Deficiency of ribosomal protein S19 in CD34⁺ cells generated by siRNA blocks erythroid development and mimics defects seen in Diamond-Blackfan anemia

Diamond-Blackfan anemia (DBA) is a hypoplastic congenital anemia that presents within the first 2 years of life.¹  In addition to symptoms of anemia, patients display various other symptoms, including growth retardation, skeletal deformities, congenital heart defects, mental retardation, or other abnormalities.^2-5  Patients show low hemoglobin and high fetal hemoglobin levels, high adenosine deaminase activity, and high mean corpuscular volume; the diagnosis is confirmed by bone marrow (BM) examination, which typically shows a normal cellularity but few erythroid progenitors.^1,6-8  The initial treatment is steroid therapy, but patients not responding to steroids require regular transfusions that over time result in lethal tissue iron overload.^1,2  A BM transplantation can cure DBA,⁹  but only a fraction of patients have an HLA-matched sibling. Most cases of DBA are sporadic, but 10% to 25% have a positive family history.^4,10  Twenty-five percent of the patients have a mutation in the ribosomal protein S19 (RPS19) gene.^11,12  We have previously shown that RPS19 gene transfer into BM cells from several RPS19-deficient DBA patients increases the number of erythroid progenitors (erythroid burst-forming unit [BFU-E] and erythroid colony-forming unit [CFU-E]) 2- to 3-fold compared with patient cells transduced with a green fluorescent protein (GFP) control vector. RPS19 transgene overexpression also improves the erythroid progenitor proliferation defect in CD34⁺ cells from patients with RPS19-deficient DBA in vitro.^13,14  A second DBA locus on chromosome 8p22-p23 shows genetic linkage in 50% of DBA patients, but the gene has not been identified yet.¹⁵  The pathophysiologic mechanism of the disease remains unknown.

To investigate the hematopoietic and molecular mechanisms responsible for pathogenesis in RPS19-deficient DBA, it is essential to develop a suitable model of RPS19 deficiency that allows extensive molecular studies. Recently, Matsson et al¹⁶  developed a mouse model with a targeted disruption of the RPS19 gene. The homozygous RPS19^-/- mice exhibit an embryonic lethal phenotype prior to placental implantation in mice with C57BL/6J background. Unlike human RPS19-deficient DBA patients, the haploinsufficient RPS19^+/- heterozygote mice have normal levels of RPS19 mRNA and display no hematologic abnormalities. Collectively, it is clear that novel murine and human model systems for DBA are required to explore the molecular mechanism for the disease, to discover the pathogenesis of the anemia, and to develop new therapeutic strategies for the disorder.

It has been shown that specific knockdown of endogenous genes can be achieved by RNA interference. Permanent, stable expression of small interfering RNAs (siRNAs) that target specific mRNAs for degradation can be generated with plasmid- or viral vector-based delivery systems.^17-20  In this study we take advantage of a lentiviral vector system that we have recently developed in which the 3′ long-terminal repeat, which duplicates during reverse transcription, contains the human polymerase III promoter H1, driving expression of an siRNA hairpin precursor sequence directed against the RPS19 mRNA.²¹  Using this system we demonstrate that down-regulation of RPS19 mRNA in primary human CD34⁺ hematopoietic cells significantly impairs erythroid development and mimics defects in erythroid differentiation seen in cells from DBA patients.

The human RPS19 cDNA sequence (NM_001022)¹²  was searched for suitable siRNA target sequences starting with aag followed by 18 nucleotides. The 21nt sense and antisense sequences were subjected to BLAST searches, eliminating sequences with more than 16 base pair (bp) homologies in the human genome.²²  A control vector was designed by scrambling the B sequence so that there was no significant homology to any sequence in the human genome.

