The Influence of Oxygen on the Proliferative Capacity and Differentiation Potential of Lacrimal Gland–Derived Mesenchymal Stem Cells

Abstract

Purpose: The application of lacrimal gland–derived mesenchymal stem cells (LG-MSC) for the regeneration of lacrimal gland tissue could result in a novel therapy for dry-eye syndrome. To optimize the culture conditions, the purpose of this study was to evaluate the influence of low oxygen on phenotype, differentiation potential, proliferative, and regenerative capacity of murine LG-MSC.

Methods: Murine LG-MSC were cultured in 21% and 5% oxygen and characterized by flow cytometry, cell sorter assisted proliferation–, and colony forming unit–assays. Reactive oxygen species (ROS) levels as well as lineage differentiation were evaluated. The effect of conditioned medium of LG-MSC from both oxygen conditions (CM^{MSC 21%}, respectively, CM^{MSC 5%}) on lacrimal gland epithelial cells (LG-EC) was examined in wound healing and proliferation assays.

Results: Cells under both culture conditions revealed differentiation potential and presented a MSC-specific flow cytometric phenotype. In 5% oxygen, cells yielded less ROS, showed a stable morphology, higher colony forming potential, and an increased proliferation capacity. Five percent oxygen significantly increased the number of CD44+ LG-MSC. Furthermore, CM^{MSC 5%} significantly enhanced migration and proliferation in LG-EC.

Conclusions: In vitro expansion in low oxygen preserves the proliferation capacity and differentiation potential of LG-MSC and increases the effects of conditioned medium on migration and proliferation in LG-EC. Therefore, expansion in low oxygen seems to be an excellent method, to obtain vital MSC. Also, an increased number of LG-MSC expressing CD44 was observed under low oxygen, which might be a valuable marker to identify a potent MSC subpopulation.

Dry-eye syndrome, also known as keratoconjunctivitis sicca, is a multifactorial chronic disabling disease, caused by qualitative or quantitative tear deficiency.¹ According to the International Dry Eye Workshop report, the global prevalence of dry eye syndrome is estimated between 5% and 35%, affecting various ages (21–65 years).² Quantitative tear deficiency due to primary or secondary lacrimal gland (LG) insufficiency is one of the major causes for the development of severe dry-eye syndrome. The patients suffer from severe pain and significant deterioration of visual acuity.³ In these severe cases, tear substitutes are not sufficient to relief symptoms and to date, no causative treatment for dry-eye syndrome exists. The understanding of mesenchymal stem cells (MSC) and the use of those cells for the regeneration of LG tissue could result in a novel therapy for severe dry-eye patients. 

Several studies showed the high therapeutic potential of MSCs: the outcome after myocardial infarction,^4–6 acute renal failure,^7–9 neonatal stroke,^10,11 osteogenesis imperfect,¹² multidrug resistant tuberculosis,¹³ fulminant hepatic failure and further diseases was improved by the application of MSC,^14,15 even in the absence of apparent MSC long-term engraftment.¹⁶ Although the precise therapeutic mechanisms still remain unclear, so far the prevailing view is, that MSCs home to sites of injury or inflammation, secrete trophic factors to promote the recovery of injured cells and support recruitment as well as proliferation of resident progenitor cells to replace damaged cells. Furthermore, they participate in regeneration through matrix remodeling, immunomodulation, and anti-inflammation.¹⁷ Also, some evidence exists that MSC might contribute to the reconstitution of epithelial tissue by differentiation, as demonstrated for the liver and the cornea.^18–22 

Due to specialization in subpopulations in different tissue, orthotopic transplantation of MSC will assumable show the best results as a possible therapy for dry-eye syndrome.²³ The existence and possibility of isolating LG–derived MSC (LG-MSC) has been shown.²⁴ And it was demonstrated that after experimentally induced inflammation of the LG, the number of MSC in the LG tissue increased, and that formation of extracellular matrix corresponds to the location of those cells.^25–27 

Prior to an application of LG-MSC for the regeneration of LG tissue in an animal model and potentially later on in clinical settings, those cells need to be characterized thoroughly. Especially factors influencing their differentiation potential and proliferation, such as isolation and culture conditions, are of great importance. Oxygen concentration is one of the most important determinants of tissue metabolism: changes in morphology and gene expression, growth rate and survival are strikingly modified by altered oxygen concentrations.^28–30 Recent studies suggest, that the oxygen concentration plays an important role especially in murine MSC cultures.³¹ The aim of this study was to evaluate the influence of oxygen concentration reduced to the physiologically relevant value on the phenotype, the proliferative capacity, and the differentiation potential of murine LG-MSC in order to improve the maintenance of LG-MSC in vitro for transplantation purposes.³² 

Materials and Methods

Mice

C57BL/6 mice (8–10 weeks old, male) were used for the study. All experiments were conducted in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and the guidelines of the Federation of Laboratory Animal Science Association (FELASA). 

