Cells with the highest Hlf-tdTomato expression levels have bone marrow reconstitution capacity

We previously reported that HSCs have higher levels of tdTomato than other hematopoietic progenitors in the fetal livers of Hlf-tdTomato KI mice25. Transplantation experiments were performed to confirm the stem cell potential of bone marrow cells with higher levels of Hlf-tdTomato in the bone marrow of the tibia. Flow cytometry analysis showed that 47.04.4% of whole bone marrow cells from adult long bones of Hlf-tdTomato KI mice were positive for CD45, which is a panhematopoietic marker (Fig.1a, left panel, n=4). Cells with the top 0.011% tdTomato intensity within the CD45-positive cells (0.0049% of the whole bone marrow cells) were defined as Hlf-tdTomatohi cells (Fig.1b). All the Hlf-tdTomatohi cells were located within the Sca-1+c-Kit+ fraction, including phenotypic HSCs (CD150+CD48) and short-term HSCs (CD150CD48) (Fig.1a, right panel and Supplementary Fig.1). To determine whether Hlf-tdTomatohi cells show a high frequency of functional HSCs within the bone marrow of the tibia, we compared engraftment capacity between the Hlf-tdTomatohi and Hlf-tdTomato cells (44.3% of the whole bone marrow cells) by the transplantation assay (Fig.1b). A total of 100 Hlf-tdTomatohi cells were capable of engraftment after primary and secondary transplantation, whereas 5000 Hlf-tdTomato cells were not capable (Fig.1c, d). These results indicate that functional HSCs were enriched within the Hlf-tdTomatohi fraction in the bone marrow of the long bones.

a Hlf-tdTomato expression in the bone marrow blood cells obtained from long bones. The black box indicates the Hlf-tdTomatohi population. b Flow cytometry sorting of Hlf-tdTomato cells. The Hlf-tdTomatohi box population was used for transplantation experiments. The tdTomato box population was used as a negative control. c Bone marrow transplantation (BMT) experiments. Irradiated mice were transplanted with 100 tdTomatohi/CD45+ cells or 5000 tdTomato/CD45+ cells. d Second BMT experiments.

Next, we developed a method to observe the dynamics of Hlf-tdTomatohi HSCs in the tibial bone marrow in vivo (Fig.2). Previous studies have used drilled tibia for intravital imaging of HSCs in the long bones14,15,16,17. However, drilling of the long bones precluded the comparison of the HSC dynamics between adults and neonates for two reasons. First, drilling could disturb the microenvironment of the bone marrow, and second, the long bone of neonates is too fragile to be drilled, and it is not possible to avoid bleeding from the blood vessels penetrating the neonatal tibia (Supplementary Fig.2a and Supplementary Movie1). A multiphoton imaging system equipped with a bone-penetrating fiber laser (average power, >2W; pulse width, 55> fs; wavelength, 1070nm) was established to overcome the limitations of the conventional methods (Fig.2a). In our system, tdTomato-positive cells were observed under the intact tibial bone tissue, which was visualized with second harmonic generation (SHG) signals in adult mice (3 months old) in vivo (Fig.2b; Supplementary Movie2). Blood capillaries in the bone marrow were visualized by intravenous injection of an infrared fluorescent dye Qtracker 655 to confirm the location of the tdTomato-positive cells within the tibial bone marrow. Intravital imaging showed that tdTomato-positive cells were located around the blood capillary network (Fig.2c, d), which are typical blood vessel patterns in the bone marrow of long bones28. These results suggest that Hlf-tdTomato-positive cells in the intact tibial bone marrow can be observed by the method developed in the present study.

a Experimental schema of intravital imaging of the tibial bone. The tibial bone was exposed, and the bone marrow was imaged without bone drilling using a high-powered infrared laser. b Z-stack images obtained via in vivo imaging of the bone marrow in the undrilled tibia of Hlf-tdTomato KI mice. Hlf-tdTomato-positive cells were observed in the bone marrow cavity. Bone surface, bone, and bone marrow cavities. The depth is indicated in the upper right corner of each panel. SHG, second harmonic generation. c Orthogonal view of 3D images confirmed that Hlf-tdTomato-positive cells were in the bone marrow under the tibial bone. Blood capillaries in the tibial bone marrow were visualized by intravenous injection of Qtracker 655. Cells indicated with arrows 1 and 2 are shown in higher magnification in (d). See also Supplementary Movie2. d Higher magnification images of Hlf-tdTomato-positive cells located around the bone marrow capillary. e Intravital imaging of a Runx1-GFP mouse before drilling. Depth is 155m from the bone surface. f Intravital imaging of the Runx1-GFP mouse after drilling. The same region in (d) was reimaged after drilling. Background signals were imaged in the red channel, suggesting that most signals in the green channel were artificial backgrounds. Arrows indicate Runx1-GFP cells that had no background signals in the red channel.

