Cellular and molecular architecture of hematopoietic stem cells and progenitors in genetic models of bone marrow failure
Stephanie Heidemann, Brian Bursic, Sasan Zandi, Hongbing Li, Sagi Abelson, Robert J. Klaassen, Sharon Abish, Meera Rayar, Vicky R. Breakey, Houtan Moshiri, Santhosh Dhanraj, Richard de Borja, Adam Shlien, John E. Dick, Yigal Dror
Stephanie Heidemann, Brian Bursic, Sasan Zandi, Hongbing Li, Sagi Abelson, Robert J. Klaassen, Sharon Abish, Meera Rayar, Vicky R. Breakey, Houtan Moshiri, Santhosh Dhanraj, Richard de Borja, Adam Shlien, John E. Dick, Yigal Dror
View: Text | PDF
Research Article Hematology Oncology

Cellular and molecular architecture of hematopoietic stem cells and progenitors in genetic models of bone marrow failure

Abstract

Inherited bone marrow failure syndromes, such as Fanconi anemia (FA) and Shwachman-Diamond syndrome (SDS), feature progressive cytopenia and a risk of acute myeloid leukemia (AML). Using deep phenotypic analysis of early progenitors in FA/SDS bone marrow samples, we revealed selective survival of progenitors that phenotypically resembled granulocyte-monocyte progenitors (GMP). Whole-exome and targeted sequencing of GMP-like cells in leukemia-free patients revealed a higher mutation load than in healthy controls and molecular changes that are characteristic of AML: increased G>A/C>T variants, decreased A>G/T>C variants, increased trinucleotide mutations at Xp(C>T)pT, and decreased mutation rates at Xp(C>T)pG sites compared with other Xp(C>T)pX sites and enrichment for Cancer Signature 1 (X indicates any nucleotide). Potential preleukemic targets in the GMP-like cells from patients with FA/SDS included SYNE1, DST, HUWE1, LRP2, NOTCH2, and TP53. Serial analysis of GMPs from an SDS patient who progressed to leukemia revealed a gradual increase in mutational burden, enrichment of G>A/C>T signature, and emergence of new clones. Interestingly, the molecular signature of marrow cells from 2 FA/SDS patients with leukemia was similar to that of FA/SDS patients without transformation. The predicted founding clones in SDS-derived AML harbored mutations in several genes, including TP53, while in FA-derived AML the mutated genes included ARID1B and SFPQ. We describe an architectural change in the hematopoietic hierarchy of FA/SDS with remarkable preservation of GMP-like populations harboring unique mutation signatures. GMP-like cells might represent a cellular reservoir for clonal evolution.

Authors

Stephanie Heidemann, Brian Bursic, Sasan Zandi, Hongbing Li, Sagi Abelson, Robert J. Klaassen, Sharon Abish, Meera Rayar, Vicky R. Breakey, Houtan Moshiri, Santhosh Dhanraj, Richard de Borja, Adam Shlien, John E. Dick, Yigal Dror

×

Figure 3

Patterns of single nucleotide and trinucleotide alterations among FA, SDS, and healthy control subjects.

Options: View larger image (or click on image) Download as PowerPoint
Patterns of single nucleotide and trinucleotide alterations among FA, SD...
(A) Average number (±SEM) of each transition (inside the CT purine group or inside the GA pyrimidine group) variant per subject among the FA, SDS, and healthy control groups. *P < 0.05 when comparing each patient group to the control group. P = 0.9313 for A>G; P = 0.0735 for G>A; P = 0.086 for C>T; P = 0.2586 for T>C by Kruskal-Wallis test with Dunn’s post hoc test when comparing the 3 subject groups. The average numbers of transversions (change from pyrine to pyrimidine or vice versa) are in Supplemental Figure 5. (B) Heatmap depicting trinucleotide SNV patterns. The heatmap depicts specific trinucleotide variants (SNV including the base immediately 3′ and 5′ to the SNV site). The 5′ base is shown on the y axis and the 3′ base on the x axis. Z score of the log-transformed values from 0 to 2 was used. To generate the heatmap, the number of each variant plus 1 was converted to log. (C) Percentage of SNVs and indels according to their damaging effects on the protein in each of the study subject groups. (D) Mean number of mutated genes in FA subjects, SDS subjects, and controls with SEM. Results of comparison between each patient group to controls by Student’s t test are shown. P = 0.069 by Kruskal-Wallis test when comparing the 3 subject groups. The box plots depict the minimum and maximum values (whiskers), the upper and lower quartiles, and the median. The length of the box represents the interquartile range.