinoblastoma proteins to recruit histone deacetylases (HDACs), thereby removing activating acetylation marks.^[26]

In contrast, activation marks—such as H3K4me1 and H3K27ac—are associated with TE-derived enhancers.^[27] Knockout of CTBP1 and KMT2D reduces H3K9me3 levels and increases H3K4me1/H3K27ac, thereby de-repressing multiple TE families.^[10] Multiple additional factors that erase or suppress activating histone modifications have been recently reviewed elsewhere, and therefore will not be discussed in detail here.^[28]

RNA interference (RNAi)

In germline cells, the PIWI-interacting RNA (piRNA) pathway serves as the primary defense mechanism against TEs. The piRNA-PIWI complex mediates TE suppression through dual mechanisms: (1) post-transcriptional cleavage of TE-derived RNAs (including L1, Alu, and ERV transcripts) and (2) transcriptional silencing at the chromatin level, thereby safeguarding genomic integrity during germ cell development.^[29,30,31] In somatic cells, TE regulation shifts to small RNA-mediated pathways, where endogenous small interfering RNAs (siRNAs) and microRNAs (miRNAs) target TE-derived RNAs for degradation.^[32,33] For a more comprehensive discussion of RNAi in transposon control, we refer readers to a recent in-depth review on this subject.^[34]

Transcription factors (TFs)

TFs such as Yin Yang 1 and Runt-related transcription factor 3 bind to L1 promoters to enhance transcription.^[10,21] Similarly, p53 regulates the expression of Alu and ERV elements in response to cellular stress and DNA damage.^[35] In addition, TEs have evolved strategies to evade epigenetic suppression—for instance, some L1s acquire mutations or 5′-untranslated region (UTR) deletions to escape Zn finger (ZNF)-mediated repression.^[36]

RNA-binding protein (RBP)

TE-derived RNAs are regulated by various RBPs that control their stability, splicing, and retrotransposition.^{[10,11,37,38,39]} For example, PTBP1, MATR3, hnRNPM, SRSF10, and SAFB bind L1 RNAs to suppress aberrant splicing into adjacent exons.^{[10,11,37,38,39]} Remarkably, genome-wide clustered regularly interspaced short palindromic repeats (CRISPR) screens targeting both endogenous and codon-optimized L1 elements identified multiple RBPs—including HUSH and SAFB,^[21] confirmed in subsequent independent studies^{[11,38,40,41,42]}—that regulate L1 activity in a sequence-dependent manner. Additional RBPs, such as condensin I/II complexes, PABPC1, and PABPN1, regulate L1 translation and RNP assembly.^[43] Furthermore, DIS3, EXOSC6, PML bodies, and nuclear cyclic GMP-AMP synthase (cGAS) contribute to depleting TE-derived RNAs and proteins.^[10,44,45]

RNA modification

N⁶-methyladenosine (m⁶A) RNA methylation can directly recruit YTH domain-containing proteins or indirectly influence RNA-protein interactions by altering RNA structure, thereby affecting multiple aspects of RNA fate, including TE-derived transcripts.^[10,40,46] Recent findings suggest crosstalk between m⁶A RNA modifications and histone/DNA modifications in regulating TE transcription, supporting the concept of m⁶A-dependent chromatin epigenetics.^{[10,28,47,48,49,50,51,52]} In addition to m⁶A, m⁵C and its reader protein MBD6 play important roles in regulating chromatin-associated TE RNAs and, consequently, influence chromatin compaction.^[53] Moreover, adenosine-to-inosine (A-to-I) RNA editing in Alu RNAs can reduce complementarity between inverted Alu repeats, unwind Alu double-stranded RNA (dsRNA) and mitigate the potential risks associated with its accumulation.^[54] Thus, TE RNAs harbor diverse RNA modifications that influence their fates.

