Identification of a novel RSPO1-NUMT insertion in a LUAD patient cohort and the challenges and insights into NUMT detection highlighting the importance of reference genomes and population databases

Introduction

Nuclear mitochondrial DNA sequences (NUMTs) represent a fascinating intersection of cellular compartments, where fragments of mitochondrial DNA (mtDNA) become integrated into the nuclear genome. This integration process occurs continuously across both evolutionary and somatic timescales, serving as “molecular fossils” that document the historical genetic transfer during mitochondrial endosymbiosis and the ongoing dynamic relationship between mitochondrial and nuclear genomes (1-3). The formation of NUMTs is primarily driven by double-strand break repair mechanisms, particularly through non-homologous end joining (NHEJ). In humans, this process occurs at an estimated rate of 5×10⁻⁶ per germ cell per generation (2,4). These insertions preferentially target open chromatin regions and fragile genomic sites, potentially influencing gene regulation and genomic stability (5,6). Recent whole-genome sequencing (WGS) studies have identified de novo NUMT formations, confirming the ongoing nature of this process (3,6).

The clinical significance of NUMTs has become increasingly apparent, particularly the impact in cancer research and the disease pathology. Studies have revealed elevated NUMT abundance in colorectal cancer genomes, correlating with increased mortality rates and mutations in NUMT-related genes such as YME1L1 (3,7). Similar patterns have been observed across various malignancies (2,4). Furthermore, NUMTs exhibit substantial polymorphisms across human populations, with some ultra-rare variants present in less than 0.1% of individuals (6).

Despite the growing research into NUMT frequency and distribution patterns (3,6), a critical knowledge gap persists in understanding their functional impact on nuclear gene expression, particularly in cancer-related genes and tumorigenic processes. Our study addresses this gap by identifying a NUMT insertion in the R-spondin 1 (RSPO1) gene (ENSG00000169218) which could play an integral role in growth and metastasis pathways of lung adenocarcinoma (LUAD) (8,9). In this study, we combine our in silico analysis with molecular-genomic experiments to further characterize and validated our findings across an interesting LUAD cohort incorporating important clinical parameters. Our study design is outlined in Figure 1. We present this article in accordance with the STROBE reporting checklist (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-586/rc).

Figure 1 Study design flowchart. Summary of the in silico (WGS dataset-The Cancer Genome Atlas) and molecular genomic profiling performed for NUMT identification and characterization in the LUAD cohort with clinical observations. LUAD, lung adenocarcinoma; NUMT, nuclear mitochondrial DNA sequence; PCR, polymerase chain reaction; WGS, whole-genome sequencing.

Methods

In silico identification and annotation of NUMT insertions

WGS data from 38 LUAD samples and 38 matched control samples (non-tumor) were obtained from The Cancer Genome Atlas (TCGA) database. The human reference genome hg38 (GRCh38) and mitochondrial genome reference sequence in hg38 were used (Appendix 1).

FASTQ files were aligned to the human reference genome build hg38, using the default setting for the GATK4 pipeline “processing-for-variant-discovery-gatk4” (https://github.com/gatk-workflows/gatk4-data-processing/blob/master/processing-for-variant-discovery-gatk4.wdl), generating BAM files. NUMT insertions were identified using the dinumt tool (1) and the published method based on paired-end read proportions (7) in both LUAD and control samples (Appendix 1).

The resulting VCF files were annotated with AnnotSV (10-12) providing functional and structural information. Custom R scripts were employed to filter and extract NUMTS that interfered with the coding regions or untranslated regions (UTRs) of genes for each sample. Finally, selected NUMTs were validated using the Manta tool for structural variant detection (13).

