Kindlin-3-mediated integrin adhesion is dispensable for quiescent but essential for activated hematopoietic stem cells

Abstract

Hematopoietic stem cells (HSCs) generate highly dividing hematopoietic progenitor cells (HPCs), which produce all blood cell lineages. HSCs are usually quiescent, retained by integrins in specific niches, and become activated when the pools of HPCs decrease. We report that Kindlin-3-mediated integrin activation controls homing of HSCs to the bone marrow (BM) and the retention of activated HSCs and HPCs but not of quiescent HSCs in their BM niches. Consequently, Kindlin-3-deficient HSCs enter quiescence and remain in the BM when cotransplanted with wild-type hematopoietic stem and progenitor cells (HSPCs), whereas they are hyperactivated and lost in the circulation when wild-type HSPCs are absent, leading to their exhaustion and reduced survival of recipients. The accumulation of HSPCs in the circulation of leukocyte adhesion deficiency type III patients, who lack Kindlin-3, underlines the conserved functions of Kindlin-3 in man and the importance of our findings for human disease.

Figure 1.

Survival of K3^−/− chimeras and distribution of K3^−/− HSPCs. (A) Kaplan-Meier survival curve of first generation K3^+/+ and K3^−/− FL chimeras. ***, P < 0.0001 by log-rank test. n = 41–47 per genotype; 15 independent experiments. (B) Representative FACS plots showing FL MNCs gated for lin⁻ cells (left), expression of AA4.1 and Mac-1 on lin⁻ cells (middle), and c-kit and Sca-1 expression on lin⁻AA4.1⁺Mac-1^med cells (right). Shown are the percentages of events within the gate ± SD. n = 8–9 per genotype. (C) Total number of FL MNCs from E14.5 embryos. n = 22–23 per genotype; four independent experiments. (D) Quantification of overall frequencies (percentage of live leukocytes) of LSK cells in E14.5 FLs. Error bars represent mean percentage ± SD. n = 8–9 per genotype. (C and D) ***, P < 0.0001 by unpaired t test. (E) Total number of LSK cells in E14.5 FLs. ns, P = 0.3527 by unpaired t test. n = 8–9 per genotype. (F) Frequency of CFU-Cs in E14.5 FLs. Error bars represent mean frequency ± SD. **, P = 0.0021 by Mann-Whitney test. n = 10 per genotype; three independent experiments. (G) LTC-IC assay performed with FL MNCs by limiting dilutions. Percentage of wells without CFUs is plotted against the MNC number. The frequency of LTC-ICs is shown. P < 0.0001 by Pearson’s χ² test. n = 4 per genotype; two independent experiments. (H) Total number of CFU-Cs in E14.5 FLs. ***, P = 0.0002 by unpaired t test. (I) Total number of LTC-ICs per FL of E14.5 embryos. ns, P > 0.05 by Mann-Whitney test. (H and I) n = 22–23 per genotype; four independent experiments. (J) Total number of MNCs from BM and whole Spl. Data are mean cell counts ± SD. ns, P > 0.05; ***, P < 0.0001 by Kruskal-Wallis test and Dunn’s multiple comparison post-test. n = 10–15 per group; 12 independent experiments. (K) Total number of B cells (B220⁺), T cells (CD3e⁺), erythroid cells (Ter119⁺), neutrophils (Gr-1^hiMac-1⁺), and monocytes/macrophages (Gr-1^intMac-1⁺) in the BM. Data are presented as box and whisker plots. The horizontal lines inside the boxes represent the median, the box edges show the lower and upper quartiles, and the whiskers indicate the minimum and maximum. *, P < 0.05; **, P < 0.01; and ***, P < 0.001 by Kruskal-Wallis test and Dunn’s multiple comparison post-test. n = 10–15 per genotype; 12 independent experiments. (C, E, H, and I) Error bars represent mean cell counts ± SD. ns, not significant.

Figure 2.

