Cell Growth Dynamics in Embryonic and Adult Mouse Thyroid Revealed by a Novel Approach to Detect Thyroid Gland Subpopulations

Abstract

Background:

The thyroid is composed of endocrine epithelial cells, blood vessels, and mesenchyme. However, no data exist thus far on absolute cell numbers, relative distribution, and proliferation of the different cell populations in the developing and mature thyroid. The aim of this study was therefore to establish a flow cytometry protocol that allows detection and quantification of discrete cell populations in embryonic and adult murine thyroid tissues.

Methods:

Cell-type anti-mouse specific antibodies were used for erythroid cells (Ter119), hematopoietic cells (CD45), epithelial cells (EpCam/CD326, E-cadherin/CD324), thyroid follicular cells and C-cells (Nkx2-1), endothelial cells (Pecam/CD31, Icam-1/CD54), and fibroblasts (PDGFRa/CD140a). Proliferating cells were detected after labeling with 5-bromo-2′-deoxyuridine (BrdU). For flow cytometry analyses, micro-dissected embryonic (E) and adult thyroids were pooled (E13.5, n = 25; E15.5, n = 15; E17.5, n = 15; adult, n = 4) in one sample.

Results:

The absolute parenchymal cell numbers per mouse thyroid (M ± SD), excluding the large number of CD45⁺ and Ter119⁺ cells, increased from 7425 ± 1338 at E13.5 to 271,561 ± 22,325 in adult tissues. As expected, Nkx2-1⁺ cells represented the largest cell population in adult tissues (61.2 ± 1.1%). Surprisingly, at all three embryonic stages analyzed, thyroid follicular cells and C-cells accounted only for a small percentage of the total thyroid cell mass (between 4.7 ± 0.4% and 9.4 ± 1.6%). In contrast, the largest cell population at all three embryonic stages was identified as PDGFRa/CD140a⁺ fibroblasts (61.4 ± 0.4% to 77.3 ± 1.1%). However, these cells represented the smallest population in adult tissues (5.2 ± 0.8%). Pecam/CD31⁺ endothelial cells increased from E13.5 to E15.5 from 3.7 ± 0.8% to 8.5 ± 3.0%, then remained stable at E17.5 and adult tissues. Proliferation rates were sizable during the entire organogenesis but differed between cell populations, with distinct proliferative peaks at E13.5 in epithelial cells (32.7 ± 0.6% BrdU⁺ cells), and at E15.5 in endothelial cells (22.4 ± 2.4% BrdU⁺ cells). Fibroblasts showed a constant proliferation rate in embryonic tissues. In adult tissues, BrdU⁺ cells were between 0.1% and 0.4% in all cell types.

Conclusions:

Using a novel flow cytometry–based method, a previously unobserved highly dynamic growth pattern of thyroid cell populations during embryogenesis was uncovered. This approach will provide a useful new tool for cell function analyses in murine thyroid disease models.

Introduction

The thyroid gland is an endocrine organ composed of different cell types (1 –3). The endocrine cell mass consists mainly of thyroid follicular cells (TFC), which are organized in the functional unit of the thyroid—the thyroid follicle. TFCs produce the thyroid hormones thyroxine (T4) and triiodothyronine (T3) (1 –3). The second endocrine cell type is parafollicular cells (C-cells), which represent a small population of the entire cell mass and secrete calcitonin (1 –4). The thyroid also contains blood vessels, which form a dense capillary net around the thyroid follicles. This structure ensures close contact with the basal membrane of the TFCs in order to provide optimal uptake of secreted thyroid hormones into the blood, the so-called angiofollicular unit (5). Finally, fibroblasts are interspersed between vessels and follicles (4).

At present, however, very few data are available with regard to the precise cellular composition and cell growth dynamics of the developing thyroid due to lack of TFC-specific surface antigens and protocols for nuclear staining with TFC-specific transcription factors such as Nkx2-1. One previous study quantified the different cell populations in the adult dog thyroid before flow cytometry was available by transmission electron microscopy (6). In that study, 70% of thyroid cells were epithelial cells, 6% endothelial cells, and 24% fibroblasts (6). However, these results were never confirmed in other species.

Flow cytometry allows for a precise quantification of relative distributions of different cell types during various stages of differentiation. Using human thyroid tissues, endothelial cells were studied by flow cytometry using factor-VII related antigen (7). However, other thyroid cells were not analyzed in that study. In another study, thyroglobulin and calcitonin antibodies were used to quantify TFC and C-cell mass in adult rat thyroids. This study showed that 40% of the cells were thyroglobulin positive, thus representing TFCs, and 4% were calcitonin positive, representing C-cells (8).

To make more precise data available on thyroid cell populations both in adults and during thyroid development, a flow cytometry protocol was developed to quantify the main cell populations in the murine thyroid gland and to provide data on the developmental growth dynamics of the embryonic thyroid gland at different consecutive stages for the first time. This protocol provides a new tool for future in-depth analyses of murine thyroid disease models.