DNA damage and repair

DNA damage response pathways are also involved in regulating TE activity. APOBEC3 cytidine deaminases hypermutate the cDNA of L1 and ERV elements to prevent successful RT and genomic integration.^[55] In addition, DNA repair factors—including PCNA, non-homologous end-joining genes, Fanconi anemia-related factors, BRCA1-mediated homologous recombination pathways, and DNA/RNA helix-resolving proteins—process L1 retrotransposition intermediates to either promote or restrict their integration.^[21,56,57] However, the precise mechanisms by which these repair factors influence retrotransposition remain unclear and warrant further investigation.

These multilayered regulatory mechanisms—including DNA methylation, histone modifications, RNA interference, TFs, RNA modifications, and DNA damage response pathways—work together to control TE activity. Yet the fundamental question remains: How are these regulatory factors coordinated to achieve spatiotemporal control of individual TE elements? Are there species- or cell-type-specific hierarchies among these pathways? Which TE subsets are most relevant to disease development? Addressing these questions will offer critical insights into the regulation of TEs across various cell types and developmental stages.

Multifaceted regulatory functions of TEs

Comprising nearly half of the human genome, TEs are dynamic regulators of chromatin architecture, transcriptional networks, RNA processing, and innate immunity [Figure 2]. Through various mechanisms, their activity has significantly influenced genomic complexity and the evolution of gene regulation.

Figure 2.

Molecular basis and clinical implications for human TEs in development and disease. TEs are originated from functional RNAs or retrovirus, continue to shape the human genome through transposition, and have been co-opted as diverse regulatory DNA elements, which are particularly active and essential for multiple key biological contexts, including embryonic development, neurodevelopmental disorders, ageing, immune responses, and cancer progression. TEs have emerged as diagnostic biomarkers and therapeutic targets due to their disease-specific activation patterns and immunogenic properties. ASO: Antisense oligo; DNMTi: DNA methyltransferase inhibitor; ERV: Endogenous retroviruse; HDACi: Histone deacetylase inhibitor; RT: Reverse transcription; SVA: SINE-variable numbers of tandem repeats (VNTR)-Alu; TEs: Transposable elements.

Exon insertion

TE insertions into coding regions or exons can have major consequences, including loss-of-function mutations, altered protein products, and increased genomic instability.^[58,59] In breast and colon cancers, L1 insertion into tumor suppressor genes or oncogenes—such as adenomatous polyposis coli—can contribute to carcinogenesis by disrupting normal gene expression and cellular regulation.^[58,59] Moreover, Alu insertions can affect gene regulation and have been implicated in diseases such as multiple sclerosis (MS).^[60] TE insertions into non-coding regions, including regulatory elements, may also lead to long-term effects on gene expression and cellular functions.^[61,62]

Chromatin compartment

TEs can act as nucleation sites for heterochromatin formation, typically within transcriptionally inactive regions of the genome.^[63,64] L1 elements frequently localize to heterochromatic domains, such as lamina-associated domains and nucleolus-associated domains, to organize the 3D genome structure.^[64] Both L1 RNA and DNA can trigger phase separation of heterochromatin protein 1α, promoting heterochromatin compartmentalization and silencing of L1-enriched genomic regions.^[64]

Topologically associating domain (TAD)

TEs also function as TAD boundaries—regions of the genome within which genes and regulatory elements interact more frequently than with regions outside the domain.^[3] HERVH and L1 elements can establish TAD boundaries independently of CTCF, and their functions are closely linked to transcriptional activity.^[11,65] Active L1s can recruit RNA Pol II, generate L1 chimeric RNAs, and help form domain boundaries within euchromatin.^[11,38] The nuclear matrix protein SAFB interacts with L1-enriched regions, modulates Pol II occupancy, and weakens L1-associated boundaries, illustrating how TE-related transcription and RNA processing shape genome architecture.^[11] Over time, L1 insertions generate lineage-specific TAD boundaries, contributing to species-specific chromatin organization and gene regulatory networks.^[11]

