Isolation and Investigation of Presumptive Murine Lacrimal Gland Stem Cells

Abstract

Purpose: Aqueous tear deficiency due to lacrimal gland insufficiency is one of the major causes of dry eye disease. In severe cases, such as Sjogren's syndrome, Stevens-Johnson syndrome, or ocular cicatricial pemphigoid, therapy with artificial tears is often insufficient to relieve severe discomfort, prevent progressive ocular surface disease, or enable visual rehabilitation by corneal transplantation. Cell or organ generation from stem cells, resulting in tear-like secretion, presents an option as a suitable alternative treatment. To obtain deeper insights into lacrimal gland stem cells we analyzed murine lacrimal glands for markers of pluripotency, self-renewal, and differentiation.

Methods: A special, patented technique with mechanical and enzymatic digestion was used to generate high numbers of cells in vitro from murine lacrimal glands. These presumptive “murine lacrimal gland stem cells” (“mLGSCs”) can be propagated as monolayer cultures over multiple passages. By means of RT-PCR, Western blot, and immunohistochemistry, markers of pluripotency and differentiation were demonstrated. Hanging drop culture was used to build organoid bodies from mLGSCs to investigate their spontaneous differentiation in three-dimensional culture with histology, immunohistochemistry, and transmission electron microscopy methods.

Results: Isolated mLGSCs were cultured over more than 65 passages. Murine lacrimal gland stem cells expressed markers of pluripotency such as Nanog, Sox2, Kruppel-like factor 4 (Klf4), as well as early-lineage markers of all three germ layers. Three-dimensional culture of these cells revealed their ability to differentiate into various cell types.

Conclusions: Our results suggest that mLGSCs were isolated and cultured successfully. These cells have the ability to differentiate into all three germ layers. The results provide further insights into lacrimal gland stem cell physiology for engineering of a lacrimal gland construct to treat severe cases of tear deficiency in the future.

Dry eye disease (DED) is a multifactorial disease of the tear film and ocular surface that results in symptoms of discomfort, visual disturbance, and tear film instability with potential to damage the ocular surface. It has a diverse etiology and the major categories are aqueous-deficient and evaporative dry eyes.¹ Lacrimal gland insufficiency is the major cause of aqueous deficiency. It can be primarily caused by congenital alacrimia and familial dysautonomia or—more frequently—secondarily caused by age-related atrophic changes, inflammatory disease, radiotherapy, ablation, and denervation.¹ The 2007 International Dry Eye Workshop (DEWS) report has estimated the global prevalence of DED to be between 5% to more than 35% at various ages (21 to >65 years of age).¹ Initially, DED can cause a drop in contrast sensitivity and visual acuity, leading to degraded performance in routine vision-related activities such as driving and reading.² The signs and symptoms include ocular dryness, grittiness, burning and foreign body sensation, redness, and blurred vision that clears on blinking.³ Over time, the loss of tear film integrity induces corneal epithelial irregularities and epithelial defects with a risk of secondary infection.⁴ In severe cases, the result can be severe ocular surface pain and blindness due to corneal scarring. It is in this latter situation that the treatment options available, such as tear substitutes, punctual plugs, or moisture-chamber spectacles either provide insufficient symptomatic relief or enable restoration of vision. Microvascular salivary gland transplantation has been used to treat patients with otherwise intractable severe DED.⁵ However, the substantial differences between the electrolyte and protein composition of tears and saliva have led to microcystic corneal epithelial edema and prevented successful visual rehabilitation with corneal transplantation. A possible long-term solution for this severely affected group of patients might be a lacrimal gland transplant.^6,7 

In other areas, artificial tissues or organs have been cultured in vitro and successfully transplanted. For example, cartilage is currently used in orthopedics and entire urethrae are used for clinical studies in urology.⁸ Recently, Eiraku et al.⁹ have demonstrated the autonomous differentiation of stem cells to an optic cup in vitro in a three-dimensional network. Because of the interesting results with cultured β cells or pancreatic stem cells in the search for new diabetes mellitus cures,¹⁰ investigators have envisaged use of stem cells for DED therapies. A lot of effort has been put into the culture of lacrimal glands in vitro, for example, studies in a micrograft environment¹¹ and transfer of amniotic membranes,¹² but these methods have not shown the expected results over an extended period of time. Another way to create an artificial lacrimal gland is forced differentiation of stem cells to adult lacrimal gland tissues. Because of the controversial ethical debate on embryonic stem cells,¹³ adult stem cells (ASCs) should be used. Adult stem cells renew different cell types of the tissue in which they originally reside. Other types of stem cells are embryonic stem cells and induced pluripotent stem cells, which are capable of differentiating into all three germ layers.¹⁴ Recent work by Avilion et al.,¹⁵ Loh et al.,¹⁶ and Karbanová et al.¹⁷ has detected a mesenchymal cell type that shows markers of “stemness,” for example, Oct-4, Nanog, and Sox2^15,16 in human and murine lacrimal glands, or typical stem cell markers in salivary glands, such as CD133.¹⁷ 

The aim of the present study was to isolate and characterize murine lacrimal gland stem cells and grow them in culture in two- and three-dimensional networks. To this end, cells showing typical stem cell characteristics were purified and cultured. They express an RNA/protein pattern specific to embryonic stem cells and progenitor cells. These cells also have the capacity to differentiate into cell types of all three germ layers as well as the potential to form multicellular three-dimensional structures in vitro. Since amounts of lacrimal gland ASCs are apparently very low in vivo, such investigations prove very difficult and unsatisfactory. We therefore used a unique patented method¹⁸ from the Fraunhofer Research Institution for Marine Biotechnology (Lübeck, Germany) to isolate progenitors of adult lacrimal glands of wild-type mice. 

