p67 is expressed on erythroid progenitors and guides their marrow homing. (A-D) BFU-Es reside in the CD34⁺p67⁺ fraction: CD34⁺p67⁺ or CD34⁺p67^– BM cells were isolated by FACS sorting and cultured in colony assays, to test frequency and lineage distribution of CFU-Cs. (A) Methyl cellulose colony assays: Most of the CFU-Cs in the CD34⁺p67^– fraction (left panel, insert) were CFU-GMs, whereas the CD34⁺p67⁺ sample contained mostly BFU-Es (right panel, insert). Total CFU-C frequency was much greater in the p67⁺ fraction. (1500 sorted cells/35-mm plate). Dark-field photography was performed with a DIX Nikon camera, and a 105AFMacroNikkor objective (Nikon, Melville, NY). Plates were photographed with 1:1 macro; the insets, each showing one representative colony, with 4:1 macro. Images were captured with Nikon Capture/View acquisition software and processed with Photoshop (Adobe Systems, San Jose, CA) for white balance adjustment. (B) Plasma-clot assays: Total BM cells were sorted based on p67 expression and plated in plasma-clot assays. BFU-Es were depleted in the p67^– and enriched in the p67⁺ BM fraction. (Note that 10 000 cells/clot were plated in unsorted sample, compared with 5000 cells/clot in p67^– and p67⁺ samples.) Images were captured with a benzidine stain, using a Leica MZ6 dissecting microscope (Leica, Heidelberg, Germany), 0.63 ×/1.0 NA objective, and a Nikon CoolPix 995 camera, 2 × zoom. Unadjusted color image gray-scaled in Powerpoint (Microsoft). For color image, see Figure S3). (C) CFU-C frequency and distribution: Among CD34⁺p67⁺ BM cells, the frequency of CFU-Cs was 2-fold higher than in unsorted CD34⁺ cells (224 ± 19 vs 114 ± 20 CFU-Cs/1000 CD34⁺ cells, P < .005). This increase of colony-forming activity in the CD34⁺p67⁺ population was restricted to BFU-Es, which were 5-fold enriched in absolute (218 ± 16 vs 43 ± 8 BFU-Es/1000 cells) and 2.5-fold enriched in relative (95% ± 4% vs 38% ± 1% of CFU-Cs) frequency compared with unsorted BM (P < .005). CFU-GMs were accordingly depleted by selecting for p67⁺ cells (P < .005). Inversely, BFU-E activity was reduced in the sorted CD34⁺p67^– population, to 13% ± 5% of all CFU-Cs. Shown on the Y-axis are CFU-Cs (mean + SEM of 3 independent experiments)/1000 sorted CD34⁺ cells. Essentially similar data were observed with sorted MPB CD34⁺ cells (Figure S2). (D) Anti-p67Ab is nontoxic for HPCs: Incubation of human CD34⁺ cells with anti-p67Ab (without washing prior to culture) did not affect CFU-C frequency and lineage distribution compared with untreated CD34⁺ cells (ctrl), indicating that anti-p67Ab is nontoxic for human CFU-Cs in vitro (2 experiments for BM CD34⁺ cells [shown here]; 1 experiment for MPB CD34⁺ cells). The cells plated for these experiments are aliquots of the cell suspensions injected for the homing assays depicted in Figure 2E and Figure S2. (E) p67 mediates marrow homing of BFU-Es: BM CD34⁺ cells, incubated with blocking anti-p67Ab, had 49.7% ± 6.6% (P < .005) reduced BM homing compared with untreated control, almost exclusively at the expense of BFU-Es (BFU-Es reduced by 86.3% ± 3.3%, P < .005) and CFU-mixed (reduced by 90.5% ± 3.7%, P < .005) (CFU-GMs: reduced by 14.7% ± 11.6%, P = .38). Similar inhibition was seen when MPB CD34⁺ cells were transplanted (Figure S2). (n = 3 independent experiments, each with 5 recipients/group. Mean + SEM of percent of total injected CFU-Cs homed to marrow 20 hours after transplantation.)