Mesenchymal cell replacement corrects thymic hypoplasia in murine models of 22q11.2 deletion syndrome

Abstract

22q11.2 deletion syndrome (22q11.2DS) is the most common human chromosomal microdeletion, causing developmentally linked congenital malformations, thymic hypoplasia, hypoparathyroidism, and/or cardiac defects. Thymic hypoplasia leads to T cell lymphopenia, which most often results in mild SCID. Despite decades of research, the molecular underpinnings leading to thymic hypoplasia in 22q11.2DS remain unknown. Comparison of embryonic thymuses from mouse models of 22q11.2DS (Tbx1neo2/neo2) revealed proportions of mesenchymal, epithelial, and hematopoietic cell types similar to those of control thymuses. Yet, the small thymuses were growth restricted in fetal organ cultures. Replacement of Tbx1neo2/neo2 thymic mesenchymal cells with normal ones restored tissue growth. Comparative single-cell RNA-Seq of embryonic thymuses uncovered 17 distinct cell subsets, with transcriptome differences predominant in the 5 mesenchymal subsets from the Tbx1neo2/neo2 cell line. The transcripts affected included those for extracellular matrix proteins, consistent with the increased collagen deposition we observed in the small thymuses. Attenuating collagen cross-links with minoxidil restored thymic tissue expansion for hypoplastic lobes. In colony-forming assays, the Tbx1neo2/neo2-derived mesenchymal cells had reduced expansion potential, in contrast to the normal growth of thymic epithelial cells. These findings suggest that mesenchymal cells were causal to the small embryonic thymuses in the 22q11.2DS mouse models, which was correctable by substitution with normal mesenchyme.

Keywords: Embryonic development; Genetic diseases; Genetics; Immunology; T cell development.

Figure 1. Hypoplastic embryonic thymuses isolated from 22q11.2DS mouse models have normal proportions of thymocytes and TEC subsets.

(A–E) E18–18.5 embryonic thymuses were obtained from Tbx1^+/neo2 intercrossed mouse lines. (A) Live cell images from the cardiothoracic regions of Tbx1^+/+, Tbx1^+/neo2, and Tbx1^neo2/neo2 embryos. Thymic lobes are indicated with black arrows. An interrupted aortic arch (white arrow) often copresented with thymic hypoplasia in the Tbx1^neo2/neo2 embryos (Tbx1^neo2/neo2, e.g. 1). Scale bars: 50 μm. (B) Thymic tissue sections were processed for H&E staining and IHC. Top row: In the H&E-stained images, the cortical and medullary regions are dark and light purple, respectively. A medullary region is indicated by the boxed area. Scale bars: μm. Original magnification, ×10. Middle and bottom rows: With IHC, staining with antibodies selective for cortical (cytokeratin 8, red; middle row) and medullary (cytokeratin 14, green; bottom row) TECs is shown; DAPI staining revealed nuclei (blue; middle row). Coexpression of both cytokeratins (green and red) represents immature TECs. Original magnification, ×20 (middle and bottom rows). (C) T cell development was assessed by staining single-cell suspensions with antibodies selective for the CD4 and CD8 coreceptor proteins. The 4 thymocyte subsets are distinguished by electronic gating for the CD4^–CD8^– (DN), CD4⁺CD8⁺ (DP), and the CD4⁺CD8^– and CD4^–CD8⁺ (SP) subsets. (D) DN cells are further categorized by CD44 and CD25 cell-surface expression. This identifies the DN1 (CD44⁺CD25^–), DN2 (CD44⁺CD25⁺), DN3 (CD44^–CD25⁺), and DN4 (CD44^–CD25^–) subpopulations in the Tbx1^+/+, Tbx1^+/neo2, and Tbx1^neo2/neo2 thymuses. (E) The total cell number and percentages of DP thymocytes, DN4 subpopulation of DN thymocytes, and cTECs were compared among the Tbx1^+/+ (n = 7–10), Tbx1^+/neo2 (n = 10–15), and Tbx1^neo2/neo2 (n = 4–7) genotypes. Statistically significant differences among the 3 groups were determined by 1-way ANOVA (Brown-Forsythe and Welch tests).

