CD41 is developmentally regulated and differentially expressed on mouse hematopoietic stem cells

CD41 (or integrin αIIb) is a marker of the megakaryocytic lineage and also of other clonogenic progenitors.^1-5  CD41 marks the onset of murine primitive and definitive hematopoiesis.^5-9  At embryonic day 7 (E7), yolk sac (YS) primitive erythroid progenitors express low levels of CD41.⁷  By E8.25/E9.5, most YS definitive hematopoietic progenitors express CD41 to high levels,⁷  and definitive hematopoietic progenitors in the intraembryonic para-aortic splanchnopleura or aorta-gonad-mesonephros (AGM) region also express CD41.^5,7-9  These embryonic CD41⁺ hematopoietic cells possess no endothelial potential.^10,11  Immunostainings show CD41⁺ hematopoietic cells within YS blood islands,^7,8,12  intra-aortic clusters,^12-14  and attached to vessel walls of the placenta labyrinth.^15,16  Real-time vital imaging demonstrates that rare phenotypically defined hematopoietic stem cells (HSCs) become CD41⁺ as soon as they emerge from the aortic endothelium.¹³  Based on these findings, CD41 has been suggested as a marker expressed by nascent HSCs, denoting the onset of hematopoietic fate. However, it is still unclear whether HSCs express CD41 throughout development.

HSCs (as defined by the ability to long-term, high-level, multilineage repopulate the hematopoietic system of irradiated adult mouse recipients) start to be detected at E10.5 in the aorta and vitelline/umbilical arteries^17-20  and emerge from hemogenic endothelium.^13,21,22  HSCs are also in the YS, placenta, and fetal liver (FL) beginning at E11.5.²³  Only few transplantation data are available for CD41-sorted AGM cells.^24,25  When sorted in the context of other markers (CD48, CD150, CD45, and endomucin), HSCs were in both CD41⁺ and CD41⁻ fractions. To clarify whether CD41 is a HSC marker during development, we performed a comprehensive analysis of CD41 expression (negative, intermediate, high) on HSCs (during and subsequent to the developmental time in which they are generated) in the various HSC-containing tissues. We show that all E11 AGM HSCs express CD41 to intermediate levels and that expression is time and hematopoietic territory dependent.

Timed matings were set up between males of transgenic mouse line Ln72²⁶  or Ly6A GFP¹⁷  and wild-type (C57BL/10 × CBA)F1 females; or C57BL6 Ly5.1 males and females. Vaginal plug day is E0. Recipients were (C57BL/10 × CBA)F1 or C57BL6 Ly5.2 (8-10 weeks old). Animals were housed according to institutional guidelines, and procedures were performed in compliance with Standards for Care and Use of Laboratory Animals with approval from the Erasmus MC ethical review board.

E11/E12 AGM and YS, and E12 placenta and E14 FL were dissected and collagenase dissociated (0.125% weight/volume, type 1, Sigma-Aldrich Chemie) or crushed (FL). Cells were washed, counted, and suspended in phosphate-buffered saline (PBS), 10% fetal calf serum, and penicillin/streptomycin. Cells were stained with antibodies (supplemental Table 1, available on the Blood Web site; see the Supplemental Materials link at the top of the online article) for 30 minutes on ice in PBS/fetal calf serum/penicillin/streptomycin, washed and stained with Hoechst 33258, before analysis or sorting on FACScan or ArialIII (BD Biosciences).

In vitro clonogenic analysis was performed on dilutions of sorted cells plated in methylcellulose (M3434; StemCell Technologies). Hematopoietic colonies were counted at day 12.

For in vivo HSC analysis, sorted cells (various cell doses) were intravenously coinjected with 2 × 10⁵ wild-type spleen cells into irradiated (9 Gy split-dose, ¹³⁷Cs-source) recipients. After 4 months, donor chimerism (Ln72 or Ly6A GFP) was analyzed by semiquantitative polymerase chain reaction (PCR). Signal quantitation was by DNA normalization (myogenin) and Ln72 or Ly6A GFP control DNA dilutions. For multilineage repopulation analysis, T, B, erythroid, and myeloid cells were sorted from recipient bone marrow (BM) and spleens after antibody staining (supplemental Table 1). Primary recipient BM (2 × 10⁶) cells were injected into secondary irradiated recipients to assess self-renewal capacity. For the Ly5.2 recipients injected with Ly5.1 cells, percentage chimerism was determined by flow cytometry on blood after erythrocyte lysis (Beckman Coulter) and antibody and 7-amino-actinomycin D staining (supplemental Table 1).

E11 Ly6A GFP embryos were fixed (2% paraformaldehyde/PBS, 4°C, 1-2 hours), cryoprotected (30% sucrose/PBS, 4°C overnight), Tissue Tek embedded, frozen (dry ice), and cryosectioned. Immunohistochemical staining was as described¹⁷  with anti-CD41 purified, anti–rat IgG1 biotin, and streptavidin-Cy5 (supplemental Table 1) and detected by laser scanning microscopy.

