Functional Differences in the Role of Ductal Stem Cells in Mouse Major Salivary Glands

Abstract

Salivary gland (SG) stem cells are the only cell population capable of extended growth in organotypic cultures, and thus they are considered a source for cell-based therapies aimed at SG regeneration. Studies in the mouse submandibular gland have identified only one population of tissue stem cells capable of salisphere formation in culture. These cells are actively dividing ductal cells that express epithelial progenitor markers keratin (K) 5/14 and normally function as lineage-restricted stem cells for differentiated ductal cells. In response to severe injury, however, these cells undergo a multipotency switch and contribute to regeneration of multiple cell lineages, including secretory units or acini. Little is known about the mechanism of cell renewal and regeneration in the other major SGs and whether comparable stem cell populations exist in the parotid (PG) and sublingual (SLG) glands. Using in vivo and ex vivo models, we show that both the PG and SLG contain a small population of K14-expressing ductal cells. Although they do not cycle frequently, K14-expressing ductal cells are the source of salisphere-forming cells in these glands. Long-term lineage tracing studies in adult mouse PGs showed a progenitor–progeny relationship between the K14-expressing ductal cells and the K19-expressing ductal cells in the striated ducts. In the SLGs, however, K14-expressing ductal cells did not generate a differentiated cell progeny for a 6-month period of observation and did not make a significant contribution to regeneration of gland after severe injury. These studies reveal the functional similarities and differences in tissue stem cells among the major SGs and have implications for developing strategies for SG regenerative therapies.

Introduction

Saliva is mainly produced by three pairs of major salivary glands (SGs), including submandibular (SMG), sublingual (SLG) and parotid (PG) glands. These glands have a common lobular structure consisting of multiple secretory end-pieces or acini connected to a branching ductal system composed of intercalated (ID), striated (SD), excretory ducts that modify and transport saliva through a main duct to the oral cavity. In addition, contraction of myoepithelial cells (MECs) that wrap around acini and ducts facilitates transport of saliva through this network [1,2]. Despite this common structure, each gland has unique histological features and cellular composition. In mouse, the SMG and PG are composed of serous acini and show more similarity in global gene expression when compared with SLG, which is composed of mucous acini [1,3]. In addition, in the SMG the ductal network is more prominent due to the presence of granular convoluted ducts (GD) [1].

Recent studies have provided new insights into the mechanism of cell renewal and regeneration in the SG. Studies in the adult mouse SMG have shown that cell lineages are maintained in a lineage-restricted manner [4–7]. The acinar cells, cKit-expressing ID cells, and MECs are maintained as self-sustained cell populations, whereas GD cells are maintained by a small number of actively dividing stem cells located at the ID/GD junction [5–10]. Phenotypically, these cells express common markers of epithelial progenitors, including K14, K5, and p63 but lack the expression of lineage-specific markers, including aquaporin (Aqp)5 (for acinar cells), cKit (for ID cells), K19 (differentiated ductal cells), and smooth muscle actin (SMA; for MECs) [5,7,9].

Although K14⁺ ductal stem cells function as lineage-restricted ductal stem cells for GDs during homeostasis [7], they undergo a uni- to multipotency switch in response to severe injury and contribute to regeneration of multiple cell types, including acinar cells [5,10]. Furthermore, genetic tracking studies in three-dimensional (3D) cultures of SMGs have identified K14⁺ ductal cells as the major source of salispheres-forming cells, consistent with their function as tissue stem cells [8,10–13].

Little is known about the mechanism of cell renewal and regeneration in the SLG and PG. Given the absence of the GD compartment in these glands, it is not clear if K14⁺ ductal cells are present and function as tissue stem cells. Although cells with high proliferative and regenerative potential could be cultivated from all three major SGs, the source of these cells in the PG and SLG have not been defined [14]. In this study, we first determined if K14⁺ ductal cells present in the PG and SLG and then used genetic labeling and cell tracking analysis to compare their function with SMG-K14 ductal stem cells. Our study revealed similarities and differences in the role of K14⁺ ductal cells as tissue stem cells among the major SGs.

