Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2021 Apr 1;12(1):2031.
doi: 10.1038/s41467-021-22200-5.

Presence of complete murine viral genome sequences in patient-derived xenografts

Zihao Yuan  1   2 , Xuejun Fan  2 , Jay-Jiguang Zhu  3 , Tong-Ming Fu  2 , Jiaqian Wu  3 , Hua Xu  1 , Ningyan Zhang  2 , Zhiqiang An  4 , W Jim Zheng  5
Affiliations

Presence of complete murine viral genome sequences in patient-derived xenografts

Zihao Yuan et al. Nat Commun. .

Abstract

Patient-derived xenografts are crucial for drug development but their use is challenged by issues such as murine viral infection. We evaluate the scope of viral infection and its impact on patient-derived xenografts by taking an unbiased data-driven approach to analyze unmapped RNA-Seq reads from 184 experiments. We find and experimentally validate the extensive presence of murine viral sequence reads covering entire viral genomes in patient-derived xenografts. The existence of viral sequences inside tumor cells is further confirmed by single cell sequencing data. Extensive chimeric reads containing both viral and human sequences are also observed. Furthermore, we find significantly changed expression levels of many cancer-, immune-, and drug metabolism-related genes in samples with high virus load. Our analyses indicate a need to carefully evaluate the impact of viral infection on patient-derived xenografts for drug development. They also point to a need for attention to quality control of patient-derived xenograft experiments.

PubMed Disclaimer

Conflict of interest statement

The authors declare no competing interests.

Figures

Fig. 1
Fig. 1. Abundant murine virus sequence reads found in RNA-Seq data generated from PDX tumor cells.
a The proportion of murine virus reads in the total virome in various PDX tumor samples and corresponding controls, wild-type mice and NSG mice. Box limits: 25th and 75th percentiles; center line: median; whiskers: 1.5x interquantile range from box limits; dots: outliers. b Major categories of murine virus as identified by sequence reads from all PDX cancer samples. c Assembly of full viral genome from RNA-Seq reads. A-G: murine leukemia virus from seven types of PDX. H. Zika-virus-infected human brain organoids as positive control. d Sequence reads from breast cancer PDX (top) but not from non-PDX samples (bottom) cover the entire murine leukemia virus genome. e Amplification of XMLV ENV, XMRV GAG, and MLV GPP gene fragments from genomic DNA preparations of two breast cancer PDX tumor samples. The human GAPDH gene was used as a control to validate the experimental system. Reaction without template is considered as negative control. f Amplification of the MLV GPP gene fragment from 4 breast cancer PDX models with duplicates (A, B). PCR amplicfication was performed twice for each experiment. Source data for (a, b, e, and f) are provided in Source Data File.
Fig. 2
Fig. 2. Murine viral sequences from PDX are not from human viruses or correlated with mouse tissue contamination.
a Phylogenetic analysis of GAG genes of murine (black) and human viruses (blue). b same analysis performed for ENV genes. c Low level of reads from PDX that can be mapped to mouse reference genome. d High-level murine virus reads mapped to the MLV viral genome. e No significant correlations were found between the number of mouse-specific (x-axis) and murine viruses reads (y-axis) across all PDX samples examined (n = 184). Source data for e are provided in Source Data File.
Fig. 3
Fig. 3. Single-cell sequencing support the presence of murine viruses inside PDX tumor cells.
a Relative proportions of murine viruses in the virome of lung cancer PDX cells are significantly higher than in the cells from the primary lung cancer samples as measured by single-cell RNA-Seq (PDX: n = 1147; non-PDX: n = 152). b, c Viral reads from single-cell sequencing of PDX-derived tumor cells cover the entire genomic region of MLV. d No single-cell sequencing reads from non-PDX control samples were observed or mapped to the viral genome. e The single-cell sequencing reads that can be mapped to the mouse reference genome are very low in both PDX-derived tumor cells and non-PDX control cells (PDX: n=9; non-PDX: n = 152). For box plots in (a) and (e), Box limits: 25th and 75th percentiles; center line: median; whiskers: 1.5x interquantile range from box limits; dots: outliers. Source data for a and e are provided in Source Data File.
Fig. 4
Fig. 4. Chimeric sequence reads containing both viral (blue) and human (orange) genome sequences were identified from PDX RNA-Seq reads for six tumor types, indicating the integration of murine viral genome.
The chromosome, the gene name and the exon within which the integration site located is also provided. a Breast Cancer. Integration site is on chromosome 4, within the fifth exon of the MSM01 gene which has five exons. b GBM. Integration site is on chromosome 8, within the 18th exon of the RECQL4 gene which has 22 exons. c Lung Cancer. d Ovarian Cancer. e Bladder Cancer. f Colon Cancer. The genomic position and the orientation from human and virus genome are marked. The depth of coverage is labeled between reads. The detailed information is provided in Supplementary Data 2.
Fig. 5
Fig. 5. Impact of murine viral infection on gene expression in PDX tumors.
Transcription profiles of lung cancer PDX samples with highest and lowest viral numbers. Pathways and functions were enriched among: a the top 200 upregulated (left) and downregulated genes (right) in the 5 PDX samples with the most murine viral infection. b Transcription profiles of pictilisib-treated bladder cancer PDX samples with highest and lowest viral numbers. Pathways and functions were enriched among the top 200 upregulated (left) and downregulated (right) genes in the samples with the most murine viral infection. c Primary tumors before creation of PDX models have consistent gene expression profiles and form a major component in principal components analysis, compared to PDX tumors with viral infection and very diverse transcription profiles. d Three genes (STAT, IRF5, and TERT) differentially expressed between PDX tumors with highest (High, n = 5) and lowest (Low, n = 5) levels of murine viruses. There genes are also differentially expressed in oncolytic virus-infected lung cancer. Box limits: 25th and 75th percentiles; center line: median; whiskers: 1.5x interquantile range from box limits; dots: outliers. Source data for (a, b, and d) are provided in the Source Data File. Datasets used to create c are listed in Supplementary Table 2.
See this image and copyright information in PMC

Similar articles

See all similar articles

Cited by

See all "Cited by" articles

References

    1. Izumchenko E, et al. Patient-derived xenografts as tools in pharmaceutical development. Clin. Pharm. Ther. 2016;99:612–621. doi: 10.1002/cpt.354. - DOI - PubMed
    1. Williams, J. A. Using PDX for preclinical cancer drug discovery: the evolving field. J. Clin. Med.7, 10.3390/jcm7030041 (2018). - PMC - PubMed
    1. Cosper PF, et al. Patient derived models to study head and neck cancer radiation response. Cancers. 2020;12:419. doi: 10.3390/cancers12020419. - DOI - PMC - PubMed
    1. The PDX finder. https://www.pdxfinder.org/ (last accessed: 18 February 2021).
    1. The Mouse Tumor Biology Database. http://tumor.informatics.jax.org/mtbwi/index.do (Last accessed: 18 February 2021).

Publication types

MeSH terms

LinkOut - more resources