Explant Culture

After euthanasia both extraorbital LGs were removed immediately. The glands were washed in culture medium (α-MEM; Biochrom, Berlin, Germany), 2 mM L-glutamine (Biochrom), 1% penicillin/streptomycin (Invitrogen, Carlsbad, CA, USA), 15% fetal bovine serum (FBS; Biochrom) and cut into small pieces (< 1 mm). The pieces were placed on a 6-cm uncoated culture dish (Sarstedt, Nümbrecht, Germany) and allowed to adhere for 5 minutes before culture medium was added. 

All experiments were set up in triplicates. Cells were grown on tissue culture treated plastic (Sarstedt) either under routine culture conditions, respectively, high oxygen (21% O2, 5% CO2, 37°C) or under low oxygen (5% O2, 5% CO2, 37°C). Media was replaced every 2 to 3 days. At approximately 80% confluence, the cells were passaged by trypsinization (Trypsin 0.05%; Invitrogen) and reseeded at a density of 2500 cells/cm² in a T25 flask. At passage 1 and 5 cells of both culture conditions were expanded for further analysis in T75 flask. 

Flow Cytometry

Cells from both culture conditions in passage 1 were trypsinized, collected, and incubated with phycoerythrin- or fluorescein isothiocyanate-conjugated antibodies against murine CD29, CD31, CD34, CD44, CD45, CD49f, CD54, CD90.2, CD105, CD106, CD117, CD166, SCA-1, and IgG (Biolegend, San Diego, CA, USA; R&D Systems, Minneapolis, MN, USA) for 30 minutes at 4°C. Excess antibody was removed by two washes with staining buffer. Fluorescent labeling (minimum 10,000 events) was detected on a calibrated Gallios cytometer (Beckman Coulter, Brea, CA, USA). For measurement of reactive oxygen species (ROS) a Total Reactive Oxygen Species Assay Kit was used (eBioscience, San Diego, CA, USA). Cells were incubated with ROS assay stain solution for 2 hours in 21% and 5% oxygen, respectively, and afterward analyzed on a flow cytometer (FCS 500; Beckman Coulter) according to the manufacturer's instructions. Histograms were created with FlowJo software (Treestar, Ashland, OR, USA). 

Cell Proliferation

Viable Cells from a single cell suspension were sorted by flow cytometry (Mo Flow XDP; Beckman Coulter) with a 100-μm nozzle at 25 PSI into 6-cm culture dishes (10,000 cells per dish). After 7 and 14 days the dishes where trypsinized, stained with propidium iodide and the viable cells were counted by flow cytometry (FCS 500; Beckman Coulter). 

Cumulative Population Doublings

In order to calculate the cumulative population doublings, in every passage up to passage 5 the cells were counted initially after trypsinization and after harvesting at the following passage. 

Colony Forming Unit (CFU) Assay

Viable cells from a single cell suspension were sorted by flow cytometry as described above into 6-well plates at two densities (100 cells/well and 1000 cells/well). Cultures were set up in triplicate and incubated at the respective oxygen level. Medium was changed every 2 to 3 days. After 2 weeks, colonies were washed with PBS and then fixed for 30 minutes with methanol at −20°. For enumeration the fixed cultures were stained with 1% Rhodamine red (Sigma-Aldrich Corp., St. Louis, MO, USA) for 30 minutes, then rinsed with water and allowed to dry. As described earlier, aggregates of greater than 50 cells were counted as CFU using an inverted microscope (DM 750; Leica Microsystems, Wetzlar, Germany).³³ 