Enhancer

TEs also function as enhancers.^{[10,66,67,68,69]} The 5′-UTRs of younger L1 subfamilies, such as L1PA1–L1PA6, show stronger STARR-seq enrichments than older L1 subfamilies, suggesting greater enhancer potential.^[10,67] This activity is supported by Hi-C and 4C analysis, which demonstrate that L1 elements can physically interact with distal target genes and modulate their expression.^[10] In human cells, various factors, including KMT2D, DIS3, ACTL6A, and m⁶A RNA modifications, are involved in modulating L1 enhancer activity.^[10,67] Alu repeats can also facilitate enhancer-promoter interactions by forming complementary RNA duplexes that bridge these genomic regions.^[70] In addition, ERVs function as enhancers in modulating immune response and tumorigenesis.^[71,72] And SVAs act as RNA-dependent enhancers to regulate both erythropoiesis and myelopoiesis.^[73] However, further research is needed to elucidate the precise mechanisms by which TEs selectively activate the transcription of one or more genes across long genomic distances.

Trans-regulation

L1-derived RNAs can serve as scaffolds for transcriptional complexes that regulate gene expression, demonstrating trans-regulatory roles in ESCs.^[74] During cellular stress—such as heat shock and viral infection—Alu transcripts are upregulated and can globally inhibit RNA Pol II-mediated transcription, acting as a regulatory brake on gene expression.^[12] Moreover, TE-derived RNAs or their RT products can activate RNA sensors (RIG-1, MAVS) or DNA sensors (the cGAS–STING pathway) to trigger type I interferon (IFN) responses.^[7,44,75,76]

Alternative initiation, splicing, and termination

TEs contribute initiation, splicing, and termination sites that give rise to TE chimeric transcripts.^[77] Evolutionarily, young L1 elements typically have repressive effects by recruiting splicing factors such as SAFB, MATR3, PTBP1, and HNRNPM, which help prevent cryptic splicing in intronic regions.^{[11,37,38,39]} In contrast, older L1s are more likely to lose binding sites for these repressive proteins, allowing them to be spliced into exons in a tissue-specific manner.^[39] In addition, the 5′-UTR of L1 exhibits bidirectional promoter activity—both sense and antisense—and can generate fusion proteins with nearby exons.^[11,78]

Overall, TEs are integral components of the genome, influencing key biological processes such as genome stability, 3D genome folding, TAD boundary formation, enhancer activity, RNA splicing, and innate immunity. Further research is needed to dissect the molecular features, mechanisms, and context-specific functions of individual TE elements, as well as how they interact with other regulatory DNA elements to ensure precise gene expression during development and disease.

Physiological Roles of TEs in Development and Disease

TEs serve as pivotal genomic architects, contributing to both evolutionary innovation and disease pathogenesis.^[1,2,3,79] Their capacity to reshape genomes and modulate gene expression makes them critical regulators in various biological contexts [Table 2].

Table 2.

Impact of representative TEs in human development and disease.

Development and disease	TE types	Impact
Embryonic development
ZGA	L1, SVA, ERVs (HERVH, LTR7)	Activates ZGA and facilitates ZGA exit; Facilitates chromatin remodeling; Activates pluripotency genes.
Placental development	MER50, RLTR13	Promotes invasive placentation.
Neuro-development disorders
Neurodevelopment	L1, HERV, Alu, SVA	Regulates neural differentiation, synaptic plasticity, and brain circuitry.
Schizophrenia	L1	Disrupts synaptic genes and neuronal signaling.
Rett syndrome	L1	Genome destabilization by retrotransposition.
Autism pectrum disorder	L1, Alu	Alters synaptic connectivity and gene expression.
Ataxia-telangiectasia	L1	Purkinje cell dysfunction and degeneration.
Alzheimer’s disease	L1, Alu, HERV	Induced accumulation of G-quadruplex structures; triggers inflammatory pathways and apoptosis in non-neural cells
Amyotrophic lateral sclerosis	HERVK	Triggers neuroinflammation, neuron spreading, and neuron apoptosis
Aicardi-Goutières syndrome	L1	cGAS–STING pathway; neuroinflammatory activation.
Immune-related disease
Systemic lupus erythematosus	L1, HERV	Triggers chronic inflammation and autoantibody production.
Rheumatoid arthritis	HERVK	Drives synovial inflammation and joint damage.
Cancer
Breast, colon, liver, prostate, skin cancer	L1, ERV, HERVK, SVA	Promotes tumorigenesis and immune evasion.