Patient cohorts and DNA samples

DNA samples were isolated from 298 LUAD tumor tissues obtained through collaborations with the accredited Lung Biobank Heidelberg at Thoraxklinik (Heidelberg, Germany), member of the BioMaterialBank Heidelberg and the platform biobanking of the German Center for Lung Research (DZL) (Figure 2). DNA extraction and isolation was performed with the AllPrep DNA/RNA/miRNA Universal Kit (QIAGEN, Germany) according to the Simultaneous Purification of Genomic DNA and Total RNA, including miRNA, from Tissues- protocol. Only samples with a tumor content >50% were considered for the analyses. All histopathological diagnoses were made according to the 2015 WHO classification for lung cancer by at least two experienced pathologists. Tumor stage was designated according to the 8th edition of the UICC tumor, node, and metastasis. All patients had signed informed consent. The study was approved the ethics committee of the Medical Faculty Heidelberg [use of biomaterial and data (S-270/2001)] and the ethics committee of the General University Hospital in Prague (proof of project 2157/19 S-IV Grant). The study was performed according to the principles set out in the Declaration of Helsinki and its subsequent amendments. For six individuals that had the homozygote-profile, DNA from tumor and matched distant (>5 cm) tumor-free lung tissue samples, as well as blood samples were tested. Clinical parameters [e.g., sex, age, smoking history, BMI, stage, grade, histological subtypes, overall survival (OS), disease-free survival (DFS)] of the LUAD cohort are shown in Table 1.

Figure 2 Schematic of locations of the lung tissue/tumor DNA sampling and the principle behind NUMT occurrences. (A) Location of the different lung tumor sections that the LUAD cohort DNA samples (n=298) were isolated from. (B) A detailed view of the principle behind the NUMT fragment/site insertion phenomenon. LUAD, lung adenocarcinoma; mtDNA, mitochondrial DNA; NUMT, nuclear mitochondrial DNA sequence.

PCR-based screening assay and Sanger sequencing of NUMT insertion in the RSPO1 gene

A PCR-assay was designed to screen for the 70 base pairs (bp) NUMT insertion located at chromosomal position chr1:37,611,748 (UCSC Genome browser on human Dec. 2013, GRCh38/hg38 assembly) within the RSPO1 gene (ENSG00000169218); primers flanking the surrounding region: forward primer: 5'- GTCACAGGGCAGCTTGTTAG-3'; and reverse primer: 5'-TTGTCCCCAACCATCACTG-3' (Generi-Biotech, Czech Republic) (Figure 3). PCRs were performed on DNA using the standard manufacturer’s protocol for Combi PPP Master Mix (TopBio, Czech Republic) with an annealing temperature of 59 ℃. In the presence of the NUMT insertion, the size of the PCR product increased to 399 bp versus the expected 329 bp. This difference was detectable by 1.5% agarose gel electrophoresis (Figure 3).

Figure 3 Schematic of the RSPO1 gene region and the molecular-genomic analysis (PCR and sequencing) in a LUAD patient cohort. (A) Representation of the RSPO1 chromosome 1 region depicting the exons, introns, and UTRs, highlighting the 3’UTR region where the specific location of the 70-base pair NUMT insertion site (primer positions highlighted with black arrows, insertion site indicated in red) (genome location: chr1:37611748, hg38). (B) The gel electrophoresis results show the distinct PCR amplification profiles observed across the tested sample sets. Three primary profiles were identified: Profile 1, representing the reference sequence without NUMT insertion (single 329 bp band); Profile 3-homozygous RSPO1-NUMT insertion (399 bp band), and Profile 2-heterozygous NUMT insertion—two bands sized 329 and 399 bp. (C) Illustrates the Sanger sequencing chromatogram confirming the presence of the RSPO1-NUMT insertion in Profile 3-homozygous. The sequence alignment reveals the precise integration site of the 70 bp NUMT within the 3'UTR region of the RSPO1 gene. The junction sequences at both the 5' and 3' ends of the insertion demonstrate clean integration. LUAD, lung adenocarcinoma; mtDNA, mitochondrial DNA; NUMT, nuclear mitochondrial DNA sequence; PCR, polymerase chain reaction; RSPO1, R-spondin 1; UTRs, untranslated regions.