Differentiation Assays

To proof the multilineage differentiation potential of LG-MSC, adipogenic, and osteogenic differentiation was induced in both culture conditions. Cells from both culture conditions were trypsinized at passage 1 and 5 and reseeded into 6-cm culture dishes in culture medium. When the cells were subconfluent, medium was changed to osteogenic medium (adipogenic medium [α-MEM]), 2 mM L-glutammine, 1% pen/strep, 10% FBS, 1 × 10⁻⁸ M dexamethasone (Sigma-Aldrich Corp.), 0.05 mM ascorbic acid 2 phosphate (Sigma-Aldrich Corp.), 10 mM Beta glycerophosphate (Sigma-Aldrich Corp.) or α-MEM, 2 mM L-glutammine, 1% pen/strep, 10% FBS, 5 μg/mL insulin (Sigma-Aldrich Corp.), 1 × 10⁻⁶ M dexamethason (Sigma-Aldrich Corp.), 50 μM indomethacin (Sigma-Aldrich Corp.), 0.5 μM 3-Isobutyl-1methylxantine (Sigma-Aldrich Corp.), and replaced every 2 to 3 days. After 7, 14, and 21 days, cultures were collected for analysis. As control, cells in standard culture medium were collected at day 0 and 21. 

The differentiation potential was quantified by quantitative real-time polymerase chain reaction (qRT-PCR). Differentiation was visualized by staining fixed adipogenic cultures with Oil-Red-O solution and osteogenic cultures with Alizarin-Red as described earlier.^34,35 

Quantitative RT-PCR

For qRT-PCR, cells were harvested from culture dishes, RNA was isolated, and cDNA was synthesized with Super Script II Reverse Transcriptase kit (Invitrogen), RNAse Out Inhibitor (Invitrogen), 100 mM dNTP-Set (Invitrogen), and Oligo(dt) 12-18 Primer (Invitrogen) according to the manufacturer's protocol. To evaluate the adipogenic differentiation the expression of fatty acid binding protein 4 (FABP4) was examined. The forward primer of FABP4 was TGAAATCACCGCAGACGACA, the reverse primer ACACATTCCACCACCAGCTT. For the evaluation of osteogenic differentiation, the expression of Osteopontin (OPN) was examined. The forward primer of OPN was CAGTGATTTGCTTTTGCCTGTTTG, the reverse primer GGTCTCATCAGACTCATCCGAATG.³⁶ A 96-well qRT-PCR system (7500 fast; Applied Biosystems) was used for amplification (95°C for 15 minutes, 45 cycles of 94°C for 15 seconds, 59°C for 20 seconds, 72°C for 35 seconds). Amplifications were carried out in triplicates using SYBR green master mix (Invitrogen). No-template-samples served as the negative controls. All measurements were normalized by ribosomal protein S6 (RPS6) expression (forward primer: CTTTTTCGTGACGCCTCCCA; reverse primer: GGGAAGGAGATGTTCAGCTTCA). For data analysis (ΔΔC_t method) SDS Version 2.0.6 (Life Technologies, Carlsbad, CA, USA) and Excel 2011 (Microsoft, Redmond, WA, USA) were used. 

Porcine Lacrimal Gland Epithelial Cell (LG-EC) Isolation and Culture

German domestic pigs (Sus scrofa) were euthanized and the LGs were extracted immediately. The LG-EC were isolated and cultured as previously described.^37,38 Briefly, after removal, the LG tissue was washed, minced, and enzymatically digested by collagenase (350 U/mL, Sigma-Aldrich Corp.), DNAse (40 U/mL, Sigma-Aldrich Corp.) and hyaluronidase (300 U/mL, Sigma-Aldrich Corp.). After digestion and filtering through a 70-μm mesh (Roth, Karlsruhe, Germany) and a ficoll gradient (Sigma-Aldrich Corp.), the cells were cultured in Dulbecco's modified Eagle's medium with glutamax (DMEM/F12) containing 10% FBS (Life Technologies), 0.4 μg/mL hydrocortisone, 0.1 nM cholera toxin, 0.18 mM adenine, 5 μg/mL transferrin, 5 μg/mL insulin (all Sigma-Aldrich Corp.), 10 ng/mL epidermal growth factor, and 1% penicillin/streptomycin (Life Technologies) at a density of 4.0 × 10⁴ cells per cm² on a 3T3/J2 mouse fibroblast feeder layer. For expansion, as well as for experiments, the porcine LG-EC were kept in standard conditions (21% O², 5% CO², 37°C) at all times. 