The RPS19 siRNA A oligo nucleotides included gatccccgcacaaagagcttgctcccttcaagagagggagcaagctctttgtgctttttggaaa and agcttttccaaaaagcacaaagagcttgctccctctcttgaagggagcaagctctttgtgcggg. RPS19 siRNA B oligo nucleotides included gatccccgtccgggaagctgaaagtcttcaagagagactttcagcttcccggactttttggaa and agctttccaaaaagtccgggaagctgaaagtctctcttgaagactttcagcttcccggacggg. RPS19 siRNA C oligo nucleotides included gatccccgagatctggacagaatcgcttcaagagagcgattctgtccagatctctttttggaa and agctttccaaaaagagatctggacagaatcgctctcttgaagcgattctgtccagatctcggg. Non-specific control siRNA Scr oligo nucleotides included gatccccgacacgcgacttgtaccacttcaagagaggtggtacaagtcgcgtgtctttttggaaa and agcttttccaaaaagacacgcgacttgtaccactctcttgaagtggtacaagtcgcgtgtcggg. Sense and antisense siRNA sequences are shown in bold, and loop sequences are underlined. Hybridized oligonucleotides were phosphorylated by T4 DNA kinase and ligated into the pSuper vector (OligoEngine, Seattle WA) into the BglII-HindIII site, downstream of the H1 promoter. The H1-hairpin-precursor cassette was excised from pSuper with EcoR1 and Cla1 and further cloned into the EcoR1-Cla1 site of the pLV-TH plasmid.²¹ 

The lentiviral construct pLV-mRIY containing an internal spleen focus-forming virus promoter that expresses high levels of a transcript coding for yellow fluorescent protein (YFP) together with a modified RPS19 cDNA sequence made resistant to silencing by RPS19 siRNA B was constructed. A cassette consisting of the cDNA for human RPS19, an internal ribosomal entry site (IRES) and YFP was cloned into the SalI-BamH1 site of the pLithmus 38i plasmid (New England BioLabs, In Vitro Sweden AB, Stockholm, Sweden). The sequence between the AgeI and MscI sites of the RPS19 cDNA was excised and replaced with the product of the hybridization between the forward primer 5′-CCGGTGCCACCATGCCAGGTGTAACAGTTAAGGATGTCAATCAACAAGAATTTGTGAGAGCT-3′ and the reverse primer 5′-CTCACAAATTCTTGTTGATTGACATCCTTAACTGTTACACCTGGCATGGTGGCA-3′. Next, the cDNA between the MscI and EcoRI sites was amplified in a polymerase chain reaction using the forward primer 5′-TGAGAGCTCTCGCTGCGTTTCTGAAGAAATCGGGCAAACTGAAAGTCCCCGA-3′ and the reverse primer 5′-CTTTGTGCTTGGCCAGCTTGACGGTA-3′. Altered bases are underlined. The modified RPS19-IRES-YFP cassette was finally cloned into the AgeI-BsrGI site of the p9960 plasmid (a generous gift from Christopher Baum, Department of Hematology and Oncology, Hannover Medical School, Hannover, Germany). The pLV-YFP control vector was constructed by cloning YFP into the AgeI-BsrGI site of the p9960 plasmid.

Lentiviral vectors were produced by transient transfection of 293T cells according to standard protocols. Briefly, 293T cells were cultured in Dulbecco modified Eagle medium (DMEM), supplemented with 10% fetal calf serum (FCS), and, when subconfluent, transfected with 7 μg pCMV-ΔR8.91, 5 μg pMDG, and 10 μg pLV-TH-Scr (Scr), pLV-TH-RPS19-A (A), pLV-TH-RPS19-B (B), pLV-TH-RPS19-C (C), pLV-mRIY (RPS19 rescue), or pLV-YFP (YFP). The transfection was performed by calcium-phosphate precipitation. Medium was changed to DMEM and supplemented with 2% FCS after 6 to 8 hours. Vector supernatants, containing viral particles, were harvested approximately 48 hours later and concentrated by ultracentrifugation (1.5 hours at 25 000g at 4°C). The viral batches were titered on HeLa cells.

