{"paper_id":"12430a27-4e40-4bbb-9e92-7cbd98a08875","body_text":"1 \n \nSendai virus persistence questions the transient naive reprogramming  \nmethod for iPSC generation \n \nAlejandro De Los Angeles1*, Clemens B. Hug2*, Vadim N. Gladyshev3, George M. Church4,5,  \nSergiy Velychko4,5* \n \n1McGovern Institute for Brain Research, Massachusetts Institute of Technology, Cambridge MA, USA  \n2Laboratory of Systems Pharmacology, Harvard Program in Therapeutic Science, Harvard Medical \nSchool, Boston, MA, USA \n3Division of Genetics, Department of Medicine, Brigham and Women’s Hospital and Harvard Medical \nSchool, Boston, MA, USA \n4Department of Genetics, Harvard Medical School, Boston, MA, USA \n5Wyss Institute, Harvard University, Boston, MA, USA \n*equal contribution as corresponding authors \ne-mail: adelosan@mit.edu; clemens_hug@hms.harvard.edu; sergiy_velychko@hms.harvard.edu \n \nAbstract \nSince the revolutionary discovery of induced pluripotent stem cells (iPSCs) by Shinya Yamanaka, the \ncomparison between iPSCs and embryonic stem cells (ESCs) has revealed significant differences in their \nepigenetic states and developmental potential. A recent compelling study published in Nature by \nBuckberry et al. 1 demonstrated that a transient-naive-treatment (TNT) could facilitate epigenetic \nreprogramming and improve the developmental potential of human iPSCs (hiPSCs). However, the study \ncharacterized bulk hiPSCs instead of isolating clonal lines and overlooked the persistent expression of \nSendai virus carrying exogenous Yamanaka factors. Our analyses revealed that Sendai genes were \nexpressed in most control PSC samples, including hESCs, which were not intentionally infected. The \nhighest levels of Sendai expression were detected in samples continuously treated with naive media, \nwhere it led to overexpression of exogenous MYC, SOX2, and KLF4, altering both the expression levels \nand ratios of reprogramming factors. Our findings call for further research to verify the effectiveness of the \nTNT method in the context of delivery methods that ensure prompt elimination of exogenous factors, \nleading to the generation of bona fide transgene-independent iPSCs. \n \nDetection of Sendai virus sequences in established pluripotent stem cell lines \nOur analysis of publicly available RNA-seq data provided by Buckberry et al. revealed significant \nexpression of Sendai virus genes in nearly all PSC samples (Fig. 1a). The highest levels of Sendai genes \nwere observed in naive hiPSCs, suggesting a possible selection of cells with persistent viral expression\n2,3. \nSimilar findings were published by Yamanaka and colleagues reporting the persistence of Cytotune 2.0 \nSendai viruses in human naive PSCs cultured in t2iLGo conditions 4.  One of the two primed hiPSC \nsamples showed expression of Sendai virus genes as late as passage 17. Traces of Sendai virus were \nalso found in naive-to-primed (NTP) hiPSCs, which were established and passaged in naive media before \nbeing transferred into primed PSC conditions, whereas TNT hiPSCs were Sendai-free. Surprisingly, \ncontrol MEL1 hESC samples from Buckberry et al. also had low but detectable levels of Sendai \nexpression. In contrast, hESC samples from another group\n5 did not express any Sendai genes, as \nexpected. hESCs were not deliberately infected with Sendai viruses, suggesting a contamination or \npossibly a mix-up of hESC and TNT hiPSC samples. \n \nNaive media selects for high levels of exogenous Yamanaka factors \nWe assessed the levels of Sendai virus sequences in RNA-seq data across human reprogramming \nintermediates and established naive and primed hiPSC bulk lines, encompassing Buckberry et al. and \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted March 12, 2024. ; https://doi.org/10.1101/2024.03.07.583804doi: bioRxiv preprint \n\n2 \n \ntwo additional studies by the same group (Liu et al., Nature Methods 2017; and Liu et al., Nature \n2020)1,6,7 (Fig. 1b). The analyses confirmed the absence of Sendai virus in control fibroblast samples as \nwell as expected pronounced expression levels in Day 3 and Day 7 reprogramming samples. For each of \nthe culture conditions tested (t2iLGoY, 5iLAF, NHSM, RSeT—naive medias, and E8—primed media), the \nlevels of Sendai viruses decreased with passaging, yet remained detectable even at late passage \nnumbers. The presence of Sendai virus in control