Niche WNT5A regulates the actin cytoskeleton during regeneration of hematopoietic stem cells

Abstract

Here, we show that the Wnt5a-haploinsufficient niche regenerates dysfunctional HSCs, which do not successfully engraft in secondary recipients. RNA sequencing of the regenerated donor Lin^- SCA-1⁺ KIT⁺ (LSK) cells shows dysregulated expression of ZEB1-associated genes involved in the small GTPase-dependent actin polymerization pathway. Misexpression of DOCK2, WAVE2, and activation of CDC42 results in apolar F-actin localization, leading to defects in adhesion, migration and homing of HSCs regenerated in a Wnt5a-haploinsufficient microenvironment. Moreover, these cells show increased differentiation in vitro, with rapid loss of HSC-enriched LSK cells. Our study further shows that the Wnt5a-haploinsufficient environment similarly affects BCR-ABL^p185 leukemia-initiating cells, which fail to generate leukemia in 42% of the studied recipients, or to transfer leukemia to secondary hosts. Thus, we show that WNT5A in the bone marrow niche is required to regenerate HSCs and leukemic cells with functional ability to rearrange the actin cytoskeleton and engraft successfully.

Figure 1.

LT-HSCs from Wnt5a^+/− mice show unchanged repopulating capacity. (A) Absolute numbers of MPs, CLPs, CD34⁺ LSKs, and CD34⁻ CD150⁺ LSKs in the BM of WT and Wnt5a^+/− mice (n = 15 each). (B) Total numbers of CFCs. (C) Experimental design. Serial transplantations of HSCs of both genotypes. (D) Donor engraftment in the BM of primary, secondary, and tertiary (1°, 2°, and 3°) recipient mice. The number of engrafted mice compared to the total number of transplanted mice is also shown. (E) Representative FACS plots of the BM from the 1° recipients (n = 8). (F–H) The lymphoid and myeloid engraftment in the PB of 1°, 2°, and 3° recipients 5, 10, and 16 wk after Tx, respectively. (I–K) The corresponding percentage of donor cells and absolute numbers of HSCs and progenitors in the BM 16 wk after transplantation. (L) The total number of transplanted and repopulated LT-LSKs in serial transplantations (n = 8) in 1° and 2° recipients. n = 4 (WT); n = 5 (Wnt5a^+/−), respectively, in the 3° recipients. (M) Numbers of donor CD34⁻ CD150⁺ LSKs regenerated in 1°, 2°, and 3° recipients. Shown are the results of two independent experiments. Column plots show the mean ± SEM. *, P < 0.05 (Student’s t test).

Figure 2.

Impaired engraftment of WT LSKs after exposure to Wnt5a-deficient environment (LSK-5a) compared with regeneration in WT recipients (LSK-WT). (A) Experimental design. Serial transplantations of WT cells into recipients of both genotypes. (B) The donor engraftment in PB, BM, and spleens of recipients (WT, n = 10; Wnt5a^+/−, n = 14). (C) Lymphoid and myeloid engraftment in the PB of 1° recipients 5, 10, and 16 wk after transplantation. (D) Representative FACS plots of BM from 1° recipients. (E) Absolute numbers of engrafted Lin⁻, MPs, LSKs, and CD34⁻ CD150⁺ LSKs in the BM of 1° recipients. (WT, n =15; Wnt5a^+/−, n = 7). Presented are the results of three independent experiments. (F) The donor engraftment in BM of 2° recipients. The engraftment with >1% of lymphoid and myeloid cells was defined as positive. (G) Limiting dilution analysis of 1,000, 3,000, and 6,000 1° LSKs in the PB of the recipient mice. (H) Lymphoid and myeloid engraftment in PB of 2° mice after 5, 10, and 16 wk after transplantation. (I) Representative FACS plots of BM from 2° recipients 16 wk after transplantation. (J) The absolute numbers of donor Lin⁻, MPs, LSKs, and LT-LSKs in BM of 2° mice. (WT, n = 15; Wnt5a^+/−, n = 7). These are the combined results of three independent experiments shown as mean ± SEM. *, P < 0.05 (Student’s t test).

Figure 3.

