In vivo time-lapse imaging shows diverse niche engagement by quiescent and naturally activated hematopoietic stem cells

The function of hematopoietic stem cells (HSCs) depends on their interaction with complex niches within the bone marrow (BM)^1,2  and must readily adjust to meet the changing needs for differentiated cells during situations of stress, such as the immune responses generated by infectious agents.^3-5  The current working hypothesis is that changes in HSC-niche interactions may cause changes in HSC function and therefore a major challenge in the field is to understand the specific cellular behaviors (eg, duration of cell-cell/cell-matrix interactions) resulting in different HSC fates (quiescence vs proliferation, self-renewal vs differentiation). Direct visualization of HSCs in vivo and over time is the ideal approach to monitor the nature of the cellular interactions between individual stem cells and their niches and can be achieved by intravital microscopy of mouse calvarium BM.^6,7 

We studied a naturally occurring, nonlethal mouse model of infection initiated by ingestion of the parasitic nematode Trichinella spiralis, the infectious agent causing trichinosis. Ingested T spiralis reside for several days in the intestinal epithelium, where they mature, proliferate, and release larvae, which migrate into the circulatory system and invade the skeletal muscle, where they form cysts and survive for the life of the host.⁸  Migration of parasites causes extensive tissue damage and intense inflammation over the early phase of infection (weeks 1-3), which is ameliorated by the induction of regulatory cytokines during the chronic phase of the disease.^9,10 

Here we show that acute T spiralis infection affects hematopoietic dynamics and boosts HSC function. Using time-lapse intravital microscopy of transplanted HSCs, we show that acute infection is accompanied by changes in HSC-niche interactions. Steady-state HSCs reside within confined niches, whereas infection-exposed HSCs become migratory and interact with wider microenvironment regions.

For information on the study design, see supplemental Methods, available on the Blood Web site.

The UK Home Office PPL No. 70/7061 approves and covers all the animal work described here.

We examined hematopoietic stem and progenitor cell (HSPC) frequencies at 3, 14, and 28 days after ingestion of T spiralis larvae to capture the consequences of the early, intense inflammatory response and of the chronic infection phase, respectively (Figure 1A-C; supplemental Figures 1 and 2). Common lymphoid progenitors (CLPs) frequency increased already 3 days after ingestion (supplemental Figure 2C); at day 14, we observed increased CLP and granulocyte/monocyte progenitors (GMPs) but decreased common myeloid progenitors (CMP; Figure 1A), and by day 28 all frequencies were back to control values (supplemental Figure 2D). We observed the same trends in the total numbers of these cell populations (data not shown). These changes were reflected by increased numbers of colonies detected in CFU-C assays performed at day 14 after ingestion (Figure 1B) and by mild peripheral blood leukocytosis at later time points (supplemental Figure 3A-F).

The frequency, total number, and cell cycle profile of more primitive HSPC populations remained unchanged (Figure 1D-E; supplemental Figure 2; data not shown). However, competitive repopulation assays showed that BM from infected animals had a consistent reconstitution advantage across all lineages both short and long term (Figure 1E-F; supplemental Figure 4). Analysis of spleens from control and infected animals revealed changes similar to those observed in the BM (supplemental Figure 3G-I) and consistent with previous observation that numbers of lymphoid and myeloid cells in secondary lymphoid organs peak at day 15 after ingestion, returning to normal by day 30.¹¹ 

T spiralis infection therefore was an appropriate model to investigate whether a transient imbalance of hematopoiesis and an increase in HSC engraftment potential would correspond to changes in HSC-niche interactions.

Stable positioning of the anesthetized mouse for long periods allowed us to monitor HSC-niche interactions in mouse calvarium through longitudinal tracking of the cells of interest.¹²  DiD-labeled LT-HSCs (Lin^low, c-Kit⁺, Sca1⁺, CD34⁻, Flk2⁻ cells, herein referred to as HSCs) were injected into lethally irradiated recipient mice expressing GFP in osteoblastic cells (Col2.3GFP reporter mice¹³ ). The next day, we performed 5-hour time-lapse imaging of identified HSCs and surrounding osteoblasts. HSCs remained within confined regions throughout the observation period; however, they often exhibited oscillatory movements as large as 1/2 cell diameters (Figure 2A-C,G; supplemental Movie 1). Only 12% of all observed HSCs exhibited a net displacement greater than the distance covered by the oscillations (Figure 2I). The observation that HSCs often physically oscillate within their niches could reflect an as yet unappreciated level of complexity in the cross-talk between HSCs and niche cells. This may be mediated not only by secreted factors but also by dynamic cell-cell contacts, where the timing of each connection could further modulate any signals triggered.

