henhaobio
Sign in

bioRxivCancer BiologyPreprint

Time-resolved lineage recording reveals a pre-existing, heritable cell state underlying metastatic potential

Jihye Park, Yoojin ChangiD, Joshua S. SchiffmaniD, Divya Koyyalagunta, Hrishikesh SomayajiiD, Caden N. McQuilleniD, Joseph Chan, Quaid MorrisiD, Dan Avi LandauiD, Hyongbum Henry Kim, Junhong ChoiiD

PDFDOICode
In-depth full-text digest available · Jump to digest ↓

Abstract

Metastasis causes most cancer deaths1,2, yet no recurrent mutation specifically drives it3,4, raising the possibility that metastatic potential is a non-genetic yet heritable cell state. Classic experiments established that metastatically predisposed subclones pre-exist within a tumor and that these predispositions are inherited over many cell divisions5, but what molecular states or factors underlie this predisposition remain unknown. While previous lineage recording studies6,7 mapped how tumors disseminate, their recording sites saturate too quickly to resolve when a lineage branched, or to attribute a state to its founder. Here we show, using a DNA Typewriter lineage recorder8 with nearly 1,000 recording sites in lung cancer cells, that metastatic potential is already present before dissemination, with colonization predicted by a pre-existing glycolytic state and further spread by expression of ENO1, a glycolytic enzyme that also moonlights as a cell-surface plasminogen receptor9. Profiling the pre-transplant cells and the post-transplantation tumors for both their transcriptomes and their lineage recordings, we reconstructed time-resolved lineage trees across three orthotopically transplanted mice. These trees trace each liver metastasis to a single founder of known pre-transplant state, dating each dissemination event from the primary lung. When every clone was scored before transplant against 349 genes recurrently heritable in vitro, both that set and the glycolytic state independently shifted a clones odds of colonizing the lung. At the gene level, sixteen genes were both heritable and predictive of colonization, and ENO1 alone also predicted which established clones spread further. Hypoxia, the program most strongly associated with phylogenetic fitness within the metastases, did not predict colonization when scored before transplant, separating niche-selected traits from the inherited cell state. Metastatic potential in this system is therefore transmitted along the lineage rather than acquired after seeding. Looking forward, we anticipate that time-resolved lineage recorders will enable the separation of the heritable and acquired components of the cellular heterogeneity seen in single-cell studies of tumor progression and drug tolerance.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (12 Aug 2026) · CC BY 4.0

Ordered lineage recording in lung cancer xenografts shows pre-transplant glycolytic and lineage-heritable expression states predict metastatic colonization, while ENO1 alone predicts onward spread.

No recurrent mutation specifically drives metastasis, and earlier Cas9 recorders saturate quickly, so they cannot date branching events or tie a metastasis to the pre-dissemination state of its founder. A prime-editing recorder with 996 sites, read together with transcriptomes from the same cells, links founder state to later metastatic outcome.

Key results

  1. ShownThree dated liver metastases in Mouse 2 and Mouse 3 were most likely each founded by a single lung-derived cell. A calibrated recorder clock placed these seedings between about day 21 and day 34.

    Founder edit depth dated GCATTAGCGA-M3 to about day 21, and TTTGGAAACC-M3 and GCATTAGCGA-M2 to near day 34. A model-free distance-ratio check gave R = 0.57, 0.44, 0.13 and kept the ordering. Clock fit A = 4.96, tau = 10.71; R^2 = 0.99 across the calibration series including held-out day-45 cells. · A recorder clock dates the liver metastases as…; Recorder-clock calibration and metastasis dating; Figure legends

  2. ShownDissemination spanned one migration to a lymph-node hub with many migrations (Mouse 1). Clone growth rate measured in vitro did not explain colonization.

    Of 33 clones with a migration out of the primary, 32 seeded lymph nodes. 51 of 75 clones disseminated from lymph nodes vs 33 from the primary and 6 from other sites; lymph nodes accounted for 76% of migrations. In vivo tissues sat near the neutral pool mean lambda (264 clones with growth rates). · Metastatic dissemination varies widely and is…; A lymph-node migration hub drives dissemination; Clonal growth rate analysis

  3. ShownAmong 1,229 pre-transplant clones, a glycolytic state and the 349-gene heritable set each shifted the odds of colonizing the primary lung in all three mice. Hypoxia did not, and only glycolysis predicted dissemination.

    Logistic models adjusted for pool abundance: glycolysis P < 0.003, heritable set P < 0.02 for lung colonization in all three mice. Dissemination among established clones: glycolysis 1.59, heritable set 1.28 (P = 0.051). Both effects persisted when fitted jointly. · Transcriptional basis of the heritable…; Single-cell lineage and transcriptomic analyses

  4. ShownA genome-wide scan found 16 genes that were both recurrently heritable in vitro and predictive of clone establishment. ENO1 alone also predicted dissemination.