Materials and Methods

Animal studies

The care and experimental use of animals, including the surgical procedures, were approved by the Institutional Animal Care and Use Committee (protocol no. 287271) of the Stony Brook University, and animals were housed in an Association for Assessment and Accreditation of Laboratory Animal Care-accredited center in a pathogen-free facility on a 12-h light/dark cycle provided with food and water ad libitum. The following transgenic strains, including K14-Cre^TRE, αSMA-Cre^ERT2+/−, and Axin2-Cre^ERT2+/−, were described previously [7,10,15]. Rosa-CAG-LSL-tdTomato (stock no. 007914) and C57Bl/6J (stock no. 00664) were purchased from Jackson Laboratory (Bar Harbor, ME).

All transgenic lines were maintained on C57Bl/6 background. Cre-recombinase was induced by replacing regular diet with doxycycline- (1 g/kg pellet, Bioserve) or tamoxifen (TAM)-containing diet (250 mg/kg pellet, Envigo) for 4–5 days. Glandular injury to the SLG was induced through unilateral ligation of Wharton's duct and periductal tissues, which include Bartholin's duct as previously described [10]. Three days later, a second surgical procedure was performed to remove the ligature. All surgical procedures were performed on adult female mice (>8 weeks of age) under general anesthesia and aseptic conditions.

SG cell preparation and spheroid cultures

Pairs of SMG, SLG, and PG were collected separately, dissected, and dissociated by mechanical/enzymatic digestion and cultured in media as described previously [8,10,11]. In brief, digested tissue passed through 70 μm strainer and cultured (20 mg starting tissue/mL) in suspension in a low-binding 12-well plate for 2–3 days. To generate salisphere from single cells, cultured cell suspensions were collected, washed, trypsinized (0.1% Trypsin-EDTA for 5 min), passed through a 40 μm filter, and resuspended to 2 × 10⁵ cell/mL. Forty microliter of this cell suspension was mixed with 60 μL of Matrigel (Cat. No. 354230; BD Biosciences, San Jose, CA) on ice and 25 μL of the mix were seeded in the periphery of each well of 48-well plate.

These cell-matrices were allowed to solidify for 20–30 min at 37°C before covered with media. Media was changed every other days and cultures were analyzed after 7 days by fluorescent phase-contrast microscopy (Nikon, Eclipse TS-100). Images were captured from the entire periphery of the wells and used to quantify the number and percentage of Tdtomato (TdT)-labeled spheres in culture. For passage, cultures were treated with dispase (0.45 mg/mL of media) for 30 min at 37°C, collected, rinsed in phosphate-buffered saline, trypsinized and resuspended to 75,000/mL before mixed in Matrigel as described earlier.

Tissue processing and immunofluorescence staining

SGs were collected, cleared of the connective tissues, and fixed in cold 4% paraformaldehyde for 1 h. Fixed tissues were soaked in 30% sucrose, embedded in optimal cutting temperature compound (OCT, Tissue Tek) and cryopreserved tissues were sectioned at 5 μm thickness. For analysis of salispheres, cultures were grown for 10 days, treated with dispase, rinsed and fixed in 4% paraformaldehyde for 20 min at 4°C. Salispheres were pelleted (300g), embedded in OCT media and cryopreserved. For immunofluorescent staining, cryosections were stained with primary antibodies to anti-Aqp5 (Rabbit, 1:1,000, Cat. No. 178615; EMD Millipore, Billerica, MA), anti-K19 (Rat, 1:10, TROMA-III, DHSB, University of Iowa), anti-K14 (Rabbit, 1:2,000, Cat. No. 905303; Biolegend), anti α-SMA (Rabbit, 1:1,000, Cat. No. A01072-1; Boster), and anti-NKCC1 (Rabbit, 1:500, Cat. No. PA2169; Boster).

In genetically labeled tissues, TdT (red fluorescent protein) signal was detected by direct fluorescence. When co-immunostaining, antibodies were added sequentially. Bound antibodies were detected with either Alexa-488 or Alexa-594 conjugated secondary antibodies (ImmunoReagents, Inc., Raleigh, NC) and sections were mounted with media containing 4′,6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA). To evaluate proliferating cells, tissue sections were boiled in the citrate buffer for 3 min and were allowed to cool.

After the antigen retrieval process, tissues were co-immunostained sequentially with anti-Ki67 antibody (Rabbit, 1:100, Cat. No. 12202; Cell Signaling Technology), and either anti-K14 or anti-SMA antibodies. Images were captures using either E800/NIS-Elements (Nikon Instrument Inc., Melville, NY) or BZ-X800E (KEYENCE Corp., Elmwood Park, NY) fluorescent microscope and quantified using image J. For quantification of Ki67- or TdT-labeling efficiency in the targeted cell population at least 15 images of 40 × magnification from each gland were used and the ratio of labeled cells to total number of target cell population (n ≥ 300 nuclei) were determined.