Materials and Methods

Mice

Embryos from Swiss albino mice at embryonic day (E) 13.5, 15.5, and 17.5, and adult female pregnant mice (8 weeks old) were obtained from Janvier Labs. All experiments were performed in accordance with federal regulations.

In vivo labeling of mice with bromodeoxyuridine

Mice were injected intraperitoneally with 2 mg 5-bromo-2′-deoxyuridine (BrdU; BD Biosciences; 1 mg/100 μL of phosphate-buffered saline [PBS] per mouse) 2 h prior to tissue isolation.

Cell suspension preparation

Embryonic thyroids from E13.5, E15.5, and E17.5 embryos and thyroids from adult female pregnant mice were micro-dissected, cleaned of fat and connective tissue, and placed in ice-cold PBS containing 2% fetal calf serum (FCS). Only entire bilobed thyroid glands with intact isthmus were used for experiments. Each experiment used pooled tissues to obtain sufficient cells for analysis: E13.5, n = 25 pooled thyroids per sample; E15.5, n = 15 pooled thyroids per sample; E17.5, n = 15 pooled thyroids per sample; adult tissues, n = 4 thyroids per sample. To reduce experimental variation, all embryonic and adult thyroid tissues were micro-dissected on the same experimental day, and all four stages were analyzed within the same experiment on the same day. Thyroids were enzymatically digested with 1 mg/mL collagenase/dispase and 2 μg/mL DNase I (Roche Diagnostics) at 37°C for 20 min until single-cell suspension was obtained. Adult thyroids were first cut into small pieces using sterile scissors before enzymatic digestion. After obtaining the single-cell suspension, cells were centrifuged with PBS containing 2% FCS for 5 min at 453 g. Cells were then stained with cell surface markers for 20 min on ice. Prior to intracellular staining, cells were fixed and permeabilized using the BD Cytofix/Cytoperm^™ Fixation/Permeabilization solution kit (BD Biosciences), according to the manufacturer's guidelines. Cells were then exposed to 10% goat serum for 45 min at room temperature. Intracellular staining with anti-Nkx2-1 antibody was done for 45 min at room temperature followed by secondary antibody for 30 min at room temperature. Non-nucleated cells were excluded by staining with DAPI; (4′,6-Diamidine-2′-phenylindole dihydrochloride; Roche Diagnostics). To analyze cell proliferation, cells were also treated with 30 μg of DNase I (BD Biosciences) followed by anti-BrdU staining and 7-AAD (7-Aminoactinomycin D; BD Biosciences) after the fixation and permeabilization step.

Flow cytometry analysis

Cells were acquired on BD FACSAria II flow cytometer (Becton Dickinson). FlowJo software (Tree Star, Inc.) was used for analyzing flow cytometry data. The following monoclonal antibodies were used: E-cadherin/CD324 (ECCD-2) from Invitrogen; EpCam/CD326 (G8.8), PDGFRa/CD140a (APA5), CD45 (30-F11), Ter119 (TER-119), Icam-1/CD54 (YN1/1.7.4), and Pecam/CD31 (390) from Biolegend; and Nkx2-1 (EP1584Y) from Abcam. As secondary antibody, goat anti-rabbit A647 was obtained from Life Technologies.

To calculate the absolute cell numbers, all events acquired were first gated to remove cell debris by using forward-scattered light (FSC) and side-scattered light (SSC; Fig. 1A). Then, cell doublets were removed by using SSC-height (SSC-H) and SSC-width (SSC-W). Organ cellularity was determined with a Z1 Coulter Counter (Beckman Coulter). The absolute cell number of each cell type was calculated by multiplying organ cellularity with the frequency of individual cell populations determined by flow cytometry.

FIG. 1.

Gating strategy for flow cytometric analysis of thyroid glands. (A) Representative FACS plot of pooled and fixed E15.5 thyroids demonstrating the gating strategy used for the identification of thyroid cell populations. CD (Cluster of differentiation) numbers are indicated for each antibody, when appropriate, following the original name (e.g., Pecam/CD31). Arrows indicate which gated population is analyzed in the next panel. Top-left panel: Scatter plot: SSC-A (side-scattered area) and FSC-A (forward-scattered area) gate. All events outside this gate are debris. Top-middle panel: Doublet discrimination: SSC-H (side-scattered height) and SSC-W (side-scattered width) gate, defining singlets. All events outside this gate are doublets or aggregates. Top-right panel: The right gate represents events that are Nkx2-1⁺ thyroid follicular cells and C-cells, the upper-left gate represents events that are CD45⁺ lymphocytes, and the lower-left gate represents remaining cell populations, that were further analyzed in the next panel. Bottom-right panel: The right gate represents events that are Pecam⁺ endothelial cells, the upper-left gate represents events that are PDGFRa⁺ fibroblasts, and the lower-left gate represents a remaining cell population, that were further analyzed in the next panel. Bottom-middle panel: The right gate represents events that are EpCam⁺ epithelial cells, which were Nkx2-1^– in the top right panel, thus representing parathyroid cells; the upper-left gate represents events that are Ter119⁺ erythrocytes, and the lower left gate represents the remaining unknown cell population(s) that is/are characterized as Nkx2-1/CD45^–/Pecam^–/PDGFRa^–/EpCam^–/Ter119^–. (B) Validation of cell types by co-labeling with two different antibodies specific for the same cell population. Left panel: The top-right gate represents events that are co-labeled with Pecam and Icam-1, representing endothelial cells. Events outside the gate represent all other cell populations. Middle panel: The top-right gate represents events that are EpCam⁺/E-cadherin⁺ epithelial cells. All events outside the gate represent all other cell populations. Right panel: The top-right gate represents events that are EpCam⁺/Nkx2-1⁺ thyroid follicular cells and C-cells; the top-left gate represents events that are Epam⁺/Nkx2-1^– parathyroid cells. All events outside the gate represent all other cell populations.

