Defective fetal liver erythropoiesis and T lymphopoiesis in mice lacking the phosphatidylserine receptor

The phosphatidylserine receptor (PSR) has been identified as a molecule expressed on macrophages, fibroblasts, and epithelial cells that specifically binds to phosphatidylserine (PS) exposed on apoptotic cells.¹  However, several molecules have been also implicated in the recognition and ingestion of apoptotic cells by macrophages. These include cell-surface molecules such as CD14, class A scavenger receptor, adenosine triphosphate binding cassette transporter 1, receptor tyrosine kinase Axl/Mer/Tyro3, and α_Vβ₃ integrin, which, in association with CD36, binds to thrombospondin recognized by an undefined ligand on apoptotic cells.^2-7  In addition, 2 soluble molecules, growth arrest-specific gene 6 product and milk fat globule-endothelial growth factor-factor 8, have been reported to bind to PS.^8,9  Therefore, although in vitro experiments clearly indicate that PSR is involved in anti-inflammatory clearance of cells undergoing apoptosis,^1,10,11  the physiologic relevance of PSR remains unclear. To address this issue, we generated PSR-deficient mice by homologous recombination in embryonic stem (ES) cells.

The targeting vector was constructed to replace a 2.8-kilobase fragment from the EcoRI site located upstream of exon 1 to the EcoRI site within exon 3 with a neomycin resistance cassette. Correctly targeted ES clones were microinjected into C57BL/6 blastocysts, and chimeras obtained were crossed with C57BL/6 mice. Mice heterozygous for the mutant allele were crossed to obtain PSR^-/- embryos.

Poly(A) mRNA was prepared from mouse pups and hybridized with a cDNA probe.

Hemoglobin (Hb) concentration and hematocrit (Ht) were determined with an automatic cell analyzer (Sysmex, Kobe, Japan). Blood smear samples were stained with Giemsa solution (Merck, Darmstadt, Germany).

Cells were stained with the relevant monoclonal antibodies (PharMingen, San Diego, CA) and analyzed on a FACScan (Becton Dickinson, Mountain View, CA).

Fresh frozen sections were stained with terminal deoxynucleotidyl transferase-mediated deoxy uridine triphosphate-fluorescein nick end labeling (TUNEL) and then incubated with phycoerythrin-conjugated monoclonal antibody to F4/80 (Caltag Laboratories, Burlingame, CA).

Total RNA samples treated with RNase-free DNase I (Invitrogen, Carlsbad, CA) were reverse transcribed using oligo(dT) and subjected to PCR with specific primers.

Although PSR^+/- mice were healthy and fertile, no PSR^-/- mice were found among the progeny of the PSR^+/- intercrosses when analyzed at 4 weeks after birth. At various embryonic stages, homozygous mutant mice were present at a frequency roughly consistent with Mendelian inheritance. Close examination at parturition revealed that some PSR^-/- pups were born alive, but died within 24 hours. Lack of PSR expression in mutant mice was confirmed by Northern blot analysis using whole neonates (Figure 1A).

Although PSR^-/- embryos were apparently normal until embryonic day 12.5 (E12.5), the mutant embryos at E15.5 were pale and small, compared with PSR^+/- littermates (Figure 1B). Therefore, we first analyzed Hb concentration and Ht of peripheral blood at E16.5. Although no difference was found between PSR^+/+ and PSR^+/- embryos, both Hb concentration and Ht were remarkably reduced in PSR^-/- embryos (Figure 1C). In peripheral blood obtained from PSR^+/- embryos, enucleated erythrocytes characteristic of definitive erythropoiesis were the preponderant cell types found (Figure 1D). In contrast, peripheral blood from PSR^-/- embryos contained predominantly larger, nucleated erythrocytes derived from the yolk sac (Figure 1D).