Materials and Methods

Tissue

Murine lacrimal, parotid, and pancreatic gland samples were obtained from cadavers of wild-type ICR/5-HT4 mice and BL6 mice not later than 30 minutes post mortem, supplied by the group of Nicolas Keller of the Department of Pharmacology and Toxicology and Uwe Rückschloss of the Department of Physiology, both at Martin Luther University of Halle-Wittenberg, Germany. Murine stomach and brain samples were obtained as positive controls as well. All animals served for other experiments and we only used their lacrimal glands after they had been killed. The animals were treated in these other experiments in compliance with institutional review board regulations for animal ethics and in accordance with the Association for Research in Vision and Ophthalmology Statement for the Use of Animals in Ophthalmic and Vision Research and the recommendations of the National Institutes of Health Guide for the Care and Use of Laboratory Animals. The experiments had received governmental approval. To remove the lacrimal glands, the skin of the mice was removed and, because the extraorbital gland is a very distinct organ, it was delimited from the parotid or other salivary gland. The tissue from each cadaver was split for stem cell isolation, immunohistochemistry, and transmission electron microscopy (TEM). Some tissue samples from each lacrimal gland were immediately snap-frozen in liquid nitrogen and stored at −80°C for further biomolecular investigation. Other samples were fixed in paraformaldehyde or ITO buffer (see below). 

In addition to murine tissues, 12 human lacrimal glands from six body donors were also used. This use was conducted in compliance with institutional review board regulations for human ethics, informed consent regulations, and the provisions of the Declaration of Helsinki. Lacrimal glands were obtained from cadavers (three males, three females, age 48–86 years) donated to the Department of Anatomy and Cell Biology, Martin Luther University Halle-Wittenberg. All tissues used were dissected from the cadavers between 4 and 12 hours post mortem. The donors were free of recent trauma, eye and nasal infections, or diseases affecting lacrimal functions. 

Isolation of Murine Lacrimal Gland Stem Cells

Extraorbital lacrimal glands were obtained and digested in an Erlenmeyer flask filled with digestion medium containing HEPES-buffered modified Eagle's medium (MEM; PAA Laboratoires, Pasching, Austria), 5% bovine serum albumin (BSA), and 0.1 M CaCl₂ plus collagenase NB8 (SERVA, Heidelberg, Germany). The glands were freed of connective tissue and fat with small surgical scissors. Afterwards the tissue was dissociated with small scissors and the resulting suspension was equilibrated with carbogen adjusted to pH 7.4 and incubated under constant shaking for 20 minutes at 37°C and 2.4g. Floated fat and residues of connective tissue were then exhausted together with the digestion medium. The gland particles were washed three times with digestion medium, but without collagenase. Subsequently, digestion/incubation was repeated, followed by treatment with medium containing collagenase for 15 minutes. To obtain only very small cell heaps, up-and-down suction was done through serologic glass pipettes (10, 5, and 2 mL). Then the emulsion was filtered through nylon nets with a 250-μm mesh. After centrifugation at 86.0g for 5 minutes, the pellet was solubilized in Dulbecco's modified Eagle's medium (DMEM) supplemented with 20% fetal calf serum (FCS; Biochrom AG, Berlin, Germany). Centrifugation and solubilization were done four times to get rid of single cells and bigger cell heaps. The purified acini were seeded in a small culture flask (5000 cells/cm² [500 cells/μL]) with DMEM and 20% FCS containing 1% streptomycin–penicillin and gentamicin (Sigma-Aldrich, Taufkirchen, Germany) to prevent contamination in a humidified 5% CO₂ incubator at 37°C for 48 hours. 

Cell Culture

After culturing for 48 hours, the culture medium was removed and the cells were washed three times with phosphate-buffered saline (PBS) to remove dead cells because they do not adhere to the bottom of the culture flask. The cells were then cultured with HEPES-buffered DMEM (high glucose) and 10% FCS, which will be referred to as “culture medium” in this article. The cells were studied daily under light microscopy until confluence was achieved (approximately 1 week). After reaching confluence, presumptive lacrimal gland stem cells (pLGSCs) were split by digestion using trypsin, then counted and reseeded. After passage number 1 the cells were split every 3 days. This procedure was repeated several times to obtain a large number of cells. For RNA and protein isolation, pLGSCs were cultured in culture flasks. For the purpose of immunofluorescence, pLGSCs were seeded on small glass cover slips in Petri dishes filled with culture medium. 

Hanging Drop Culture

Cells were counted after digestion with trypsin and reseeded in a 20-μL drop of culture medium containing approximately 1000 cells. Approximately 50 drops were placed on the inner side of Petri dish lids. The lids covered Petri dishes containing PBS to generate a humidified atmosphere. Hanging drops were cultured for 48 hours at 37°C and 5% CO₂, and then rinsed with culture medium in new Petri dishes coated with 0.1% gelatin for culture in suspension.¹⁹ 

Transmission Electron Microscopy

For TEM, organoid bodies and outgrowing cells cultured for 6 to 7 days were fixed at 4°C overnight in ITO buffer²⁰ containing 25% glutaraldehyde, 25% paraformaldehyde, and cacodylate buffer, adjusted to pH 7.3 by adding picric acid. The culture flasks were then washed two times in PBS for 60 minutes. After an ascending alcohol series (70% two times for 60 minutes, 80%–95% for 30 minutes, and 100% two times for 30 minutes), the samples were embedded in Epon. Using a microtome (Ultracut E; Reichert-Jung, Munich, Germany) we prepared sections. To obtain an overview, semithin sections (0.5-μm thickness) were stained with toluidine blue and other sections (70 nm) were used for TEM. Afterwards the slices were contrasted with citrate and uranyl acetate. All TEM investigations were performed with a Zeiss TEM 902 ESI (Carl Zeiss, Jena, Germany). 