Further confirmation and validation of the 70 bp RSPO1-NUMT insertion was performed by Sanger sequencing using the 3500xL Genetic Analyzer (Applied Biosystem, USA) on PCR products from homozygous tumor DNA samples.

NGS sequencing and data analysis in LUAD homozygote samples

Final DNA libraries were sequenced primarily using the Illumina NextSeq 550 Dx instrument. Raw data from sequencing in FASTQ format underwent quality control and standard bioinformatic pipelines for DNA-Seq panel analysis, which were developed in Genovesa, (a web-based platform with comprehensive tools for analyzing genomic data, clinical interpretation, etc. compatible with sequencing technologies e.g., Illumina, MGI, Oxford Nanopore Technologies…) (Appendix 1). The DNA pipeline included standard sequencing variant detection, CNV, MSI, and tumor mutational burden (TMB) analysis. After final data analysis and evaluation, external databases—OncoKB, ClinVar, Varsome and COSMIC and pathway analyses (Ingenuity Pathway) were used to obtain disease-related pathways and to model relevant biological processes.

Molecular and clinical data correlation and statistical analysis

Statistical analysis of associations between molecular profiles and clinical characteristics was performed using Fisher’s exact test for categorical variables, Welch’s t-test for numerical variables, and Multiple Linear Regression Model. All reported P values were considered statistically significant at P≤0.05. The analysis results created using statistical Program R (version 4.2.2).

Results

In silico identification and analysis of NUMTs

The NUMT presence in LUAD was investigated, by analyzing the TCGA WGS dataset from LUAD samples (n=38) and matched controls (n=38) using a computational pipeline. A ~2-fold increase in NUMT counts were identified in tumor samples compared to control samples (tumor: average 28-median 12 NUMT per sample vs. control average 13.47-median 13 NUMT) from lung tissue or peripheral blood. The observed difference in mean NUMT counts was primarily attributed to a subset of tumor sample displaying exceptionally high NUMT numbers. Our systematic analysis of NUMT insertion sites aimed to assess the potential functional impact on gene expression and disease pathology. The focus was on identifying NUMTs that disrupted either coding sequences or located within UTRs of genes. One of the significant findings was a 70 bp NUMT is part of the MT-ATP6 gene (chrM:8527-9207) encoding ATP6 subunit of F₁F_O-ATPsynthase chrM:8933-9008 (NC_012920.1) was inserted in the 3'UTR region of the RSPO1 gene in four patients (approximately 10% of the study cohort). To investigate the molecular drivers of LUAD, we analyzed targetable mutations in all available in silico LUAD samples, with a focus on the identification of tumor driver genes. WGS data analysis revealed a potential association between NUMT insertions and known oncogenic pathways in LUAD development. Specifically, two patients with NUMT insertions exhibited concurrent KRAS mutations (c.34G˃T p.Gly12Cys and c.35G>T p.Gly12Val), and one patient harbored an EGFR mutation (c.2156G>C p.Gly719Ala)—both well-established drivers of LUAD tumorigenesis. No additional targetable mutations were detected in the analyzed samples.

Molecular findings: RSPO1-NUMT PCR-assay in the LUAD cohort

Our RSPO1-NUMT PCR-screening assay successfully identified three distinct PCR amplification profiles across the tested sample set. Profile 1 representing the RSPO1 reference (no NUMT insertion) sequence (single amplicon band of 329 bp). Profile 2 showed two bands: a heterozygous RSPO1-NUMT insertion (399 bp) and the reference sequence (329 bp). Profile 3, indicating homozygous RSPO1-NUMT insertion resulted in a 399 bp single band (Figure 3).

The frequency of these profiles varied across the study LUAD samples (n=298), the profiles were as followed: Profile 1 observed in 68% (203/298), Profile 2-heterozygotes were detected in 29.8% (89/298), while Profile 3-homozygotes was observed in 2% (6/298) of cases.