Statistical analysis

All values are given as M ± SD. The data represent three independent experiments. For multiple group comparisons, one-way analysis of variance with Tukey's multiple comparison post-test was performed using GraphPad Prism v5 for Mac OS X (GraphPad Software; www.graphpad.com). p-Values of <0.05 were considered statistically significant.

Results

Establishment of a combination of specific antibodies suitable for the evaluation of the thyroid gland

First, a protocol of intranuclear staining with Nkx2-1, a transcription factor expressed in TFCs and C-cells (9), was established. Using this approach, it was possible to identify the TFC/C-cell population (Fig. 1A and B). Second, the use of established surface markers to identify epithelial cells (EpCam/CD326⁺, E-cadherin/CD324⁺), endothelial cells (Pecam/CD31⁺, Icam-1/CD54⁺), fibroblasts (PDGFRa/CD140a⁺), erythroid cells (Ter119⁺), and hematopoietic cells (CD45⁺) in the thyroid tissue was validated (Fig. 1A and B). The final gating strategy consisted of the combined use of Nkx2-1, CD45, Pecam, PDGFRa, EpCam, and Ter119 (Fig. 1A). Nkx2-1^–/EpCam⁺ cells represent parathyroid cells (Fig. 1A and B). The used gating strategy allowed characterizing between 82% and 92% of cells in the embryonic thyroid and 93% of cells in the adult murine thyroid gland (Fig. 1A and Table 1). However, a population ranging from 8% to 20% of cells remained negative for all six antibodies in the different embryonic and adult tissues.

Table 1.

Relative Frequencies of Cell Populations in the Developing and Adult Thyroid Organ After Exclusion of Erythyroid (Ter119 ⁺ ) and Hematopoietic Cells (CD45 ⁺ )

	E13.5	E15.5	E17.5	Adult
Nkx2-1⁺ cells: thyroid follicular cells and C-cells	9.4 ± 1.6%	4.7 ± 0.4%	8.7 ± 0.9%	61.2 ± 1.1%
Nkx2-1^–/EpCam⁺ cells: parathyroid cells	1.5 ± 0.4%	0.8 ± 0.2%	2.5 ± 0.6%	17.8 ± 3.9%
Pecam⁺ cells: endothelial cells	3.7 ± 0.8%	8.5 ± 3.0%	8.5 ± 0.2%	8.4 ± 1.1%
PDGFRa⁺ cells: fibroblasts	77.3 ± 1.1%	66.0 ± 2.5%	61.4 ± 0.4%	5.2 ± 0.8%
All^– unknown cells: (Nkx2-1^–/CD45^–/Pecam^–/PDGFRa^–/EpCam^–/Ter119^–)	8.2 ± 0.9%	20.0 ± 0.7%	18.9 ± 1.0%	7.5 ± 2.9%

Data represent results from three independent experiments (M ± SD) with pooled tissues (n = 25 bilobed thyroids at E13.5, n = 15 bilobed thyroids at E15.5 and E17.5, and n = 4 thyroids in adult samples for each experiment).

E, Embryonic day.

Assessment of absolute thyroid population cell numbers

The mean absolute total cell number (Table 2 and Fig. 2) of one thyroid gland was 7425 ± 1338, 18,444 ± 1259, 31,804 ± 2610, and 271,561 ± 22,325 cells at E13.5, E15.5, and E17.5, and in adult tissues, respectively. Different growth patterns were observed in endocrine cells, endothelial cells, and fibroblasts. First, Nkx2-1⁺ TFC/C-cell and Nkx2-1^–/EpCam⁺ parathyroid cell numbers were 684 ± 91, and 109 ± 33 cells/thyroid at E13.5, respectively, and increased 243-fold and 438-fold at the adult stage (166,208 ± 17,334 and 47,598 ± 7008 cells/thyroid). Second, the cell number of Pecam⁺ endothelial cells was 274 ± 75 at E13.5 and reached 22,813 ± 4027 in adult tissues (83-fold increase with a peak between E13.5 and E15.5 during angiogenesis of the embryonic thyroid). Lastly, PDGFRa⁺ fibroblasts showed highest cell numbers in embryonic tissues, although only with a slight increase from E13.5 to adult stage (2.0×), and surprisingly with a decrease by 28% between E17.5 and the adult stage. Nkx2-1^–/CD45^–/Pecam^–/PDGFRa^–/EpCam^–/Ter119^– cells showed a 34-fold increase from 609 ± 154 cells/thyroid at E13.5 to 20,837 ± 9,899 cells/thyroid in the adult. The absolute cell numbers of each cell type are summarized in Table 2 and Figure 2.