hESCs from Liu et al, 2017 was also noted. Finally, a \nnearly 1,000-fold higher level of Sendai expression was observed in t2iLGoY-cultivated hiPSCs relative to \nprimed hiPSCs at similar high passage numbers, which amounts to approximately two-fold enrichment for \nSendai-expressing cells during each passage in naive media (every 3 days). \nA significant presence of Sendai virus sequences in control naive hiPSCs enabled us to evaluate the \nexpression level of each Yamanaka factor. Endogenous transcripts include both coding sequences (CDS) \nand untranslated regions (UTRs), while the Sendai RNA contains only the CDS sequences allowing \ndiscrimination of the endogenous and exogenous RNAs. Examining the RNA-seq read coverage for the \nYamanaka factors revealed a striking difference in expression patterns between naive cells and other \nsamples: naive samples showed significantly lower UTR expression while maintaining high CDS \ncoverage, suggesting that most transcripts are derived from loci lacking UTRs ( Fig. 1c). To quantify the \nextent of exogenous expression, we evaluated the ratio of reads mapping to CDSs versus UTRs. Our \nanalysis found higher CDS/UTR ratios for MYC, SOX2, and KLF4 (SKM) in naive hiPSCs compared to \nthe expected CDS/UTR ratio found in samples with low or absent transgene expression, such as \nfibroblasts and hESCs, indicating a selection for exogenous SKM expression in the naive hiPSCs ( Fig. \n1d-e). Exogenous MYC was the most enriched, with a CDS/UTR ratio greater than 100, followed by \nSOX2 and KLF4 with ratios exceeding 10. Exogenous OCT4, however, was not noticeably enriched, \nwhich likely reflects relatively high levels of endogenous OCT4 expression in primed PSCs.  \nThe resulting overexpression of SKM in reprogramming samples treated with naive media echoes our \nstudies showing that omitting OCT4 from the Yamanaka cocktail could improve the quality of iPSCs\n8 and \nreset PSCs across species9, hinting at the possible mechanism underlying TNT reprogramming. \n \nImplications and recommendations for future studies \nPrior research emphasized the need for transgene clearance to ensure reactivation of the endogenous \npluripotency network and to avoid issues like abnormal differentiation. Even minimal leakage of \nexogenous OSKM from a tet-inducible promoter compromises the developmental potential of iPSCs, \nrendering tet-inducible OSKM iPSCs incapable of producing healthy animals\n8-9. Our results suggest that \nthe t2iLGoY naive medium, used in the TNT reprogramming protocol, promotes the retention of \nexogenous reprogramming factors and changes their ratios ( Fig. 1e ). This raises concerns about \napplications of such media in reprogramming protocols.  Yamanaka factors can induce a transient naive-\nlike state even in primed media\n10. Identification of the optimal reprogramming factor ratio could eliminate \nthe need for the naive medium treatment with its associated risks of genetic and epigenetic instability 11. \nRefinements of the Cytotune 2.0 Sendai kit, such as an all-in-one design or the addition of microRNA-\nbinding sites, might promote efficient virus elimination. Meticulous assessment for exogenous factor \nelimination is crucial. Furthermore, we recommend isolating and characterizing clonal iPSC lines rather \nthan bulk passaging of reprogramming cultures, to reduce heterogeneity and the potential for selection.  \n \nConclusion \nThe study by Buckberry et al. suggests that a specific naive media regimen can boost the hiPSC \ntechnology. Our finding of Sendai virus sequences in control hiPSC and hESC lines, coupled with naive \nmedia favoring high Sendai expression suggests that further work needs to be done to support the \neffectiveness of the TNT reprogramming method, underscoring the necessity for a refined delivery \nsystem.