The technical advances of our method were evaluated by comparing it with the conventional method. For the conventional method, intravital imaging of Runx1-GFP transgenic mice, in which HSCs and progenitor cells strongly express GFP29, was conducted using a short wavelength (920nm) laser which does not easily penetrate bone. Intravital imaging showed that Runx1-GFP labeled cells in the bone marrow were blurred without drilling (Fig.2e), and GFP signals were only clearly observed after drilling (Fig.2f), indicating that drilling is essential for the conventional method. As previously reported11, artificial background signals were observed in the green channel (Fig.2f), whereas these were rarely seen using our imaging method (Fig.2b). These results highlight the advances of the intravital imaging developed in the present study.

There were 18.21.0 tdTomato-positive cells in the volume of the intravital images (around 600m600m100m=3.6107m3; n=5 volumes from five mice). Immunohistochemical staining of the tibial bone sections obtained from Hlf-tdTomato KI mice was conducted (Fig.3a) to determine which tdTomato-positive cells in the intravital images corresponded to the Hlf-tdTomatohi HSCs that were identified in the flow cytometry analysis (Fig.1). In the histological sections of the tibial bone, there were 18,201933 CD45-positive cells and 1,630269 tdTomato-positive cells within the same volume as the in vivo-imaged volume (n=3 sections from three mice; Fig.3b, c). Since the cells with the top 0.011% tdTomato intensity in CD45-positive cells were defined as Hlf-tdTomatohi cells (Fig.1), those with the top two tdTomato fluorescent intensities in the in vivo images corresponded to Hlf-tdTomatohi cells (18,201 cells0.011%=2 cells; Fig.3c). We defined the remaining Hlf-tdTomato-positive cells with lower intensity in the intravital images that include differentiated progenitor cells25 as Hlf-tdTomatolow cells (16 cells in Fig.3c), which had lower stem cell potential (Supplementary Fig.3). Most of the tdTomato-positive cells in the histological sections were not visible using intravital imaging due to the very low intensity (1612 cells in Fig.3c). These results suggested that the cells with top two tdTomato fluorescent intensities in the in vivo image were HSCs.

a Immunohistochemical staining of tibial bone sections was obtained from an adult Hlf-tdTomato KI mouse. Ter-119, red blood cell marker; CD45, blood cell marker except for mature red blood cells and platelets; DAPI, nuclear marker. b Fluorescence distribution of tdTomato signals in CD45-positive cells in the histological sections of the tibial bone. The number of CD45-positive cells within the same volume as the in vivo-imaged volume was 18,201933 cells. Error bars indicate standard error. c The number of CD45 and tdTomato-double-positive cells within the same volume as the in vivo-imaged volume was 1630269 cells. The number of tdTomato-positive cells was 18.21.0 cells, suggesting that the remaining 1612 cells were not visible in the intravital images due to the very low fluorescent intensity. The cells with the top 0.011% tdTomato intensity in CD45-positive cells were defined as Hlf-tdTomatohi cells (Fig.1); therefore, the cells with the top two tdTomato fluorescent intensities in the in vivo images correspond to the Hlf-tdTomatohi cells (18,201 cells0.011%=2 cells). We defined the remaining Hlf-tdTomato-positive cells with lower fluorescent intensity (16 cells in average) as Hlf-tdTomatolow cells for quantitative analysis of Hlf-tdTomato-positive cell dynamics in Figs.4 and 5.

Three-dimensional time-lapse imaging of undrilled tibial bone marrow was performed to observe the in vivo dynamics of Hlf-tdTomatohi HSCs (Fig.4a and Supplementary Movie3). Artifactual movement of the image area, mainly caused by the heartbeat, was corrected using image registration. Hlf-tdTomatohi HSCs were stationary (Fig.4b), although they showed oscillatory movements in the restricted area. In contrast, Hlf-tdTomatolow cells migrated (Fig.4c). Quantitative analysis of the HSC migration using TrackMate30 revealed that the velocity of the Hlf-tdTomatohi HSCs (0.0960.019m/min, 10 cells from five mice) was significantly lower than that of Hlf-tdTomatolow cells (0.1690.017m/min, 81 cells from five mice; p=0.008; t=2.886; g=0.505; Fig.4d). We also showed long-term engraftment of Hlf-tdTomatohi cells (Fig.1c, d and Supplementary Fig.1), indicating that HSCs were stationary but oscillatory. However, differentiated cells were motile in the bone marrow of adult long bones. Therefore, our findings demonstrate that Hlf-tdTomato KI mice, the inside-bone intravital imaging system and quantitative bioimaging analysis facilitate the evaluation of the migration dynamics of endogenous HSCs in the native microenvironment of long bones.

a Hlf-tdTomato-positive cells in the tibial bone marrow were imaged for 2h in vivo. Arrows indicate Hlf-tdTomatohi HSCs and arrowheads indicate Hlf-tdTomatolow cells. See also Supplementary Movie3. b Higher magnification time-lapse images of Hlf-tdTomatohi HSCs. c Higher magnification time-lapse images of Hlf-tdTomatolow cells. d Quantitative comparison of the migration dynamics between Hlf-tdTomatohi cells (ten cells from five mice) and Hlf-tdTomatolow cells (81 cells from five mice).