RNA Isolation and Complementary DNA (cDNA) Synthesis

For reverse transcription–polymerase chain reaction (RT-PCR), murine stomach, brain, and liver, as well as lacrimal and salivary glands, were frozen immediately after extraction under liquid nitrogen and pounded with mortar and pestle. Further tissue samples of six human lacrimal glands (from three women and three men, body donors) were removed and treated as described above. After this step, a homogenizer (Polytron, Norcross, GA, USA) was used to enlarge the surface, so total RNA was isolated from the lysates by using the RNeasy Mini Kit (Qiagen, Hilden, Germany) according to manufacturer's instructions. 

Total RNA from cultured pLGSCs and murine embryonic stem cells (mESCs), a gift from Insa Schroeder, was extracted by using TRIZOL reagent (Invitrogen, Karlsruhe, Germany). The amount and purity of all samples of total RNA were measured photometrically. Reverse transcription of all RNA samples to first-strand cDNA was performed by using the RevertAid H Minus Reverse Transcriptase Kit (Fermentas, St. Leon-Rot, Germany) according to manufacturer's instructions. Two micrograms of total RNA (A_260/280 ratio from 2.0–2.3) were used for each reaction. To prove integrity and amplification, tubulin was used as the housekeeping gene and was detected in all synthesized samples of cDNA. To avoid contamination with genomic DNA, reverse transcriptase was replaced by water free of RNase and DNase during cDNA synthesis as a negative control. 

Polymerase Chain Reaction

Each reaction was primed with 2 μL cDNA (from each sample), 9.8 μL H₂O, 2 μL 50 mM MgCl₂, 2 μL 10 mM dNTPs, 2 μL 10x PCR buffer, 0.2 μL (1.25 U) Taq polymerase (Invitrogen) and 2 μL intron-spanned primer mix (each 10 pmol; Table 1) for RT-PCR. All PCRs underwent specific steps, that is, with an initial cycle at 95°C for 5 minutes followed by 35 to 40 cycles of 95°C for 15 seconds, specific primer annealing temperature for 30 seconds, 72°C for 30 to 40 seconds, and a final elongation at 72°C for 5 minutes. Samples were stored at 4°C until gel electrophoresis was done. All of the specific primers used were synthesized at Eurofins MWG Operon (Ebersberg, Germany). Twelve microliters of the samples were loaded on 2% agarose gel together with 3 μL 6X Orange DNA Loading Dye (Fermentas), and the amplified DNA bands were visualized under ultraviolet illumination. All PCR products were also confirmed by sequencing with BigDye (Applied Biosystems, Foster City, CA, USA) and the results were compared with GenBank (BLAST) data. 

Table 1

 

Reverse Transcription–PCR Primers

 

Protein Isolation and Western Blot Analysis

Protein of pLGSCs, mESCs, and lysates of murine stomach and brain, as well as lysates of human lacrimal glands, was extracted with Triton buffer and homogenized on ice with an Ultra-Turrax T25 homogenizer (Janke & Kunkel, Staufen i. Br., Germany). After centrifugation for 30 minutes at 16,000g, the supernatant was transferred to Eppendorf tubes. The total protein content was measured by using Bradford reagent, BIO-RAD Protein Assay (Bio-RAD, Munich, Germany). For Western blot analysis, 20 μg total protein of each sample was heated at 96.6°C for 5 minutes in 10 μL RSB buffer, after which the samples were stored at 4°C, then subjected to SDS-PAGE. Proteins were transferred to Hybond-ECL nitrocellulose membranes (Amersham Life Technologies, Piscataway Township, NJ, USA) by using Semidry-Blot (Roth, Karlsruhe, Germany) according to manufacturer's instructions. This was followed by incubation overnight in Tris-buffered saline and 0.05% Tween (TBST) containing 5% dried milk with antibodies (Table 2) at 4°C. Secondary antibodies were used for 1.5 hours in TBST with 5% dried milk at room temperature, followed by chemiluminescence detection for visualization (Amersham ECL Prime Western Blotting Detection Reagent; Amersham Biotech, London, UK). Control standard molecular weight markers (Prestained Protein Ladder; Fermentas) were used to determine the molecular weights. 

Table 2

 

Antibodies for Western Blot and Immunofluorescence

 

Immunohistochemistry

Extracted murine lacrimal and salivary glands were fixed in 4% paraformaldehyde, after which the tissues were watered overnight. The tissues were passed through an ascending alcohol series before embedding in paraffin, then cut with a microtome (MICROM, Walldorf, Germany) into 7-μm slices and mounted on slides. The slides were incubated overnight at 60°C followed by deparaffinization and rehydration. The slides were put through hydrogen peroxide for 30 minutes to block endogenous peroxidases and then cooked in a microwave containing citrate buffer (C₆H₅Na₃O₇) (pH 6.0) to unmask the epitopes. For staining, the slides were blocked with secondary antibody donor sera. The slides were then treated with primary antibodies diluted in TBS buffer overnight, then with the respective secondary antibody coupled to biotin for 1 hour (Table 2). Before washing the slides with distilled water three times they were incubated with a peroxidase-coupled streptavidin–biotin complex. Finally, the slides were blued with hemalaun and embedded in Aquatex (Roche, Mannheim, Germany). 