Flow cytometric analysis of CD41 expression was performed on cells from E11 and E12 AGM and YS, E12 placenta, and E14 FL. Time points correspond to organ-specific peaks of HSC activity. In all tissues, AGM (Figure 1A,H), YS (Figure 1E), placenta (Figure 2A), and FL (Figure 2E), 3 distinct cell populations were observed: CD41⁻, CD41^{intermediate (int)} and CD41^high, in agreement with YS and AGM data.⁷  The highest percentage of CD41^{int + high} cells is found in placenta, compared with AGM, YS, and FL. This is consistent with the high number of hematopoietic progenitors and HSCs in E12 placenta compared with the other tissues as previously described^15,27,28  and the high proportion of megakaryocytic lineage (CD41^high) cells as determined by coexpression of Gp1bβ (not shown). In all tissues, the frequency of CD41^int cells is greater than that of CD41^high cells.

To test which CD41 fractions contain hematopoietic progenitors, sorted cells were plated in methylcellulose and colonies counted. Hematopoietic progenitors were found almost exclusively in the CD41^int fraction of E11 AGM and YS cells (Figure 1B,F). Although the majority of E12 placenta and E14 FL hematopoietic progenitors were present in the CD41^int fractions, some were also CD41⁻ (Figure 2B,F). Together, these data suggest that CD41 expression on hematopoietic progenitors is developmentally regulated and/or dependent on the specific tissue microenvironment.

To examine whether HSCs during ontogeny express CD41, each sorted cell fraction from the distinct embryonic tissues was transplanted and donor cell engraftment examined at 4 months after transplantation. As expected, no mice were reconstituted with the CD41^high fraction. E11 AGM HSCs were found exclusively in the CD41^int fraction (Figure 1C) and were bona fide self-renewing HSCs as shown by high-level, multilineage engraftment of primary and secondary adult recipients (Figure 1J). CD41 expression colocalizes with expression of the Ly6A green fluorescent protein (GFP) HSC marker in some cells closely associated with the ventral aortic endothelium (Figure 1D), and multilineage self-renewing HSCs are enriched in the CD41^intLy6A GFP⁺ fraction (supplemental Figure 1A-B). These data indicate, together with previous live imaging data,¹³  that CD41 is a marker for the earliest emerging aortic HSCs as they are transiting from endothelial to hematopoietic fate. In contrast, all E12 AGM HSCs were restricted to the CD41⁻ fraction (Figure 1I), suggesting that CD41, as an AGM HSC marker, is developmentally time dependent. This is consistent with the data of McKinney-Freeman et al²⁵  showing that, at E11.5, AGM HSCs are in both CD41⁺ and CD41⁻ (CD150⁻CD48⁻) fractions.

CD41 expression on HSCs is not AGM restricted because, when sorted E11 (or E12, not shown) YS cells were transplanted, HSCs were found in both CD41⁻ and CD41^int fractions (Figure 1G). However, in E12 placenta (Figure 2C,H) and E14 FL (Figure 2G,I), HSCs were exclusively in the CD41⁻ fraction. The presence of HSCs in the E12.5 placenta CD41⁺ fraction was previously reported, but, as stated by the authors, this was probably because of contaminating CD41⁻ cells.²⁵  At E11.5, no repopulation was found with either the CD41^int or CD41⁻ placenta cell fractions (n = 2) because of the extremely low HSC frequency at this time point in the placenta. Thus, we cannot conclude whether the first placental HSCs are CD41 expressing. Immunostainings of E11 FL sections (and flow cytometric analysis of E14 FL) show no overlap in expression of CD41 and Ly6A GFP (Figure 2D; supplemental Figure 1C-D). Thus, CD41 is expressed on HSCs in a time-dependent manner. CD41 expression now adds a further possibility for HSC enrichment during development when used in combination with other known HSC markers (Ly6A, c-kit, CD45, CD34, and CD150; supplemental Figures 1-2).

CD41 as a marker of the earliest HSCs generated in the AGM, together with the finding that some YS HSCs are CD41 expressing, may indicate HSC generation in the YS or HSC migration from the early AGM. CD41 expression on HSCs is lost with developmental time and HSC amplification in placenta, and FL is independent of CD41 expression. Hence, these data, showing that CD41 is a temporally restricted early HSC marker, are pivotal to future studies of its transcriptional regulation and role in HSC generation/migration.

The authors thank Henk Dronk and Experimental Dieren Centrum (Erasmus MC) for mouse care and Reinier van der Linden and Fredrik Wallberg for cell sorting.

This work was supported by The Netherlands Organisation for Scientific Research (VIDI grant 917-76-345, C.R.; and VICI grant 916-36-601, E.D.) and the National Institutes of Health (Merit Award R37 DK51077, E.D.).

National Institutes of Health

Contribution: C.R. performed experiments and wrote the paper; K.O., J.-C.B., and A.O. performed experiments; and E.D. wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

The current affiliation of K.O. is Department of Haematology, Cambridge Institute for Medical Research, University of Cambridge, Cambridge, United Kingdom.

Correspondence: Elaine Dzierzak, Erasmus Medical Center, Erasmus MC Stem Cell Institute, Department of Cell Biology, PO Box 2040, 3000 CA Rotterdam, The Netherlands; e-mail: e.dzierzak@erasmusmc.nl.