C57BL/6J (CD45.2), B6.SJL-Ptprca Pepcb/BoyJ (CD45.1), CD45.2/CD45.1 F1 hybrid mice (F1), and C57Bl/6-Tg(UBC-GFP)30Scha/J transgenic mice (UBC-GFP) of both sexes were used. The mice were bred in the Center of Experimental Models of the First Faculty of Medicine, Charles University.

Whole-body irradiation from a ⁶⁰Co source of gamma rays (approximately 0.40 Gy/min) was used.

Bone marrow cells from one femur (approximately 29 million cells) was transplanted via the retro-orbital route to mice conditioned with irradiation (9 Gy) 2─3 hours prior. Blood samples were obtained at various time points after transplantation and analyzed by the Auto Hematology Analyser BC-5300 Vet (Mindray, China). The weight of wet spleens was determined. Bone marrow was flushed from femurs into one mL of phosphate buffered saline, and cells were counted by the Auto Hematology Analyser BC-5300Vet (Mindray, China).

Bone marrow cells from CD45.2 mice, either untreated or transplanted with syngeneic bone marrow from one femur 30, 40, or 60 days prior, were mixed with the same number of bone marrow cells from untreated CD45.1 mice. The cell mixture was transplanted to irradiated F1 hybrid mice, and the proportion of CD45.2 and CD45.1 cells was determined in blood after six months.

Bone marrow cells lacking lineage markers and expressing the Sca-1 antigen and c-Kit receptor (LSK cells), subdivided according to the expression of CD150 and CD48 markers, were determined in normal and regenerating bone marrow. Cells were stained for 30 minutes by fluorescently labeled antibodies on ice in the dark and analyzed by the flow cytometer BD FACSAria IIu (BD Biosciences, USA). All antibodies are listed in Supplementary Material .

Mice (CD45.2, CD45.1, F1, or UBC-GF) conditioned by irradiation were transplanted with bone marrow cells from normal CD45.2 or CD45.1 mice (1^st BMT) within 2─3 hours after irradiation. The second transplant (2^nd BMT) of bone marrow from congenic CD45.1 or CD45.2 mice was administered after 2 hours and for various days up to 60 days after the 1^st BMT.

Blood samples were stained with the anti-CD45.2 and anti-CD45.1 antibodies, and antibodies labeling granulocytes and monocytes (GM cells), B cells (B220 cells), and T cells (CD4 and CD8).

Chimeric bone marrow resulting from two successive BMTs was transplanted to lethally irradiated (9 Gy) CD45.2 or CD45.1 secondary recipients. In one experiment, the chimeric bone marrow from the secondary recipients was transplanted into the tertiary UBC-GFP recipient mice. The peripheral blood was analyzed after four and six months.

For multi-group comparisons, one-way ANOVA was used. The Student´s t-test with a two-tailed distribution was used to compare results from two experiments. The significance level is indicated as follows: * p < 0.05, ** p < 0.01, *** p < 0.001. The GraphPad Prism software (GraphPad San Diego, USA) was used.

Clinical BMT aims for early robust recovery of blood cell production by transplanting a high number of hematopoiesis-reconstituting cells, including stem cells and progenitor cells. Before conducting experiments with two transplants applied successively at various periods of hematopoiesis regeneration, we analyzed the time course of hematopoiesis recovery after a single transplant. Robust hematopoiesis regeneration was induced in lethally irradiated mice by transplanting bone marrow cells from one femur. The size of the graft exceeded approximately 200 times the minimum number of cells needed for survival and long-term hematopoiesis reconstitution (13). Bone marrow cellularity significantly increased after day 5 ( Figure 1G ) and spleen weight was normal when determined on day 5 and later ( Figure 1F ). Blood cell production resumed by day10, as shown by the increasing number of white blood cells, red blood cells, and platelets in the peripheral blood ( Figures 1A–E ). Granulocytes and monocytes were mainly produced in the first 12 days, but then lymphocytes also began to be produced in significant quantity ( Figure 1B ). Bone marrow collected 30, 40, and 60 days after BMT was tested for its capacity to compete with normal bone marrow in the reconstitution of damaged hematopoiesis by the competitive transplantation assay. Bone marrow collected 30, 40, and 60 days after BMT produced only 8.9%, 7.0%, and 10.4% cells ( Figure 1H , Days 30, 40, 60), significantly less than ≈ 50% in mice competitively transplanted with bone marrow from both untreated CD45.2 and CD45.1 mice ( Figure 1H , Ctrl). Hence, the pool of the cells providing long-term hematopoiesis has only partly recovered 30–60 days after BMT.

The delayed recovery of long-term acting stem cells after BMT corresponds to the slow recovery of immature lineage negative, Sca-1, and c-Kit positive (LSK) cells lacking the CD48 marker (LSK CD48^─ cells) in the bone marrow ( Supplementary Figure 1 ), which contain hematopoiesis-reconstituting cells with middle to long-term functioning after transplantation (8, 32, 34).

Results presented in this section demonstrate the robust recovery of hematopoiesis and blood cell production in the first two weeks after transplantation and the significantly delayed replenishment of the stem cell pool.

Our previous results with split BMT (32, 33) demonstrated the engraftment of stem and progenitor cells during hematopoiesis regeneration and showed a faster decline in the engraftment of progenitors to stem cells. We have verified the previous results in new experiments and show all results in Figure 2 and Supplementary Figure 2 .