Umbilical cord blood and bone marrow samples were collected from healthy donors after informed consent under protocols approved by the Lund University Hospital Ethics Committee. CD34⁺ cells were isolated as previously described,¹³  using a Lymphoprep density gradient (Nycomed, Oslo, Norway) and Midi MACS (magnetic-activated cell sorting) LS⁺ separation columns and isolation kit (Miltenyi Biotec, Auburn, CA), then frozen in DMEM supplemented with 20% FCS and 10% dimethyl sulfoxide (DMSO) and stored in liquid nitrogen. On the day of transduction, cells were thawed and cultured in serum-free medium X-Vivo 15 (Bio Whittaker, Walkersville, MD) supplemented with 1% bovine serum albumin (BSA; Stem Cell Technologies, Vancouver, BC, Canada), 2 mM l-glutamine (Gibco BRL, Grand Island, NY), 10^-4 mM 2-mercaptoethanol (Sigma Chemicals, St Louis, MO), 100 IU/mL penicillin, 100 μg/mL streptomycin (Gibco BRL), and the following cytokines: stem cell factor (SCF; 100 ng/mL), thrombopoietin (100 ng/mL), and Flt-3 ligand (100 ng/mL). Cells were transferred to 24-well plates (2.5 × 10⁵ cells/well), precoated with fibronectin fragment CH-296 (Retronectin; Takare Shuzo, Otsu, Japan), and blocked with 2% BSA in phosphate-buffered saline (PBS). The cells were incubated at 37°C in a humidified atmosphere with 5% CO₂ for approximately 12 hours and, thereafter, transduced with viral vectors at a multiplicity of infection of 10 (5 × 10⁶ HeLa transducing units/mL). Cells were sorted 48 hours later for green fluorescence using a fluorescence-activated cell sorting (FACS) Vantage cell sorter (Becton Dickinson, San Jose, CA). The transduction efficiencies were 25% to 30% in CD34⁺ cord blood (CB) cells and 20% to 25% in CD34⁺ BM cells. In the rescue experiment sorted GFP⁺ cells were transduced with the pLV-mRIY and pLV-YFP vectors at a multiplicity of infection of 5, resulting in 50% transduction efficiency 2 days later. The cells were cultured in erythroid differentiation medium and analyzed by FACS 7 days later to determine YFP and Glycophorin A expression.

Total RNA was isolated from 1 × 10⁵ GFP⁺CD34⁺ cells using the RNeasy micro kit as described by the manufacturer (Qiagen GmbH, Hinden, Germany), and cDNA was reverse transcribed using Superscript II (Invitrogen, Carlsbad, CA). The expression level of RPS19 was analyzed by quantitative reverse transcriptase-polymerase chain reaction (Q-RT-PCR) using a LightCycler instrument (Roche, Idaho Falls, ID). The cDNA was used for quantitative PCR using Sybr GreenI (Sigma) for detection of PCR products. cDNA (2 μL) was used in a 15-μL final volume reaction containing 1 U Platinum Taq DNA Polymerase, 1 × buffer (provided with the enzyme), 0.8 mM dNTP (deoxyribonucleoside triphosphate), 3 mM MgCl₂, 0.5 mg/mL BSA, 5% DMSO, 0.5 μM RPS19 forward (fw; 5′-GCC TGG AGT TAC TGT AAA AGA CG-3′), 0.5 μM RPS19 reverse (rev; 5′-CCC ATA GAT CTT GGT CAT GGA GC-3′), and 1:20 000 dilution of Sybr GreenI. The RPS19 values were normalized against human β-actin (huAktEx4 F, 5′-CCA TTG GCA ATG AGC GGT T-3′; huAktEx6R, 5′-GCG CTC AGG AGG AGC AA-3′).