Immunofluorescence Staining

Porcine LG-EC were fixed with methanol at −20°C for 5 minutes, washed in PBS and nonspecific binding sites were blocked for 1 hour at room temperature with 5% normal goat serum (VWR) in PBS. The cells were incubated with an anti-pan-cytokeratin antibody (antipan-cytokeratin [AE1/AE3] conjugated with Alexa Fluor 488; dilution 1:500; eBioscience, San Diego, CA, USA) and anti-Rab3D antibody (dilution 1:500; Abcam, Cambridge, UK) overnight at 4°C and washed with PBS. For Rab3D-staining, cells were washed again, incubated with the appropriate secondary antibody (dilution: 1:1000; Alexa Fluor) for 1 hour at room temperature and washed again. Nuclear staining and mounting was performed with Prolong Gold Antifade mounting medium with DAPI (Life Technologies). Images were taken with an immunofluorescence microscope (DFC 450; Leica Microsystems). 

Periodic Acid Schiff (PAS)-Alcian Blue Reaction

Porcine LG tissue was embedded in paraffin and cut into 6-μm sections. After deparaffinization, rehydration, and incubation in alcian blue solution for 5 minutes, the sections were washed with running tab water for 3 minutes and rinsed with distilled water. Then the sections were incubated with Schiff's reagent (Merck, Darmstadt, Germany) for 15 minutes and nuclear staining was performed with hematoxylin solution (Roth). 

β-Hexosaminidase Assay

At day 14 porcine LG-EC were seeded in 24 well plates (5 × 10⁵ cells/well) and allowed to adhere overnight. After washing and incubation in serum-free DMEM for 2 hours, a medium sample was removed for the baseline measurement. Afterward, carbachol was added for 15 minutes (100 μM), and a stimulated sample was removed. 4-methylumbelliferyl N-acetyl-β-D-glucosaminide (Sigma-Aldrich Corp.) was used as a substrate in order to measure β-hexosaminidase activity. Fluorescence intensity (360-nm excitation, 450-nm emission) was determined on a microplate reader (FLUOstar Omega; BMG LABTECH, Ortenberg, Germany). The experiment was performed with porcine LG-EC from three different animals. 

Transmission Electron Microscopy (TEM)

The cell pellet was fixed in 2.5% glutaraldehyde with 4% PFA in 0.1 M cacodylate buffer, embedded in Spurr media (Serva, Heidelberg, Germany). Sections of 60 to 80 nm were cut and analyzed with a Hitachi H600 transmission electron microscope (Krefeld, Germany). The images were taken with a Gatan BioScan camera (München, Germany). 

Conditioned Medium

Lacrimal gland derived–MSCs were isolated in both 5% and 21% oxygen as described above and expanded on a T75 flask. When subconfluent, medium was removed, cells were washed with PBS and medium was changed to 20 mL DMEM/F12 medium without any supplements. After 24 hours the conditioned medium (CM) was collected, centrifuged at 188 g for 5 minutes to remove detached MSCs and stored at −80°C until further use. 

Wound Healing Assay

Porcine LG-EC were seeded in migration assay inserts (ibidi, Martinsried, Germany) according to the manufacturer's instructions. Briefly, one insert per well was placed in a 24-well plate and firmly attached. Into each side of the insert, 7 × 10⁴ porcine LG-EC were seeded and allowed to adhere for 24 hours. Medium was removed, cells were washed two times with PBS, and the insert was carefully removed. Afterward, the cells were washed again two times and incubated with CM from both culture conditions (CM^{MSC 5%}, respectively, CM^{MSC 21%}) and DMEM/F12 medium without any supplements (DMEM/F12). In the gap area a position was randomly marked and pictures of the gap at this position were taken at 0, 5, 10, and 20 hours post insert removal. Experiments were performed independently with three different CM per condition. 