FIG. 2.

Comparison of absolute cell numbers of the whole thyroid and of each subpopulation at embryonic stages and in adult tissues. One thyroid; data are representative for the whole cell mass of one entire bilobed thyroid gland after exclusion of Ter119⁺ erythroid cells and CD45⁺ hematopoietic cells; Nkx2-1⁺ cells, thyroid follicular cells, and C-cells; Nkx2-1^–/EpCam⁺ cells, parathyroid cells; Pecam⁺ cells, endothelial cells; PDGFRa⁺ cells, fibroblasts; Nkx2-1^–/CD45^–/Pecam^–/ PDGFRa^–/EpCam^–/Ter119^– cells, unknown cell type. Data represent results from three independent experiments (M ± SD) with pools of n = 25 bilobed thyroids at E13.5, n = 15 bilobed thyroids at E15.5 and E17.5, and n = 4 thyroids in adult samples for each experiment.

Table 2.

Absolute Cell Numbers of the Different Cell Populations in the Developing and Adult Thyroid Organ

Cell populations	E13.5	E15.5	E17.5	Adult
Nkx2-1⁺ cells: thyroid follicular cells and C-cells	684 ± 91	865 ± 28	2746 ± 290	166,208 ± 17,334
Nkx2-1^–/EpCAM⁺ cells: parathyroid cells	109 ± 33	152 ± 27	784 ± 138	47,598 ± 7008
Pecam⁺ cells: endothelial cells	274 ± 75	1582 ± 663	2704 ± 238	22,813 ± 4027
PDGFRa⁺ cells: fibroblasts	5748 ± 1107	12,151 ± 453	19,541 ± 1744	14,104 ± 2346
All^– unknown cells: (Nkx2-1^–/CD45^–/Pecam^–/PDGFRa^–/EpCam^–/Ter119^–)	609 ± 154	3694 ± 291	6028 ± 670	20,837 ± 9899
CD45⁺ cells: hematopoietic cells	329 ± 58	2518 ± 314	6353 ± 442	31,330 ± 12,808
Ter119⁺ cells: erythroid cells	28 ± 14	156 ± 29	523 ± 100	1442 ± 500
Total cell number of one thyroid gland	7782 ± 1395	21,118 ± 983	38,681 ± 3093	304,333 ± 24,732
Total cell number of one thyroid gland without Ter119⁺ and CD45⁺ cells	7425 ± 1338	18,444 ± 1259	31,804 ± 2610	271,561 ± 22,325

Relative frequencies of cell types

For calculation of relative frequencies of parenchymal cell populations, the number of Ter119⁺ erythroid cells and CD45⁺ hematopoietic cells were subtracted from the absolute total cell number of one thyroid gland. TFCs and C-cells (Nkx2-1⁺ cells) were surprisingly low in relative frequency (4.7–9.4%) in embryonic tissues (Table 1, Fig. 3). In contrast, in adult tissues, they represented as expected the predominant population (61.2% of all cells). Parathyroid cells (Nkx2-1^–/EpCam⁺ cells) were low (0.8–3.0%) in embryonic tissues, but increased to 17.8% in adult thyroid tissues. Endothelial cells (Pecam⁺) doubled their frequency between E13.5 and E15.5 from 3.7% to 8.5% and remained at 8.5%, and 8.4% at E17.5 and adult stage, respectively. Fibroblasts (PDGFRa⁺) were the predominant cell type in embryonic thyroid glands (61–77%), while they formed the smallest cell population in adult tissues with 5.2% (Figs. 3 and 4A). Finally, the cell population negative for all used antibodies (Nkx2-1^–/CD45^–/Pecam^–/PDGFRa^–/EpCam^–/Ter119^–) increased from 8% at E13.5 to 20% and 19% at E15.5, and E17.5, respectively, but was low in adult tissues (8%). The origin of this cell population remains to be characterized.

FIG. 3.

Relative frequencies of cell populations characterized by flow cytometry in embryonic and adult thyroid tissues. Relative frequencies were calculated after exclusion of erythroid (Ter119⁺) and hematopoietic cells (CD45⁺). Results represent data from three independent experiments (M ± SD) with pooled thyroid tissues (n = 25 bilobed thyroids at E13.5, n = 15 bilobed thyroids at E15.5 and E17.5, and n = 4 thyroids in adult samples for each experiment).