BM homing of K3^−/− LSK cells. (A) Short-term homing assay with labeled BM LSK cells; mean percentages ± SD. **, P = 0.0046 by unpaired t test. n = 5–6 per genotype; 10 independent experiments. (B) Quantification of extravasated cells per 10⁵ injected LSK cells in BM by intravital two-photon microscopy 18 h after transplantation, given as mean value ± SD. *, P = 0.0159 by Mann-Whitney test. n = 4–5 per genotype; nine independent experiments. (C) Representative images illustrating transmigration of LSK cells up to 6 h after injection in Col2.3-GFP recipients. White (CMTMR), LSK cells; green (GFP), osteoblasts; blue (second harmonic signal), bone; and red (Q-Tracker 695 nm), blood. Arrowheads indicate LSK cells transmigrating across endothelium (white dashed lines); left panels, before transmigration; right panels, after transmigration. (D) LSK cells visualized in BM microvessels for up to 6 h after injection. Mean ± SD. ***, P = 0.0006 by unpaired t test. n = 5 per genotype; 10 independent experiments. (E) Firm adherent cells shown in D grouped according to their residence time on the BM endothelium and shown as frequencies ± SEM of the absolute number of visualized cells. ***, P < 0.0001 for <1 min; **, P = 0.0037 for >1 min; ***, P = 0.0008 for >10 min, by Fisher’s test. (F) Representative images of adherent LSK cells at the indicated time after transfer. Red (CMTMR), LSK cells; blue (second harmonic signal), bone; and green (FITC dextran), blood. Arrows, arrowheads, and asterisk indicate different single LSK cells, and the white dashed lines outline the endothelium. (G) Surface expression of integrins on LSK cells isolated from BM (top) or FL cells (bottom) is presented as histograms of the mean fluorescence intensity for the indicated anti-integrin mAb on K3^+/+ (blue lines) or K3^−/− cells (red lines). Isotype controls are indicated as shaded histograms. n = 3–4 per group; three to four independent experiments. (H and I) Adhesion analysis of sorted K3^+/+, with or without preincubation with an anti–α4 integrin–blocking mAb, and K3^−/− BM (H) or E14.5 FL LSK (I) cells under shear in microflow chambers precoated with rmE-selectin alone (E), rmE-selectin and VCAM-1, or rmE-selectin, CXCL12, and VCAM-1. Error bars represent the mean percentage of adherent cells to total LSK cells ± SEM. ns, P > 0.05; *, P < 0.05; **, P < 0.01 by Kruskal-Wallis test and Dunn’s multiple comparison post-test. n ≥ 4 per group; 10 independent experiments. ns, not significant. Bars, 50 µm.

Figure 3.

Kindlin-3–deficient HSCs home to the BM. (A) Frequency of CFU-Cs in sorted CD45.2⁺ BM cells from chimeras 4 mo after transplantation. Mean cell counts ± SD. P = 0.8857 by Mann-Whitney test. n = 4 per genotype; three independent experiments. (B) LTC-IC assay performed with donor-derived BM-MNCs from chimeras 4 mo after transplantation. The percentage of wells without CFUs is plotted against the MNC number. The frequency of LTC-ICs is shown. P = 0.0965 by Pearson’s χ² test. n = 4 per genotype; three independent experiments. (C) Representative FACS plots of BM HSPCs 1–13 mo after transplantation analyzed with signaling lymphocytic activation molecule (SLAM) markers. The left panels show c-kit and Sca-1 expression on lin⁻ cells. Numbers refer to the percentages of live leukocytes ± SD. n = 24–27 per genotype; 10 independent experiments. (D) Quantification of C. LSK (P > 0.05 by unpaired t test); LSK CD150⁻CD48⁺ (*, P = 0.0121 by unpaired t test); LSK CD150⁺CD48⁺ (***, P < 0.0001 by unpaired t test); LSK CD150⁺CD48⁻ (***, P < 0.0001 by Mann-Whitney test); LSK CD150⁺CD48⁻CD34⁺ (***, P = 0.0003 by unpaired t test); and LSK CD150⁺CD48⁻CD34⁻ cells (***, P < 0.0001 by Mann-Whitney test). Mean cell numbers ± SD are given. (E) Representative FACS plots of BM LSK cells at the indicated time after transplantation gated for CD150 against CD48. Numbers represent mean percentage ± SD of events within the LSK population. n = 3–11 per genotype and time point; three independent experiments. (F and G) Quantification at the indicated time after transplantation of BM LSK CD150⁺CD48⁻ (1 mo: P = 0.4278; 4 mo and 9–13 mo: ***, P < 0.0001 by unpaired t test; F) and LSK CD150⁺CD48⁻CD34⁻ cells (1 mo: P = 0.4642 by Mann-Whitney test; 4 mo: ***, P = 0.0005 by unpaired t test; 9–13 mo: **, P = 0.002 by unpaired t test; G). Cells are shown as mean percentage ± SD. n = 6–10 per genotype; four independent experiments. ns, not significant.