We then imaged HSCs from mice at day 14 post ingestion (hereafter called exp.HSCs for infection-exposed HSCs). Surprisingly, 57% of exp.HSCs appeared migratory and moved within the field of view, generating distinct 3-dimensional tracks that occupied a volume of ∼4 times the size of that of HSC tracks (Figure 2D-F,H-I; supplemental Movie 2). A detailed comparative analysis of the tracks revealed that little or no total displacement was detected for HSCs, whereas a considerable proportion of exp.HSCs covered a wider range of net distances (Figure 2J). Despite the difference in total displacement, HSCs and exp.HSCs moved at a similar mean speed; however, the variability of individual speeds was significantly larger for exp.HSCs compared with HSCs (Figure 2K).

The major difference between HSCs and exp.HSCs was the shape of their tracks. Whereas HSCs generated confined circular tracks, exp.HSCs often alternated oscillatory movement and net migration, generating tracks similar to those of leukocytes scanning peripheral tissues (supplemental Movie 2).¹⁴  These observations were quantified by assessing the cells' linear progression (ratio between displacement and total track length, yielding a maximal value of 1 for objects moving in a straight line). HSC tracks had lower values (up to 0.35; median, 0.1) and exp.HSC tracks had higher values (up to 0.87; median, 0.21; Figure 2L). Notably, HSCs showed lower variability in individual linear progression values.

Given the nature of our in vivo imaging assay, the molecular foundations underpinning migratory behavior must be acquired by exp.HSCs in situ, prior to harvest, transplantation, and imaging. We tested this hypothesis by examining the expression levels of molecules known to be involved in HSC adhesion and migration and observed that exp.HSCs had consistently increased levels of CD44 and CXCR4 (supplemental Figure 5A-D).

CXCR4 guides to and retains HSCs within the BM, and administration of its antagonist AMD3100 leads to HSPC mobilization into peripheral blood within 1 hour.¹⁵  Injection of AMD3100 during time-lapse imaging of exp.HSCs did not significantly change the parameters describing cell migration in the subsequent 3 hours (data not shown); however, the length of the tracks was reduced overall (supplemental Figure 5E). This was likely due to intra-/extravasation events, resulting in abrupt ending/beginning of tracks and reflecting increased cell trafficking stimulated by the drug.

Available reports indicate that chronic and lethal infections/cytokine stimulation lead to HSC exhaustion,^16-19  acute infection to transient effects,²⁰  and poly-IC administration to differential HSC localization.²¹  Here, we demonstrate that acute T spiralis infection changes the nature of HSC-niche interaction, which correlated with a marked engraftment advantage. The observed migratory behavior may be an essential component of the signaling cascade maintaining the HSC pool during a nonlethal infection.

The authors thank Nicola Ruivo, Rhiannon White, Catherine Simpson, Jane Srivastava, Dr Viktor Janzen, Samuel Jones, Dr Corinna Schnoeller, Dr Simon Schultz, and Dr Martin Spitaler for technical assistance and advice with mouse husbandry, T spiralis harvest and administration, flow cytometry, mouse holder production, and imaging and Dr Edwin Hawkins, Dr Peter O'Donovan, and Prof Robert Sinden for critical feedback on writing the manuscript.

This work was supported by Human Frontiers Science Program (grant RGP0051/2011 to C.L.C. and I.R.), Biotechnology and Biological Sciences Research Council, Cancer Research United Kingdom, Kay Kendal Leukaemia Fund (grants BB/I004033/1, C36195/A11831, KKL460 to C.L.C.) and Wellcome Trust (PhD fellowship to J.S.K.).

Contribution: N.M.R., M.K.S., N.S., A.K., J.S.K., K.G., M.E.S., I.R., and C.L.C. designed/performed experiments; analyzed data; and contributed to the writing of the manuscript.

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

Correspondence: Cristina Lo Celso, Sir Alexander Fleming Building, Imperial College London, South Kensington Campus, London, SW7 2AZ United Kingdom; e-mail: c.lo-celso@imperial.ac.uk.