WNT5A expression and analysis of the niche cells after transplantation. (A) Western blots and protein content in cultured stromal cells from WT and Wnt5a^+/− mice quantified with ImageJ software and presented as relative values to GAPDH (n = 6). (B) Representative FACS plots of intracellular WNT5A expression in MSCs, OBCs, and ECs. (C) Laminin and SCA-1 expression in bone sections of mice from both genotypes (n = 3). (D) Experimental workflow. RNA-Seq on OBCs and MSCs isolated form the 1° recipients of both genotypes (n = 3). (E) Representative FACS gating for sorting of niche cells. (F) The percentages of ECs, OBCs, and MSCs in 1° recipients of both genotypes. (G) The two-dimensional representation of RNA-Seq probes by PCA computed using the 400 most variable gene expression values. (H) Volcano plot comparison of MSCs and OBCs irrespective of recipient genotype. (I–K) Expression of niche and perivascular genes compiled from published reports (Khan et al., 2016; I and J), and our own previous work (K). (L) Volcano plot of the differential gene expression in MSCs from WT and Wnt5a^+/− recipients. For each detected gene, differential gene expression testing is shown as log₂ fold-change against −log₁₀ (p-adj). Significantly (FDR < 0.1) DEGs are shown in red.

Figure 4.

The actin-regulatory pathway is dysregulated in LSK-5a cells. (A) Experimental workflow. The RNA-seq analysis of donor LSKs from 1° recipients of both genotypes (WT, n = 6; Wnt5a^+/−, n = 5). (B) Plot showing the two-dimensional representation of RNA-Seq samples by PCA computed using the 500 most variable gene expression values. (C) Shown are expression values of significantly (FDR < 0.05) DEGs in LSK-WT and LSK-5a samples. Rows denote genes and columns denote samples of donor LSKs isolated from individual recipient mice. Values are normalized to the mean log-expression in LSK-WT samples. (D) The 10 KEGG pathways with the strongest significant (FDR < 0.05) enrichment of DEGs. Bars show the number of significantly up- (red) or down-regulated (blue) genes in the corresponding pathway and are sorted according to the enrichment score (ES). Pathways containing <5 DEGs were discarded (Table S6). (E) Graph showing the comparison of the ratio of the mean expression of differentially up-regulated genes in the regulation of the actin cytoskeleton pathway between LSK-WT and LSK-5a. (F) The DOCK2, CDC42, WAVE2, and F-actin protein expression in LSK-WT and -5a counter stained with DAPI. The fluorescence intensities were quantified with ImageJ software. (G) The percentage of cells with apolar expression of respective proteins in 1° LSKs from WT or Wnt5a^+/− recipients. (H) F-actin expression in CD34⁻ subfractions of 1° LSKs and the percentage of cells with apolar distribution of the F-actin complex. (I) The β−catenin and c-MYC expression in 1° LSKs of both genotype recipients. (J) The protein content of the active CDC42-GTP in donor Lin⁻ cells of 1° recipients relative to the total CDC42. The intensities on Western blot calculated with ImageJ software. (K) Transcription factor binding motifs with the highest influence on gene expression changes between LSK-WT (blue boxes) and LSK-5a (pink boxes). Shown are the changes in motif activities with corresponding z-scores of the averaged ISMARA analysis. (L) Enrichment analysis for the KEGG pathway actin regulation in genes with promoters containing transcription factor binding motifs identified by ISMARA. The log-transformed enrichment scores are plotted and shaded according to significance (Fisher’s exact test). (M) The mRNA content of ZEB1 target genes in in Zeb1 mutant and WT MEFs relative to the Gorasp2. *, P < 0.05 (E, G, J, and M, Student’s t test; F and I, MWU test).

Figure 5.