\n  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted March 12, 2024. ; https://doi.org/10.1101/2024.03.07.583804doi: bioRxiv preprint \n\n3 \n \nFig. 1: Persistence of Sendai virus and selection for exogenous factor expression by naive media. \na, Read coverage of the Sendai virus genome. Two replicates per condition are shown (in red and blue). \nCoverage was binned at 100 bp resolution and is shown as counts per million reads (CPM). The F gene \nwas deleted in the commercial kit used for reprogramming in these studies. b, Time-course of Sendai \nexpression across reprogramming and hPSC samples. The y-axis shows the mean expression as \ntranscripts per million (TPM) of all Sendai genes – except for F. Symbols indicate the study from which \nsamples were sourced and colors indicate the cell culture media that the samples were grown in. The \ndashed line at the bottom corresponds to the location of samples with zero TPM. c, Read coverage of \nYamanaka factors. The CPM values are rescaled such that each sample has a range from exactly zero to \none. Large boxes in the gene models indicate exons, small boxes correspond to 5’ and 3’ untranslated \nregions (UTRs), and lines with arrows indicate introns. UTRs are highlighted using gray shading. d, \nQuantification of the ratio between coding sequence and untranslated region expression (CDS/UTR \nratio). For each Yamanaka factor we quantified the number of reads aligning to CDS and UTR regions. \nThe ratio of normalized counts CDS/UTR serves as a proxy for the proportion of transcripts originating \nfrom the viral transgene vs the endogenous locus. The horizontal black line is drawn at the mean ratio of \nthe hESC control samples from Buckberry et al., which serve as baseline for the expected ratio in the \nabsence of a transgene. The horizontal dashed line is drawn at the location of samples with a ratio of \nzero. Samples with low expression below <30 raw counts are drawn in a lighter shade. e, Graphical \noverview and mechanism of TNT method: the study utilizes t2iLGoY naive media, which preferentially \nselects for cells expressing high levels of exogenous reprogramming factors from Sendai virus RNA \n(illustrated in dark red). This selective expression of exogenous factors may significantly contribute to the \noutcomes observed in the TNT reprogramming.\n  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted March 12, 2024. ; https://doi.org/10.1101/2024.03.07.583804doi: bioRxiv preprint \n\n4 \n \nMethods \n \nProcessing of RNA-seq data \n \nRaw sequencing reads were downloaded from SRA accession numbers SRP286549 (Buckberry et al. 1), \nSRP259918 (Liu et al. 2020 7), SRP115256 (Liu et al. 2017 6), SRP059279 (Ji et al. 2016 12), and \nSRP068579 (Pastor et al. 20165). \n \nRNA-seq alignment and quantification \n \nHuman transcripts were quantified using human genome version GRCh38 and version 110 of the \nEnsembl gene annotations. For Sendai virus transcripts we used sequences and gene annotations from \nNCBI reference sequence NC_075392.1 ( Respirovirus muris). We quantified transcript expression using \nSalmon\n13 1.10.2 using the options -l A --seqBias --gcBias --posBias --softclip. The \nSalmon index was created by concatenating the CDS fasta file from Ensembl with the Sendai transcripts \nobtained from NC_075392.1. The raw transcript counts from Salmon were imported into R 4.3.1 using the \ntximport package\n14 and aggregated to gene-level counts. For each RNA-seq sample, a library specific \ncorrection factor accounting for differences in sequencing depth was calculated using \nestimateSizeFactorsForMatrix() from DESeq215. \n \nAdditionally, all reads were aligned to the human genome using STAR 16 2.7.9a with default settings. The \ngenome index was prepared by concatenating the unmasked primary DNA assembly from Ensembl with \nthe whole Sendai genome. \n \nRead coverage plots \n \nRead coverage plots for the Sendai genome and Yamanaka factors were generated based on our STAR \nalignments using the ggcoverage R package. Reads were counted in evenly spaced 100 bp bins, \nnormalized using the previously calculated size factors, divided by the mean number of reads per sample, \nand then multiplied by a million to get Counts Per Million (CPM). For comparing coverage of coding \nsequences (CDS) to 5’ and 3’ untranslated regions (UTR) we rescaled raw CPM values for each sample \nso that their minimum and maximum are zero and one, respectively. \n \nCalculation of CDS to UTR ratios \n \nThe ratio between expression of CDS to UTRs in naturally occurring mature mRNAs is expected to be \nclose to one, given how they are usually transcribed and spliced as a single unit. mRNA transcribed from \nexogenous viral sequences do not contain UTRs, therefore shifting the CDS/UTR ratio up the more they \nare expressed. Read counts for the coding sequences (CDS), as well as 5’ and 3’ untranslated regions \n(UTR) for each Yamanaka factor were quantified based on our STAR alignments