Total RNA was isolated from GFP⁺ CB cells using TRIzol reagent, as described by the manufacturer (Invitrogen, Carlsbad, CA). RNA from each sample was loaded on a 1% agarose gel, transferred onto a Hybond N⁺ membrane (Amersham Pharmacia, Buckinghamshire, United Kingdom), and hybridized using a full-length human RPS19 probe obtained by digesting the pSkRIGw plasmid with AgeI and EcoRI. The probe was labeled with (³²P) deoxycytidine triphosphate using a random priming kit (Amersham Pharmacia). The membrane was washed twice with 2 × standard saline citrate (SSC) at 42°C, twice with 2 × SSC + 1% sodium dodecyl sulfate (SDS) at 65°C and twice with 0.1 × SSC + 1% SDS at 65°C and subsequently exposed to Kodak (Rochester, NY) x-ray film.

The chicken polyclonal anti-RPS19 antibody was derived from eggs immunized with 2 RPS19-derived peptides (C)GVMPSHFSRGSKSVA and (C)VEKDQDGGRKLTPQG. Immunoglobulin Y was extracted²³  and loaded on a HiTrap normal human serum (NHS)-activated column (Amersham Biosciences, Buckinghamshire, England) containing covalent coupled Escherichia coli produced His-tagged RPS19 for affinity purification. Total cellular proteins were extracted by boiling in sample buffer containing 60 mM Tris (tris(hydroxymethyl)aminomethane) HCl, pH 6.8, 2% SDS, 5% vol/vol glycerol, 2% β-mercaptoethanol, 20 mM dithiothreitol. Proteins were separated by 15% SDS-polyacrylamide gel electrophoresis and blotted to polyvinyliden fluoride membrane (Hybond-P; Amersham Bioscience, Uppsala, Sweden) and analyzed with anti-RPS19 and anti-Actin antibodies (BD Biosciences, Lexington, KY). Proteins were visualized using chemiluminescence reagents (Western Lightning; Perkin Elmer Life Science, Boston, MA) according to the manufacturer's protocol. Densitometry was performed on developed film by using Image J 1.30v software (http://rsb.info.nih.gov/ij/) to quantify the data.

After each sorting, 2000 GFP⁺ BM or CB cells transduced with each vector were plated in quadruplicate 35-mm dishes in 1.1 mL H4230 methylcellulose (Stem Cell Technologies) with erythropoietin (Epo; 5 U/mL; Janssen-Cilag, Sollentuna, Sweden), interleukin 3 (IL-3; 100 ng/mL; a gift from Novartis, Basel, Switzerland), and granulocyte-macrophage colony-stimulating factor (GM-CSF; 200 ng/mL; a gift from Novartis). The cells were cultured at 37°C in a humidified atmosphere with 5% CO₂. BFU-E and CFU-G/GM (colony-forming unit-granulocyte/granulocyte-macrophage were counted on day 14, and the average of the 4 plates was determined. Each colony experiment was repeated 5 times except for the DBA patient cells that were done twice.

To assay the developmental potential of RPS19-silenced cells, 2000 CD34⁺GFP⁺ BM cells were cultured in 96-well microtiter plates containing medium supporting either erythroid, myeloid, or erythroid and myeloid (mixed) differentiation as described before.^24,25  Erythroid liquid culture medium included the following: Iscoves modified Dulbecco medium (IMDM) containing FCS (30%), penicillin-streptomycin (100 U/mL), hydrocortisone (10^-6 M), 2-mercaptoethanol (2-ME; 10^-4 M), Epo (10 U/mL), IL-3 (0.01 ng/mL), and IL-6 (0.001 ng/mL).²⁵  Myeloid liquid culture medium included the following: X-vivo-15 containing FCS (20%), penicillin-streptomycin (100 U/mL), SCF (20 ng/mL), IL-3 (30 ng/mL), granulocyte colony-stimulating factor (G-CSF; 100 ng/mL).²⁴  Myelo-erythroid liquid culture medium included the following: X-vivo-15 containing FCS (20%), penicillin-streptomycin (100 U/mL), SCF (20 ng/mL), IL-3 (30 ng/mL), G-CSF (100 ng/mL), Epo (2 U/mL).²⁴  Cell count and FACS analysis were done on days 7, 14, and 21. DAF (4,5-diaminofluoroene) staining and cytospins (Wright-Giemsa) were done at day 7.

