The p67 laminin receptor identifies human erythroid progenitor and precursor cells and is functionally important for their bone marrow lodgment

Homing of mature immunoregulatory cells to sites of inflammation follows a well-characterized sequence of rolling, firm adhesion, and transmigration. Although some of the same major players in this cascade are also involved in hematopoietic progenitor cell (HPC) homing to bone marrow (BM), differences in the hierarchy of molecular pathway usage exist, likely because of different tissue microenvironments and receptor repertoires.^1,2  α4/β1-integrin/VCAM-1 and CXCR4/SDF-1 are dominant pathways of hematopoietic–stromal cell interactions, important for retention of transplanted HPCs in BM.^3-5  In addition, an array of disparate dominant and ancillary molecules,^1,2  which work within a complex, interactive network, is involved in BM homing.^6,7  Within BM, lineage-committed HPCs and stem cells are likely partitioned in different functional/spatial “niches,”^8,9  to encounter an environment accommodating their specific functional requirements. How these niches are defined and the specific molecular interactions between seeded HPCs and their supportive microenvironmental niches are incompletely understood.

Laminins are expressed in BM, and cognate receptors are expressed on HPCs, as demonstrated by HPC adhesion to laminin.^10-12  Instrumental roles of the nonintegrin laminin receptor p67 were described in metastasis, embryo implantation, and lymphocyte trafficking.^13,14  p67-mediated laminin binding was also reported for subsets of normal and malignant hematopoietic cells.^15-18  Expression of p67 on HPCs and its role in hematopoiesis were not previously addressed. Our present study shows that p67 recognizes erythroid progenitors/precursors, and that its inhibition blocks BM homing of BFU-Es.

NOD/SCIDβ2microglobulin^–/– mice (Jackson Laboratories, Bar Harbor, ME), housed and bred in the Helicobacter-free section of the Specific-Pathogen-Free vivarium at University of Washington, with water and chow ad libitum, served as recipients. Mice were lethally irradiated and received CD34⁺ cell transplants as described,⁷  in accordance with procedures approved by the University of Washington IACUC.

Cryopreserved cadaveric BM-derived or healthy volunteer-donor mobilized peripheral blood (MPB)–derived CD34⁺ cells were received from the NIH repository (Fred Hutchinson Cancer Research Center) with permission of the University of Washington IRB. Cells were thawed and incubated overnight with 25 ng/mL recombinant human stem cell factor (rhuSCF; Peprotech, Rocky Hill, NJ).⁷ 

Fluorescence-activated cell sorter (FACS) analysis was performed according to standard protocols, using anti-p67 antibody (anti-p67Ab) (MLuC5; NeoMarkers, Fremont, CA) with Alexa 488–labeled rat anti–mouse IgM secondary antibody (MolecularProbes, Eugene, OR) or secondary antibody alone, and PE- or APC-labeled anti–glycophorin A, anti-CD34, anti-CD45, anti-CD11a, anti-CD71, or isotype-control antibodies (BD Pharmingen, San Diego, CA). Antigen expression was quantified using the FACSCalibur (BD Immunocytometry Systems, San Jose, CA). p67⁺ and p67^– cell populations were isolated on the FACSAria (BD Immunocytometry Systems). Flow data were analyzed with CellQuest software (BD Immunocytometry Systems) for OS10.

Colony assays were performed as described and evaluated for CFU-GMs, CFU-mixed, and BFU-Es after 14 days.¹⁹  Methyl cellulose colony-assay medium was from StemCell Technologies (Vancouver, BC) or Miltenyi Biotech (Auburn, CA). Colonies were picked on days 7 and 14 to evaluate p67 expression in early and late erythroid and nonerythroid precursor cells.²⁰  Plasma-clot assays were done to enumerate BFU-Es after benzidine staining, as described.²¹ 

Cells were incubated with or without anti-p67Ab (8 μg/2 × 10⁶ cells/1 mL) for 20 minutes at 4°C prior to injection. BM homing of CFU-Cs was quantified as described.^19,22 

Descriptive statistics and t tests were calculated using Excel (Microsoft, Redmond, WA). A P value less than .05 was considered statistically significant.