Immunofluorescence

Presumptive lacrimal gland stem cells cultured on glass cover slips were fixed for 5 minutes in methanol:acetone (7:3) at −20°C and washed three times in PBS. After incubation with goat serum at room temperature for 20 minutes, the cells were incubated with 0.1% BSA in TBST containing primary antibodies (Table 2) for 1 hour at 37°C in a humid chamber. The cells were then washed three times in PBS before incubation with specific Alexa 488–labeled secondary antibodies (Table 2) respectively, diluted in PBS for 1 hour at 37°C, also in a humid chamber. After additional washing with PBS three times, the cells were covered with PBS/glycerol solution containing 4′,6-diamidino-2-phenylindole (DAPI; 1:1000) and were analyzed by using fluorescence microscopy (Keyence BZ 8000; Keyence, Osaka, Japan). 

Results

Stemness Markers Are Detected in Adult Murine and Human Lacrimal Gland

Typical markers for stemness of embryonic and ASCs were detected in adult murine lacrimal gland and in human lacrimal gland. Reverse transcription–PCR revealed mRNA expression of Nanog, Oct4, Sox2, Pax6, and Nestin in these tissues (Fig. 1). With immunohistochemistry and immunofluorescence, positive cytoplasmic reactivity for Nanog, Oct4, and Sox2 was visible in the ductal epithelium of murine and human lacrimal glands, but also in salivary glands of human and mice used as positive controls (Figs. 2, 3). To ensure that the extraorbital lacrimal gland was prepared and used for investigation, and not a salivary gland, three main genes characteristic of the murine lacrimal gland were analyzed: odorant binding protein 1a (ODR1a), lactotransferrin (LTF), and lacrein (Fig. 4); they were detected at the mRNA level. The results of investigation of human lacrimal gland for Nanog and Oct4 at the mRNA level were verified by Western blot analyses and immunofluorescence. Immunostaining revealed positive cytoplasmic signals for Nanog and Oct4; Western blot analyses showed positive signals for Nanog and Oct4 (Fig. 5). 

Figure 1

 

Reverse transcription–PCR analyses of Oct4, Sox2, Nanog, Pax6, Nestin, and tubulin in murine lacrimal and salivary glands and human lacrimal glands. Polymerase chain reaction controls were performed with cDNA from embryonic stem cells (esc) as positive control and without cDNA template (Ø) as negative control. hlg, human lacrimal gland; mlg, murine lacrimal gland; msg, murine salivary gland.

Figure 2

 

Immunohistochemical analyses of murine and human lacrimal glands for Oct4. (a) Positive cytoplasmic red staining for Oct4 in acinus cells (arrow) with negative reactivity of connective tissue (arrowheads) in human adult lacrimal gland (hlg). (b) After use of trypsin-positive cytoplasmic and membrane immunoreactivity for Oct4 (red staining) of the ductal epithelium in hlg. (c) Positive cytoplasmic red immunoreactivity for Oct4 in the excretory duct epithelium (arrow) in murine adult lacrimal gland (mlg). (d) Without use of hemalaun, positive nuclear staining for Oct4 (arrows) in acinus cells of mlg. (e, f) Control sections of human salivary gland (hsg) and msg reveal similar positive (red) immunoreactivity for Oct4 in excretory duct epithelium (arrows in [e] and [a] section through an excretory duct in [f]).

Figure 3

 

Immunofluorescence analyses of stem cell markers Nanog, Oct4, and Sox2 in lacrimal glands of mice. Positive cytoplasmic antibody reactivity in cells of the lacrimal gland for Nanog (a), Oct4 (b), and Sox2 (c). Control sections from parotid gland (d–f) and pancreas (g–i) for the same markers; green staining indicates the positive antibody reaction. Scale bar: 50 μm.

Figure 4

 

Messenger RNA expression of investigated pLGSCs. Investigated passage numbers 0, 1, 2, 5, 10, 20, and 30 for pluripotency, endodermal, ectodermal, and mesodermal development, as well as lacrimal gland–specific markers and markers specific for lacrimal gland development; tubulin was used as positive control. P, passage number.

Figure 5

 

Immunofluorescence and Western blot analyses of Nanog and Oct4 in adult human lacrimal gland and embryonic stem cells. (a, d) Positive blue staining for DAPI of nuclei is demonstrated in slices of adult human lacrimal gland. (b) Green cytoplasmic staining of cells at the ductal epithelium positive for Nanog in the middle of the micrograph. (c) Merging of positive green-staining ductal cells or negative parenchyma cells. (e) Green cytoplasmic staining of cells at the ductal epithelium positive for Oct4 in the middle of the micrograph. (f) Merging of positive green-staining ductal cells and negative parenchyma cells. Control sections (secondary antibody only) were negative (unstained) for each tissue. (g) Positive results in Western blot investigations of adult hlgs for Nanog (upper row) and Oct4 (lower row). esc, positive control, embryonic stem cells.