FIG. 4.

Cell cycle analysis at all stages in cell populations. (A) Analysis of BrdU incorporation 2 h after intraperitoneal injection of pregnant mice. Flow cytometry plots depict intracellular expression of BrdU and 7-AAD (7-Aminoactinomycin D), which marks the 2n or 4n DNA. (B) The numbers represent frequencies of cells in G0/1 phase (BrdU^–/7-AAD²ⁿ), S-phase (BrdU⁺/7-AAD^2n/4n), and G₂/M phase (BrdU^–/7AAD⁴ⁿ). Flow cytometry plots depict one representative experiment. The frequencies represent data from three independent experiments (M ± SD) with pooled thyroid tissues (n = 25 bilobed thyroids at E13.5, n = 15 bilobed thyroids at E15.5 and E17.5, and n = 4 thyroids in adult samples for each experiment).

Cell cycle analysis of discrete thyroid cell types

The different phases of the cell cycle in the identified cell populations were assessed by BrdU and 7-AAD labeling. Cells in G₀/G₁ phase were BrdU^–/7-AAD²ⁿ, cells in S-phase were BrdU⁺/7-AAD^2n/4n, and cells in G₂/M phase were BrdU^–/7-AAD⁴ⁿ (Fig. 4A). Embryonic stage-specific frequencies of cells in S-phase were significantly different between E13.5-E15.5-E17.5 tissues for EpCam⁺ epithelial cells (p < 0.01) and Pecam⁺ endothelial cells (p < 0.05), while in PDGFRa⁺ fibroblasts, values were not statistically different in embryonic stages (p = 0.07). In contrast to epithelial cells, endothelial cells, and fibroblasts, the unknown cell population (Nkx2-1^–/CD45^–/Pecam^–/PDGFRa^–/EpCam^–/Ter119^–) harbored markedly stable frequencies of BrdU⁺ cells between E13.5 and E17.5 (p = 0.6). Hence, significant differences of the relative frequencies of cells in S-phase were observed between the different cell types at each embryonic stage (E13.5, p < 0.01; E15.5, p < 0.05; E17.5, p < 0.05), while the BrdU-positive cell frequency did not differ in adult tissues between cell types (p = 0.55).

Frequencies of proliferating cells among EpCam⁺ epithelial cells were highest at E13.5 (32.7 ± 0.6%). The fraction of proliferating cells then decreased to 13.6 ± 4.1% at E15.5 and to 6.7 ± 1.5% at E17.5 (Figs. 4B and 5B). In adult thyroid glands, the frequency of BrdU⁺ cells was 0.1 ± 0.2%. In contrast, the fraction of proliferating cells among Pecam⁺ endothelial cells increased in their proliferation from 17.2 ± 5.8% (E13.5) to 22.4 ± 2.4% at E15.5. Late in embryonic development (E17.5), a fraction (1.1 ± 0.4%) of Pecam⁺/BrdU⁺ double-positive cells was present. In adult tissues, the proliferation rate of Pecam⁺ cells was 0.1 ± 0.1%. Finally, PDGFRa⁺ fibroblasts showed stable proliferation rates at E13.5 and E15.5 (9.0 ± 1.8%, and 8.4 ± 2.1%, respectively; Figs. 4B and 5B). They decreased to 3.0 ± 0.6% at E17.5, and to 0.4 ± 0.3% in adult thyroids. The cell population negative for all used surface and intracellular markers (Nkx2-1^–/CD45^–/Pecam^–/PDGFRa^–/EpCam^–/Ter119^–) showed proliferation rates of 14.8 ± 1.9% at E13.5, 16.8 ± 5.5% at E15.5, and 11.4 ± 9.1% at E17.5, and 0.0 ± 0.1% in adult tissues.

FIG. 5.

Summary of cell growth dynamics of the three main populations of the thyroid gland during late embryonic stages and in adult tissues. (A) Representation of the fundamental switch from the fibroblast to the TFC/C-cell population between embryonic and adult stage. (B) Cell type–specific distinct proliferative waves during embryonic development.

Thus, our data indicate distinct cell population–specific proliferative waves during thyroid development, with a peak proliferation at E13.5 in the epithelial population and at E15.5 in endothelial cells. In contrast, fibroblasts showed more stable proliferation rates during thyroid differentiation without clear peaks, and the unknown cell population is characterized by a markedly stable but high frequency of BrdU⁺ cells in embryonic tissues.

Discussion

Improved tools for suitable cell analysis are required to obtain more in-depth knowledge with regard to thyroid development and disease. Flow cytometry has not been a standard protocol to date for thyroid research. The use of flow cytometry in this study made it possible to quantify absolute cell numbers, relative frequencies, and the proliferation rate of endocrine epithelial cells, endothelial cells, and fibroblasts in embryonic and adult murine thyroid tissues (summarized in Fig. 5A and B). The protocol provides a new tool for research in thyroid development and disease models in the mouse.