Wnt5a-deficient niche modulates actin-dependent cellular responses. (A) Experimental workflow. 16 wk after primary transplantation, LSKs were sorted from the BM of the 1° recipients of both genotypes and analyzed for mechanical cell properties, adhesion, migration, homing, and single cell cultures. (B) Cell diameters of 1° LSKs determined from phase-contrast images and presented as boxplots with 25th, 50th, and 75th percentiles denoted as horizontal lines, and 10th and 90th percentiles shown as whiskers. (C) The 1° LSKs adhering to a VCAM-1–coated surface were probed using an AFM cantilever equipped with 2.5-µm-radius bead. Experiments were conducted in the presence of 150 ng/ml rmCXCL12. Apparent Young’s Moduli for individual cells (averaged for repeated measurements) are presented as dot plots. The number of probed cells is indicated. Mean apparent Young’s Moduli of the probed cell population are denoted by horizontal lines. (D) The F-actin distribution in 1° LSK treated with 150 ng/ml rmCXCL12. (E) The percentages of cells with apolar distribution of F-actin. Presented are the combined results of two independent experiments. (F) The pictures of adhered 1° LSKs on VCAM-1–coated slides stained with DAPI and counted. Presented are the results of three independent experiments. (G) FACS plots of 1° donor-type Lin⁻ cells (dot plots, left) migrated toward 150 ng/ml rmCXCL12 in the lower compartment of a Boyden camber (column graphs). These are the results of two independent experiments. (H) FACS plots of homed 1° LSKs cells in the BM 2° WT mice. (I) The total number of homed 1° LSKs the BM of 2° WT recipients. (J) The mean clone size of CD34⁻ LSK-WT and 5a cells cultured as single cells in conditioned medium from the UG26-1B6 cell line, supplemented with SCF and IL-11, with or without 500 ng/ml rWNT5A, counted each day for 7 d under the light microscope. (K) The total number of wells with more than two cells, shown over time. (L) FACS plots of Lin⁻ cells from pooled clones after 7 d culturing (Fig. S2). (M) The quantified percentage of lymphoid, myeloid, and LSK fractions of live cells. *, P < 0.05 (E, F, and G, Student’s t test; I, MWU test).

Figure 6.

Transplantation of donor Wnt5a^+/− HSCs in WT and Wnt5a^+/− recipient mice. (A) Experimental design. Donor BM cells from Wnt5a^+/− were transplanted into WT or Wnt5a^+/− recipient mice. LSK cells were sorted from primary recipients and transplanted further into 2° WT animals. (B) The engraftment of 1° and 2° recipients. (C) Lymphoid (dotted line) and myeloid engraftment (solid line) for Wnt5a^+/− into WT (HET/WT, 1°, n = 15; 2°, n = 6) and Wnt5a^+/− into Wnt5a^+/− (HET/HET,1°, n = 6; 2°, n = 4). (D) Expression of DOCK2, WAVE2, ACTB, and F-actin in 1°CD34⁻ LSKs recovered from both genotype recipients. (E) The total pixel number of immunofluorescent staining quantified from 1° CD34⁻ LSKs with ImageJ software. (F) Assessment of polarized expression of the indicated proteins using ImageJ. *, P < 0.05 (MWU test).

Figure 7.

Reduced Wnt5a in the niche affects development of BCR-ABL^p185 expressing cells. (A) Experimental design. The BCR-ABL^p185 fusion protein was stably inserted via retroviral infection into the genome of BM cells from mice pretreated with 5FU. Further, 10⁵ of GFP⁺ cells were transplanted into 1° recipients of both genotypes. The mice were sacrificed when disease was apparent. All healthy recipient mice were sacrificed 120 d after Tx. 10⁶ of 1° GFP⁺ spleen cells were transplanted into 2° WT recipients. (B) Survival curve of 1° recipients (WT, n = 12; Wnt5a^+/−, n = 28). (C) Representative pictures of spleens from 1°recipients. (D) The blood cell number and the spleens size of 1° recipient mice. (E) Survival curve of 2° recipients (these are the results of three independent experiments; 2° from both WT and Wnt5a^+/− recipients, n = 11). (F) Expression of DOCK2, CDC42, WAVE2, and F-actin protein in 1° GFP⁺ B220⁺ B cells, counterstained with DAPI. The graphs represent fold change of the corresponding protein expression. (G) The numbers of cells with apolar protein distribution in 1° GFP⁺ B220⁺ cells. (H) The relative immunofluorescent staining for DVL, GSK3β, phospho-β-catenin, and β-catenin in 1° GFP⁺B220⁺ cells. (I) Representative pictures of F-actin distribution in 1° GFP⁺ B220⁺ of both genotypes treated or not with 150 ng/ml rmCXCL12. (J) The 1° GFP⁺ B220⁺ VCAM adhered to VCAM-coated slides and stained with DAPI. The graph displays the total cell nuclei counted per field, at least 10 fields were counted. (K) Representative FACS plots of migrated 1° GFP⁺ B220⁺ spleen cells toward 150 ng/ml rmCXCL12 in the lower chamber of a Boyden chamber. The graphs show absolute number of migrated 1° GFP⁺ cells. Shown are the mean results of three independent experiments. *, P < 0.05 (D, G, I, and K, Student’s t test; F and H, MWU test).