Messenger RNA of Stem Cell Markers and Early Lineage Markers of All Three Germ Layers Are Expressed in All Investigated Passage Numbers

Samples were investigated for mRNA expression of typical embryonic stem cell markers via RT-PCR analysis. Specific cDNA amplification products (Sox2, Nanog, Nestin, Klf4, c-Myc) were detected in samples (n = 5) of passages 0, 1, 2, 5, 10, 20, and 30 (Fig. 4). GATA4 and GATA6 (endodermal development), Pax6 and neurofilament (ectodermal development), bone morphogenetic protein (BMP) 4 and BMP7 as well as α smooth muscle actin (α-SMA; mesodermal development) were detected in all investigated samples. The markers BMP7, fibroblast growth factor 10 (FGF10), and fibroblast growth factor receptor type 2 (FGFR2), described by Zoukhri et al.^21,22 (early development of lacrimal gland), were detected in all samples but not in adult lacrimal gland tissue (Fig. 4). 

All cells of passages higher than number 2 (after 14 days of culture) did not express mRNA of lacrein, lactotransferrin, and ODR1a (Fig. 4). Investigation for late-lineage markers as a sign of progressed development showed negative results for albumin, somatostatin, and acidic glial fibrillary peptide (data not shown). As a negative control of each PCR, cDNA synthesis was performed without reverse transcriptase. As positive controls we used murine liver for endodermal development, murine brain for ectodermal development, and stomach for α-SMA, FGF10, and BMP7. Embryonic stem cells were used as the positive control for all markers of stemness, BMP4, and FGFR2. All base pair values had the expected fragment size. Sequencing of the bands showed agreement with the corresponding murine DNA sequence of 90% to 100% (sequence alignment with BLAST, GenBank; data not shown). 

Proteins of Stem Cell Markers, Lineage Markers, and Markers of Early Lacrimal Gland Development Were Detected in Adult Murine Lacrimal Glands and Cultured pLGSCs

Protein samples of passage numbers 1, 7, 12, and 22 showed clear results for typical stem cell factors such as Nanog (35 kDa), Sox2 (34 kDa), Nestin (200 kDa), and c-Myc (51 kDa). Western blot also revealed positive results for α-SMA (43 kDa), FGF10 (20 kDa), and its receptor FGFR2 (120 kDa) (Fig. 6). Tubulin was used as a control marker. 

Figure 6

 

Protein expression of markers for stemness and differentiation in pLGSCs of different passage numbers. Western blots of passage number 1, 7, 12, and 22: Proteins were separated by SDS-PAGE so distinct marks for (a) Nanog (35 kDa), (b) c-Myc (51 kDa), and (c) Sox2 (34 kDa), (d) Nestin (200 kDa), and also (e) FGFR2b (130 kDa), (f) FGF10 (19 kDa), and (g) α-SMA (43 kDa) were detected in all investigated passages. Tubulin was used as a control (h).

Positive nuclear staining in cells of passage number 10 was detected for Nanog, Sox2, and c-Myc by using immunofluorescence. Among positively stained cells there were also nonreactive cells. Positive cytoplasmic staining for FGF10, FGFR2, and α-SMA verified the previously mentioned results (Fig. 7). 

Figure 7

 

Immunofluorescence analyses for different markers of stemness and differentiation of cultured pLGSCs (passage number 10); merge of DAPI (blue) and Alexa 488 (green) staining. Positive Alexa 488–stained nuclei merged with DAPI staining appear as cyan staining. (a) Positive cyan staining of nuclei for Nanog is observed for several pLGSCs in between negative cells stained blue. (b) Cyan-stained nuclei positive for Sox2 in the middle of the micrograph; other cells (blue) are negative for the respective antigen. (c) c-Myc–positive nuclei are located on the right side of the micrograph. (d) Positive green cytoplasmic staining for Nestin as a neurofilament marker around blue nuclei. (e) Positive cytoplasmic reactivity of pLGSCs for FGF10. (f) Discrete cytoplasmic reactivity of pLGSCs for FGFR2b. (g) Positive green cytoplasmic reactivity for α-SMA filaments is visible. Control sections (secondary antibody only) were negative (unstained) for each tissue (not shown).

Light Microscopic Analysis of Isolated pLGSCs Show Their Changing Phenotype in Two- or Three-Dimensional Growth

Growth of pLGSCs isolated from adult murine lacrimal glands succeeded over 65 passages, for approximately 1 year. After 48 hours of culturing, the first medium change was performed. At this point, stages of early development were observed, revealing colonies of small cells with a high nuclear to plasma ratio and a few pseudopodia in a very small amount per square centimeter. These cells appeared with a small, spherical, fibroblast-like phenotype. After 96 hours of proliferation, the cells had grown larger and changed their phenotype to large and flat. First digestion with trypsin led to loss of many differentiated cells. The adherent pLGSCs stopped growing or revealed delayed growth, compared to cells of the following passage numbers. Between passages 2 and 3, the cells were classified in two main types. The first cell type was distributed all over the bottom of the culture flask. These cells were very big and flat with a low nuclear to plasma ratio. Distributed among these cells, very small islands of the second type of cells were visible (Fig. 8). After the next passage this cell type became the main type. Only a few scattered cells with a round, fried-egg–like appearance remained. They appeared only as a minority of the population during the monitored passages. After culturing of the pLGSCs (passage number 5) in hanging drops, the organoid bodies also adhered to the culture flask bottom. After 24 hours an outgrowth of cells was visible. Occasionally, these cells formed new three-dimensional networks or cell colonies without additional hanging drop culture (Fig. 9). Cultured pLGSCs exhibited the same growth rate after cryopreservation in liquid nitrogen without loss of viability. Lacrimal gland stem cells did not show any signs of malignant transformation, so there were no atypical nuclei and contact inhibition was apparent. 