Quantification and statistical analyses

Images captured from ≥4 sections taken at least 50 μm apart were analyzed for each gland. For lineage-labeled acini, the area of the TdT⁺Aqp5⁺ cells to total Aqp5⁺ areas are expressed in pixel density. All statistics were performed using SPW12 software (Systat Software Inc., San Jose, CA). Differences among means were evaluated by one-way analysis of variance and the Tukey's HSD post hoc test or by unpaired two-sided Student's t-test. Significance was set at P < 0.05.

Results

K14-expressing ductal cells are present in the adult mouse PG and SLG

To determine whether K14⁺ ductal cells are present in SLG and PGs, glands were collected from adult C57Bl/6 mice (8-week-old females, n = 2 mice) and cryosections were co-immunostained with antibodies against K14 and SMA. The latter is used to distinguish between ductal stem cells and MECs, which also express K14 [7,9]. Immunofluorescent analysis showed that although the majority of K14⁺ cells co-expressed SMA (Fig. 1A, B, arrowheads), a subset of K14⁺ cells located in the basal position within the intralobular ducts in both PGs and SLGs did not express SMA (Fig. 1A, B, arrows). Quantification of K14⁺SMA⁻ and K14⁺SMA⁺ cells indicated that about 20% of the K14⁺ cells did not express SMA and thus represented the K14⁺ ductal cells (Fig. 1C). This was consistent with the ratio of ductal stem cells to MECs in the SMG [10], verifying the presence of K14⁺ ductal cells in the lobular compartments of all three major SGs.

FIG. 1.

Characterization of K14-expressing ductal cells in PG and SLG. (A, B) Representative fluorescent images of mouse PG and SLG co-immunostained with antibodies against K14 and SMA (n = 2 mice). Arrows point to ductal cells and arrowheads point to MECs. Blue nuclear staining is DAPI. Scale bars = 50 μm. (C) Quantification of MECs (K14⁺SMA⁺) and ductal cells (K14⁺SMA⁻) in PG and SLGs expressed as mean ± standard deviation, using 35 images from two female mice. (D) Proliferation rate of K14⁺ ductal cells and MECs in PG, SLG, and SMG of adult mice assessed by Ki67 immunostaining. Values are expressed as mean ± standard deviation using 45 images of magnification 40 × from two female mice. (E) Representative images of sections of SMG, PG, and SLG co-stained for Ki-67 and K14 or Ki-67 and SMA. Scale bars = 50 μm. DAPI, 4′,6-diamidino-2-phenylindole; K, keratin; MECs, myoepithelial cells; PG, parotid gland; SLG, sublingual gland; SMA, smooth muscle actin; SMG, submandibular gland.

The SMG-K14⁺ ductal cells are actively cycling stem cells with a proliferation index that is severalfold higher than MECs or any other cell types in the SMG [7]. To determine the cell cycle kinetics of SLG- and PG-K14⁺ ductal cells, we compared the Ki-67 (a nuclear protein associate with cell cycle) labeling indices of K14⁺ ductal cells among the major SGs. Tissue sections were co-immunostained for Ki-67 and either K14 or SMA, and the percentage of Ki-67 positive nuclei in each cell population was quantified. The number of Ki-67⁺SMA⁺ cells was deducted from that of Ki-67⁺K14⁺ cells to determine the proliferation index for K14⁺ ductal cells (Fig. 1D, E). This analysis showed that K14⁺ ductal cells in the PG and SLG cycled at a significantly lower rate than those in the SMG (Fig. 1D). Intra-gland comparative analysis revealed that K14⁺ ductal cells cycled at a slightly but significantly higher rate than MECs in the PG but not in the SLG (Fig. 1D). These data suggest functional differences in K14⁺ ductal cells among major SGs.

K14⁺ ductal cells function as ductal stem cells in the PG but not in the SLG

In the SMG, there is a progenitor–progeny relationship between K14⁺ and K19⁺ ductal cells [7,16]. To determine if the K14⁺ ductal cells function as progenitors for differentiated ductal cells, we used inducible lineage tracing in K14Cre^TRE:ROSA-tdT (K14-TdT) mice to assess the fate of K14⁺ ductal cells in these glands for a long period of time. K14-TdT mice (8-week-old females and males, n = 4 mice/time point) were fed with doxycycline-containing diet for 5 days and SGs were collected either after the end of the pulse period (T0) or after a 28-week chase period (T28) (Fig. 2A). Analysis of glands collected at T0 and stained for K14 confirmed selective labeling of K14⁺ cells with TdT (red fluorescence signal) with an efficiency of about 40% (Fig. 2B, C).

