Haploinsufficiency of the essential gene Rps12 causes defects in erythropoiesis and hematopoietic stem cell maintenance

Ribosomal protein (Rp) gene haploinsufficiency can result in Diamond-Blackfan Anemia (DBA), characterized by defective erythropoiesis and skeletal defects. Some mouse Rp mutations recapitulate DBA phenotypes, although others lack erythropoietic or skeletal defects. We generated a conditional knockout mouse to partially delete Rps12. Homozygous Rps12 deletion resulted in embryonic lethality. Mice inheriting the Rps12KO/+ genotype had growth and morphological defects, pancytopenia, and impaired erythropoiesis. A striking reduction in hematopoietic stem cells (HSCs) and progenitors in the bone marrow (BM) was associated with decreased ability to repopulate the blood system after competitive and non-competitive BM transplantation. Rps12KO/+ lost HSC quiescence, experienced ERK and MTOR activation, and increased global translation in HSC and progenitors. Post-natal heterozygous deletion of Rps12 in hematopoietic cells using Tal1-Cre-ERT also resulted in pancytopenia with decreased HSC numbers. However, post-natal Cre-ERT induction led to reduced translation in HSCs and progenitors, suggesting that this is the most direct consequence of Rps12 haploinsufficiency in hematopoietic cells. Thus, RpS12 has a strong requirement in HSC function, in addition to erythropoiesis.


Introduction 39
In the cell, protein synthesis is one of the most energetically expensive processes, and 40 both the specificity and overall level of translation are tightly regulated. The ribosome is 41 the macromolecular machine tasked with translating mRNAs into proteins and, as such,  KO/+ and RpS12 +/+ littermates starting at day 5 of age (+/+ n=39 and KO/+ n=60, log-rank Mantel-Cox test p=0.012). (H) Embryo genotype segregation from crosses between RpS12 KO/+ male and female. Graph represents percentage of developed embryos and the table shows the total numbers (E%=expected percentages, E#=expected numbers, O=observed numbers). (I) Representative pictures of E13.5 embryos with their placentas. (J) E13.5 embryo weights (n=10 on each genotype, unpaired ttest p=0.0420).
among the embryos obtained (Fig. 1H), which led us to conclude that this genotype must 165 be lethal prior to stage E13.5. Furthermore, RpS12 KO/+ embryos are smaller in size 166 compared to their wildtype counterparts (Fig. 1I, J). Therefore, these results indicate that 167 RpS12 is an essential gene, whose homozygous loss leads to early embryonic lethality, 168 and heterozygous loss causes reduced growth starting in embryogenesis, in addition to 169 other defects recognized post-partum. 170 171
(K) Total number of colonies per plate (1x10 4 BM cells from 6-7-month-old mice plated in round 1 and 1x10 4 cells plated from previous plate on each re-plating round) on each round of re-plating in complete methylcellulose media (+/+ n=5 and KO/+ n=5, 2 replicates per biological sample). Statistical analysis: quantifications represent mean +/-SEM, unpaired t-tests were performed to established significance among populations between genotypes *p < 0.05, **p < 0.01 , ***p < 0.001, ****p < 0.0001  (Fig. 3B). In addition, in RpS12 KO/+ bone marrow, the numbers of all 211 hematopoietic progenitor populations were significantly reduced (Fig. 3C). Accordingly, 212 compared to the WT littermates, young (6-8-week-old) RpS12 KO/+ mice had lower bone 213 marrow cellularity and decreased spleen weights (Fig.3D, E). Additionally, older (6-7-214 month-old) RpS12 KO/+ mice also had lower HSC numbers (Fig. 3F). Interestingly, we 215 observed a partial recovery of some of the HSPC populations with age, such as multi- There are several important factors that maintain the HSC pool, including 279 quiescence, low translation levels, and cell survival. We therefore analyzed the 280 distribution of HSPCs among the cell cycle stages defined by the DNA content (Hoechst) 281 and the levels of Ki67 (Fig. 6A). We observed a lower proportion of RpS12 KO/+ HSCs in 282 the G0 stage of the cell cycle, and a significantly increased proportion in the actively 283 cycling phases G1 and S/G2/M (Fig. 6B). Similar results were observed in MPP2/3 (LSK, 284 increased levels of global translation (Fig. 6C, D). The difference was especially 297 remarkable in HSCs. Interestingly, compared to the controls, RpS12 KO/+ myeloid 298 progenitors did not exhibit differences in OPP intensity, and among different myeloid 299 progenitor populations, only the megakaryocyte-erythrocyte progenitors (MEP) had a 300 significant increase in OPP incorporation (Fig. 6E). Thus, these data suggest that a 301 Cell death can deplete the HSC pool, and can result from chronic HSC activation. 304 We asked whether this reduction of HSCs in RpS12 KO/+ animals is due to an increase in 305 apoptosis. Interestingly, compared to controls, RpS12 KO/+ animals have an increased 306 number of apoptotic cells in bone marrow cytospins (Fig. 6F). To quantify the level of 307 apoptosis in the immunophenotypic populations of the bone marrow cells, we used the 308 flow cytometry markers PI and Annexin V together with population-specific cell surface 309 markers (Fig. 6G). Our flow analysis confirmed a significant increase in apoptosis in 310 Lineage -Sca1 + c-Kit + (LSK) cells, a population that contains HSCs and MPPs, but not in 311 more mature myeloid progenitors (Fig. 6H, I). wild-type littermates, not only upon stem cell factor (SCF) stimulation, but even at the 323 non-stimulated baseline. Indeed, the RpS12 KO/+ genotype alone was a more potent activator than SCF (Fig. 7A-C). Interestingly, this was not the case for the more mature 325 myeloid progenitor cells, where the levels of p-AKT, p-S6 and p-4E-BP1 are comparable 326 to the controls in both non-stimulated and SCF-stimulated conditions, and these cells also 327 exhibited more normal translation rates (Fig. 7D-F). Since phosphorylation of S6 and 328 4EBP1 leads to increased translation, this data corroborates the increase in translation 329 observed in the RpS12 KO/+ LSK population (Fig. 6D). Interestingly, more mature MPROG 330 do not have increased translation (Fig. 6E)  Peripheral blood smears and cytospins from RBC lysed bone marrow samples were 567 stained using the Hema 3 System (Fisher) following the manufacturer's instructions. The 568 images were acquired using a Zeiss Axiovert microscope with a digital camera. 569 570