    43 of 16,221 genes passed 5% FDR for establishment against label-reshuffled backgrounds (200 rounds). The heritable set (349 genes) came from Hotspot (396 genes) and PATH (Spearman rho = 0.86). Heritability and establishment were near-independent (rho = 0.12). · Transcriptional basis of the heritable…; Single-cell lineage and transcriptomic analyses

  5. SuggestedThe programs most strongly induced in the liver (proliferation, OXPHOS) were not those associated with phylogenetic fitness. Fitness tracked a smaller hypoxia, glycolysis and partial-EMT axis, separating niche-driven plasticity from selection.

    Liver vs lung: MYC and E2F targets NES = 3.3, OXPHOS NES = 2.8; 4,669 genes at 5% FDR. After covariate control, proliferation and OXPHOS fitness correlations fell to zero or below; hypoxia rose 0.14 to 0.19, glycolysis fell 0.30 to 0.14. Only three clones (n = 3). · Metastatic cells converge on a reproducible…; Fitness is decoupled from the strongest liver…

Methods, data and reproducibility

A 996-site DNA Typewriter prime-editing recorder plus static clonal barcodes was placed in NCI-H1299 cells and transplanted orthotopically into three mice. 10x scRNA-seq captured the transcriptome, clonal barcode and recorder edits per cell. Trees (sequential parsimony), a calibrated clock, Metient migration inference, Hotspot/PATH heritability and logistic models were applied.

  • Single-cell RNA, clonal barcode and TAPE libraries from pre-transplant pool, subclonal colonies and tumors of three orthotopic mice · human NCI-H1299 cells (xenografted in NOG mice) · n = 105,260 single cells; 16 libraries · BioProject PRJNA1495924; GEO GSE339406
  • Pre-transplant clones scored for programs and heritable gene set to predict lung colonization and dissemination · human (NCI-H1299) · n = 1,229 clones with at least 10 cells (31,838 cells)
  • Eight monoclonal in vitro colonies with lineage trees and transcriptomes for gene heritability · human (NCI-H1299) · n = 295 to 11,081 cells per colony
  • Mouse 1 clones used for Metient migration-history inference · human (NCI-H1299 in mouse) · n = 75 clones (5,551 cells)
  • Published Cas9 lineage-tracing A549 xenograft allele table reanalyzed for comparison of seeding modality · human (A549 in mouse) · n = 79 clones retained of 100 · GSM4905334
  • Parallel in vitro culture with weekly bulk amplicon profiling for recorder calibration and clone growth rates · human (NCI-H1299) · n = five time points, day 17 to day 45; 264 clones with growth rates
Methods & tools
DNA Typewriter (prime-editing lineage recorder), Cell Ranger v9.0.1, Seurat, Scrublet, Cassiopeia, Metient v1.0.0, Hotspot, PATH (Phylogenetic Analysis of Transcriptional Heritability), UCell, edgeR, fgsea, decoupleR (CollecTRI)
Software & environment
Python 3.10 (Cassiopeia 2.1.0, ete3 3.1.3, numpy 2.4, scipy 1.17, pandas 2.3); scanpy 1.10.3, decoupleR 1.8.0, hotspotsc 1.1.3; R 4.3.3 with Seurat 5.2.1, UCell 2.6.2, edgeR 4.0.16, fgsea 1.28.0.
Reproducibility gaps
GEO access was described with a reviewer token at the time of the preprint. Seurat v5.4.0 and R v4.5.2 are named in the preprocessing methods, versus Seurat 5.2.1 and R 4.3.3 in the statistics section.
Limitations (5)
  • Recorder readout by amplicon capture scales with transcriptional output, so a large low-RNA, non-proliferative population (77% of tumor cells without barcodes) is excluded from lineage reconstruction. · stated by the authors
  • Timed disseminations rest on three reconstructable liver clones (n = 3), so adaptation dynamics cannot be fully resolved. · stated by the authors
  • A single lung cancer line (NCI-H1299) in immunodeficient hosts was used, so tumor-immune interaction is absent. · stated by the authors
  • The earliest regional spread in Mouse 1 may have been seeded at or soon after orthotopic injection rather than by later migration; the glycolytic link is an association, not a tested causal dependency. · stated by the authors
  • The authors call the effects modest (dissemination odds ratios of 1.59 for glycolysis and 1.28 for the heritable set), and all come from one shared transplanted pool in three mice, so replication in independent pools or with ENO1 perturbation is untested. · noted in this summary

Read this if you want to see how ordered prime-editing lineage recorders combined with single-cell transcriptomes can separate heritable from niche-induced cell states in metastasis.

Citations
0

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISQuinn '21Simeonov '20Moorman '23Jakab '22Handler '25Kaur '22
65% semantic matchCancer Biology
J. J. Quinn, M. G. Jones, R. A. Okimoto+5 more·Science·Feb 26, 2021