FIG. 2.

Long-term lineage tracing of K14⁺ ductal cells in the PG and SLG. (A) Timeline of induction of genetic labeling and tracking of K14-expressing cells in the adult mice. (B, C) Distribution and quantification of TdT-labeled K14⁺ cells in SLG, PG, and SMG collected after a 5-day pulse period (T0) and stained for K14. Data are mean ± standard deviation (n = 4 mice). (D) Comparative analysis of the size of lineage-traced cell clusters in SMG, PG, and SLG in males and female mice after a 28-week period of chase (T28). Values are expressed as mean ± standard deviation (n = 4 mice). Data were analyzed by ANOVA with P is indicated in the figure. (E, F) Representative images of SLG, PG, and SMG collected at T28 and stained for K14 or K19 showing distribution of lineage-labeled cells in the ducts. Blue nuclear staining is DAPI. Scale bars = 50 μm. Arrows in (B) and (E) point to TdT-labeled K14-expressing ductal cells. ANOVA, analysis of variance; NS, not significant; TdT, tdTomato.

Staining of the glands collected at the end of chase period (T28) for K14 and K19 showed organization of lineage traced cells into cell clusters that included both K14⁺ and K19⁺ cells in the PG and SMG, although the average size of clusters was smaller in the PG (Fig. 2D–F). Analysis of the SLG, however, showed no significant difference in the distribution pattern of TdT-labeled cells at T0 compared with T28 (Fig. 2B, E). Even after >6 months, TdT expression was restricted to K14⁺ cells and did not expand into the K19⁺ ductal cells (Fig. 2E, F). This was more consistent with the behavior of self-sustaining cells such as MECs rather than ductal stem cells [9,10]. Overall, these data indicated that K14⁺ ductal cells function as a lineage-restricted ductal stem cell population in the PG but a similar function for the SLG-K14⁺ ductal cells could not be defined.

K14⁺ ductal cells are the source of salisphere-forming cells in all three major SGs

When any of the major SGs are cultivated in 3D cultures, a small number of cells that are thought to represent tissue stem cells form salispheres [11,14,17] (Xiao, 2014 No. 5670; Lombaert, 2008 No. 4955]. Ex-vivo cell tracking analysis in cultures of the adult mouse SMG have shown that the K14⁺ ductal stem cells are the only cell population with a robust salisphere-forming capacity [8,10]. To assess the salisphere-forming capacity of K14⁺ ductal cells in the PG and SLG, we induced TdT label in K14-TdT mice and tracked the fate of TdT-labeled cells in 3D cultures established from all three major SGs (Fig. 3A).

FIG. 3.

K14+ ductal cells are the source of salispheres in all major SGs. (A) Schematic representation of the strategy used for tracking of TdT-labeled SMA⁺ or K14⁺ cells in spheroid cultures of SMG, SLG, and PG. (B) Percentage of TdT-labeled K14-expressing cells (K14-TdT) and MECs (SMA-TdT) in SMG, PG, and SLG tissues stained for K14 or SMA (n = 3 mice). (C–E) Quantification of images of TdT-labeled salispheres in cultures of SMG, SLG, and PG (>2,000 organoids counted/each SG type, from three independent experiments from a total of six mice), and representative phase-contrast images of spheroid cultures of glands from K14-TdT (D) or SMA-TdT (E) mice after 7 days in culture with TdT-labeled salispheres expressing red fluorescence. (F) TdT-labeling efficiency in salispheres derived from K14-TdT⁺ cells during three passages. Values are mean ± standard deviation from three independent experiments. Data were analyzed by ANOVA. (G) Representative images of salispheres derived from TdT-labeled SLG-K14⁺ cells immunostained for K14 (progenitors), K19 (ductal cells), or NKCC1 (acinar cells). Red is TdT and antibody staining is in green. Scale bars = 50 μm.