Previously, sorting of human thyroid endothelial cells and TFC precursors expressing green fluorescent protein (GFP) in vitro had been reported (7,10). Both protocols did not allow analysis of the in vivo frequencies and cell growth dynamics of cell subpopulations in the developing thyroid. One study quantified TFCs in adult rat thyroids in vivo by flow cytometry (8). These investigators used intracellular thyroglobulin (TG) and calcitonin staining to identify TFCs and C-cells. They observed 40% TG-positive TFCs, and 2–4% calcitonin-positive C-cells. The remaining cells were not characterized. In the present study, the frequency of TFCs in adult tissues was higher (61%; see Table 2 and Fig. 2), which is likely explained by two technical differences. First, the exclusively intranuclearly expressed TFC and C-cell transcription factor Nkx2-1 was used in the current study instead of TG to characterize TFCs. TG can be detected in the cytoplasma of TFCs, especially in embryonic tissues (11 –13). However, synthesized TG is excreted and accumulates within the follicular lumen and might be a less specific marker for TFCs than Nkx2-1 is. Second, Moerch et al. used a much longer dissociation protocol of thyroid tissue compared with the current protocol, possibly leading to a considerable loss of TFCs by dissociation (8).

Using morphological analysis of adult dog thyroids by transmission electron microscopy, Dow et al. quantified the different cell populations in the adult dog thyroid and found a frequency of 70% for TFCs, 26% of fibroblasts, and 6% of endothelial cells (6). These results were published in 1986, and are the only data characterizing all three main populations of the thyroid gland so far. The present results in the adult mouse thyroid are in accordance with these data concerning the percentage of TFCs (Nkx2-1-positive cells in the present study; 61% vs. 70% in the dog) and endothelial cells (8% vs. 6% in the dog), since these two cell types can be clearly differentiated from interstitial cells by morphological criteria. The relative abundance of fibroblasts was clearly lower in the current study (Fig. 3). It is hypothesized that by morphological analysis, all interstitial cells were classified as fibroblasts, as morphological analysis might not be specific enough to differentiate between fibroblasts, C-cells, and the population of unknown cells detected by flow cytometry in this study. Interspecies differences might be a further reason for differences with published data.

Quantitative data on the relative distribution in embryonic thyroids were not available thus far. A surprisingly low frequency of Nkx2-1-positive TFC/C-cells was found at all three embryonic stages analyzed, ranging from 4.7% to 9.4%, in contrast to their high abundance (61%) in adult tissues (Fig. 3 and Fig. 4A). These results in the embryonic tissues could have been limited by the gating strategy, only using Nkx2-1 and not adding further specific TFC markers such as Pax8. However, the epithelial origin and the low relative frequency of Nkx2-1⁺ cells was validated by double staining with EpCam, an epithelial surface marker known to be expressed on embryonic and adult TFCs (Fig. 1B). It is therefore thought that the strategy identified bona fide TFCs.

Importantly, the present data indicate a switch of relative tissue composition between embryonic and adult tissues. The high percentage of mesenchyme at E13.5, E15.5, and E17.5 is in accordance with immunohistochemical data in the embryonic mouse as well as human thyroid tissue (11,13,14). In both species, Nkx2-1-positive TFCs form strands of epithelial cells embedded in or surrounded by mesenchyme. This close contact of mesenchyme and epithelial cells during development is in accordance with data from other endoderm-derived organs, for example the lung and the pancreas (15,16). The observation that mouse models with deletion of mesenchymal genes (e.g. Fgf10^–/– , or Shh^–/– ) suffer from thyroid dysgenesis further supports the hypothesis of a functional interaction between mesenchyme and epithelial cells (17,18). Finally, thyroidal fibroblasts showed growth-promoting effects on a thyroid carcinoma cell line but not on differentiated FRTL-5 cells injected intrathyroidally, providing further arguments for epithelial–mesenchymal interactions in an undifferentiated state (19).

The mechanisms by which the Nkx2-1-positive TFC/C-cell population becomes so predominant in the adult tissue remain unclear. The important switch between fibroblast and TFC populations (60-fold increase between E17.5 to adult) occurred after E17.5, the latest stage that was analyzed. For a better understanding of developmental processes, late embryonic, neonatal, and postnatal stages until the adult age need to be analyzed by the same approach. One key observation can be made between the E17.5 and the adult thyroid: the fibroblast population is the only cell population showing a significant decrease not only in relative frequency (61% to 4%) but also in absolute cell numbers (–28%; Table 2 and Fig. 2). Whether this reduction in fibroblast/mesenchymal cell numbers could be explained by apoptosis or by processes of cell fate reprogramming or inter-lineage transdifferentiation, as described recently for cardiac endothelial cells, Schwann cells, or hepatocytes, remains speculative (20 –23).