Figure 2. Hypoplastic embryonic thymuses have proportions of mesenchymal cells and TECs similar to those of normal control tissues.

(A–G) E13–13.5 embryonic thymuses from Tbx1^+/neo2 intercrossed time pregnant mice were genotyped and analyzed by live cell imaging, IHC, and flow cytometry. (A) Live cell images reveal the size and location of the developing thymus in Tbx1^+/+, Tbx1^+/neo2, and Tbx1^neo2/neo2 embryos (demarcated with dotted lines). In the Tbx1^neo2/neo2 genotype, an IAA-B was routinely visualized (white arrow). Scale bars: 0.5 mm. (B) Transverse sections comprising the thymus region were processed for H&E staining. Black arrows point to the thymus, with trachea (Tr) and esophagus (Eso) locations shown. Scale bars: 50 μm. In the Tbx1^neo2/neo2 genotyped lines, the thymic lobes were not always in the same plane of the transverse section. (C) IHC was performed with antibodies selective for neural crest cell–derived mesenchymal cells, marking the thymus capsule (Pdgfrα, red) and thymus vasculature (Pdgfrβ, yellow), along with antibodies specific for thymic epithelial cells (EpCAM, green). Two examples of hypoplastic thymuses from Tbx1^neo2/neo2 embryos are shown (e.g. 1 and 2). Scale bars: 50 μm. (D and E) Flow cytometric analyses of single-cell suspensions revealed the percentages of (D) mesenchymal (Pdgfrα⁺) and epithelial (EpCam⁺) cells and (E) ETPs coexpressing CD117 (c-kit) and CD45. (F) Enumeration of the total number of thymic cells and the specific numbers of mesenchymal and epithelial cells from multiple Tbx1^+/+, Tbx1^+/neo2, and Tbx1^neo2/neo2 embryos. In addition, the ratio of mesenchymal and TECs (Mes/TECs) is shown. The Tbx1^+/+ (n = 17), Tbx1^+/neo2 (n = 32), and Tbx1^neo2/neo2 (n = 28) genotyped embryos were used to determine cell numbers. (G) Percentages of mesenchymal cells, epithelial cells, and ETPs in the same thymic tissues characterized in F. TEC and ETP percentages were determined from a smaller number of Tbx1^+/+ (n = 9–15), Tbx1^+/neo2 (n = 9–20), and Tbx1^neo2/neo2 (n = 6–17) mice. Statistically significant differences were established by 1-way ANOVA (Brown-Forsythe and Welch tests).

Figure 3. Hypoplastic fetal thymic lobes from 22q11.2DS mouse models have diminished thymopoiesis potential in culture.

(A) Paired normal-sized (Tbx1^+/+ or Tbx1^+/neo2) and hypoplastic (Tbx1^neo2/neo2) thymic lobes (E13–E13.5) were cultured for 4 days and 8 days. Live cell imaging revealed changes in thymus size, which were limited in the Tbx1^neo2/neo2 22q11.2DS mouse model. Scale bars: 1 mm. (B) T cell development was assessed by comparing the percentage of DN, DP, and SP thymocytes using electronic gating following antibody staining for surface CD4, CD8, and the TCR-β subunit expression. (C) The percentages of mesenchymal cells (Pdgfrα⁺) and TECs (EpCAM⁺) were determined after 4- and 8-day cultures via flow cytometric analyses. (D) After 4 and 8 days of FTOC, thymic lobes were processed, and total cell numbers along with the percentages of mesenchymal cells (Mes), TECs, and DP thymocytes were determined. Tbx1^+/+ (n = 8), Tbx1^+/neo2 (n = 14), and Tbx1^neo2/neo2 (n = 12) embryonic thymuses were used. (E) Eight days after FTOC, the total cell numbers and percentages of live cells, DP thymocytes, and TECs were determined. Note that by day 8, relatively few Pdgfrα⁺ cells remained due to the differentiation of these cells. Tbx1^+/+ (n = 4), Tbx1^+/neo2 (n = 6), and Tbx1^neo2/neo2 (n = 10) embryonic thymuses were used. Statistically significant differences were established by 1-way ANOVA (Brown-Forsythe and Welch tests).