cGAS-STING: Cyclic GMP-AMP synthase - stimulator of interferon genes; ERV: Endogenous retroviruse; HERVH: Human endogenous retrovirus H; HERVK: Human endogenous retrovirus K; L1: Long interspersed nuclear elements 1; LTR: Long terminal repeat; SLE: Systemic lupus erythematosus; SVA: SINE-variable numbers of tandem repeats (VNTR)-Alu; TE: Transposable element; ZGA: Zygotic genome activation.

Embryonic development

TEs exhibit dynamic, stage-specific regulatory roles in embryonic development.^[19,80] Distinct TE families show precise temporal activation patterns: the LTR7Y (HERVH) subfamily is active during blastocyst formation, while MLT2A1 (HERVL) peaks at the eight-cell stage.^[81] These stage-specific activities highlight TEs as critical developmental regulators.

During zygotic genome activation (ZGA)—the transition from maternal to zygotic control of gene expression—LTRs and L1 elements are reactivated and function as enhancers or alternative promoters to modulate the expression of genes essential for embryogenesis.^[80] L1 is highly expressed in preimplantation embryos and is subsequently downregulated during differentiation; its repression impairs chromatin accessibility and blastocyst formation.^[82] Percharde et al^[59] demonstrated that L1 RNA acts as a nuclear scaffold in mouse embryonic stem cells (ESCs), facilitating ribosomal DNA expression and repressing DUX, a master regulator of the ZGA program. In humans, L1 RNAs repress gene activation in 8-cell-like cells, which correspond to the ZGA stage.^[83] Recently, Li et al^[10] found that transcriptionally active L1s function as enhancers that selectively contact and activate minor ZGA genes in early mouse embryos. Depletion of L1 RNA abolished the activation of nearly half of these genes, resulting in embryo arrest at the 2–4-cell stage. Collectively, these findings suggest that L1 transcription and RNA initially act in cis to promote ZGA gene expression and initiate ZGA and subsequently act in trans to repress ZGA genes and facilitate ZGA exit.

Beyond ZGA, TEs also play essential roles in placental development.^[84,85] In mammals, MER50 and RLTR13 ERVs provide key enhancers that regulate genes involved in invasive placentation—a defining trait of eutherian mammals.^[86] The ERV-derived envelope (env) gene encodes proteins, such as syncytin, which are crucial for trophoblast fusion, proliferation, and antiviral defense.^[87] Besides, a mouse-specific retrotransposon, MT2B2, provides an alternative promoter that drives expression of an N-terminally truncated isoform of Cdk2ap1, which plays a key role in embryo implantation.^[88] Moreover, a systematic scan has identified 1507 human env-derived ORFs, including truncated forms, such as Suppressyn and Refrex-1, which protect placental cells via receptor interference.^[89]

While TEs clearly contribute to developmental regulation, their evolutionary importance requires nuanced interpretation. Recent studies challenge the widespread assumption that TE-derived regulatory elements are universally essential. Instead, many may represent lineage-specific adaptations with limited functions or evolutionarily neutral “genomic noise” lacking conserved biological impact.^[90] For example, the MER50 endogenous retrovirus-derived enhancer plays a critical role in placental development, specifically in eutherian mammals, displaying strict trophoblast-specific activity and being entirely absent in non-placental species.^[91] These findings underscore the need for a more nuanced framework to assess TE contributions to gene regulation—one that considers both their biological relevance and the frequent lack of conservation across species.