Figure 4.

K3^−/− HSCs remain hyperactive and accumulate in the PB. (A and B) PolyI:C and 5-FU double treatment of K3^+/+ and K3^−/− chimeras. (A) Diagram showing drug combinations and timing. (B) Kaplan-Meier survival curves. ***, P < 0.0001 by log-rank test. n = 16–22 per treatment group; five independent experiments. (C) In vivo BrdU uptake assay shown as mean percentage ± SD of BrdU⁺ cells. BrdU levels were measured 12–14 h after injection in LS⁻K⁺ (LK; P = 0.4078 by unpaired t test), LSK (***, P = 0.0009 by unpaired t test), LSK CD150⁺CD48⁻ (**, P = 0.0014 by Mann-Whitney test), and LSK CD150⁺CD34⁻ cells (***, P = 0.0004 by Mann-Whitney test). n = 6–14 per genotype; six independent experiments. (D) Percentage of Ki67⁺ BM LSK cells indicated as mean values ± SD. **, P = 0.0022 by paired t test. n = 18 per genotype; 12 independent experiments. (E) Relative expression of p21, p27, and p57 mRNA in sorted CD150⁺ LSK cells analyzed by quantitative PCR (mean ± SD). ***, P = 0.0007 by unpaired t test. n = 6–7 per genotype; six independent experiments. (F) Mean percentage ± SD of BrdU LRCs in the indicated populations after 45 d of chase phase: total LSK cells (***, P < 0.0001 by unpaired t test), LSK CD150⁻CD48⁺ (***, P < 0.0001), LSK CD150⁺CD48⁺ (***, P = 0.0006), LSK CD150⁺CD48⁻ (**, P = 0.0014), and LSK CD150⁺CD34⁻ (*, P = 0.0350) by Mann-Whitney test. n = 7–8 per genotype; three independent experiments. (G) Representative FACS plots showing the expression of c-kit and Sca-1 on lin⁻ cells from Spl and PB. The frequencies of LS⁻K⁺ (orange boxes) and LSK cells (green boxes) are shown (percentages of live leukocytes ± SD). n = 24–26 per genotype; 10 independent experiments. (H) Quantification of LS⁻K⁺ (LK) and LSK cells in Spl. (ns, P > 0.05; **, P = 0.0031) and PB (***, P < 0.0001). Statistics by Mann-Whitney test. Error bars represent mean cell counts ± SD. n = 25–26 per genotype; 10 independent experiments. (I) Frequency of CFU-Cs shown as mean cell counts ± SD in CD45.2⁺ cells from Spl (*, P = 0.0286) and PB (*, P = 0.0294). Statistics by Mann-Whitney test. n = 4 per genotype; three independent experiments. (J) Multilineage engraftment potential of PB cells isolated from chimeras. Shown are percentages of donor-derived whole leukocytes (CD45⁺), myeloid cells (Gr-1⁺, Mac-1⁺, and Gr-1⁺Mac-1⁺), B cells (B220⁺), and T cells (CD3e⁺) from PB at the indicated times after transplantation. Data are plotted for individual recipients. n = 5 for K3^+/+ (blue lines) and n = 19 for K3^−/− (red lines); three independent experiments. Differences in mean lineage engraftment 16 wk after transplantation are indicated (leukocytes: *, P = 0.0157; myeloid cells: **, P = 0.0085; B cells: **, P = 0.0028; and T cells: ***, P = 0.0008). Statistics by Mann-Whitney test. (K) Mobilization of LSK cells in chimeras shown as LSK cells per milliliter of blood after G-CSF or PBS treatment. Significance at day 6 is indicated (K3^+/+ vs. K3^−/− with G-CSF: ***, P < 0.0001; K3^+/+ vs. K3^−/− with PBS: **, P = 0.0049 by unpaired t test). n = 3–5 per genotype; four independent experiments. ns, not significant.

