Molecular map of chronic lymphocytic leukemia and its impact on outcome

Authors

Binyamin A. Knisbacher

Ziao Lin

Cynthia K. Hahn

Ferran Nadeu

Martí Duran-Ferrer

Kristen E. Stevenson

Eugen Tausch

Julio Delgado

Catherine J. Wu

Elias Campo

Gad Getz

Doi

10.1038/s41588-022-01140-w

PMID: 35927489 · DOI: 10.1038/s41588-022-01140-w · Journal: Nature Genetics (2022)

TL;DR

Knisbacher et al. assembled the largest integrated CLL cohort to date (1,148 patients: 1,095 CLL + 54 MBL) and combined WES/WGS (n=1,074), RNA-seq (n=712), and DNA methylation (n=999) to build a “CLL map.” They nominate 202 candidate genetic drivers (109 novel), refine molecular profiles of IGHV-mutated (M-CLL) vs IGHV-unmutated (U-CLL) subtypes, discover 8 gene-expression clusters that serve as independent prognostic factors, and integrate genetic, epigenetic, and transcriptomic features into a combined outcome model PMID:35927489.

Cohort & data

  • 1,148 patients (1,095 CLL, 54 MBL); WES/WGS n=1,074 (984 WES CLL, 177 WGS), RNA-seq n=712 (603 treatment-naive used for clustering), DNA methylation n=999 (450k array n=490, RRBS n=509).
  • Sampling contexts: active surveillance (n=680), post-treatment (n=52), clinical trial enrollment (n=416; 371 treatment-naive, 45 relapsed/refractory) PMID:35927489.

Key findings

  • 82 putative CLL drivers from recurrent sSNV/indels (q<0.1) across the full cohort, 37 not previously reported as significantly altered in CLL. Six additional drivers (incl. MAP2K2, DIS3, DICER1) found via 3-D clustering with CLUMPS PMID:35927489.
  • Top known drivers: SF3B1 17.5%, NOTCH1 12.3%, ATM 11.2%, TP53 9.1%; IGLV3-21 R110 9.5%; U1 snRNA g.3A>C 3.8%. 59 of 82 drivers mutated in <2% of patients; 24.2% of patients carried at least one novel driver mutation PMID:35927489.
  • Power analysis: drivers discovered nearly doubled from ~38.8 (n=500) to 74.5 (~n=1,000); also identified 5 novel focal sCNA gains and 30 novel focal losses PMID:35927489.
  • IGHV-stratified analysis (M-CLL n=512, U-CLL n=459): U-CLL had substantially more drivers than M-CLL (54 vs 25; ratio 2.16, Binomial p=0.0015; treatment-naive-only ratio 2.82, p=5×10^-11). Only 16 drivers were significant in both subtypes, and 10 of those were twice as frequent in U-CLL PMID:35927489.
  • SV findings (177 WGS): 681 breakpoints in 141 patients (~4.8/patient); BCL2 translocations predominantly in M-CLL (5/88, 5.7%) via aberrant V(D)J; a recurrent 37-Mb chr14 deletion disrupting ZFP36L1 (and DICER1, TRAF3) in U-CLL (4/87, 4.6%) via class-switch recombination PMID:35927489.
  • Mutational signatures across 177 WGS: aging, canonical AID (SBS84) enriched in clustered mutations in U-CLL, non-canonical AID (SBS85) enriched in M-CLL (p=1.6×10^-9), plus SBS18 (reactive oxygen species) PMID:35927489.
  • Temporal ordering (PhylogicNDT): Trisomy 12 is early; TP53/NOTCH1 intermediate in both subtypes; BRAF is early in M-CLL but late in U-CLL (q<0.1); MYD88 early in M-CLL; 20p loss and FUBP1 potentially initiating in U-CLL PMID:35927489.
  • Unsupervised clustering of treatment-naive RNA-seq (n=603) identified 8 robust expression clusters that are independent prognostic factors beyond IGHV/epitype PMID:35927489.
  • Epitype framework (n-CLL, i-CLL, m-CLL) and the epiCMIT mitotic clock extended via RRBS; DNA methylome variation dominated by cell-of-origin and proliferative history PMID:35927489.
  • Fraction of patients with no identifiable driver dropped from 8.9% in prior work to 3.8%, concentrated in M-CLL (6.6% vs 0.6% in U-CLL; Fisher p=1.04×10^-7) PMID:35927489.

Genes & alterations

Clinical implications

  • U-CLL harbors substantially more prognostic genetic events than M-CLL (41 vs 18 linked to FFS/OS), and 18 of those are novel PMID:35927489.
  • Gene expression clusters are independent prognostic factors beyond IGHV and epitype, motivating integrated genetic + epigenetic + transcriptomic risk stratification PMID:35927489.
  • In the era of ibrutinib/venetoclax, isolated TP53 mutation (without 17p loss) does not confer adverse prognosis in U-CLL, consistent with prior reports the authors cite PMID:35927489.
  • Novel prognostic markers (ZC3H18, 5q32 loss, 15q25.2 loss in M-CLL; RFX7, NFKB1 in U-CLL) are candidate additions to CLL risk models PMID:35927489.

Limitations & open questions

  • Many newly nominated drivers are mutated in <2% of patients, so validation in independent cohorts is needed before clinical use PMID:35927489.
  • SV analysis relied on only 177 WGS; SV landscape of the full cohort is necessarily extrapolated from WES/IgCaller PMID:35927489.
  • Treatment heterogeneity across cohort/trial subsets complicates survival inference; authors partially mitigate with treatment-naive subset analyses PMID:35927489.
  • How expression clusters relate mechanistically to driver genotype and epitype — and whether they are stable over time/treatment — is left open PMID:35927489.
  • Functional validation of the 109 novel candidate drivers (e.g., MAP2K2, DIS3, INO80, RFX7) is deferred to future work PMID:35927489.

Citations from this paper used in the wiki

  • “we integrated genomic, transcriptomic, and epigenomic data from 1148 patients. We identified 202 candidate genetic drivers of CLL (109 novel)” (Abstract).
  • “SF3B1, NOTCH1, ATM, and TP53 (mutated in 17.5%, 12.3%, 11.2%, and 9.1% of patients, respectively)” (Results, Identification of novel CLL drivers).
  • “the number of significant putative drivers was much higher in U-CLL (54 versus 25 genes, respectively; ratio 2.16, Binomial test p=0.0015)” (Results, Molecular profiles of IGHV subtypes).
  • “TP53 mutation in the absence of 17p deletion was not associated with adverse outcomes in U-CLL” (Results, clinical impact section).
  • “the percent of patients lacking at least one potential driver was reduced to 3.8% … predominantly M-CLL (Fisher’s Exact test p=1.04×10−7; 6.6% relative to 0.6% in U-CLL)” (Results).

This page was processed by crosslinker on 2026-04-08.