Neural development and disorders

TEs play crucial regulatory roles in brain development through multiple mechanisms: (1) Somatic L1 retrotransposition occurs frequently in neural precursor cells, potentially contributing to genome plasticity and gene regulatory programs.^[92,93] (2) L1 elements regulate genes involved in neurogenesis, shaping the transcriptional landscape of neural progenitors and stem cells.^[94] (3) Alu elements modulate the expression of neurotransmitter receptors and growth factors, thereby influencing synaptic plasticity.^[95] These findings suggest that TEs actively participate in regulating key genes involved in neural differentiation, synaptic plasticity, and neurogenesis.

Dysregulation or over-activation of TEs contributes to the pathogenesis of neurological disorders through multiple pathways: (1) Somatic L1 insertions in the prefrontal cortex and hippocampal neurons can disrupt synaptic genes (e.g., DISC1) and alter neuronal signaling pathways in patients with schizophrenia.^[96] (2) Alu elements regulate neuroligins and neurexins, whose dysregulation is implicated in autism spectrum disorders (ASDs), characterized by impaired social communication and repetitive behavior.^[97] In addition, TE-induced genomic instability may cause copy number variations (CNVs), which are frequently observed in patients with ASD and are believed to contribute to disease etiology.^[98] (3) In Rett syndrome, mutations in MeCP2 lead to L1 hypomethylation and increased L1 retrotransposition, resulting in genomic instability and disease progression.^[99] (4) L1 activation contributes to Purkinje cell dysfunction and degeneration in the mouse cerebellum, driving the onset of ataxia.^[100] This finding is supported by increased L1 mobility observed in patients with ataxia-telangiectasia.^[101] (5) In Alzheimer’s disease (AD), active L1 elements promote G4 structure accumulation, potentially disrupting gene expression in neurons.^[102] Alu RNAs have also been shown to activate inflammatory pathways and induce apoptosis in non-neural cells, including retinal pigment epithelial cells.^[103] In addition, cytoplasmic nucleic acids derived from HERVs stimulate innate immune sensors, leading to type I IFN production and neuroinflammation in AD.^[104] (6) In amyotrophic lateral sclerosis (ALS), elevated TDP-43 expression enhances HERVK activity, triggering neuroinflammation and motor neuron apoptosis through neuronal spreading.^[105] (7) In Aicardi–Goutières syndrome, the accumulation of L1-derived single-stranded DNA activates the cGAS–STING pathway, leading to neuroinflammation.^[106] Overall, these findings highlight the conserved role of TE dysregulation in neurological disorders and support the potential for therapeutic targeting of L1 activity, including epigenetic modulation, anti-inflammatory strategies, and genome stabilization therapies.

Immune response

TEs influence both innate and adaptive immunity, as well as autoimmune disease pathogenesis, highlighting their complex and multifaceted roles in human health.^[62,107,108]

Innate immunity, the body’s first line of defense against pathogens, is activated in part by TEs.^[108] For example, L1-derived dsRNA activates RIG-I and MDA5, triggering a type I IFN response.^[75] In parallel, L1 cDNA can stimulate the cGAS pathway.^[109] HERVK also engages RNA sensors, such as TLR7 and TLR8, and DNA sensors, including cGAS and AIM2.^[110,111] Beyond these trans-effects, ERVs and L1 elements can act in cis as IFN-inducible enhancers that activate IFN-stimulated genes and modulate the IFN pathway.^[69,107] In addition, ERVs recruit TFs such as IRF3, STAT1, and NF-κB, promoting immune activation and inflammation.^[71] Although these interactions remain context-dependent and mechanistically underexplored, further research is required to clarify the specific pathways involved and their implications in various disease settings.