Using FACS, we found p67 expression on 34% ± 4% of primary CD34⁺ BM cells (mean ± SEM), with a distinct negative fraction, and a wide shoulder of p67⁺ cells (Figure 1A). The curve shape for p67 expression on MPB CD34⁺ cells was similar, but the percentage of p67⁺ cells was greater than in BM cells (Figure 1B). Further FACS analysis revealed that although the number of glycophorin A⁺ cells among the CD34⁺p67⁺ cells was small, p67 expression on this population (CD34⁺ glycophorin A⁺) reached almost 100% (Figure 1C). Similarly, almost all CD34⁺CD71^high BM cells (93.9%) expressed p67 (not shown). Compared with MPB cells, the number of CD34⁺ glycophorin A⁺ cells was low in BM cell samples studied, which can explain the lower percentage of p67 expression found in the BM CD34⁺ population. To test whether p67 recognized determined progenitors of erythroid lineage, CD34⁺p67⁺ and CD34⁺p67^– cells were isolated by FACS sorting and plated in colony assays. Most of the BFU-E–forming activity resided within the p67⁺ fraction. Whereas among total CD34⁺ cells only 1 in 25 cells was a BFU-E, BFU-E frequency was increased to almost 1 in 5 cells in the CD34⁺p67⁺ population. The CD34⁺p67^– population was significantly depleted of BFU-Es (P < .005 for all differences) (Figure 2A-C). To study p67 expression at various stages of erythroid maturation, “white” (ie, nonhemoglobinized) day-7 BFU-E–derived cells were picked from methyl cellulose cultures. Early BFU-Es were identified by colony morphologic appearance. Most of the cells were proerythroblasts, as determined by cell morphology on Wright-Giemsa–stained cytospins, glycophorin A expression on FACS, and negative benzidine staining (Figure S1, available on the Blood website; see the Supplemental Figures link at the top of the online article). Similarly, erythroblasts were picked from day-14 methyl cellulose plates from mature, visibly hemoglobinized BFU-Es. Their identity was additionally confirmed by evaluation of the same parameters (ie, morphology, glycophorin A positivity, and benzidine staining). The day-7 population, which contained mostly proerythroblasts, overwhelmingly expressed p67 (Figure 1D), as did a subset (20%-30%) of late erythroblasts (Figure S1), suggesting some attenuation of p67 expression with differentiation. Myeloid cells derived from CFU-GMs were p67 negative (Figure 1E).

These studies thus identify p67 as a novel lineage marker among adult human erythroid HPCs. A transferrin receptor isoform was previously described to be selectively expressed in erythroid HPCs.²³  However, unlike transferrin receptor, which is clearly related to erythroid-specific proliferative/metabolic functions, the functional role of p67 was not immediately apparent. The hypothesis was entertained that p67 might interact with BM laminins, to mediate homing/retention of BFU-Es. MLuC5 antibody was previously demonstrated to block p67-mediated adhesion, thereby attenuating laminin binding.²⁴  Blockade of laminin binding by MLuC5 antibody was incomplete, seemingly due to HPC expression of additional laminin receptors, such as VLA-6.^14,25  Therefore, to test the effect of blocking p67-mediated binding on BM homing, BM or MPB CD34⁺ cells were incubated with anti-p67Ab and injected into lethally irradiated NOD/SCIDβ2m^–/– recipients. Control mice received untreated CD34⁺ cells from the same cell sample. Total colony number and frequency of erythroid or mixed colonies were the same for anti-p67Ab–incubated and control-incubated cells, indicating absence of in vitro toxicity or antiproliferative effects of the antibody (Figure 2D). The fraction of BM-derived CFU-Cs homed to BM was 50% reduced by anti-p67Ab, predominantly at the expense of BFU-Es (Figure 2E). The effect of anti-p67Ab treatment on BM homing of MPB CD34⁺ CFU-Cs was similar in pattern and magnitude to that of BM CD34⁺ CFU-Cs (see Figure S2 for comparison).

Inhibition of homing interactions between HPCs and BM stroma was previously reported.^1,2,4,6,25  However, a lineage-specific effect was not previously observed. As shown, the laminin-binding receptor p67 is preferentially expressed on human erythroid progenitors and precursors, and its blockade preferentially restricts BM homing/retention of BFU-Es. This specificity is likely explained by low expression of p67 on nonerythroid CFU-Cs. The existence of lineage-specific homing requirements may not be unexpected, given the different microenvironmental needs for specialized lineage development, but they were not previously documented. These data, albeit associated with the known caveats of antifunctional antibodies, suggest that laminin may be a functional component of the erythroid niche. Although the p67 dependence was observed in a xenogeneic setting, we speculate that it may be also operative within human BM. These studies do not exclude the possibility that nonerythroid HPCs might interact with laminins through other receptors, such as VLA-6.²⁵ 

The generous gift of primary cells by Shelly Heimfeld (FHCRC) and Joanna Reems (Puget Sound Blood Center) is gratefully acknowledged.