Since K14-expressing cells include MECs, we used SMACre^ERT2:ROSA-TdT (SMA-TdT) mice in which TAM-inducible Cre expression is targeted to MECs [10,15], to genetically label and track the fate of MEC, in addition to the K14⁺ cells. Cre-recombinase was induced in the adult K14-TdT or SMA-TdT mice (8- to 10-week-old males, n = 4 glands/experiment) and 1 week later, SLGs, PGs, and SMGs were collected separately. One gland was used to assess TdT-labeling efficiency of targeted cell populations and the remaining three glands were pooled, dissociated into single cell suspensions and cultured to allow formation of salispheres as described in the Materials and Methods section (Fig. 3A).

Immunostaining of the tissue sections for K14 or SMA verified comparable TdT-labeling efficiencies ranging from 31% to 45% among all SGs (Fig. 3B). Analysis of cultures showed that despite a comparable number of total salispheres among SGs established from K14-TdT and SMA-TdT mice, there was a significant difference in the number of TdT-labeled salispheres derived from K14- and SMA-expressing cells (Fig. 3C–E). In cultures established from K14-TdT mice, nearly half of the SLG-, PG-, or SMG-derived salispheres were uniformly expressing TdT (Fig. 3C, D).

On the contrary, in cultures established from SMA-TdT mice, only a few TdT-labeled salispheres were detected (Fig. 3C, E). A similar ratio for the initial TdT-labeling efficiency in vivo and the TdT-labeled salispheres in culture indicated that, similar to the SMG, nearly all PG- and SLG-derived salispheres were originated from K14⁺ ductal cells. Subsequent passage of these cultures did not affect the ratio of TdT-labeled salispheres indicating self-renewal capacity of SLG- and PG-K14⁺ cells in culture (Fig. 3F).

Immunostaining of SLG-derived salispheres for K19 and NKCC1 (a marker for acinar cells) verified expression of both ductal and acinar markers in salispheres (Fig. 3G), consistent with previous studies that have shown no significant difference in the expression of ductal and acinar markers in salispheres derived from major SGs [14]. Overall, these results indicated that despite differences in the in vivo function of K14⁺ ductal cells in the SLG from those in the SMG and PG, SLG-K14⁺ ductal cells retain a robust proliferative and differentiation potential suggesting that they function as reserved stem cells.

Contribution of SLG-K14⁺ ductal cells to acinar regeneration after a severe injury

In various epithelial tissues, quiescent stem cells function as reserved stem cells that are mobilized in response to severe injury [18]. To investigate if SLG-K14⁺ cells function as quiescent stem cells and mobilized in response to injury, we characterized the dynamics of ligature-induced injury to SLG. We have previously used a model of reversible obstruction-induced injury to map the fate of various cell populations during regeneration of the mouse SMG.

In this model, a 3-day period of ligation of the Wharton's duct and periductal tissues results in extensive degeneration of acini followed by a progressive regeneration of the secretory complex when the ligature is removed [10]. In rodents, the Bartholin's duct that runs parallel to the Wharton's duct [1] is also obstructed by the ligature thus inducing injury to the SLG. To evaluate the extent of injury to the SLG in this model, C57Bl/6 mice (8-week-old females, n = 2/time point) were subjected to unilateral ligation as described earlier and the SLGs were collected at days 0, 5, 10, and 20 after removal of the ligature (Fig. 4A).

FIG. 4.

Histological analysis of SLG injury and regeneration in a model of severe injury. (A) Timeline of unilateral ligature-induced injury and recovery in 8-week-old female C57Bl/6 mice. (B, C) Representative images of control (noninjured) and injured SLGs collected either immediately after deligation (T0) or at 5-, 10-, and 20-day postdeligation stained either with hematoxylin and eosin or with antibody against Aqp5, a specific marker for acini (n = 2/group). Arrowheads point to acini/proacinar cells and arrow points to duct-like structures. Scale bars = 50 μm. Aqp, aquaporin.

Tissue sections were stained either with hematoxylin and eosin or immunostained for Aqp5 (Fig. 4B, C). Histological analysis of the injured SLGs verified an extensive damage to acinar cells at T0 (Fig. 4B, arrowheads), followed by gradual reappearance of acini with mucous-type morphology by T10 (Fig. 4B, arrowheads). Immunofluorescent analysis of the injured SLGs confirmed the loss of Aqp5 signal at T0 and gradual recovery of the signal after deligation. By day 20, most regions of the glandular compartment recovered as shown by the extent of regenerated acinar and ductal cells (Fig. 4B, C, T20).