Further characterization of the homozygous RSPO1-NUMT insertion profile

The RSPO1-NUMT PCR-screening was performed on the homozygous individuals’ DNA samples from blood, non-tumor (‘normal’) lung tissue, and lung tumor to examine whether the NUMT insertion was a somatic or germline occurrence. The same Profile 3-homozygous RSPO1-NUMT insertion, was detected across normal-lung, blood and tumor, indicating a germline origin in all six cases (Figure 3).

The RSPO1-NUMT insertion was further investigated via Sanger sequencing in the homozygous samples which confirmed the presence of the NUMT insertion in Profile 3. The Sanger sequencing alignment revealed the precise integration site of the 70-bp NUMT within the 3'UTR region of the RSPO1 gene. The junction sequences at both the 5' and 3' ends of the insertion demonstrate clean integration without additional modifications, suggesting a precise insertion event. The chromatogram shows clear peak separation and high signal-to-noise ratio, validating the sequence authenticity (Figure 3). Blasting the sequenced in the USCS genome browser identified the insertion site to be at chr1:37,611,748 (hg38-genome browser build).

Following in silico molecular characteristics for targetable somatic tumor mutations, we confirmed only one pathogenic KRAS mutation through additional validation of six RSPO1-NUMT homozygotes. Furthermore, TMB was found to be low, and microsatellite stability (MSS) was observed across the cohort. These findings suggest a relatively limited mutational landscape in RSPO1-NUMT homozygotes, suggesting presence of NUMT insertions in the context of a genetically stable tumor environment (Table 2).

Correlation of RSPO1-NUMT profiles with clinical data: clinical implications

Our statistical analysis has unveiled compelling associations between RSPO1-NUMT profiles and key clinical parameters in our patient cohort. These parameters include age at diagnosis, BMI, smoking status, pack-years, pathological stage, grade, histology subgroups, OS, and DFS. Notably, age at diagnosis emerged as a potential significant factor, with Profile 3-homozygous (RSPO1-NUMT insertion) showing a markedly earlier disease onset compared to both Profile 1 (no NUMT insertion) and Profile 2-heterozygous carriers (P values =0.04 and, 0.041, respectively). The mean age at diagnosis for Profile 3-homozygous carriers was 54 years, in contrast to 63.3 years for Profile 1 and 2 carriers.

Furthermore, our research suggests that RSPO1-NUMT profiles, particularly Profile 3, might have significant prognostic implications. Patients with Profile 3-homozygotes were more likely to be diagnosed at an advanced pathological stage compared to both healthy individuals and heterozygotes for NUMTs in the RSPO1 gene (P=0.07 and P=0.086, respectively). No significant differences were observed between profiles for sex, BMI, smoking status, pack-years, pathological stage, grading, OS, or DFS. These findings hint at a potential link between NUMT profiles and more aggressive disease progression and earlier onset of LUAD, further underscoring their potential prognostic significance.

Discussion

The RSPO1 gene encodes one of the most potent ligands of the WNT1 receptor (14). The WNT signaling pathway maintains lung homeostasis and plays a key role in LUAD carcinogenesis (14,15). In LUAD, two functionally distinct cell populations exist: one producing WNT receptor ligands and another responding through pathway activation. We hypothesize that the identified NUMT insertion increases RSPO1 expression by affecting silencer regions, potentially promoting carcinogenesis, tumor vascularization, drug response, and disease progression through the WNT signaling pathway (15,16).

Impact of reference genome versions and technical challenges in NUMT detection

A critical finding is the difference in reported frequency of the RSPO1-NUMT insertion between reference genome versions (17,18). In hg19/GRCh37, where the insertion is present in the reference, gnomAD v2.1.1 reports an 81% deletion frequency. Conversely, in hg38/GRCh38, where it is absent from the reference, gnomAD v3.1.2 and v4.0.0 report very low frequency (0.00079) (Appendix 1).