Proliferation Capacity After Ethanol Injury

The cell proliferation after injury was measured by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay as described before.³⁹ Briefly, porcine LG-EC were seeded at a density of 1 × 10⁴ cells in a 96-well plate and allowed to adhere for 24 hours. Cells were washed with PBS and incubated with 20% ethanol for 1 minute for injury. After another washing step with PBS, the cells were incubated in CM from both culture conditions (CM^{MSC 5%}, respectively, CM^{MSC 21%}) and DMEM/F12 medium without any supplements as control medium (DMEM/F12) up to 48 hours. Afterward, medium was replaced by 100 μL MTT solution (0.5 mg/mL) and incubated for 1 hour at 37°C and 5% CO₂. MTT solution was discarded and 100 μL of dimethyl sulfoxide (DMSO) was added to the cells to dissolve the formazan crystals. Measurement was done at 540 nm using an ELISA reader upon 0, 6, 24, and 48 hours after treatment. Experiments were performed independently three times with CM^{MSC 5%}, CM^{MSC 21%}, and DMEM/F12 as control. Ten wells were evaluated for each experimental condition. 

Statistical Analysis

The experimental data was analyzed with Prism 6 (Graphpad, La Jolla, CA, USA). All values are shown as mean ± SD. P values were determined using Student's t-test with two-tailed distribution. Differences were considered significant at P less than or equal to 0.5. For the migration and regeneration assays a two-way ANOVA with a Bonferroni post hoc test (P < 0.05) was performed. 

Results

Cell Morphology

After isolation, LG-MSC in both culture conditions presented an MSC-specific elongated fibroblast-like and spindle-shaped morphology (Fig. 1A). Under high-oxygen conditions the MSC underwent morphologic changes over the culture period. In passage 2 these cells developed stress fibers and an increase in cell-size was found compared with cells under low oxygen (Figs. 1B, 1C). Furthermore, under these conditions cells in two cultures became very large with many stress fibers and ceased growth in passages 4 and 5 (Fig. 1D). 

Figure 1

 

Morphology and proliferation of cultured LG-MSC under high and low oxygen. Cells growing out of the explant piece (asterisk) with a fibroblast-like, spindle-shaped morphology in both culture conditions (A). Passage 2 cells with a stable spindle-shaped morphology under low oxygen (B). Passage 2 cells under high oxygen with an increase in cell size, a high cytoplasmic to nuclear ratio and stress fibers (red arrow [C]). Large cells with stress fibers and ceased growth under high oxygen in passage 4 (D). Scale bars: 200 μm.

Flow Cytometric Cell Characterization

Analyzed cells in passage 1 showed a MSC-typical marker expression. Cells in both conditions were positive for CD29, CD49 (+/−), CD90, CD105 (+/−), CD106, CD166, and SCA-1, and negative for CD31, CD34, CD45, CD 54 (+/−), and CD117 as shown in Figure 2A. The marker expression did not differ between the two culture conditions, except for a significant difference in CD44. This marker increased under low oxygen (Fig. 2B). Increased fluorescence intensity, indicating higher ROS levels was detected in cells cultured under high oxygen (Fig. 2C). 

Figure 2

 

Flow cytometric characterization of the phenotype and ROS measurement. Mesenchymal stem cells–typical marker expression of LG-MSC in both culture conditions (A). Increase of CD44 positive cells under low oxygen (B). Higher ROS levels under high oxygen, measured as a shift to an increased fluorescence intensity (C).

Differentiation Assays

In both culture conditions, adipogenic differentiation resulted in the formation of lipid vesicles, microscopically visible approximately 3 days after the initiation of differentiation. Those lipid vesicles could be stained with Oil-Red-O (Fig. 3A). After induction of osteogenic differentiation, calcium deposition was microscopically detectable using Alizarin Red (Fig. 3D). Macroscopically and microscopically no difference between the two culture conditions was noted. The differentiation potential was quantified by qRT-PCR. As an indicator for adipocyte differentiation FABP4 expression was measured, respectively, OPN expression as an indicator for osteocyte differentiation. Under low as well as under high oxygen a significant fold change of the measured genes was examined in comparison to the control (Figs. 3B, 3C, 3E, 3F). The fold change after induction of differentiation was higher for FABP4 under high oxygen, with the basic expression of FABP4 8.8-fold elevated under low oxygen (±1.76, P = 0.01). The fold change of OPN was significantly higher under low oxygen, without a significant difference in the basic expression between the two conditions. 

Figure 3

 

Adipogenic and osteogenic differentiation of LG-MSC under high and low oxygen concentration. Formation of lipid vesicles after adipocyte differentiation stained with Oil-Red-O (A). Calcium deposits stained with Alizarin-Red (D). Significant increase of FABP4 (B, C) and OPN expression (E, F) after induction of differentiation under high and low oxygen. Scale bars: 200 μm.