Figure 4. Tissue expansion is restored for hypoplastic thymuses by replacement of Tbx1^neo2/neo2-derived mesenchymal cells with normal control cells.

(A) Depiction of RTOC using flow-sorted cells. Single-cell suspensions from E13–E13.5 fetal thymic lobes were prepared, and mesenchymal cells (Pdgfrα⁺), TECs (EpCam⁺) and the remaining unstained cells (Pdgfrα^–Epcam^–, which includes ETPs, other hematopoietic cells, and endothelial cells) were sorted by flow cytometry. These 3 subgroups were reaggregated at cell ratios established with control fetal thymuses and placed onto membranes and cultured. A minimum of 30,000 cells/aggregate was needed to sustain RTOC growth with normal cells (Supplemental Figure 5). The aggregates appear as a small dot in the yellow circled area. Endothelial cell replacements required sorting of CD31⁺ cells from the remaining cell subsets prior to reaggregate culturing. (B) Live cell imaging was used to visualize RTOCs after 10 days of culturing. The control corresponds to the 3 subgroups of cells from Tbx1^+/+;+/neo2 thymic lobes. In the first column, control thymuses were a combination of cells from either Tbx1^+/+ and/or Tbx1^+/neo2 embryos. In the second column, 22q11.2DS hypoplastic thymuses were from Tbx1^neo2/neo2 embryos. In the third column, normal mesenchymal cells were used as substitutes for those in the 22q11.2DS tissues (Sub Tbx1^neo2/neo2 Mes). In columns 4 and 5, normal TECs or endothelial cells were used as substitutes for Tbx1^neo2/neo2 TECs (Sub Tbx1^neo2/neo2 TECs) or endothelial cells (Sub Tbx1^neo2/neo2 Endo), respectively. Scale bars: 1 mm. (C) Cell viability (top row) and thymopoiesis (DN to DP and then SP progression, bottom row) are shown for the cells after 10 days of RTOC. (D) Cumulative cell numbers are shown for a representative RTOC experiment. (E–G) The fold increase in cell numbers following 10 days of RTOC along with cell viability and the percentage of DP cells developing over this period. n = 37, 28, 13, 8, and 5 experiments per group, respectively, for E–G. Statistical analyses done with 1-way ANOVA (Brown-Forsythe and Welch tests).

Figure 5. scRNA-Seq reveals distinct transcript levels in mesenchymal cells, TECs, and endothelial cells in embryonic thymuses from control, Tbx1^neo2/neo2, and Foxn1-mutant mouse models.