Figure 5.

Accelerated loss of K3^−/− HSCs under hematopoietic stress. (A) Kaplan-Meier survival curves after serial BM transplantations. ***, P < 0.0001 by log-rank test. n = 42–55 recipients per genotype; nine independent experiments. (B) Second generation recipients transplanted with BM LSK CD150⁺ cells and host-type whole Spl cells. Mean percentages ± SD of donor-derived whole leukocytes (CD45⁺), myeloid cells (Gr-1⁺, Mac-1⁺, and Gr-1⁺Mac-1⁺), B cells (B220⁺), and T cells (CD3e⁺) shown in PB at the indicated times after transplantation. ***, P < 0.0001 by Mann-Whitney test. (C) Mean percentages ± SD of donor-derived HSPCs in BM of second generation recipients 11 wk after transplantation. ***, P < 0.0001 by Mann-Whitney test. (B and C) n = 23–28; four to five independent experiments. (D) Kaplan-Meier survival curve showing response of control and K3^fl/flRosa26^Cre-ERT2 chimeras (n = 7) to tamoxifen treatment. The dashed arrow indicates the median survival of tamoxifen-treated K3^fl/flRosa26^Cre-ERT2 chimeras. The control group consisted of K3^fl/flRosa26^Cre-ERT2 (without TAM, n = 4) and K3^fl/fl (with TAM, n = 7; or without TAM, n = 3) chimeras. *, P = 0.0385 by log-rank test. (E) Kaplan-Meier survival curve of K3^fl/flRosa26^Cre-ERT2 (n = 10) and control mice to 5-FU treatment. The control group contained K3^fl/flRosa26^Cre-ERT2 (without TAM, n = 7), K3^+/+Rosa26^Cre-ERT2 (with TAM, n = 6; or without TAM, n = 5), and K3^fl/fl (with TAM, n = 2; or without TAM, n = 2) mice. ***, P < 0.0001 by log-rank test.

Figure 6.

Kindlin-3 retains active HSCs and HPCs in the BM. (A) Design showing the generation of mix chimeras. E14.5 FL cells (CD45.2⁺) were mixed with WT congenic B6.SJL E14.5 FL cells (CD45.1⁺) and transplanted to lethally irradiated (C57BL/6 × congenic B6.SJL) F1 generation recipient mice (CD45.1⁺/CD45.2⁺). (B) Representative FACS plots of PB from mix chimeras 4 mo after transplantation. The top panels show whole leukocytes, and the bottom panels show lin⁻ cells from CD45.2⁺ and CD45.1⁺ populations gated for c-kit against Sca-1. Numbers represent mean overall frequencies (percentage of live PB leukocytes) ± SD. n = 6–8 per genotype; six independent experiments. (C) G-CSF–induced mobilization of LSK cells in mix chimeras. The histogram shows the mean number of LSK cells per milliliter PB ± SD. ns, P > 0.05; **, P < 0.01; and ***, P < 0.001 by a one-way analysis of variance followed by Tukey’s multiple comparison test. n = 5–6 per genotype and treatment. (D–F) Percentage of donor-derived LSK CD150⁺CD48⁻CD34⁻ (D), LSK CD150⁺CD48⁻ (E), and LSK CD150⁻CD48⁺ populations (F) in the BM of mix chimeras at the indicated times after transplantation. Dots show the mean percentage ± SD. ns, P > 0.05; ***, P < 0.001 by unpaired t test. n = 4–11 per genotype and time point; two independent experiments. (G) LRC assay with mix chimeras. Percentages of BrdU LRCs in the indicated populations after 45–50 d of chase phase. Total LSK (*, P = 0.0286), LSK CD150⁻CD48⁺ (*, P = 0.0286), LSK CD150⁺CD48⁻ (ns, P = 0.0571), and LSK CD150⁺CD34⁻ cells (ns, P = 0.1143) by Mann-Whitney test. Error bars represent mean percentage ± SD. n = 4 per genotype; two independent experiments. (H) Percentage of Ki67⁺ donor cells (CD45.2⁺) in the indicated populations. ns, P > 0.05 by unpaired t test. n = 8 per genotype; three independent experiments. (I–K) Percentage of donor-derived LSK CD150⁺CD48⁻CD34⁻ (I), LSK CD150⁺CD48⁻ (J), and LSK CD150⁻CD48⁺ (K) populations in BM from K3^+/+Rosa26^Cre-ERT2 (blue) and K3^fl/flRosa26^Cre-ERT2 (red) mix chimeras untreated (Ctrl.), treated with tamoxifen (TAM PBS), or treated with tamoxifen and 5-FU (TAM 5-FU). The histogram shows the mean percentage ± SD. ns, P > 0.05; *, P < 0.05; and ***, P < 0.0006 by unpaired t test. n = 4–6 per genotype and treatment. (L) Competitive repopulation assay with BM LSK CD150⁺CD45.2⁺ cells from mix chimeras. Recipients received LSK CD150⁺CD45.2⁺ cells and host-type whole Spl cells. Shown are percentages of donor-derived whole leukocytes (CD45⁺), myeloid cells (Gr-1⁺, Mac-1⁺, and Gr-1⁺Mac-1⁺), B cells (B220⁺), and T cells (CD3e⁺) from individual recipients at the indicated times after transplantation. Significant differences in mean lineage engraftment are indicated (4 wk after transplantation: leukocytes [ns, P = 0.8707] by unpaired t test; myeloid cells [**, P = 0.0093], B cells [*, P = 0.0337], and T cells [**, P = 0.0074] by Mann-Whitney test; 8 and 12 wk after transplantation: leukocytes: ***, P < 0.0001 by unpaired t test; myeloid, B, and T cells: ***, P < 0.0001 [or P = 0.0006 for myeloid cells after 8 wk] by Mann-Whitney test). M) Mean percentage ± SD of donor-derived HSPCs in BM of second generation recipients 12 wk after transplantation. ***, P < 0.0001 by Mann-Whitney test. (L and M) n = 14–17; four independent experiments. ns, not significant; Wt, wild type.