Figure 8

 

Morphology of putative cultured lacrimal gland stem cells at passages 0, 1, and 2 after extraction of an adult murine extraorbital lacrimal gland. (a) Passage 0 after 48 hours of culture, small adherent cell heaps. (b) Passage 0 after 4 days, cells became larger, 80% confluence is reached. (c) Passage 0 after 7 days, 95% confluence is reached, a lot of differentiated cells are present between apoptotic cells (black clouds) (cells were split). (d) Passage 1 after 48 hours, cells adherent again are larger (black edging) than in passage 0. (e) Passage 2, large and flat cells (black edging) have the greatest amount. (f, g) Passage 2, between large and flat cells (described in [d, e]) heaps containing small cells with low nuclear plasma ratio were detectable. (h) Passage 3, small fusiform cells are the dominant cell type after a further digestion step using trypsin, and for the rest of the culture period.

Figure 9

 

Fried-egg–like cells and hanging drop observations. (a, b) Round, fried-egg–like cells (exemplarily shown at passage number 10, arrows), which occurred in every passage number (normal cell culture and hanging drop culture). (c) Passage number 5; adherent organoid body (black cloud) in the middle of the micrograph circled emigrated cells. (d) New three-dimensional formation of pLGSCs without additional hanging drop culture (arrow).

Transmission Electron Microscopy Investigations Reveal Mostly Undifferentiated Cells, Although a Few Signs of Differentiation Are Visible

Most of the isolated pLGSCs (passage number 5) showed uniform appearance regardless of whether they were cultured in culture flasks or in hanging drops. Neither the mitochondria counts nor the size of the rough endoplasmic reticulum was increased. Even though we observed no nuclear abnormalities, different types of chromatin could be detected. Cells containing high levels of heterochromatin were detected at the same level of frequency as cells containing euchromatin. Surprisingly, even at a high passage number (passage numbers 15 and 35) a few cells were detected that showed signs of high differentiation. Cells with intermediate-like filaments or large numbers of small electron-dense bodies, such as glycogen granules or actin filaments, appeared. In other cells, lamellar bodies were detected that are characteristic for pneumocytes. Small spherical cells similar to smooth muscle–like cells or fibroblasts were also found, as well as cells containing structures resembling thyroid follicles. Cells with lysosomes and with a low nuclear to plasma ratio indicate the existence of macrophages, but they have not yet been investigated or verified. Tight junctions as a sign of epithelia or extracellular matrix–like structures were observed but not investigated further at this point (Fig. 10). 

Figure 10

 

Transmission electron microscopic analyses of various morphologies of pLGSCs of passage number 35. (a) Four different cell types (upper left arrow: resembling macrophage-containing lysosomes [lower left arrow] and euchromatin; two unspecialized cells appear in the lower part of the micrograph; right arrow: pLGSC core containing much heterochromatin). (b) Magnified lysosomes of micrograph (a). (c) Filaments (arrows) range from outer nuclear membrane (star) to inner plasma membrane. (d, e) Oblique cut filaments (arrows). (f) Small black vesicles resembling glycogen granules (arrows). (g, h) Thyroid-like follicles (arrows) of different sizes. (i, j) Multivesicular (-lamellar) bodies (arrows) at different stages of development regarding their filling level. (k) Example of observed tight junction (arrow). (l) Extracellular congregation between two pLGSCs (star). (m) Spherical fibroblast-like smooth muscle cells. Star indicates the nucleus.

Discussion

Progenitor cells are necessary to regenerate sick or destroyed tissue in vivo. Their high level of ability to divide and give rise to higher specialization is useful in creating tissue-engineered surfaces and transplants. To date, according to the literature, pluripotency is limited to embryonic stem cells²³ and induced pluripotent stem cells but there is still an ethical debate about their extraction or usage. An “ideal” stem cell type should be easily detectable, isolable, and able to regenerate all cell types of the tissue of origin from all germ layers. Glands such as testes or pancreas consist of tissue with a high potential of regeneration.^24,25 As a first step in the present study, markers for “stemness” were characterized in lacrimal and salivary glands of mice and human lacrimal glands. Of these, Nanog and Sox2 are typically expressed in pluripotent stem cells. Oct4, a common transcription factor detectable in embryonic stem cells,^26,27 is expressed in isolated mesenchymal stem cells of bone marrow²⁸ but also in malignant cells.^29,30,31 The transcription factor Pax6, which is involved in the early development of the eye,^32,33 and Nestin—an intermediate filament protein—were measured in the previously mentioned tissues. There is thorough discussion of Nestin expression, but in different contexts. Nestin is designated as a marker for neuroectodermal differentiation; further, it is discussed not only as a stem cell marker for neural tissue³⁴ but also as a marker for stem cells in general.³⁵ According to the literature, two types of lineage-restricted stem cells exist in mammary glands. These cells are able to develop into either myoepithelial or luminal progenitors.³⁶ In sweat glands, others have also described unipotent myoepithelial and luminal progenitors capable of wound repair, in contrast to transplantation assays where they are able to build or support generation of glands, ducts, or stratified epidermis.³⁷ The mitotic activity is not limited to the basal layer of the epithelium in sweat glands.³⁷ This may support our findings of positively immunolabeled cells in the epithelium and the lacrimal duct in adult murine and human lacrimal glands. Literature is lacking that describes the localization of cells with a higher regenerative potency in lacrimal glands. Thus, further investigations with cross-labeling and immunolabeled cell sorting need to be performed to answer the question of whether pLGSCs are localized basally or suprabasally in the glandular epithelium or in the excretory duct system and whether these cells are able to regenerate glandular tissue in a unipotent or pluripotent manner. 