(A) Fetal thymuses, obtained from normal, Tbx1^neo2/neo2, and Foxn1^1089/1089 E13–E13.5 embryos, were used for scRNA-Seq. t-Distributed stochastic neighbor embedding (tSNE) plots reveal 17 distinct cell subgroups for all 3 paired thymic lobes (Tbx1^+/+, Tbx1^neo2/neo2, and Foxn1^1089/1089 genotypes), with the relative percentages of these subgroups differing among the 3 genotypes. Five distinct mesenchymal cell clusters (M-1 to M-5), 6 epithelial cell groups (E-1 to E-6), an endothelial cell population (En-1), 4 hematopoietic cell types (H-1 to H-4), a RBC (U-1), and a mitochondrial signature are present in each of the thymuses. The tSNE plot for the Foxn1 hypoplastic lobes (Foxn1^1089/1089) was generated by changing the total number cells to 6,000. (B) Transcripts that defined the cell subsets were compared among the 5 mesenchymal, 6 epithelial, and 4 hematopoietic cell clusters. A dot plot comparison revealed key gene expression differences among the various cell populations. (C) Heatmaps show the differential expression of transcripts of biological importance for mesenchymal and epithelial cell clusters along with the 1 endothelial cell cluster, respectively. Regions boxed in red represent the Tbx1^neo2/neo2 thymus. (D and E) Pathway enrichment analyses of mesenchymal (D) and endothelial (E) clusters with DEGs revealed key distinctions between control, Tbx1^neo2/neo2, and Foxn1^1089/1089 fetal thymuses.

Figure 6. Elevated deposition of collagen is evident in hypoplastic thymuses from Tbx1^neo2/neo2 embryos.

(A) E13–13.5 thymuses from the indicated embryos were prepared for IHC. Staining was done with antibodies detecting collagen I (green) and a combination of CD31 and endomucin (red). Sections were prepared from Tbx1^+/+, Tbx1^+/neo2, and Tbx1^neo2/neo2 embryonic thymuses. Two different Tbx1^neo2/neo2 embryos are shown for comparative purposes. The merged image combines the collagen, CD31/endomucin, and DAPI staining (nuclei) stains. Blue arrow points to the thymus. Scale bars: 50 μm. (B) IHC was performed on embryos from mice of the Tbx1^+/+, Tbx1^+/neo2, and Tbx1^neo2/neo2 genotypes. Antibodies selective for Cspg4 (green), Mcam (purple), and CD31/endomucin are independently shown along with a merged image comprising all the stains. The blue arrow indicates the thymus, the yellow arrow the carotid artery, and the light gray arrowhead the vagal trunk. Scale bars: 50 μm.

Figure 7. The presence of minoxidil in RTOC cultures restores tissue growth for hypoplastic thymuses.

RTOC assays were performed using cell suspensions generated from E13–E13.5 fetal thymic lobes. Cells from either normal or Tbx1^neo2/neo2 thymuses were reaggregated with equivalent starting clusters of approximately 30,000 cells/group. Cultures were maintained in media alone or supplemented with 3 μM minoxidil. (A) Live cell imaging revealed cell expansion after 10 days of culturing. Scale bars: 1 mm. (B) Thymopoiesis was compared using antibodies specific for CD4 and CD8. (C) Cell numbers, cell viability, and the percentage of DP cells are shown. Note that the number of cells in Tbx1^neo2/neo2 thymuses was severely limited, as established in Figure 4, B, D, and E. n = 10, 10, 3, and 3 for the indicated groups, from left to right, in each panel. Statistical significance was determined by 1-way ANOVA. (D) Control FTOCs were grown in the absence or presence of minoxidil. On day 3 and day 4 after culturing, the cells were processed for qRT-PCR using probes detecting 2 Plod and 2 Col1a genes, along with GAPDH for normalization. Day 3, n = 5; day 4, n = 4. (E) Mesenchymal cells and TECs from E13–E13.5 embryonic thymuses from Tbx1^+/+ or Tbx1^neo2/neo2 embryos were flow sorted. Mesenchymal sorted cells were grown in MesenCult differentiation media. After 15 days of culturing, the cells were fixed, and live cell images were obtained. The well was from a 6-well tissue culture plate. Bottom image: A representative cluster of cells was imaged following crystal violet staining. Scale bars: 1 mm. (F) Total number of pixels in the images in E in conjunction with 5 additional independent experiments were calculated. These values were divided by the total number of mesenchymal cells seeded in each experiment and plotted as pixel area divided by the total cell number. This was compared with TECs grown in EpiCult. These cells were enumerated by cell counting, as shown. Statistical significance was determined by Student’s t test.