A Zeiss Axioplan microscope (Jena, Germany) was used with a 20 ×/0.60 lens. The picture files, taken with an Olympus C-3040 Zoom camera (Olympus, Melville, NY), were pasted in Microsoft PowerPoint X for Mac (Redman, WA) slides and converted to TIFF-formatted picture files. The figure files were cropped in Adobe Photoshop 6.0 (San Jose, CA) and formatted to jpg format.

As described before, 2,7-diaminofluroene (DAF; Sigma) stains hemoglobin-containing cells blue because of the hemoglobin-induced pseudoperoxidase reaction.²⁶  Microscopically dark blue cells were counted as DAF positive.

Cultured CB and BM cells were analyzed by FACS Calibur and CellQuest software (Becton Dickinson). The CD34, CD33, CD13, CD71, Glycophorin A, and CD41 antibodies were purchased from Becton Dickinson. Cells were sorted on the basis of their expression of GFP or CD71/GlyA (Glycophorin A) using a FACS Vantage cell sorter (Becton Dickinson). FACS analysis for simultaneous detection of GFP and YFP was performed using special filters as previously described.²⁷ 

Results are expressed as mean values plus or minus standard errors of the mean (SEMs). Comparison among groups was made using the two-tailed nonparametric Mann-Whitney test.

To study the effects of RPS19 mRNA deficiency in hematopoiesis we asked whether a DBA phenotype could be induced in healthy cells by permanent knockdown of RPS19 mRNA expression using siRNA-expressing lentiviral vectors. Three siRNA sequences against RPS19 mRNA (RPS19-A, RPS19-B, and RPS19-C) and 1 scrambled (Scr) control sequence were inserted into a lentiviral construct (pLV-TH)²¹  to generate 3 separate siRNA vectors and 1 control vector (Figure 1A). CD34⁺ CB cells were transduced with the siRNA-expressing vectors, and GFP⁺ cells were sorted. Northern blot analysis of GFP⁺ cells shows that vectors LV-TH-RPS19-A, -B, and -C knock down RPS19 mRNA expression to various levels, while the LV-TH-Scr control vector did not affect the level of RPS19 mRNA, compared with mock-transduced cells (Figure 1B). RPS19 mRNA levels were further quantified by Q-RT-PCR, and, as indicated in the Northern blot, the silencing effect was mild in LV-TH-RPS19-A-transduced cells and most pronounced in LV-TH-RPS19-C-transduced cells (Figure 1C). Repeated Western blots of sorted cells, 3 (data not shown) and 5 days (Figure 1D) after transduction, show a reduction in RPS19 protein levels whereby the RPS19 protein is gradually decreasing over time because the reduction in protein levels was less pronounced and took longer time to develop. At 5 days following transduction, the siRNA vectors have generated haploinsufficient levels (40%-60% of normal) in the transduced cells (Figure 1E). The reduction in RPS19 protein levels reflected the RPS19 mRNA levels and was most pronounced in cells treated with LV-TH-RPS19-C vector and least affected in cells treated with the LV-TH-RPS19-A vector.

We asked whether healthy CD34⁺ BM cells expressing siRNA against RPS19 mRNA would exhibit a defect in erythroid development, reminiscent of the impaired erythroid development seen in BM cells from patients with DBA. To study clonogenic capacity of hematopoietic progenitors, transduced and sorted CD34⁺GFP⁺ CB cells were plated in semisolid medium supporting BFU-E and CFU-G/GM colony formation. BFU-E colony formation capacity was significantly reduced in RPS19-silenced CB cells, and the severity of the defect correlated with the level of RPS19 mRNA deficiency (Figure 2A). The CFU-G/GM colony formation capacity was also decreased, although the difference was not statistically significant (Figure 2B). The colony formation experiments were repeated with BM-derived CD34⁺ cells, in which the BFU-E defect was similar to that in CB cells (Figure 2C). CFU-G/GM formation was significantly decreased in BM cells transduced with the LV-TH-RPS19-C vector (Figure 2D). Myeloid colony formation was only significantly affected with the vector that generated the lowest levels of RPS19. The colony-forming defect in RPS19-silenced cells was compared with Scr-transduced BM cells from a previously studied DBA patient lacking 1 allele of the RPS19 gene.¹³  Normal CD34⁺ BM cells transduced with the LV-TH-RPS19-B vector exhibited a similar decreased level in colony-forming capacity as with the LV-TH-Scr-transduced DBA cells.