The reactivity with the Oct4 antibody should show nuclear pattern to verify functionality of the Oct4 transcription factor. However, Oct4 can also occur cytoplasmically in embryonal stem cells, cultured progenitors, as well as in differentiated adult cells.^38,39–41 This is because Oct4 is translated in the cytoplasm and staining in cytoplasm depends on the function of the four Oct4 isoforms. The diverse functions of cytoplasmic Oct4 protein localization have been well summarized in the review by Wand and Dai.⁴² Further, the scattered and weak/almost-absent nuclear staining is plausible in ASCs and progenitor cells because uniform strong nuclear staining of Oct4 contributes to the Oct4 expression pattern in pluripotent stem cells. There are multiple Oct4-like genes and proteins in both mice and humans that have the same DNA binding domain but have differences in the transcription activation domain.⁴³ Moreover, Oct4 is described in somatic cells as a stress protein.⁴² Both the amount and the time-dependent expression of Oct4 staining are variable.⁴¹ When functioning as a stress protein or in isolated progenitors of the pancreas, a cytosolic immunoreactivity is possible.^42,38 

Evidence of just one marker of stemness should be discussed with caution; further investigations should show concomitant expression of markers for stemness with double or multiple staining. In recent years, there have been advances in the discovery of potential markers of stemness in lacrimal glands.⁴⁴ However, none of the current discussions establish a marker of stemness unique to lacrimal glands. This is not the case for other eye tissues or the ocular surface. For example, ABCG2, an ATP-binding cassette (ABC) transporter, is a confirmed limbal stem cell marker,^45,46 but whether it is a reliable marker for lacrimal gland stem cells has not been adequately confirmed. ΔNp63, a transcription factor of the p53 family, is also a well-described and reliable marker for limbal stem cells.^47,48 ΔNp63 is also discussed as a marker for lacrimal gland stem cells.⁴⁷ The pluripotency markers used in this study are also used in other current studies as evidence for lacrimal gland progenitors.^45,49 So in a first summary, cells showing signals of stemness and early gland development were detected, and evidence of progenitors is possible within excretory epithelial cells in adult murine and human lacrimal glands as well as murine salivary glands. 

In the second part of this study, three-dimensional clusters were only observed after hanging drop culture of pLGSCs and were not detected in further passages. Taken together, multiple divisions, the formation of clusters, and evidence of Nestin have been designated as a sign for stemness.^45,49 In this regard, it would be of interest to analyze organoid bodies containing pLGSCs to detect all three germ layers of teratoma-like morphology. However, hanging drops from lacrimal stem cells should not necessarily show a teratoma-like morphology because this is a unique property of pluripotent stem cells and a few other ASC types. The generation of embryonic organoids is the property of ASCs and progenitor cells such as neural stem cells that form neurospheres. Nevertheless, it would be interesting to check whether derivatives of three different germ layers can be generated from organoid bodies to further characterize the differentiation capacity of lacrimal stem cells, but we think that such an analysis is not necessary to verify the stemness of those lacrimal stem cells. 