using the Rsubread\n17 R \npackage. For each of the four factors we picked a representative transcript, annotated as “MANE select” \nin Ensembl. These were ENST00000325404, ENST00000259915, ENST00000621592, and \nENST00000374672 for SOX2, POU5F1, MYC, and KLF4, respectively. Raw counts for each feature type \nand transcript were added up, normalized using DESeq2 size factors, and divided by the total feature \nlength. These normalized counts for CDS and UTR for each transcript were then divided by each other to \nyield CDS/UTR ratios. \n  \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted March 12, 2024. ; https://doi.org/10.1101/2024.03.07.583804doi: bioRxiv preprint \n\n5 \n \nReferences \n1. Buckberry, S. et al.  Transient naive reprogramming corrects hiPS cells functionally and \nepigenetically. Nature 620, 863–872 (2023). \n2. Takashima, Y. et al.  Resetting Transcription Factor Control Circuitry toward Ground-State \nPluripotency in Human. Cell 158, 1254–1269 (2014). \n3. Guo, G. et al. Naive Pluripotent Stem Cells Derived Directly from Isolated Cells of the Human Inner \nCell Mass. Stem Cell Rep. 6, 437–446 (2016). \n4. Kunitomi, A. et al. Improved Sendai viral system for reprogramming to naive pluripotency. Cell Rep. \nMethods 2, 100317 (2022). \n5. Pastor, W. A. et al.  Naive Human Pluripotent Cells Feature a Methylation Landscape Devoid of \nBlastocyst or Germline Memory. Cell Stem Cell 18, 323–329 (2016). \n6. Liu, X. et al.  Comprehensive characterization of distinct states of human naive pluripotency \ngenerated by reprogramming. Nat. Methods 14, 1055–1062 (2017). \n7. Liu, X. et al. Reprogramming roadmap reveals route to human induced trophoblast stem cells. Nature \n586, 101–107 (2020). \n8. Velychko, S. et al. Excluding Oct4 from Yamanaka Cocktail Unleashes the Developmental Potential \nof iPSCs. Cell Stem Cell 25, 737-753.e4 (2019). \n9. MacCarthy, C. M. et al. Highly cooperative chimeric super-SOX induces naive pluripotency across \nspecies. Cell Stem Cell 31, 127-147.e9 (2024). \n10. Cacchiarelli, D. et al.  Integrative Analyses of Human Reprogramming Reveal Dynamic Nature of \nInduced Pluripotency. Cell 162, 412–424 (2015). \n11. Bar, S., Schachter, M., Eldar-Geva, T. & Benvenisty, N. Large-Scale Analysis of Loss of Imprinting in \nHuman Pluripotent Stem Cells. Cell Rep. 19, 957–968 (2017). \n12. Ji, X. et al. 3D Chromosome Regulatory Landscape of Human Pluripotent Cells. Cell Stem Cell 18, \n262–275 (2016). \n13. Patro, R., Duggal, G., Love, M. I., Irizarry, R. A. & Kingsford, C. Salmon provides fast and bias-aware \nquantification of transcript expression. Nat. Methods 14, 417–419 (2017). \n14. Soneson, C., Love, M. & Robinson, M. Differential analyses for RNA-seq: transcript-level estimates \ni\nmprove gene-level inferences [version 2; peer review: 2 approved]. F1000Research 4, (2016). \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted March 12, 2024. ; https://doi.org/10.1101/2024.03.07.583804doi: bioRxiv preprint \n\n6 \n \n15. Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq \ndata with DESeq2. Genome Biol. 15, 1–21 (2014). \n16. Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013). \n17. Liao, Y., Smyth, G. K. & Shi, W. The R package Rsubread is easier, faster, cheaper and better for \nalignment and quantification of RNA sequencing reads. Nucleic Acids Res. 47, e47 (2019). \n \nAcknowledgements \nWe thank Caitlin MacCarthy for editing the manuscript. We acknowledge support from the NIA grant R01 \nAG058063 (C.H.) and NCI U54-CA225088 (C.H.). \n \nAuthor contributions \nS.V. and A.A. conceived the study. A.A. performed initial analysis. C.H. performed an independent full \nanalysis and generated the figures. A.A., S.V. and C.H. interpreted the results and wrote the manuscript. \nV.N.G. and G.M.C. advised on the study. \n \nCompeting interests \nS.V. is listed as an inventor of a submitted patent on SK/SKM naive reset. All other authors declare no \ncompeting interests. \n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted March 12, 2024. ; https://doi.org/10.1101/2024.03.07.583804doi: bioRxiv preprint \n\n.CC-BY 4.0 International licenseavailable under a \n(which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made \nThe copyright holder for this preprintthis version posted March 12, 2024. ; https://doi.org/10.1101/2024.03.07.583804doi: bioRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}