This discrepancy stems from significant challenges in detection and mapping large insertions (17-19). In hg38/GRCh38, reads containing the insertion are often discarded during mapping due to the absence of the corresponding sequence, while in hg19/GRCh37, they map accurately. These challenges are compounded when the insertion size approaches the detection limits of tools like GATK’s haplotype caller, highlighting current mapping algorithm limitations.

Our PCR findings show a significant proportion of the population carrying the insertion. Population-level database comparisons further complicate interpretation: the Enigma database indicates a 16% frequency of a shortened variant (~10 bp), while TCGA reports 0.000492 for a ~40 bp variant. These differences likely result from sequencing technologies, detection thresholds, or population-specific genetic variation (17,19-22).

Insights from PCR analysis and oncogenic relationships

Our PCR-screening detected three distinct profiles (no-insertion NUMT, heterozygous, and homozygous carriers), identifying potentially functionally relevant homozygous NUMT insertions in LUAD pathogenesis. Homozygous carriers for the RSPO1-NUMT insertion showed trends toward earlier disease onset and advanced-stage LUAD at diagnosis, suggesting these possible hereditary insertions may represent previously unrecognized predisposing factors.

We observed co-occurrence of NUMT insertions with established oncogenic drivers like KRAS and EGFR mutations in silico. The limited mutational landscape in RSPO1-NUMT homozygote samples—only one confirmed pathogenic KRAS mutation among six patients and no additional targetable alterations (23)—suggests possible connections between genomic stability and NUMT presence which warrant further investigation.

Conclusions and future direction

While NUMTs pose technical challenges for mitochondrial genetics, they provide valuable insights into evolutionary biology and population genetics (3,6,24), serving as molecular markers for mitochondrial ancestry and nuclear-mitochondrial genome interactions.

Future research should examine NUMT frequency and functional impact on gene expression in cancer, using matching control tissues to differentiate germline from somatic alterations. Optimizing computational pipelines across genome builds is essential, as is leveraging long-read sequencing to improve large insertion detection in diverse genetic backgrounds.

Acknowledgments

We thank all study participants and the Translational Research Unit staff at Thoraxklinik, Heidelberg University Hospital for recruitment and sample collection support. Special thanks to J. Votruba and M. Koziar Vasakova for research facilitation.

Footnote

Reporting Checklist: The authors have completed the STROBE reporting checklist. Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-586/rc

Peer Review File: Available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-586/prf

