Sec23b deficiency In Mice Results In Pancreatic Destruction and Defective long Term Hematopoietic Stem Cell Function

Abstract 2038

Congenital Dyserythropoietic Anemia type II (CDAII) is an autosomal recessive disorder caused by mutations in the gene SEC23B. SEC23B is a component of the COPII coat complex that transports proteins from the endoplasmic reticulum (ER) to the Golgi apparatus. CDAII is characterized by mild to moderate anemia and >10% bi- or multi- nucleate erythroblasts in the bone marrow. We generated Sec23b deficient mice (Sec23b gt/gt) from ES cells with a genetrap cassette inserted into the last intron of Sec23b. Sec23b gt/gt mice die at birth with destruction of the exocrine pancreas. Eight to 12 week old lethally irradiated C57BL/6J recipients were transplanted with 10⁶ E17.5 fetal liver cells from either wildtype or Sec23b gt/gt mice. Mice were bled at 6, 8 and 12 weeks post transplant and no significant difference was observed in either hematocrit or hemoglobin (N=3-7 mice per group). Red blood cell (rbc) ghosts prepared from peripheral blood of transplanted mice showed no shift of Band 3 on SDS-PAGE analysis. Ghost fractions from Sec23b gt/gt recipients analyzed by western blot for the presence of residual ER proteins showed no difference between wildtype and Sec23b gt/gt fetal liver cell recipients. No significant difference was found in bi-nucleate erythroblasts in the bone marrow, myeloid:erythroid ratio or spleen mass. We quantitatively analyzed the proteome of mature rbc ghosts from Sec23b gt/gt recipients by Stabile Isotope Labeling of Amino acids in Cells (SILAC) proteomics using rbc ghosts from mice fed lysine labeled with ¹³C (SILAC chow). When SILAC-labeled rbc ghosts were compared by mass spectrometry analysis with ghosts from Sec23b gt/gt fetal liver cell recipients, no significant differences in abundance of rbc membrane proteins were observed. To more stringently test the ability of Sec23b gt/gt fetal liver cells to engraft, 5×10⁵ wildtype or Sec23b gt/gt fetal liver cells were co-transplanted with 5×10⁵ wildtype fetal liver cells expressing a GFP transgene. FACS analysis of peripheral blood obtained at 6 and 8 weeks post transplant showed significant out-competition of Sec23b gt/gt cells by GFP+ progenitors in Ter119+ rbcs (p<0.01, at all time points), CD3+ T cells (p<0.01 at 8 weeks), B220+ B cells (p<0.01, at all time points) and Mac1+Gr1+ myeloid cells (p<0.01, at all time points) compared to wildtype controls (N=5-7 mice per group). Fetal livers from wildtype and Sec23b gt/gt mice showed no difference in the total number of CD150+CD48-Sca1+cKit+(lineage-) long term hematopoietic stem cells (LT-HSCs) per fetal liver (p=0.45). In conclusion, Sec23b deficient humans and mice exhibit disparate phenotypes, apparently restricted to CDAII in humans and a prominent neonatal pancreatic insufficiency in mice. These differences could be due to evolutionary changes in relative expression and/or function of the Sec23a and b paralogs in humans and mice. However, we cannot exclude an allele-specific defect due to residual or aberrant function of the Sec23b gt allele. Analyses of a deleted and a conditional floxed Sec23b null allele are currently in progress. Finally our transplant data suggest that Sec23b traffics cargoes important for the function of murine LT-HSCs. This finding contrasts with the apparently erythroid-specific human phenotype, raising the possibility of a requirement for Sec23b function in the marrow microenvironment not tested by these transplant studies. Tissue specific deletion of Sec23b is currently in progress to address this question.