To investigate different stages of erythroid development, we cultured CD34⁺GFP⁺ BM cells in liquid culture under conditions that are known to produce unilineage erythroid differentiation as previously described.^14,25  In RPS19-deficient cells erythroid proliferation decreased (Figure 3A). At days 7 and 14 cells were taken from the erythroid culture to determine the stage of erythroid maturation. At day 7, DAF staining^2,28  showed a reduced fraction of hemoglobin-containing cells in cultures transduced with vectors LV-TH-RPS19-B and LV-TH-RPS19-C (Figure 3B). To monitor the stages of erythroid development by flow cytometry analysis, cells were stained with Glycophorin A and CD71 (Figure 3C), as described by Rogers et al,²⁹  and sorted from gates R1, R2, and R3. The 3 distinct cell populations morphologically agree with Rogers et al²⁹ ; R1 (CD71^loGlyA^lo/-) cells contained 25% blasts and 75% myeloid cells, R2 (CD71⁺GlyA^int) cells mostly contained immature erythroid cells, while the R3 (GlyA^hi) cells were almost entirely composed of more mature erythroid cells (Figure 3; Tables 1 and 2). Untransduced cells sorted out from gate R1 (CD71^loGlyA^-) at day 7 formed 115 CFU-G/GM and 3 BFU-E colonies per 1000 plated cells, while cells from gate R2 (CD71⁺GlyA^int) formed 20 CFU-E and no BFU-E or CFU-G/GM colonies, showing that cells in gate R2 represent mostly erythroid progenitors (Table 2). Comparisons of the 3 fractions in RPS19-silenced BM cells showed a relative increase of the non-erythroid-committed cell fraction (R1) and a decreased fraction of erythroblasts and mature erythroid cells (R2 and R3) day 7 (data not shown) and day 14 (Figure 3E). This indicates a failure in erythroid maturation at a differentiation stage earlier than the cells that were gated in R2. Q-RT-PCR from cells in gates R1 and R2 at day 7 showed that RPS19-silenced cells maintained RPS19 expression at levels within 30% to 60% of mock- and LV-TH-Scr-transduced cells in 2 separate experiments (data not shown).

To confirm that the erythroid defect was specifically caused by RPS19 deficiency we designed a lentiviral vector expressing YFP together with RPS19. To rescue the RPS19 deficiency, the coding sequence of RPS19 was changed so that it codes for the normal amino acid sequence but is no longer silenced by the LV-TH-RPS19-B vector (Figure 4). CD34⁺ BM cells were transduced with the siRNA vectors and after 2 days sorted for GFP and again transduced with the RPS19-rescuing vector LV-mRIY or the YFP control vector LV-YFP. After 7 days of culture in erythroid differentiation medium the cells were analyzed for YFP and Glycophorin A expression. In RPS19-rescued LV-TH-RPS19-B-transduced cells, the YFP⁺ fraction contained the same percentage of GlyA⁺ cells as the LV-TH-Scr controls (Figure 4B). As expected YFP-control-transduced and YFP^- LV-TH-RPS19-B-transduced cells had a reduced fraction of GlyA⁺ cells. Although YFP expression was stable at 50% in the other groups, the rescued YFP⁺ fraction of LV-TH-RPS19-B-transduced cells increased from 50% on day 2 to 70% on day 7 (data not shown), indicating a rescue of the proliferative defect.