In addition to phenotype data, this work provides information about cell type-specific proliferation rates. Distinct cell type–specific proliferative waves were observed during thyroid development, with a highest and earliest peak in epithelial cells at E13.5. A second but lower peak was observed in endothelial cells at E15.5. In contrast, mesenchymal cells showed a rather stable proliferation rate throughout development (Fig. 4B and 5B). So far, few data on proliferation of Nkx2-1⁺ cells during thyroid development are available (14,24). The present flow cytometry results (Fig. 4B and 5B) are in general accordance with BrdU immunohistochemistry, revealing a high proliferation rate during thyroid morphogenesis as reported previously (E11.5: 23%/40%, E13.5: 40%, E15.4: 32%/32%) (14,24).

The observed proliferative wave of endothelial cells at E15.5 is in line with the strictly descriptive data without BrdU labeling of peak angiogenesis at E15.5 during folliculogenesis in the murine thyroid (14,25). Finally, the very low turnover rate in adult thyroid tissues in this study for all cell types is in accordance with the landmark data by Dumont's group in dog and human thyroids, suggesting about five renewals per lifetime (4,26).

It was initially hypothesized that the thyroid is composed of three major cell populations: endocrine epithelial cells, endothelial cells, and mesenchymal cells. However, even after exclusion of circulating Ter119⁺ erythroid cells and CD45⁺ hematopoietic cells, a cell population ranging from 8% in adult tissues to 20% in embryonic tissues remained negative for all antibodies used in the protocol. Although the gating strategy does not use a second TFC-specific marker, the negative staining for EpCam and for Nkx2-1 argues against an epithelial or even TFC origin of these cells. Further, this cell population showed a clearly distinct proliferation pattern (with constantly high levels of BrdU⁺ cells) compared with TFCs, endothelial cells, and fibroblasts, suggesting not being synchronized with the fundamental processes of angiogenesis and folliculogenesis of the differentiating thyroid.

The presence of stem cells in the adult human and murine thyroid at very low frequency (0.3–1.4%) has been suggested recently (27 –30). This percentage is clearly lower than the relative distribution of the uncharacterized cell population in the current experiments (8%; Fig. 3). It therefore remains to be determined whether (i) the Nkx2-1^–/CD45^–/Pecam^–/PDGFRa^–/EpCam^–/Ter119^– cell population represents a homogenous population or several yet unidentified cell populations, and (ii) they would stain positive for stem-cell markers.

In conclusion, a novel validated protocol has been provided for flow cytometric analysis of the different cell populations of the thyroid gland. These results add to current quantitative data on in vivo frequencies and cell growth dynamics during normal murine thyroid development. Finally, a relevant cell population besides epithelial, endothelial, and mesenchymal cells has been identified that needs to be further characterized. This approach will provide a useful tool for cell function analysis in murine thyroid disease models.

Footnotes

Acknowledgments

The project was supported by a research grant from the Swiss National Science Foundation (31003A-130458), and a grant from the Gedächtnisstiftung Suzy Rückert zur Krebsbekämpfung.

Author Disclosure Statement

No competing financial interests exist.

References

De Felice

, Di Lauro

. 2003. Thyroid development and its disorders: genetics and molecular mechanisms. Endocr Rev, 25:722–746.

Nilsson

, Fagman

. 2013. Mechanisms of thyroid development and dysgenesis: an analysis based on developmental stages and concurrent embryonic anatomy. Curr Top Dev Biol, 106:123–170.

Szinnai

. 2014. Genetics of normal and abnormal thyroid development in humans. Best Pract Res Clin Endocrinol Metab, 28:133–150.

Dumont

, Lamy

, Roger

, Maenhaut

. 1992. Physiological and pathological regulation of thyroid cell proliferation and differentiation by thyrotropin and other factors. Physiol Rev, 72:667–697.

Colin

, Denef

, Lengelé

, Many

, Gérard

. 2013. Recent insights into the cell biology of thyroid angiofollicular units. Endocr Rev, 34:209–238.

Dow

, Dumont

, Ketelbant

. 1986. Percentage of epithelial cells, fibroblasts and endothelial cells in the dog thyroid. C R Seances Soc Biol Fil, 180:629–632.

Patel

, Logan

, Watkinson

, Uz-Zaman

, Sheppard

, Ramsden

, Eggo

. 2003. Isolation and characterization of human thyroid endothelial cells. Am J Physiol Endocrinol Metab, 284:E168–176.

Moerch

, Nielsen

, Lundsgaard

, Oleksiewicz

. 2007. Flow sorting from organ material by intracellular markers. Cytometry A, 71:495–500.

Kusakabe

, Hoshi

, Kimura

. 2006. Origin of the ultimobranchial body cyst: T/ebp/Nkx2.1 expression is required for development and fusion of the ultimobranchial body to the thyroid. Dev Dyn, 235:1300–1309.

10.

Arufe

, Lu

, Kubo

, Keller

, Davies

, Lin

. 2006. Directed differentiation of mouse embryonic stem cells into thyroid follicular cells. Endocrinology, 147:3007–3015.

11.