In the previous and new experiments, regeneration of hematopoiesis was induced in lethally irradiated mice by transplanting syngeneic bone marrow cells corresponding to half, one, or two femurs (1^st BMT). Two hours later or after 1–60 days, we challenged the transplanted cells with the second BMT (2^nd BMT) from congenic mice in the amount equal (higher in one experiment) cell number ( Figure 2A, B ). Figure 2C and Supplementary Figure 2 summarize results from seven experiments comparing blood cell production from the 2^nd BMT after one and six months. The cell production from the 2^nd BMT determined after one month declined rapidly after the 1^st BMT. The decline was delayed when blood cell production from the 2^nd BMT was determined after six months. These results indicate preferential engraftment of HSCs compared to progenitors by regenerating hematopoiesis.

Because the results read one month after transplantation could be confounded by the host T cells surviving in the peripheral blood after conditioning irradiation (35), we used UBC-GFP mice as recipients of the two successive BMTs in the next five experiments. The GFP-positive cells of the host were gated out in the peripheral blood analysis of cells derived from the 1^st BMT and 2^nd BMT. In these experiments, we induced hematopoiesis regeneration in irradiated UBC-GFP mice by transplantation of 20 million bone marrow cells from either CD45.1 or CD45.2 mice (1^st BMT). After two hours or after 7–18 days, the mice were transplanted with the same number of bone marrow cells from congenic mice with different CD45 allotype (2^nd BMT). The combination of CD45.2 and CD45.1 donors, their sex, and mice numbers in the five experiments are in Table 1 . One and six months after the 2^nd BMT, the impact of the 1^st BMT on the outcome of the 2^nd BMT was analyzed in the peripheral blood.

Each of the five experiments included a group of mice to whom the 2^nd BMT was administered 2 hours after the 1^st BMT. The aim of the 2-hours experimental groups was testing whether 20 million cells from the first transplant would not acutely reduce the available space for another 20 million cells from the 2^nd BMT. The outcome of the 2^nd BMT applied 2 hours after the 1^st BMT was not affected by the previous BMT when evaluated after one and six months, and CD45.2 and CD45.1 nucleated blood cells were equally represented in all cell types in the peripheral blood ( Supplementary Figure 3 ). Therefore, the 1^st BMT has not limited the engraftment of the 2^nd BMT by occupying a significant portion of the space created for transplanted stem and progenitor cells by conditioning irradiation.

The contribution of the 2^nd BMT to nucleated blood cells in the peripheral blood was less than 50% in the five experiments in which the 2^nd BMT was administered at intervals of 7 days (Experiments 1 and 2), 10 days (Experiments 3 and 4), and 13 and 18 days (Experiment 5) after the 1^st BMT. However, the efficacy of the 2^nd BMT was significantly higher when evaluated after six months compared to one month ( Figure 3 ; Supplementary Figure 4 ). After one month, cell production from the 2^nd BMT was myeloid-biased primarily due to very low production of T cells (CD4 plus CD8) ( Figure 3 ; Supplementary Figure 4 ). The production of lymphoid cells (B and T cells; B220 and CD4 plus CD8 cells) is compared to the production of myeloid cells (granulocytes and monocytes; GM cells) in Figure 4 and Supplementary Figure 5 . After one month, B cells (B220) and granulocytes and monocytes (GM cells) were produced equally in four out of six groups and less in two groups. After six months, the production of B cells (B220) exceeded the production of GM cells in four out of six groups and was insignificantly higher in two groups. The production of T cells from the 2^nd BMT became significant after six months ( Figure 4A ). CD4 and CD8 T cells derived from the 2^nd BMT were separately evaluated after six months in Experiments 3–5. CD4 and CD8 were produced equally when the 2^nd BMT was administered 2 hours after the 1^st BMT ( Supplementary Figure 5 ). When the 2^nd BMT was administered 10–18 days after the 1^st BMT, CD8 T cells were produced less than CD4 T cells ( Figure 4B ).

Results from this section demonstrate that BMT can be effectively administered in separate doses, and regenerating hematopoiesis preferentially engrafts stem cells to the progenitors from the 2^nd BMT. HSCs and progenitors from the 2^nd BMT rapidly resumed production of B cells and later became a significant source of T cells as well.

The ultimate indication of HSC stemness is the capacity to reconstitute depleted hematopoiesis over the long term. The longevity assay of HSC functionality involves re-transplantation of reconstituted hematopoiesis to secondary recipients with depleted hematopoiesis, and eventually to tertiary recipients. To conduct these functional tests assessing the durability of HSC pools established from the 2nd BMT, chimeric bone marrow from Experiments 2 to 5 was re-transplanted to secondary recipients. Blood cell production derived from the 2^nd BMT was assessed in the peripheral blood after 4 months ( Figure 5B ; colored columns) and compared to that in the primary transplanted mice at the time of bone marrow re-transplantation ( Figure 5B ; gray columns). This demonstrated sustained production of blood cells derived from the 2^nd BMT after bone marrow re-transplantation, whether administered 2 hours ( Supplementary Figure 6 ) or 7–18 days ( Figure 5 ) after the 1^st BMT. The BMT administered 7–18 days after the 1^st BMT resulted in a higher production of lymphoid cells compared to granulocytes and monocytes in secondary recipients, significantly 6 times and insignificantly 4 times ( Figure 6A ), with the proportion of CD4 and CD8 T cells being equal ( Figure 6B ). The production of lymphoid cells was balanced with the production of GM cells when the 2^nd BMT was administered 2 hours after the 1^st BMT ( Supplementary Figure 7 ).

Chimeric bone marrow from the secondary recipients was re-transplanted to tertiary recipients in Experiment 5, and their peripheral blood was analyzed after four and six months. Blood cell production from the 2^nd BMT remained steady from the fourth month after the primary transplantation ( Supplementary Table 1 ).

Results from the re-transplantation experiments demonstrated that the 2^nd BMT, administered either 2 hours or 7–18 days after the 1^st BMT, established durable pools of HSCs which became significant sources of granulocytes and monocytes, as well as B and CD4 and CD8 T cells.