TEs also influence adaptive immunity by regulating T-cell development and function.^[112] L1, HERV, and SVA elements contribute to T-cell receptor signaling and differentiation by serving as regulatory sequences for genes essential to T-cell activation.^[83,113] HERV-derived sequences have been shown to regulate the expression of key cytokines and TFs such as FOXP3, which is critical for developing regulatory T cells that maintain immune tolerance and prevent autoimmunity.^[114]

Although TEs are crucial for normal immune function, their dysregulation can contribute to immune-related diseases, including autoimmunity and chronic inflammation.^[108] For instance, TEs have been linked to T-cell exhaustion in chronic infections, impairing immune surveillance and response.^[58,115] TE overactivation also induces genomic instability and persistent inflammation, contributing to systemic lupus erythematosus (SLE), rheumatoid arthritis (RA), and MS.^[116] In SLE, aberrant activation of L1s and HERVs leads to the production of dsRNA and retrotransposon-derived transcripts that amplify type I IFN response.^[117] Dysregulated TE expression in immune cells, including plasma and dendritic cells, exacerbates autoimmunity and drives tissue damage and disease progression.^[118] HERVK has been implicated in RA-associated synovial inflammation, where its expression in synovial fibroblasts promotes cytokine production and immune cell infiltration into joints.^[118] Interestingly, TE activation is also observed during acute viral infections,^[119] although its functional significance remains unclear. Is TE activation a programmed component of antiviral defense or merely a byproduct of infection-induced transcription? Further mechanistic studies are needed to resolve these context-dependent roles and assess their therapeutic potential.

Aging and aging-related disease

As organisms age, TEs become more active and contribute to genomic instability.^[120] Somatic mutations arising from TE-induced genomic instability accumulate over time and are associated with age-related diseases, such as neurodegeneration and cancer.^[121] In aging neurons, L1 activation can lead to increased DNA damage and the formation of double-strand breaks, which are inefficiently repaired in aging cells.^[122] This accumulation of DNA damage contributes to cellular dysfunction and the aging process.^[122]

Chronic, low-grade inflammation is another well-documented feature of aging.^[121] TEs play a significant role in driving this inflammation by activating immune responses. For example, derepressed HERVK elements promote cellular senescence through activating the cGAS-STING pathway in premature aging models.^[123] Moreover, cytoplasmic accumulation of L1 DNA during aging can trigger type I IFN responses, thereby promoting age-related inflammation.^[121] Notably, this effect can be reduced by nucleoside reverse transcriptase inhibitors, suggesting that L1 represents a potential target for mitigating age-associated inflammatory conditions.^[124]

A recent study revealed a mechanism by which derepressed TEs contribute to aging: the SVA composite transposon promotes myeloid-biased hematopoiesis via its enhancer activity.^[73] This evidence suggests that the age-associated increase in TE activity and their accompanying enhancer functions constitute a potent force driving the aging process.

Cancer

TEs are frequently reactivated in cancer cells, leading to somatic retrotransposition and genomic instability.^[58,59] TE retrotransposition can insert new genetic materials into coding or regulatory regions, resulting in mutations that disrupt oncogenes or tumor suppressor genes.^{[1,59,125,126]} TE activation in tumor cells is also associated with CNVs, which are prevalent in cancer and linked to poor prognosis.^[59]

Despite their tumor-promoting effects, reactivated L1s can also induce genome instability, activate DNA damage response, and exert tumor-suppressive roles, particularly in myeloid leukemia.^[115] In addition, TEs influence tumor progression by modulating immune responses that may suppress cancer development.^{[7,44,76,127]} TE-derived RNAs and their RT products (complementary single-stranded DNA or DNA/RNA hybrids) can activate nucleic acid-sensing pathways, including RNA receptors like RIG-I-like receptors, Toll-like receptor 3, and the DNA sensor cGAS enzyme.^[115,125] Moreover, the translation of TE-host chimeric RNAs can generate novel peptides, broadening the range of tumor antigens.^{[58,128,129,130]} Thus, TEs’ ability to stimulate innate immune sensors and generate tumor-specific antigens makes them promising targets for cancer immunotherapy.^[131]

TEs in Diagnostics and Therapy

TEs have emerged as valuable tools in precision medicine, serving as diagnostic biomarkers and therapeutic targets due to their disease-specific activation patterns and immunogenic properties [Figure 1D, 1E and Figure 2].

