CD47^−/− mice were generated in the C57BL/6J background using a CRISPR/Cas9 system as previously described [13]. For this study, heterozygous breeding pairs were used to produce CD47^−/− mice and their wildtype (WT) littermates (n=7–9 per group). Ear tissues were used for genotyping in a conventional PCR with specific primers; forward; 5′-GGT CGG TCG TTT CCC TTG AA-3′ and reverse; 5′-GAT CCC CGA GCC ACT CAC-3′. Mice were housed in an AAALAC International accredited specific pathogen free (SPF) facility (#001169) at Seoul National University Hospital under standard housing conditions 12 h dark/light cycle, 40–60% relative humidity and 22±2℃ with free access to food (LabDiet 5002 Certified Rodent Diet, PMI Nutrition International, USA) and water. All experiments were approved by the Institutional Animal Care and Use Committee in Seoul National University Hospital and animals were maintained in accordance with the Guide for the Care and Use of Laboratory Animals [14].

From 4 weeks after birth to 26 weeks, body weight and food/water intake of mice were measured once a week with daily observation of clinical signs. Measurement of food and water intake was carried out in the day cycle and calculated as the amount consumed per mouse during 24 hours.

Hematology, serum biochemistry and urinalysis were carried out as described [15]. Briefly, at the age of 16 and 26 weeks, all mice were deeply anesthetized with isoflurane and peripheral blood was collected from the abdominal vein in a K₂EDTA contained tube (BD Microtainer, USA) for hematologic analysis using an ADVIA 2021i (Siemens Healthcare, USA). Serum biochemistry and urinalysis were carried out using a Hitachi 7070 (Hitachi, Tokyo, Japan) and a CLINTEK Adventus analyzer (Siemens Healthcare), respectively.

At necropsy, gross examination was performed for all organs. Then, major organs, except for eyes, testis and epididymis which were fixed in Davidson solution and Bouin's solution, respectively, were fixed in 10% neutral buffered formalin and embedded in paraffin wax after dehydration and clearing process. After sectioning in a thickness of 4–6 µm, tissues were stained with hematoxylin and eosin. To examine the levels of intracellular accumulation of iron in the spleens, we stained the tissue using Prussian blue reaction. Breifly, the paraffin-embedded spleen was sectioned and re-hydrated after deparaffinization. The tissue section was stained in freshly prepared 5% potassium ferrocyanide (Sigma Aldrich) in 10% HCl for 20 min and then counterstained with neutral red (Sigma Aldrich) for 5 minutes. All stained tissues were observed under a microscope (BX61, Olympus).

Cells were isolated from the spleens, bone marrow and thymuses of 16-week-age WT and CD47^−/− mice (male; n=3 and female; n=3 per genotype) and stained with antibodies for flow cytometric analysis as described [15]. In this study, we used V450-anti-CD3e (500A2), FITC-anti-CD4 (RM4-5), APC-anti-CD5 (53-7.3), BV605-anti-CD8a (53-6.7), Alexa700-anti-CD19 (1D3), and PerCP-anti-CD45R (RA3-6B2) to label distinct lymphocyte populations. All antibodies employed in this study were from BD Bioscience and used at 1:200 except anti-CD4 antibody (1:500). After staining and washing, 20,000 cells were analyzed using a FACS Calibur (BD).

All data were expressed as mean±SD, and statistical analysis was performed using a two-tailed t test in a SPSS software (version 22.0). P<0.05 was considered to be statistically significant.

To evaluate the role of CD47, we generated a novel CD47^−/− mouse line by introducing a double-strand DNA break of CD47 exon 1 using a CRISPR/Cas9 system and inducing non-homologous end joining repair, which resulted in a deletion of 26 base pairs which created a frameshift mutation and a premature stop codon (Figure 1A).

To characterize the phenotype of CD47^−/− mice, we first examined their development with monitoring of main physiological parameters including sex ratio, genotype, body weight gain, food and water intake for 26 weeks. At birth, there were no changes in the sex ratio of littermates (male; n=45, female; n=43 out of 88 siblings; CD47^−/− male; n=13, CD47^−/− female; n=12 out of 25 CD47^−/−mice), and the ratio of genotypes of mice born from CD47^+/− parents was 1:1.6:0.8 for WT, CD47^+/− and CD47^−/−, indicating a Mendelian inheritance pattern of the knockout alleles with no embryonic lethality. For 26 weeks, body weight gain and food/water consumption of CD47^−/− mice was comparable to WT (Figure 2), suggesting that loss of the CD47 gene did not affect the postnatal development.

When peripheral blood was collected from 16-week-old mice and analyzed, male CD47^−/− mice showed a significant reduction of RBC counts (11.4 %, P<0.01) and hemoglobin (10.1%, P<0.01) at the age of 16 weeks, while no abnormality was observed in the blood of female CD47^−/− mice (Figure 3A). The anemia-like feature observed in male CD47^−/− mice became more evident at 26weeks with increase of reticulocytes (136.2%, P<0.01) and MCV (102.5%, P<0.05) in addition to the reduction of RBC counts (13.1%, P<0.01), hemoglobin (10.9%, P<0.01) and hematocrit (10.7%, P<0.01) as shown in Figure 3B. 26-week-old female CD47^−/− mice also showed similar changes in the parameters for erythrocytes; reduction of RBC counts (8.0%, P<0.05), hemoglobin (11.1%, P<0.01) and hematocrit (7.5%, P<0.05) with increased reticulocytes (116.0%, P<0.05). These findings indicate the development of anemia in our CD47^−/− mice, which is progressively aggravated over aging.