Shepard

, Andersen

. 1964. Histochemical studies of the human fetal thyroid during the first half of fetal life. Anat Rec, 149:363–379.

12.

Olin

, Ekholm

, Almquist

. 1970. Biosynthesis of thyroglobulin related to the ultrastructure of the human fetal thyroid gland. Endocrinology, 87:1000–1014.

13.

Szinnai

, Lacroix

, Carré

, Guimiot

, Talbot

, Martinovic

, Delezoide

, Vekemeans

, Michiels

, Caillou

, Schlumberger

, Bidart

, Polak

. 2007. Sodium/iodide symporter (NIS) gene expression is the limiting step for the onset of thyroid function in the human fetus. J Clin Endocrinol Metab, 92:70–76.

14.

Fagman

, Andersson

, Nilsson

. 2006. The developing mouse thyroid: embryonic vessel contacts and parenchymal growth pattern during specification, budding, migration, and lobulation. Dev Dyn, 235:444–455.

15.

Serup

. 2012. Signaling pathways regulating murine pancreatic development. Semin Cell Dev Biol, 23:663–672.

16.

McCulley

, Wienhold

, Sun

. 2015. The pulmonary mesenchyme directs lung development. Curr Opin Genet Dev, 32:98–105.

17.

Ohuchi

, Hori

, Yamasaki

, Harada

, Sekine

, Kato

, Itoh

. 2000. FGF10 acts as a major ligand for FGF receptor 2 IIIb in mouse multiorgan development. Biochem Biophys Res Commun, 227:643–649.

18.

Fagman

, Grände

, Gritli-Linde

, Nilsson

. 2004. Genetic deletion of sonic hedgehog causes hemiagenesis and ectopic development of the thyroid in mouse. Am J Pathol, 164:1865–1872.

19.

Saitoh

, Mitsutake

, Nakayama

, Nagayama

. 2009. Fibroblast-mediated in vivo and in vitro growth promotion of tumorigenic rat thyroid carcinoma cells but not normal Fisher rat thyroid follicular cells. Thyroid, 19:735–742.

20.

Crespo

, Del Sol

. 2013. A general strategy for cellular reprogramming: the importance of transcription factor cross-repression. Stem Cells, 31:2127–2135.

21.

Ubli

, Duan

, Pillai

, Rosa-Garrido

, Wu

, Bargiacchi

, Lu

, Stanbouly

, Huang

, Rojas

, Vondriska

, Stefani

, Deb

. 2014. Mesenchymal–endothelial transition contributes to cardiac neovascularization. Nature, 514:585–590.

22.

Thoma

, Merkl

, Heckel

, Haab

, Knoflach

, Nowaczyk

, Flint

, Jagasia

, Jensen Zoffmann

, Truong

, Petitjean

, Jessberger

, Graf

, Iacone

. 2014. Chemical conversion of human fibroblasts into functional Schwann cells. Stem Cell Reports, 3:539–547.

23.

Zhu

, Wang

, Ding

. 2015. Reprogramming fibroblasts toward cardiomyocytes, neural stem cells and hepatocytes by cell activation and signal-directed lineage conversion. Nat Protoc, 10:959–973.

24.

Carré

, Rachdi

, Tron

, Richard

, Castanet

, Schlumberger

, Bidart

, Szinnai

, Polak

. 2011. Hes1 is required for appropriate morphogenesis and differentiation during mouse thyroid gland development. PLoS One, 6:e16752.

25.

Hick

, Delmarcelle

, Bouquet

, Klotz

, Copetti

, Forez

, Van Der Smissen

, Sonveaux

, Collet

, Feron

, Courtoy

, Pierreux

. 2013. Reciprocal epithelial:endothelial paracrine interactions during thyroid development govern follicular organization and C-cells differentiation. Dev Biol, 381:227–240.

26.

Coclet

, Foureau

, Ketelbant

, Galand

, Dumont

. 1989. Cell population kinetics in dog and human adult thyroid. Clin Endocrinol, 31:655–665.

27.

Thomas

, Nowka

, Lan

, Derwahl

. 2006. Expression of endoderm stem cell markers: evidence for the presence of adult stem cells in human thyroid glands. Thyroid, 16:537–544.

28.

Lan

, Cui

, Nowka

, Derwahl

. 2007. Stem cells derived from goiters in adults form spheres in response to intensive growth stimulation and require thyrotropin for differentiation into thyrocytes. J Clin Endocrinol Metab, 92:3681–3688.

29.

Zane

, Scavo

, Catalano

, Bonanno

, Todaro

, De Maria

, Stassi

. 2015. Normal vs cancer thyroid stem cells: the road to transformation. Oncogene [Epub ahead of print]; DOI: 10.1038/onc.2015.138.

30.

Hoshi

, Kusakabe

, Taylor

, Kimura

. 2007. Side population cells in the mouse thyroid exhibit stem/progenitor cell-like characteristics. Endocrinology, 148:4251–4258.