Figure 7.

Kindlin-3 regulates HSPC homeostasis in an integrin-dependent manner. (A) Relative mean fluorescence intensity of 9EG7 binding on BM LSK CD150⁻CD48⁺ and LSK CD150⁺CD48⁻ cells from chimeras. Error bars indicate geometric mean fluorescence intensity (MFI) ± SD. **, P < 0.01 by paired t test. n = 3 per genotype; two independent experiments. (B) Kaplan-Meier survival curve of lethally irradiated recipients transplanted with LSK cells transduced with lentiviral expression constructs as indicated together with WBM cells from a K3^−/− 100% chimera. ***, P < 0.0001 by log-rank test. n = 9–10 recipients per group; two independent experiments. (C) Mean percentages ± SD of donor-derived whole leukocytes (CD45⁺), myeloid cells (Gr-1⁺, Mac-1⁺, and Gr-1⁺Mac-1⁺), B cells (B220⁺), and T cells (CD3e⁺) in PB from a WT C57BL/6 mouse (wt-Bl6), recipients of K3^+/+ LSK cells transduced with a GFP expression construct (K3^+/+-GFP; n = 5), recipients of K3^−/− LSK cells transduced with a K3-WT expression construct (K3^−/−-K3^wt; n = 4), and a single surviving recipient of K3^−/− LSK cells transduced with a K3-QWAA expression construct (K3^−/−-K3^mut) 3 mo after transplantation. Two independent experiments. (D) FACS plots of PB from a LAD-III patient and an age-matched healthy child. Shown are MNCs gated for CD34 against CD45 and CD38 against CD34. The blue boxes depict the CD45⁺CD34⁺ cells (left), CD38⁺CD34⁺ cells (top right), and CD38⁻CD34⁺ cells (bottom right). The numbers indicate the percentage of gated cells. (E) CFU assay with PB MNCs containing 100 CD34⁺ cells (calculated according to the FACS frequencies) from a LAD-III patient and an age-matched healthy child. Shown are the numbers of colonies per blood MNC containing 100 CD34⁺ cells. BFU-E, burst forming unit–erythroid; CFU-GM, CFU-granulocyte/macrophage; wt, wild type.