To determine the contribution of SLG-K14⁺ ductal cells to acinar regeneration in this model of reversible injury, we used K14-TdT and SMA-TdT mice (10- to 12-week-old females, n = 3/group) to genetically label K14⁺ and MECs before injury and map the fate of labeled cells after recovery (Fig. 5A). In the contralateral gland that was used as control, immunostaining for SMA verified expression of TdT in both MECs (SMA⁺) and ductal cells (SMA⁻) in the K14-TdT mice, and its restriction to MECs in the SMA-TdT mice (Fig. 5B).

FIG. 5.

Contribution of SLG-K14⁺ ductal cells and MECs to glandular regeneration after injury. (A) Timeline of genetic labeling and lineage tracing of targeted cell populations in three transgenic lines after the unilateral ligature-induced injury and recovery of SLG. (B) TdT expression in unligated contralateral SLG of K14-TdT and SMA-TdT mice stained for SMA. Arrow notes the TdT-labeled K14⁺SMA⁻ cells in the K14-TdT mice. (C) Representative images of lineage-traced cells clusters in the regenerated SLGs of K14-TdT and SMA-TdT mice, immunostained for K14 and Aqp5 (n = 3 mice). Ac denotes TdT-labeled acini and arrows point to the TdT-labeled ducts. (D) Fluorescent images of control (CNT) and regenerated (REG) SLGs of Axin2-TdT mice (n = 2) immunostained for K14 and Aqp5, respectively. Arrow points to a TdT-labeled K14⁺ ductal cells and arrowheads points to MECs. Nuclear blue is DAPI. Scale bars = 50 μm. (E) The percentage of TdT-labeled acinar cells in the regenerated glands K14-TdT, SMA-TdT, and Axin2-TdT mice. Data are presented as mean ± standard deviation (n ≥ 2 mice), two-tailed t-test was used to compare differences between SMA-TdT and K14-TdT mice.

Analysis of regenerated SLGs in both mouse models showed organization of lineage-labeled cells into large cell clusters composed of acinar-like structures (Fig. 5C). Immunostaining for Aqp5 verified that the initially labeled K14⁺ or SMA⁺ cells contributed to regeneration of acini in this model of injury (Fig. 5C). Quantification of lineage-labeled acini, however, showed no significant difference in the percentage of TdT-labeled acini between the two mouse models (Fig. 5D), indicating that the TdT-labeled acini originated from MECs rather than K14⁺ ductal cells.

To further confirm the minimal contribution of K14⁺ ductal cells to acini regeneration in the SLG, we used Axin2Cre^ERT2:Rosa-TdT (Axin2-TdT) mice in which Axin2-Cre has been shown to specifically targeted to the K14⁺ ductal stem cells in SMGs [5,10]. Axin2-TdT mice were treated as aforementioned (Fig. 5A) and both SLGs were collected for analysis. Immunostaining of control glands for K14 verified TdT expression in K14⁺ ductal cells (Fig. 5E, arrow) but not in the MECs (Fig. 5E, arrowheads). Analysis of regenerated SLGs showed a few TdT-labeled acini indicating limited contribution of SLG-K14⁺ ductal cells to acinar regeneration confirming that the majority of TdT-labeled regenerated acini originated from MECs (Fig. 5D). Overall, these data indicated that SLG-K14⁺ ductal cells do not make a significant contribution to regeneration of SLG in this model of severe injury.

Discussion

Our studies have identified differences in the proliferation dynamics and function of K14-expressiong ductal stem cells in the major SGs. We found that SLG-K14⁺ ductal cells do not contribute to renewal of any differentiated cell types under steady-state conditions; however, they retain a high proliferation and differentiation potential and serve as reserved stem cells.

The PG showed greater similarity in stem cell function to the SMG than the SLG. The low proliferation rate of K14⁺ ductal cells in the PG is likely related to slow turnover of the SD. Previous studies in the mouse SMG have shown that ductal cells in the GD turnover at a substantially higher rate than those in the SD [16]. Similarly, cell kinetic analyses in the rat PG and SLG during postnatal development have shown that SD cells have the lowest turnover rate when compared with all other epithelial and mesenchymal cell in these glands [19,20]. In our study, the size of lineage-traced clusters, which reflects the cumulative cell division for the initially labeled K14⁺ ductal cell and its progeny, showed that despite the low proliferative index of the PG-K14⁺ ductal cells, their progeny expanded substantially during the chase period to form a cell cluster that included differentiated SD cells.