Proliferation Assay and CFU Assay

Cumulative population doublings are shown in Figure 4A. The average population doubling time was significantly shorter under low oxygen compared with high oxygen (low oxygen: 1.8 d ± 0.18; high oxygen 2.7 d ± 0.37; P = 0.043). Lacrimal gland–MSC cultured in low oxygen resulted in a total cumulative cell numbers of 4.4 × 10⁶ ± 3.5 × 10⁵, compared with 1.29 × 10⁶ ± 1.44 × 10⁵ cells in high-oxygen cultures (P = 0.026) at the end of five passages. 

Figure 4

 

Proliferation of LG-MSC under high and low oxygen concentrations. Cumulative population doublings of LG-MSC in high and low oxygen up to passage 5 (11.7 ± 2.5, respectively, 19.3 ± 0.24; P = 0.015 [A]). Flow cytometric proliferation assay with a significant difference in the total cell number between the two culture conditions in passage 1 day 7 (P = 0.0475), day 14 (P = 0.0276), and passage 5 day 14 (P = 0.0006 [B]).

The flow cytometric proliferation assay (Fig. 4B) revealed a difference between the two culture conditions in passage 1, with a cell count of 1796 ± 387 after 7 days and 2985 ± 432.7 after 14 days in high oxygen, compared with 27789 ± 9189 after 7 days and 132455 ± 38241 after 14 days in low oxygen (day 7: P = 0.0475, day 14: P = 0.0276). In passage 5 the difference in the cell count at day 7 was notable, although slightly not significant (high oxygen: 1528 ± 461.9; low oxygen: 12820 ± 4049; P = 0.0503). At day 14 in passage 5 the flow cytometric cell count showed a highly significant difference between the two culture conditions (high oxygen: 8091 ± 7426; low oxygen: 87006 ± 3103; P = 0.0006). 

The CFU assays showed a highly significant difference between the two culture conditions in passage 1 (high oxygen: 0.2889 ± 0.07536; low oxygen 8.556 ± 1.215; P < 0.0001), but no significant difference in passage 5 (high oxygen: 0.7778 ± 0.2778; low oxygen 1.978 ± 0.8005; Fig. 5). 

Figure 5

 

Colony forming efficiency assay under high- and low-oxygen concentrations. Significantly restricted colony forming capacity under high oxygen (1000 cells seeded in the upper and 100 cells in the lower wells; P = 0.0001 [A, C]). Single colonies under high oxygen, consisting of few very large cells with stress fibers. In contrast colonies consisting of many cells with a small nucleus-plasma-relation under low oxygen. Scale bar: 500 μm (B). Significant decrease of colony forming capacity from passage 1 to 5 under low oxygen (P = 0.0003 [C]).

Epithelial Cell Morphology and Immunocytochemistry

After 3 to 5 days colonies formed, reaching subconfluency after approximately 9 days (Fig. 6A). The porcine LG-EC presented a typical cobblestone-like morphology and an epithelial cell phenotype could be confirmed by positive pan-cytokeratin staining in the flow cytometry and immunofluorescence (Fig. 6B). 

Figure 6

 

Porcine LG-EC. Confluent monolayer of LG-EC with a cobblestone-like morphology at day 5 after isolation (P0 [A]). Immunofluorescence staining of LG-EC (P0) with antipan-cytokeratin antibody (green). Nuclear staining with DAPI (blue [B]). Scale bars: 50 μm.

Secretory Activity of Porcine LG-EC

In the PAS-alcian blue reaction, a purple-colored cytoplasmic staining, typical for neutral mucous secretions could be shown (Fig. 7A). Furthermore, the cells showed positive immunofluorescence for Rab3D, a marker for secretory vesicles (Fig. 7B) and the TEM analysis revealed nonelectron-dense vesicles inside the cells, typical for mucous secretion (Fig. 7C). Also LG-EC revealed a significant increase in secretory activity after parasympathetic stimulation with carbachol, measured in the β-hexosaminidase assay (before stimulation: 51.2 ± 17.9 nm; after stimulation: 116.7 ± 36.8 nm, P = 0.004). 