Neonatal HSCs are characterized by fast cell cycling and higher mitochondrial membrane potential4, indicating changes in the cellular properties between adult and neonatal HSCs. Gene expression patterns were compared between developing HSCs and matured HSCs using RNA-seq to evaluate changes in the properties of neonatal HSCs.

Gene set enrichment analysis (GSEA) showed significant enrichment in cell migration-related genes in neonatal HSCs (Supplementary Fig.4a, b). Consistently, changes in the expression of genes related to the cytoskeleton and cell adhesion were observed in neonatal HSCs (Supplementary Fig.4cf). Differences in the cell migration-related genes indicated differences in the migration dynamics of neonatal and adult HSCs in the tibial bone marrow. From these results, we focused on the migration dynamics in subsequent experiments for intravital imaging of neonatal mice.

Migration of HSCs into the bone marrow from other hematopoietic organs has been hypothesized since adult-type definitive HSCs are generated from the aortagonadmesonephros region8,25. However, the migration dynamics of HSCs during development is unclear. Therefore, three-dimensional time-lapse imaging of the bone marrow was performed in the undrilled tibia in neonates (postnatal day 2; Fig.5a, left). The orthogonal view confirmed that the intravital imaging of the intact tibial bone marrow enabled the observation of endogenous tdTomato-positive cells in neonatal Hlf-tdTomato KI mice (Fig.5a, right; Supplementary Movie4). Quantitative analyses showed that the velocity of Hlf-tdTomatohi cells (1.5161.010m/min, six cells from three mice; top two fluorescent intensities) was much higher than that of Hlf-tdTomatolow cells (0.0780.010m/min, 24 cells from three mice; p=0.002; Z=3.059; r=0.558) in neonates (Fig.5b). Furthermore, the velocity of Hlf-tdTomatohi cells in neonates was much higher than that in adults (p=0.017; Z=2.386; r=0.597, Supplementary Fig.2b). These results suggest that HSCs are motile in the tibial bone marrow of neonates.

a Tibial bone was exposed on postnatal day 2 (left). Orthogonal view image of the tibial bone marrow (right). See also Supplementary Movie4. b Quantitative comparison of the migration dynamics between Hlf-tdTomatohi cells (six cells from three mice) and Hlf-tdTomatolow cells (24 cells from three mice) in neonates. c Time-lapse images of developing (P2) bone marrow showing migration of an Hlf-tdTomatohi cell in the blood vessels. White arrows indicate migrating cells in the blood vessel and yellow arrows indicate stable cells outside the blood vessels. See also Supplementary Movie5. d Another example of an Hlf-tdTomatohi cell migrating in the bone marrow blood vessels in a neonatal mouse (P1).

Intravital imaging was conducted after visualizing blood vessels by injecting Qtracker 655 via the superficial temporal vein31 to check whether the motile Hlf-tdTomatohi cells were extravascular or intravascular. Some of the Hlf-tdTomatohi cells rapidly migrated in the blood vessels (Fig.5c, d, white arrows). Moreover, Hlf-tdTomatohi cell attached to the vessels during the imaging session, indicating extravasation and homing to the bone marrow (Fig.5c, 63 and 100min; Supplementary Movie5). In contrast, Hlf-tdTomatohi cells located outside the capillaries, which appeared to be extravasated prior to the imaging session, were stationary (Fig.5c, yellow arrows). These results suggest that motile Hlf-tdTomatohi cells migrate in the blood vessels of the neonatal tibia.

Finally, migration of HSCs from outside the bone marrow was observed using time-lapse imaging of the tibial bone cavity, where the blood vessels that penetrate the bone are located (Fig.6a). An Hlf-tdTomatohi cell was observed to rapidly migrate within the bone cavities (Fig.6b and Supplementary Movie6). Interestingly, the distance between the observed cell and the inner bone surface appeared to be increased (Supplementary Fig.5a, b), suggesting migration of the cell to the deeper part of the bone. Taken together, these results indicate that HSCs migrate in the bone cavities and bone marrow during tibial bone marrow formation.

a Two-photon images showing the blood vessels penetrating the bone cavity in the tibial bone in a neonatal mouse (P3). Yellow arrows indicate bone cavities. b Time-lapse images of Hlf-tdTomatohi cells migrating from the bone surface to the bone marrow cavity. White arrows indicate a migrating cell in the bone cavity. See also Supplementary Movie6. c Transient vessel penetration model for HSC homing during bone marrow formation. HSCs migrate from outside the tibia to inside via transient blood vessels penetrating the bone during bone marrow formation.

Read more:
Bone marrow imaging reveals the migration dynamics of neonatal hematopoietic stem cells | Communications Biology - Nature.com

Comments are closed.