TEs as biomarkers

TEs are often reactivated in cancer and other diseases, making them promising diagnostic biomarkers.^[58,79] Hypomethylation of L1, and L1 ORF1, are associated with worse clinical outcomes in cancers, indicating its potential as an early biomarker for cancer detection.^[126] Somatic retrotransposition of L1 and other TEs is linked to tumorigenesis in breast, colon, and liver cancers.^[59] L1-derived sequences can be detected in circulating tumor DNA (ctDNA), exosomes, and plasma, offering a noninvasive approach for tumor detection and monitoring.^[132] In addition, Alu elements are upregulated in specific cancers and may serve as cancer-specific biomarkers.^[133]

Beyond oncology, TEs show diagnostic potential for neurodegenerative diseases, including Alzheimer’s disease (AD) and Parkinson’s disease (PD).^[134,135] Reactivation of L1 and HERVs in neurons correlates with neuroinflammation and neurodegeneration, suggesting that TE activity in cerebrospinal fluid or serum could track disease progression and therapeutic response.^[136]

Tumor-specific antigens and immunotherapy

TE reactivation in cancer cells generates immunogenic neoantigens, which are absent in healthy cells due to tight epigenetic silencing.^{[58,115,128,129,130]} These TE-derived peptides are presented via MHC molecules, triggering T-cell and antibody responses against tumors.^{[58,115,128,129,130]} For example, L1-encoded proteins (ORF1 and ORF2) are highly expressed in cancers (e.g., breast, colon) and recognized by cytotoxic T lymphocytes.^[58,115,125] The L1 antisense promoter-derived L1PA2–GABRG2 fusion peptides, presented on the surface of cancer cells, represent new potential targets for cancer therapy.^[128] Besides, HERVK and Alu produce tumor-specific peptides, expanding the pool of targets for cancer vaccines and personalized immunotherapy.^[58,115,125] Moreover, a recent study showed that HIF2α upregulates intact ERVs (e.g., ERVE-4, ERV3.2) in clear cell renal carcinoma, whose peptides elicit T-cell responses post-immunotherapy or transplantation.^[137] Pharmacological HIF2α stabilization induces similar ERV expression in other cancers, highlighting broader therapeutic potential.^[137]

Epigenetic modulation of TEs for therapy

Epigenetic drugs can reactivate silenced TEs, enhancing type I IFN responses, and expose TE-derived neoantigens to improve immune recognition.^[79,129] For example, DNMT/HDAC inhibitors promote TE expression, boost type I IFN responses, and expose TE neoantigens for immune recognition.^[79,129] Besides, H3K27M mutations in pediatric gliomas redistribute H3K27 acetylation, derepress ERVs, and sensitize tumors to combined treatment with DNMT/HDAC inhibitors.^[138] Moreover, depleting or inhibiting histone modifiers (e.g., KDM1A, KDM5B, SETDB1) reactivates TEs, triggers interferon signaling, and synergizes with immune checkpoint blockade.^{[139,140,141]} These approaches hold promise for advancing epigenetic therapies and improving cancer treatment outcomes. However, challenges such as tumor heterogeneity in TE expression and immune evasion mechanisms remain barriers to efficacy.

TE depletion via genome or RNA therapy

Strategies aimed at silencing TE can reduce genomic instability and impede disease progression.^[58,142] For example, L1 knockdown has been shown to suppress tumor growth, highlighting the therapeutic potential of TE silencing in cancer and other disorders associated with genomic instability.^[143,144] Moreover, antisense oligonucleotides targeting L1 have been shown to extend the lifespan of mice by approximately 25%, suggesting their potential as antiaging interventions.^[145]