Definitive erythropoiesis in the liver starts around E11.5 to E12.5. At E13.5, the total number of fetal liver cells in PSR^-/- embryos decreased to less than 20% of the wild-type levels (Figure 1E). While the proportion of Ter119⁺c-kit^- cells significantly decreased in PSR^-/- embryos, PSR^-/- fetal livers showed a 4-fold increase in the proportion of Ter119⁺c-kit⁺ cells, compared with that of PSR^+/- embryos (Figure 1F). Hence, although the number of Ter119⁺c-kit^- erythroid cells was remarkably reduced in PSR^-/- fetal livers, the number of Ter119⁺c-kit⁺ immature erythroid cells was almost comparable between PSR^-/- and PSR^+/+ or PSR^+/- embryos (Figure 1G). All erythroid cells after the colony-forming unit-erythroid (CFU-E) stage become Ter119-positive,¹²  whereas c-kit is expressed on the cell surface up to the basophilic erythroblast stage.¹³  Therefore, it is suggested that in PSR^-/- fetal livers, erythroid differentiation is blocked at an early erythroblast stage, most likely the proerythroblast or basophilic erythroblast stage.

Another feature of PSR^-/- embryos was impaired thymic development. Histologic analysis revealed that both the cortex and the medulla were poorly developed in the PSR^-/- thymus (Figure 1H). At E18.5, the total number of thymocytes in PSR^-/- embryos was reduced to less than 20% of the wild-type levels (Figure 1E). While more than 60% of the PSR^+/- thymocytes were CD4⁺CD8⁺ thymocytes, this population was reduced to less than 15% in the PSR^-/- thymus (Figure 1I), indicating that thymocyte differentiation is blocked at the CD4^-CD8^- stage in the PSR^-/- thymus.

We then examined whether clearance of apoptotic cells was impaired in PSR^-/- embryos. While most TUNEL-positive cells colocalized with F4/80⁺ macrophages in E13.5 fetal liver sections of PSR^+/- embryos (Figure 2A), PSR^-/- fetal livers exhibited a 3-fold increase in the frequency of unphagocytosed TUNEL-positive cells (Figure 2A-B). Similar results were obtained when E16.5 thymus sections were analyzed (Figure 2E-F). The number of F4/80⁺ macrophages was reduced in both liver and thymus of PSR^-/- embryos (Figure 2C,G). However, it is unlikely that this caused the increased frequency of remnant apoptotic cells, because the number of TUNEL-positive cells was also reduced in the PSR^-/- tissues (Figure 2D,H). Therefore, it is suggested that clearance of apoptotic cells by macrophages is impaired in the absence of PSR.

Semiquantitative RT-PCR analysis revealed that inflammatory cytokines, tumor necrosis factor α (TNF-α), interferon β (IFN-β), and IFN-γ, as well as transforming growth factor β (TGF-β) were comparably expressed in PSR^+/- and PSR^-/- thymuses (Figure 2I). In the fetal liver, the expression of TNF-α and TGF-β was comparable between PSR^+/- and PSR^-/- embryos, but no expression was detected for IFN-β and IFN-γ in both embryos (data not shown). It is thus clear that the developmental defects of erythroid and T-lymphoid cells in PSR^-/- embryos do not result from inflammatory responses. The lack of inflammatory responses in the PSR^-/- fetal liver and thymus might reflect the fact that the absolute number of remnant apoptotic cells in these tissues was unchanged between PSR^+/- and PSR^-/- embryos (Figure 2B,D,F,H).

We have shown that PSR deficiency causes severe developmental defects of erythroid and T-lymphoid cells in mouse embryos. These phenotypes are similar to those of mice lacking DNase II, a lysosomal DNase in macrophages,^14-16  suggesting that PSR-mediated uptake of apoptotic cells is a major pathway leading to DNA degradation by DNase II in macrophages. At this stage, we do not know the precise mechanism by which PSR regulates erythroid and T-lymphoid differentiation. However, we found that PSR deficiency causes repression of apoptosis in several tissues including fetal liver and thymus, but not in intestine whose epithelial cells are removed independently of phagocytosis (data not shown). This raises the possibility that PSR-mediated uptake of apoptotic cells regulates the cell-death machinery in developing cells through a feedback mechanism between phagocytes and developing cells, as described in Caenorhabditis elegans.^17,18  This feedback mechanism may play an important role in terminal differentiation of erythroid and T-lymphoid cells, in which apoptosis-related signaling is required.^19-23 

Note added in proof. After submission of this manuscript, Li et al²⁴  reported that PSR is required for normal development of lung and brain.

We thank Dr Lurie Erickson for comments on this manuscript.