Statistical methods 571
Two-tailed Student's t-tests were performed to compare statistical significance between 572 two samples. When comparing more than 2 groups, one-way ANOVA tests were performed with the Turkey's multiple comparison test. For presence/absence of 574 phenotype, statistical significance was calculated with Fisher's exact test. Analysis was 575 done using GraphPad Prism v9. 576 577

Reagents 578
All antibodies used for flow cytometry assays, and primers used for CRISPR gene editing 579 and PCR can be found in Supplementary Tables 1, 2, and 3. All other reagents are 580 mentioned in the methods section.  Non-competitive transplants were performed twice, using different controls: RpS12 flox/+ or RpS12 flox/flox . The competitive transplant was performed once, using RpS12 flox/+ mice as a control group Statistical analysis: data represent mean +/-SEM, unpaired t-tests were performed to assess significance among populations between genotypes *p < 0.05, **p < 0.01 , ***p < 0.001, ****p < 0.0001 Figure 1. CRISPR gene editing and genotyping strategy for the generation of RpS12 Flox and RpS12 KO (A) Diagram of the WT, Flox and KO alleles of RpS12 generated in this study indicating the position of Snord100 and Snora33 (snoRNA),Cas9 gRNAs target locations, and primers used for genotyping. The homology arms starting sites are indicated and the ends fall outside of the RpS12 locus. To identify the first transformants, two pair of primers were used for PCR amplification: F2/R2 and F3/R3. F2 and R3 fall outside of the sequence covered by the homology arms, to ensure the inserts are on the correct location. The presence of LoxP sites was confirmed by Sanger sequencing using primers F1 and F4 for F2/R2 fragments, and with F3 and R3 for F3/R3 fragments. To determine excision of exon 2 and 3 by Cre recombination primers F1 and R1 were used, which generate a 900bp fragment in RpS12 + and a 300bp fragment in RpS12 KO (B).