RT inhibition

RT inhibitors—such as 3TC, originally developed for the human immunodeficiency virus—have been repurposed to block retrotransposition in clinical trials for AD, PD, ALS, cancer, lupus, and aging.^{[146,147,148]} Although these drugs can slow tumor progression, they typically fail to eliminate malignant cells entirely, likely because TEs are not the sole drivers of these complex diseases.^[147] Consequently, combination therapies are often required. In addition, given the physiological roles of TEs, there are concerns about off-target effects from RT inhibitors.^[148]

Harnessing technological innovations for TE-targeted therapies

TE research has long been hindered by their repetitive nature and tight epigenetic silencing, which obscure detection of their mobilization, expression, and function. Recent advances in single-cell RNA sequencing (scRNA-seq), long-read sequencing and machine learning are now overcoming these barriers, revealing unprecedented insights into TE biology and their contributions to health and disease.

scRNA-seq: mapping TE dynamics at cellular resolution

scRNA-seq enables high-resolution profiling of gene expression, revealing how TEs influence critical biological processes such as cell differentiation, immune responses, and tumor progression.^[149,150] Complementary single-cell epigenomic techniques, such as single-cell ATAC-seq, further elucidate the chromatin dynamics governing TE activation or silencing.^[151] These approaches offer valuable perspectives on the contribution of TEs to aging, cancer, and immune dysfunction.

Long-read sequencing: overcoming the repetitive challenge

Conventional short-read sequencing struggles to resolve repetitive TE sequences, whereas long-read platforms (e.g., PacBio and Oxford Nanopore) can accurately assemble full-length TE sequences and detect mobilization events and genomic rearrangements.^[152,153] For example, long-read sequencing of ctDNA could enable non-invasive detection of TE-driven mutations in cancer. However, long-read sequencing is often non-quantitative, expensive, and difficult to resolve full-length L1 insertions. Overcoming these challenges will be essential to fully leverage its potential for TE research and precision medicine.

Machine learning for TE analysis

The development of computational tools for TE analysis from long-read data represents an emerging frontier. machine learning approaches trained on multi-omic datasets show particular promise for identifying immunogenic TE-derived epitopes, which have been notoriously difficult to predict due to the complexity of immune recognition.^[154] Tools such as NeoFuse and nextNEOpi utilize machine learning algorithms, including linear regression and artificial neural networks, to predict neoantigens from TE-containing RNAs.^[155,156] While machine learning enhances high-throughput screening for TE-based immunotherapies, further validation, such as cross-referencing with experimental data, is needed to refine these models for clinical use.

Expanding the therapeutic toolkit

The recent identification of 40 novel integration-competent DNA TEs in the human genome, along with development of associated molecular toolkits, highlights the growing potential for TE-based therapeutic applications.^[157] Thus, as we deepen our understanding of fundamental TE biology, these discoveries are catalyzing innovative approaches to genome engineering and targeted therapies, presenting exciting opportunities for novel prognostic markers and targeted therapies.

Conclusions

TEs, once dismissed as “junk DNA”, are now recognized as dynamic regulators of human biology. Their tightly regulated activity shapes neural development, immune responses, and aging, while their dysregulation contributes to diverse diseases, including cancer, autoimmune disorders, and neurodegenerative diseases. Advances in single-cell omics, long-read sequencing and machine learning have enabled unprecedented resolution in mapping TE activity and interactions, paving the way for novel TE-based biomarkers and targeted therapies. While clinical translation faces challenges, such as cell-type specificity and off-target effects, TEs offer immense potential for advancing precision medicine by bridging genetic variation, gene regulation, and disease mechanisms.

Funding

This work was supported by grants from Beijing Natural Science Foundation (No. JQ24033), the National Key Research and Development Program (No. 2022YFA1302700), the National Natural Science Foundation of China (Nos. 32350011, 32270593, 32070631, and 32400441), and the China Postdoctoral Science Foundation (No. 2024T170499). We thank the support from the Benyuan Fund-Young Investigator Exploration Grant in Life Sciences, Tsinghua University Initiative Scientific Research Program and Pillars of the Nation Funding for Life Sciences, Tsinghua university.

Conflicts of interest

None.