Funding: This study was supported by the PRIMUS Research Programme, Charles University (PRIMUS/20/MED/007); institutional project MH CZ–DRO–VFN00064165; Fond of EXCELLENCE1-223011/1LF. EP’s work partially funded by MH CZ – DRO – TUH 00064190.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tlcr.amegroups.com/article/view/10.21037/tlcr-2025-586/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All patients had signed informed consent. The study was approved the ethics committee of the Medical Faculty Heidelberg [use of biomaterial and data (S-270/2001)] and the ethics committee of the General University Hospital in Prague (proof of project 2157/19 S-IV Grant). The study was performed according to the principles set out in the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Dayama G, Emery SB, Kidd JM, et al. The genomic landscape of polymorphic human nuclear mitochondrial insertions. Nucleic Acids Res 2014;42:12640-9. [Crossref] [PubMed]
2. Singh KK, Choudhury AR, Tiwari HK. Numtogenesis as a mechanism for development of cancer. Semin Cancer Biol 2017;47:101-9. [Crossref] [PubMed]
3. Xue L, Moreira JD, Smith KK, et al. The Mighty NUMT: Mitochondrial DNA Flexing Its Code in the Nuclear Genome. Biomolecules 2023;13:753. [Crossref] [PubMed]
4. Ju YS, Tubio JM, Mifsud W, et al. Frequent somatic transfer of mitochondrial DNA into the nuclear genome of human cancer cells. Genome Res 2015;25:814-24. [Crossref] [PubMed]
5. Liang BC. Evidence for association of mitochondrial DNA sequence amplification and nuclear localization in human low-grade gliomas. Mutat Res 1996;354:27-33. [Crossref] [PubMed]
6. Wei W, Schon KR, Elgar G, et al. Nuclear-embedded mitochondrial DNA sequences in 66,083 human genomes. Nature 2022;611:105-14. [Crossref] [PubMed]
7. Srinivasainagendra V, Sandel MW, Singh B, et al. Migration of mitochondrial DNA in the nuclear genome of colorectal adenocarcinoma. Genome Med 2017;9:31. [Crossref] [PubMed]
8. Lou X, Wang Y, Deng Y, et al. Prognostic and immunological roles of RSPO1 in pan-cancer and its correlation with LUAD proliferation and metastasis. Am J Cancer Res 2024;14:3800-15. [Crossref] [PubMed]
9. Wu L, Zhang W, Qian J, et al. R-spondin family members as novel biomarkers and prognostic factors in lung cancer. Oncol Lett 2019;18:4008-15. [Crossref] [PubMed]
10. Geoffroy V, Guignard T, Kress A, et al. AnnotSV and knotAnnotSV: a web server for human structural variations annotations, ranking and analysis. Nucleic Acids Res 2021;49:W21-8. [Crossref] [PubMed]
11. Geoffroy V, Herenger Y, Kress A, et al. AnnotSV: an integrated tool for structural variations annotation. Bioinformatics 2018;34:3572-4. [Crossref] [PubMed]
12. Geoffroy V, Lamouche JB, Guignard T, et al. The AnnotSV webserver in 2023: updated visualization and ranking. Nucleic Acids Res 2023;51:W39-45. [Crossref] [PubMed]
13. Chen X, Schulz-Trieglaff O, Shaw R, et al. Manta: rapid detection of structural variants and indels for germline and cancer sequencing applications. Bioinformatics 2016;32:1220-2. [Crossref] [PubMed]
14. Tammela T, Sanchez-Rivera FJ, Cetinbas NM, et al. A Wnt-producing niche drives proliferative potential and progression in lung adenocarcinoma. Nature 2017;545:355-9. [Crossref] [PubMed]
15. Stewart DJ. Wnt signaling pathway in non-small cell lung cancer. J Natl Cancer Inst 2014;106:djt356. [Crossref] [PubMed]
16. Rapp J, Jaromi L, Kvell K, et al. WNT signaling - lung cancer is no exception. Respir Res 2017;18:167. [Crossref] [PubMed]
17. Larson DE, Abel HJ, Chiang C, et al. svtools: population-scale analysis of structural variation. Bioinformatics 2019;35:4782-7. [Crossref] [PubMed]
18. Pan B, Kusko R, Xiao W, et al. Similarities and differences between variants called with human reference genome HG19 or HG38. BMC Bioinformatics 2019;20:101. [Crossref] [PubMed]
19. Triant DA, Pearson WR. Comparison of detection methods and genome quality when quantifying nuclear mitochondrial insertions in vertebrate genomes. Front Genet 2022;13:984513. [Crossref] [PubMed]
20. Gao GF, Parker JS, Reynolds SM, et al. Before and After: Comparison of Legacy and Harmonized TCGA Genomic Data Commons' Data. Cell Syst 2019;9:24-34.e10. [Crossref] [PubMed]
21. Karimzadeh M, Ernst C, Kundaje A, et al. Umap and Bismap: quantifying genome and methylome mappability. Nucleic Acids Res 2018;46:e120. [Crossref] [PubMed]
22. Lyu H, Liu E, Wu Z. Comparison of normalization methods for Hi-C data. Biotechniques 2020;68:56-64. [Crossref] [PubMed]
23. Araghi M, Mannani R, Heidarnejad Maleki A, et al. Recent advances in non-small cell lung cancer targeted therapy; an update review. Cancer Cell Int 2023;23:162. [Crossref] [PubMed]
24. Ramos A, Barbena E, Mateiu L, et al. Nuclear insertions of mitochondrial origin: Database updating and usefulness in cancer studies. Mitochondrion 2011;11:946-53. [Crossref] [PubMed]