Interestingly, the proliferation index of SLG-K14⁺ ductal cells was only slightly lower than that in the PG, yet no lineage-labeled cluster was detected in the SLG. This suggests that SLG-K14⁺ ductal cells divide to self-renew without committing to generate a differentiated progeny. It is worth noting that during postnatal development, K14⁺ ductal progenitors contribute to generation of the SDs (data not shown) [4]. The lack of contribution of SLG-K14⁺ ductal cells after the gland is fully developed (8 weeks of age) suggests that either the SD is a static compartment or self-replication of differentiated SD cells is sufficient to maintain homeostasis in this compartment.

Our data demonstrated that despite differences in the outcome of lineage tracing of the PG- and SLG-K14⁺ ductal cells in vivo, they showed a comparable salisphere-forming capacity in culture. The presence of stem cell in all major SGs have been postulated based on salisphere-forming capacity of a small subset of cells in culture [12,14,21,22]. We have previously identified the K14⁺ ductal cell population as the major source of salisphere-forming cells in the murine SMG, which is consistent with its function as tissue stem cells; neither MECs nor cKit⁺ ductal cells, which maintain an undifferentiated morphology, display a robust salisphere-forming capacity in culture [8,10].

Although the salisphere-forming capacity of PG-K14⁺ cells is consistent with their in vivo behavior as stem cells for K19⁺ SD cells, it is difficult to attribute the salisphere-forming capacity of the SLG-K14⁺ ductal cells to their function as tissue stem cells. In addition to the long-term self-renewal capacity, a major criterion for stem cells is their ability to generate a progeny that is committed to differentiation under steady-state conditions [23]. The basal position of SLG-K14⁺ ductal cells, their infrequent cell division, and lack of contribution to other cells under steady-state conditions are comparable with the behavior of MECs rather than ductal stem cells [7,9,10].

Although SLG-K14⁺ ductal cells may be defined as a quiescent stem cell population, this type of stem cells are mostly defined in rapidly renewing tissues where they coexist with actively dividing progenitors in a hierarchical relationship [18,24–26]. Such hierarchies have not been detected in the SGs where all cell lineages are maintained by either self-duplication or by a lineage-restricted stem cell population [6,7,16,27]. Whether SLG-K14⁺ ductal cells are a small subset of embryonic ductal progenitors that are retained in the adult tissue and function as reserved stem cells needs further investigations. Regardless of the SLG-K14⁺ cell function in vivo, their existence provides a cellular source for cell-based therapies of degenerative SGs as a recent study has shown a comparable regenerative capacity for salispheres derived from all three SGs when transplanted to the irradiated SMG [14].

Our fate mapping analysis revealed that MECs make a significant contribution to acinar regeneration after a severe injury to the SLG. Although SLG-K14⁺ ductal cells are also activated and contribute to acinar regeneration, their contribution was minimal when compared with MECs. Therefore, the major mechanism of SLG-acinar regeneration after a severe damage to the secretory complex relies on the plasticity of differentiated cells rather than activation of quiescent stem cells. Induction of duct-ligation injury in mouse PG is technically challenging [28]; therefore, a direct comparison between contribution of PG- and SLG-K14⁺ ductal cells to gland regeneration could not be made.

However, our previous studies in the mouse SMG have shown that even though SMG-K14⁺ ductal cells undergo a unipotency to multipotency switch in response to injury to generate both ductal and acinar cells, >60% of acini are regenerated by de-differentiation of MECs [10]. Therefore, in both SMG and SLG, plasticity of MECs is the major mechanism for regeneration of both serous and mucus acini in this model of injury. It is worth noting that plasticity of MECs does not account for all the regenerated acini in the SLG. The survival of some acinar cells on the periphery of the gland after the ligature-induced injury and the more pronounced regeneration of acini in the gland periphery (Fig. 4) suggest that the surviving acinar cells are another likely source of this regeneration [10].

Conclusions

Our data have underlined the similarities and differences in the role of ductal stem cells and mechanism of cell renewal and regeneration among SGs. The presence of K14⁺ ductal cells exhibiting a robust proliferative capacity in cultures in all SGs provide a common cell source for cell-based regenerative therapeutic strategies for these glands; however, therapeutic approaches aimed at endogenous regeneration of SGs should focus on induction of cell plasticity rather than activation of stem cells.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This study was supported by the NIH-NIDCR (R21DE030653 to S.G.).