The morphology of the cultured pLGSCs resembles a mesenchymal or myoepithelial phenotype. These spindle-shaped, fibroblast-like cells have been previously described by You et al.⁴⁹ and Shatos et al.⁴⁵ as potential progenitors in glands, more specifically, in lacrimal glands. After two passages of the cultured cells, different phenotypic features without senescent cells were detected. Spindle-shaped, spherical small cells were verified by light microscopy. Interestingly, large round cells resembling fried eggs appeared in every passage number in very low counts. Such fried-egg–like cells are also mentioned in other studies.⁵⁰ Their function is still unknown. Hypothetically, they could exert regulatory or protective influence on the cultured progenitors. The TEM results showed no signs of higher differentiation or specialization of the cultured cells with the exception of only a few more highly developed cells. They were detected after hanging drop culture. These few cells, or rather cell organelles, were not characterized at the mRNA or protein level, so these considerations remain speculative and their accuracy will require confirmation by further investigations to exclude surviving contaminants from the initial isolation process. Morphology is of course less conclusive than the expression pattern of the cultured cells in identification of progenitors, or even stem cells. Markers for stemness, but also the transcription factors c-Myc and Klf4, familiar from embryonic stem cell research, were detected in every investigated passage number. c-Myc^51,52 and Klf4⁵³ have been used to induce pluripotency in reprogrammed somatic cells (IPS cells) in other investigations.^14,54 Owing to simultaneous detection of different markers for stem cells, it can be assumed that the cultured cells are more closely related to progenitors or cells with stem cell properties than the cells in other investigations in which only one or two markers for stemness have been detected.⁴⁹ The nuclear, but not the cytoplasmic evidence (with the exception of Nestin), of these transcription factors indicates correct gene activity. This was confirmed by immunofluorescence. All investigated passages contained high counts of cells positive for c-Myc, Nanog, and Sox2 among negative cells. This phenomenon—staining an active, nuclear form of markers of stemness of pLGSCs—contradicts our results of immunostained murine and human lacrimal glands, where cytosolic staining appeared. We speculate that the isolation method and the in vitro culture of pLGSCs itself reactivates progenitors, so more cells with nuclear staining of markers for stemness would then be detected. Nestin expression was detected with cytoplasmic staining in a few cells as well. Possibly not all cells retain stem cell character, so that unregulated spontaneous differentiation is initiated in many cells. This hypothesis is underlined by the expression of differentiation markers from all three germ layers. As the results of the PCR investigations showed, markers of early differentiation were detected at the mRNA level. Evidence of the transcription factor Pax6, as well as neurofilaments of different molecular masses, suggests ectodermal development. Additionally, mRNA of GATA proteins 4 and 6^55,56 was detected as markers for early endodermal development. Transcripts of BMP4 and BMP7,^21,22,57,58 which play crucial roles in mesodermal development, were detected in all investigated passages, as well as α-SMA,⁵⁹ which is expressed by adult mesenchymal cells, in addition to the evidence of markers for stemness. To sum up, markers of stemness and of all three germ layers were detected simultaneously by RT-PCR. However, these data must be interpreted with caution, as RT-PCR can yield false-positive results and α-SMA is expressed not only by adult mesenchymal cells, but also by myoepithelial cells of lacrimal and salivary glands; moreover, we neither analyzed serial sections nor performed coexpression studies to confirm that the markers are coexpressed by the same cells. Further, the main features of our pLGSCs—preservation of a state capable of division and differentiation into cells of three germ layers—have not been previously described for cultured cells extracted from murine lacrimal glands. Our isolation method for pLGSCs was methodically easier and more practicable than those used in other current studies,^22,49 since there was no need for externally induced gland inflammation before isolation of potential lacrimal gland progenitors as described by other work groups.^22,59 The interaction of potential niche cells that are progenitors, or stem cells, is maintained by the use of collagenase^60–63 during the isolation procedure. Maybe this is the key to long-term growth of progenitors and differentiated cells together over such a long period. The “stem cell niche,” as it is known from the corneal limbus,^60–63 is also supposed to exist in the lacrimal gland, but has not yet been detected. The long-term goal of our investigations is to stimulate adult lacrimal gland stem cells in vivo or generate lacrimal gland-like tissues of cultured progenitors by means of tissue engineering. Reverse transcription–PCR investigation based on detection of mRNA of the three main genes⁶⁴ of murine lacrimal gland (see above) in the first two passages was used to exclude misisolation of pLGSCs of glands other than the lacrimal gland. The fact that there was no further expression of these markers during the entire culture process confirmed the thesis that adult, differentiated cells could not adhere in culture flasks after multiple splits. Therefore, adult cells could not be cultured over a long period and disappeared. Surprisingly, markers of early lacrimal gland development were detected in all investigated passages. Bone morphogenetic protein 7 was detected at the mRNA level. Together with the detected marker, BMP4 plays a crucial role in the branching morphogenesis of the lacrimal gland.^{21,22,57,58,65} Subsequently, positive results for FGF10 attributed to its receptor type 2b were recorded with PCR. As already mentioned in the studies of Zoukhri et al.,²¹ FGF10 and FGFR2b are required to renew damaged gland tissue in lacrimal glands and the pancreas.^66,67 Furthermore, FGF10 and Pax6 interact during the embryonic development of the lacrimal gland.^68,69 Western blot analyses and immunofluorescence investigations verified the protein expression of FGF10 and FGFR2b. It will be necessary to examine the environment of embryonic development of the lacrimal gland and the dependent steps of differentiation, and to search for the putative stem cell niche in further investigations. In particular, the interplay of mesenchymal cells and potential stem cells remains poorly understood and requires further studies. 

In summary, our investigations provided an initial overview of the possibility of culturing lacrimal gland progenitor cells and their potential development. In this connection, the existence of progenitors in murine and human lacrimal glands was demonstrated. We suggest that these cells do not arise, or do not arise exclusively, by transdifferentiation as has been observed in lacrimal gland and pancreas.^70–74 This means that in an artificial environment such as an in vitro culture, a change from epithelial to mesenchymal cells is possible. In addition, Rapoport et al.³⁸ have proposed dedifferentiation from adult epithelial to stem cell–like cells as a consequence of the isolation process also used in our investigation. Transdifferentiation and dedifferentiation could explain the higher amount of cultured pLGSCs in contrast to the low amount of pLGSCs in vivo. Further investigations will be required (1) for lineage tracing to search for different unipotent or one multi/pluripotent stem cell in vivo and (2) to provide sufficient verification of their potential stem cell character. Therefore, three-dimensional culture using hanging drop or 3D-Matrigel cultures, as well as coculture to force specific differentiation, are needed. The isolated lacrimal gland progenitor cells could possibly be used in artificial lacrimal glands by means of tissue engineering,⁶ or LGPCs could be used in development of new models to treat cases of aqueous tear deficiency such as occur during DED. 

Acknowledgments

We thank Susann Möschter, Gerti Link, and Marco Gößwein for excellent technical assistance as well as Nicolas Keller and Uwe Rückschloss for providing mESCs and the wild-type mice. 

Supported in part by Deutsche Forschungsgemeinschaft (DFG) Grants PA738/9-1 and PA738/9-2 as well as Sicca Forschungsförderung of the professional Association of German Ophthalmologists. The authors alone are responsible for the content and writing of the paper. 

Disclosure: P. Ackermann, None; S. Hetz, None; J. Dieckow, None; M. Schicht, None; A. Richter, None; C. Kruse, P; I.S. Schroeder, None; M. Jung, None; F.P. Paulsen, None