Since some patients develop neutropenia, thrombocytopenia, or both during the course of DBA, we asked whether the myeloid proliferation and differentiation is altered in cells transduced with siRNA against RPS19. Transduced GFP⁺ BM cells were cultured in medium containing IL-3, SCF, and G-CSF for 21 days. Decreased proliferation capacity in cells transduced with vectors LV-TH-RPS19-B and -C is shown in Figure 5A. Transduction with vector LV-TH-RPS19-A did not affect the proliferation in myeloid culture conditions. RPS19 expression in silenced cells was 20% to 60% of normal in cells from gates 1 to 3 at day 14 as determined by Q-RT-PCR (data not shown). To study the effects of RPS19 silencing on granulocytic development, cells from day 14 were stained with CD33 and CD13, and 3 distinct CD33⁺ cell populations were studied as sorted: CD13^- (gate 1), CD13⁺ (gate 2), and CD13⁺⁺ (gate 3) (Figure 5B). Differential counts of the different populations are shown in Table 3. In myeloid culture conditions RPS19-silenced cells contained a higher fraction of CD13^+/hi cells (Figure 5C), showing that even though proliferation is reduced, the myeloid maturation is not arrested.

To compare the effects of RPS19 silencing on proliferation and maturation of myeloid and erythroid cells, sorted GFP⁺CD34⁺ BM cells were cultured under conditions that enable differentiation into both myeloid and erythroid lineages (IL-3, SCF, G-CSF, and erythropoietin). Figure 6A shows the number of GFP⁺ cells at different time points during the liquid culture expansion. The efficient RPS19 knockdown by vectors LV-TH-RPS19-B and -C clearly reduced the proliferation capacity, while cells transduced with the least effective vector LV-TH-RPS19-A show no significant difference compared with the cells transduced with the Scr control vector. The erythroid versus myeloid lineage distribution of cultured cells changes in RPS19-silenced cells. The ratio of cells positive for CD41 or CD13 divided by cells positive for Glycophorin A increased in RPS19-silenced cells (Figure 6B), indicating that erythroid differentiation is more sensitive to RPS19 silencing than the myeloid differentiation.

Since RPS19 is a ubiquitously expressed protein required for efficient translation in all cells, very low RPS19 expression is likely to hamper various cell functions, leading to cell cycle arrest or apoptosis, which is seen in the homozygous RPS19 knock-out mice that die prior to implantation.¹⁶  To create a valid model for DBA pathogenesis, RPS19 expression must be reduced to levels that induce an erythroid defect without interfering with general cell functions. Using 3 different siRNA sequences against RPS19, which show a gradual increase of RPS19 mRNA down-regulation efficiency, we show for the first time that a lowered RPS19 level in hematopoietic progenitor cells causes a defect in hematopoiesis similar to that seen in BM cells from DBA patients.

To compare sensitivity to RPS19 deficiency in BM and CB cells, the colony formation experiments were performed with both CD34⁺ BM and CB cells. BFU-E colony formation capacity was significantly reduced in RPS19 silenced CD34⁺ cells, and the severity of the defect correlated with the level of RPS19 deficiency. RPS19-silenced CD34⁺ cells from BM and CB both formed fewer BFU-E colonies. Significant reduction in myeloid colony formation was also seen with the vector that generated the most severe down-regulation of RPS19, but no statistically significant reduction in myeloid colony formation could be observed with the other vectors. These findings suggest a more general hematopoietic defect than the classic definition of DBA. These findings are consistent with reports from the literature. Santucci et al³⁰  observed an impairment along the granulocyte-macrophage pathway in addition to reduced erythroid capacity in cells from 10 of 12 DBA patients, pointing toward a bilineage progenitor defect. Similarly, Giri et al³¹  followed 28 steroid-refractory DBA patients of which 21 developed a general BM hypoplasia. Additional reports of reduced leukocyte and altered thrombocyte counts in DBA patients^32-34  also indicate that the underlying hematologic defect in DBA can affect other lineages than the erythroid.

