Molecular convergence in ex vivo models of Diamond-Blackfan anemia

Abstract

Diamond-Blackfan anemia (DBA) is a congenital bone marrow failure syndrome characterized by erythroid hypoplasia, usually without perturbation of other hematopoietic lineages. Approximately 65% of DBA patients with autosomal dominant inheritance have heterozygous mutations or deletions in ribosomal protein (RP) genes while <1% of patients with X-linked inheritance have been identified with mutations in the transcription factor GATA1 Erythroid cells from patients with DBA have not been well characterized, and the mechanisms underlying the erythroid specific effects of either RP or GATA1 associated DBA remain unclear. We have developed an ex vivo culture system to expand peripheral blood CD34+ progenitor cells from patients with DBA and differentiate them into erythroid cells. Cells from patients with RP or GATA1 mutations showed decreased proliferation and delayed erythroid differentiation in comparison with controls. RNA transcript analyses of erythroid cells from controls and patients with RP or GATA1 mutations showed distinctive differences, with upregulation of heme biosynthesis genes prominently in RP-mediated DBA and failure to upregulate components of the translational apparatus in GATA1-mediated DBA. Our data show that dysregulation of translation is a common feature of DBA caused by both RP and GATA1 mutations. This trial was registered at www.clinicaltrials.gov as #NCT00106015.

Figures

Figure 1.

Reduced proliferation of erythroid precursors of patients with DBA. Growth curves of CD34+ cells from control (n = 24) and DBA patients with RP gene mutations: RPS17 (red; n = 4) or RPL5 (black; n = 3) (A); control (n = 24) and DBA patients with unidentified mutations: unknown patient 1 (red; n = 3) and unknown patient 2 (black; n = 3) (B); and control (n = 24) and DBA patients with a GATA1 mutation: GATA1 patient 1 (red; n = 3) or GATA1 patient 2 (black; n = 1) (C). Data are presented as the mean ± standard error of the mean.

Figure 2.

Characterization of erythroid differentiation during in vitro culture. Flow cytometric analysis of cultured CD34+ cells labeled with CD235, CD44, and CD41 antibodies. (A) Control donor and a patient with an RPL5 mutation at day 8 and day 13 of terminal erythroid differentiation. (B) Control donor and a patient with a RPS17 mutation at day 14 of culture. (C) Control donor and a patient with a RPS24 mutation at day 14 of culture.

Figure 3.

Improved proliferation and erythroid differentiation in a steroid-treated patient. (A) Growth curves of CD34+ cells from control donors (n = 24) and a DBA patient with a deletion in chromosome 3q29 containing the RPL35A gene before (magenta; n = 2) and after (red; n = 3) treatment with prednisolone. The data are presented as the mean ± standard error of the mean. (B, C) Flow cytometric analysis of terminally differentiated erythroid cells from a control donor and from a patient with a deletion in chromosome 3q29 using antibodies against CD41, CD44, and CD235 before steroid treatment (B) and after steroid treatment (C).

Figure 4.

Analysis of erythroid cells from patients with a mutation in GATA1. (A) Flow cytometric analysis of CD34+ cells from the carrier mother and a DBA patient with a GATA1 mutation cultured for 14 days. The cells were labeled with CD235, CD44, and CD41 antibodies. (B) Lysates from differentiating erythroid cells from a patient with a GATA1 mutation, the carrier mother, and a control donor were immunoblotted with antibodies against GATA1 and α-tubulin (loading control).

Figure 5.

Unsupervised hierarchical clustering of normal and DBA erythroid microarray profiles. Unsupervised hierarchical clustering of specimens by expression levels of the most variant features (coefficient of variation is ≥30% across all DBA [blue] and normal [green] control samples) shows partitioning (top dendrogram split) predominantly by differentiation stage rather than by disease status. The cell populations assayed are CD41−/CD44+/CD235− (pink) and CD41−/CD44+/CD235+ (burgundy).

Figure 6.

GSEA of CD235− cells. (A-I) Comparison of the transcriptome of normal CD41−/CD44+/CD235− cells to those from GATA1- or RP-mutated DBA patient CD41−/CD44+/CD235− cells, as well as differences in direct comparison of RP and GATA1 specimens, by GSEA. Genes are ranked by signal-to-noise ratio according to differential expression, with the position of genes comprising each set (heme, GATA1, snoRNA) illustrated by the vertical black bars below the plot in each comparison. Overrepresentation of genes from a set at the top (left) or bottom (right) of the ranked list indicates over- and underexpression and is illustrated by a peak (positive) or valley (negative) in the running enrichment score (green) shifted to the left or right, respectively. The gray line represents no change. The normalized enrichment score and false discovery rate q-value are shown for each comparison illustrated. Evaluation of heme synthesis genes (from the MSigDB hallmark set) demonstrated (A) modest downregulation in GATA1-mutated samples when compared with normal controls that was not statistically significant but showed significant overexpression in RP-mediated DBA in comparison with normal controls (D) and GATA1-mutated patients (G). Patients with GATA1 mutations did not show underexpression of any tested GATA1 gene set (B; GATA1 TRANSFAC set from Ref. illustrated), whereas patients with RP mutations showed marked and statistically significant GATA1 target gene overexpression in comparison with both normal controls and patients harboring GATA1 mutations (E, H; GATA1 GSE628 set from Ref. illustrated). Similarly, evaluation of small nucleolar RNA transcription (gene set derived from all snoRNA annotated on the array) showed highly significant downregulation in patients with GATA1 mutations in comparison with normal controls (A) and DBA patients with RP mutations (I), but no significant changes in expression of RP-mediated DBA patients when compared with normal controls (F). NES, normalized enrichment score.

Figure 7.

A convergent model for the effect of RP and GATA1 mutations on erythroid differentiation in DBA. (A) Under conditions of normal hematopoiesis, rising GATA1L levels (gold stars, nuclear) in CD41−/CD44+/CD235− cells promote the erythroid program by stimulating the expression of canonical GATA1 targets, such as heme synthesis genes, as well as enhancing expression of components of the translational apparatus, including ribosomal RNA, snoRNA, and ribosomal proteins. The net effect of these activities is the accumulation of sufficient translational competence (cytoplasmic, illustrated as mature 80S ribosomes) to support normal differentiation and proliferation from proerythroblasts into early and late basophilic erythroblasts. (B) In RP-mutated erythropoiesis, GATA1L is present to promote heme synthesis as well as upregulation of the translational apparatus; however, abortive ribosomal assembly (illustrated as a large-subunit defect with accumulation of free cytoplasmic 40S ribosomes) limits accretion of translational competence, with reduced proliferation of erythroid progenitors, delays in the maturational program, and anemia. (C) In GATA1-mediated DBA, the absence of GATA1L limits upregulation of the translational apparatus without abortive ribosome assembly, but with similar consequences to proliferation and differentiation of erythroid progenitors. This model places GATA1 as a master regulator in DBA that operates upstream from RP gene alterations. EB, early basophilic erythroblasts; H, heme synthesis genes; LB, late basophilic erythroblasts; Pro, proerythroblasts.