Serum biochemistry (Table 1) and urinalysis (data not shown) showed no changes between CD47^−/− mice and their WT littermates, suggesting that lack of CD47 may not affect the function of other organs, especially liver and kidney.

Based on the function of CD47 in RBCs, anemia observed in our CD47^−/− mice was possibly caused by increased hemolysis. As spleen is commonly enlarged in hemolytic anemia due to accumulation of RBCs and macrophages [1617], we examined the organs of CD47^−/− mice with a particular focus on their spleens. While no changes were found in other organs of CD47^−/− mice, spleens of CD47^−/− mice were found to be significantly larger than those of WT littermates (Figure 4, Table 2). When the organ per body weight ratio was calculated, both genders of 16-week-old CD47^−/− mice exhibited enlarged spleens (Figure 4A; 169.3% increase compared to WT male, 139.0 % for female mice, P<0.05 for both). In line with the progression of anemia observed in hematological analysis, splenomegaly was consistently observed in both genders of 26 week-old CD47^−/− mice (Figure 4B; 143.5 % increase in male, P<0.05 and 161.8 % in female, P<0.001). Except for the spleen, no other organs in CD47^−/− mice showed notable changes (Table 2).

When the spleen was histopathologically examined, cells containing brown-gold pigments were more frequently observed in CD47^−/− mice compared to WT mice (Figures 5A, 5B). As loss of CD47 was previously associated with increased destruction of RBCs by macrophages, these pigments were likely hemosiderins derived from increased degradation of iron-containing hemoglobins in RBCs and therefore we visualized intracellular iron in the tissue using Prussian blue reaction to determine the identity of the pigments. Staining of iron revealed noticeable increase in the number of dark blue-stained cells with a tendency of higher intracellular staining levels in CD47^−/− mice than WT mice (Figures 5C, 5D), indicating the increased intracellular accumulation of hemosiderns. It is tempting to postulate that these hemosiderin-containing cells were macrophages due to their function in hemolysis, but this proposition may need to be confirmed. There are several types of anemia reported in humans. Among them, the anemia identified in our CD47^-−/− mice is likely hemolytic based on the following observations; CD47^−/− mice showed concurrent reduction of RBC counts, hemoglobins and hematocrit with increased reticulocytes, indicating loss of mature RBCs with increased release of immature RBCs into the blood as a compensatory response. Although the anemic feature was not evident in 16-week-old female mice, both genders of CD47^−/− mice had enlargement of the spleen which continued to the age of 26 weeks with concurrent aggravation of hematological parameters regarding RBCs. Furthermore, the elevated destruction of RBCs was demonstrated by the increased number of hemosiderin-loaded cells with higher amount of hemosiderins accumulated in each cells of the CD47^−/− spleens, strongly supporting our hypothesis. Despite these findings, our data however do not exclude the possibility of other types of anemia with similar features. Therefore, further investigation is warranted to characterize the type and clinical manifestation of the anemia in our CD47^−/− mice.

The findings on the development of hemolytic anemia in our CD47^−/− mice have not been reported in the previously generated CD47^−/− mice [4], but were closely in line with AIHA in CD47^−/− NOD mice [12] with several different features noted; CD47^−/− NOD mice showed much greater reduction of hematocrit than our CD47^−/− mice, while jaundice was only observed in CD47^−/− NOD mice, suggesting a mild degree of hemolytic anemia in our mice. The mouse strains used in generating models (NOD vs. C57Bl/6) may have played a role in causing these discrepancies in manifestation of hemolytic anemia between the two mouse lines.

Next, we examined changes in subpopulations of lymphocytes including T and B cells in CD47^−/− mice as inactivation of CD47 previously resulted in reduction of T cell populations in the spleen [18]. To assess the impact of CD47 deficiency on those lymphocytes, we performed flow cytometry on the cells isolated from 16-week-old mice spleens, thymuses and bone marrow. Flow cytometry analysis on splenic lymphocytes revealed a remarkable reduction of CD3⁺CD5⁺, CD3⁺CD4⁺ and CD3⁺CD8⁺ T cells (Figures 6A, 6C, P<0.01 for all), while no change was found in CD19⁺B220⁺ B cell population. The number of CD19⁺B220⁺B cells from the bone marrow of CD47^−/− mice was also comparable to WT (Figures 6B, 6D). Our findings recapitulated the crucial role of CD47-SIRPα interaction on T cell proliferation as similar findings on the reduction of CD4⁺ T cells have been reported in the spleen of the CD47 binding partner SIRPα-deficient mice [18]. Contrary to the reduction of T cells in the spleen, no change was detected in thymic T cell populations when compared to WT (Figures 6C, 6F), suggesting differential roles of CD47 in T cells depending on their developmental stage; CD47 functions in the proliferation of T cells which occurs in the spleen, but not at the stage of initial development through negative selection in the thymus [19].

In this study, we generated a novel CD47^−/− C57BL/6J mouse line using a CRISPR/Cas9 system and found a progressive hemolytic anemia as well as profound reduction of mature T cell populations in the spleen. Intriguingly, despite its ubiquitous expression, loss of CD47 mainly resulted in hematological phenotypes without causing embryonic lethality, developmental defects or any noticeable changes in the structure and function of other tissues, indicating its preferred function in blood cells. These results, however, do not exclude the possible role of CD47 in other types of cells and further investigation may be required to reveal cell-specific impact of CD47 deficiency. In conclusion, CRISPR/Cas9-mediated generation of CD47^−/− mice confirmed the function of CD47 in the maintenance of RBC and T cell populations, demonstrating their usefulness as an in vivo platform to study the function of CD47.