To study early and late stages of erythroid differentiation, cells were cultured in liquid medium culture strongly supporting erythroid differentiation. After 7 days RPS19-silenced cell cultures contained a reduced fraction of hemoglobin-containing cells. After 14 days 3 populations, 1 representing myeloid cells and 2 different stages of erythroid development, were analyzed. Gate R1 (CD71^lo/GlyA^lo/-) contains a DAF-negative, myeloid cell population with some BFU-E but mostly CFU-G/GM colony-forming potential. Gate R2 (CD71⁺/GlyA^int) contains DAF-positive, erythroid progenitor cells, mostly with CFU-E potential, and gate R3 (GlyA^hi) contains more mature erythroid cells. RPS19-silenced erythroid cell cultures contain a higher fraction of cells in gate R1 and a lower fraction of cells in gates R2 and R3, displaying similar ratios as CD34⁺ BM cells from a DBA patient lacking 1 RPS19 allele cultured under the same conditions. Since CD34⁺ BM cells from RPS19-deficient DBA patients and from healthy RPS19-silenced BM cells fail to form normal ratios of erythroid progenitor cells, we suggest a significant block in erythroid development prior to the formation of CFU-Es as a major cause of erythroblastopenia and anemia in DBA patients. Our findings do not rule out additional reduction in erythroid differentiation at the CFU-E stage and beyond. This failure of early erythroid development can be caused by a requirement of relatively high RPS19 expression during a critical stage of erythroid development. Previous studies show that RPS19 expression is higher in early hematopoietic progenitor cells than in more differentiated erythroid cells.^14,35  In a careful in vitro study using peripheral blood cells from DBA patients, Ohene-Abuakwa et al³⁶  found that the main defect in DBA erythropoiesis is somewhat later (after the Epo-triggered onset of erythroid differentiation). However, no erythropoietin was used during the first 7 days of culture, and this difference may account for the differences in our and their findings.

RPS19 silencing decreased myeloid proliferation in myeloid liquid culture. However, the cells that succeeded to proliferate were mature myeloid cells, even to a larger extent than the control cells. In addition, culture of transduced cells in liquid medium that enables both erythroid and myeloid differentiation increased the fraction of cells belonging to the myeloid (CD13⁺) or megakaryocyte (CD41⁺) lineage compared with the erythroid lineage (GlyA⁺). Cell proliferation was always lower in RPS19-silenced cells, pointing toward a general progenitor proliferation defect in RPS19-deficient DBA, that agrees with the proliferation deficiency of early (CD34⁺/CD38^-) progenitor cells from DBA patients in liquid culture.¹⁴ 

The hematopoietic system has a great proliferative potential, capable of increasing cell production 5- to 10-fold with a 2.5-fold higher relative turnover of erythroid cells compared with myeloid cells. Reduced proliferation capacity is a possible explanation to the hematopoietic and developmental defects in DBA. The increased sensitivity to RPS19 silencing during erythroid compared with myeloid differentiation could be due to an impairment of the translational machinery or a disruption of an unknown extraribosomal function of RPS19. For example, ribosomes lacking RPS19 or containing mutated forms may alter the translation rate constants for transcripts important for erythroid development. An example of this mechanism of translational regulation is the interferon-γ-induced phosphorylation and release of ribosomal protein L13a that specifically inhibits ceruloplasmin translation.³⁷  Another possibility is that RPS19 has an extraribosomal function interacting with factors involved in erythropoiesis. RPS19 has been shown to interact with fibroblast growth factor 2 in NIH3T3 cells,³⁸  and it is possible that RPS19 also interacts with factors in erythroid development.

To understand the pathogenesis in DBA it is important to have access to cells in which the pathologic process can be induced and studied. Our model system presented in this paper significantly simplifies future studies on ribosomal and extraribosomal mechanisms involved in RPS19 deficient DBA.

We thank Zhi Ma for assistance in cell sorting; Ingbritt Åstrand-Grundström and Dr Per-Gunnar Nilsson for expert evaluation of differential counts; Dr Sarah Ball for advise on DAF staining; and Karin Olsson, Eva Nilsson, and Dr Ann Brun for help with the Northern blot and Q-RT-PCR analyses.