Figure 7

 

Secretory activity of porcine LG-EC was confirmed after expansion in passage 0: PAS-alcian blue reaction revealed multiple purple colored secretory granula (red arrows). Scale bar: 50 μm (A]). Rab3D staining depicted a granular staining, confirming the evidence of secretory vesicles (green). Scale bar: 50 μm. (B) Transmission electron microscopy analysis shows several vesicles with low electron density typical for mucous substances (red arrows). Scale bar: 1 μm (C).

Wound Healing Assay

The wound healing assay showed a significant difference between the three culture conditions (P < 0.004). Cells incubated in CM^{MSC 5%} displayed a significantly enhanced migration capacity compared with cells incubated in CM^{MSC 21%} and in DMEM/F12 at all time points, whereas cells incubated in CM^{MSC 21%} MSC did not differ from DMEM/F12 incubated cells at 5 and 10 hours. At 15 and 20 hours cells incubated in CM^{MSC 21%} migrated significantly faster than cells in DMEM/F12 (Figs. 8, 9). 

Figure 8

 

Images of porcine LG-EC migration assay. Lacrimal gland–EC incubated in CM^{MSC 5%}_, CM^{MSC 21%}, and unconditioned DMEM/F12. Significantly faster gap closure in CM^{MSC 5%} compared with CM^{MSC 21%}. In contrast very slow closure in DMEM/F12. Scale bar: 200 μm.

Figure 9

 

Quantification of the gap closure. Significantly faster gap closure with CM^{MSC 5%} (5 hours: 17.08 ± 5.22; 10 hours: 32.91 ± 4.36; 15 hours: 49.77 ± 2.92; 20 hours: 76.02 ± 7.38) compared with CM^{MSC 21%} (5 hours: 0.47 ± 0.82; 10 hours: 7.36 ± 5.65; 15 hours: 24.59 ± 10.29: 20 hours: 50.08 ± 20.57) at all time-points. In the graph only significance levels for CM^{MSC 5%} versus CM^{MSC 21%} are shown (*P ≤ 0.05, ***P ≤ 0.001).

Regeneration Capacity After Ethanol Injury

After ethanol injury, the proliferation assay revealed a significant difference between the three culture conditions (P < 0.02). At 48 hours cells incubated in CM^{MSC 5%} proliferated significantly faster than cells incubated in both, CM^{MSC 21%} and DMEM/F12 and cells incubated in CM^{MSC 21%} proliferated faster than cells in DMEM/F12 (Fig. 10). 

Figure 10

 

Regeneration capacity. Significantly higher proliferation capacity after ethanol injury of cells in CM^{MSC 5%} (1.27 ± 0.28) compared with CM^{MSC 21%} (1.00 ± 0.08) at 48 hours (P ≤ 0.05). In the graph only significance levels for CM^{MSC 5%} versus CM^{MSC 21%} are shown.

Discussion

In vitro culture conventionally uses the atmospheric oxygen concentration (21% O₂), widely defined as normoxia, whereas oxygen concentrations below 21% are traditionally called hypoxia in laboratory settings. In vivo the niche of MSCs displays oxygen concentrations between 2% and 7%, therefore the so called hypoxic conditions are rather physiological.^40,41 

Beneficial effects of low oxygen on MSC from different origin have been indicated by several studies.^29,42–47 However, a study by Hung et al.⁴⁸ described a decrease of proliferation rate in human bone marrow–derived MSC (BM-MSC) exposed to 1% oxygen. In this study LG-MSC showed significantly higher population doublings under 5% oxygen compared with 21% oxygen up to passage 5. Additionally, LG-MSC expanded under low oxygen yielded a significantly higher colony forming capacity as well as a higher proliferation rate in the flow cytometry assisted proliferation assay, indicating a beneficial effect of low oxygen on the maintenance of the MSC population during in vitro expansion. It has to be noted that the very low colony forming capacity under high oxygen might be influenced by flow cytometry assisted sorting and that those comparatively larger cells are more susceptible to the damage induced by the sorting procedure.^49,50 

In accordance with the proliferative capacity also the morphology of the cells differed between the two culture conditions in this study, displaying rather large and flattened cells with stress fibers under high oxygen and small, compact spindle shaped cells with no or few stress fibers in low oxygen. In later passages (∼passages 5–8), foci with small spindle-shaped cells without stress fibers appeared under high oxygen and with ongoing passages the phenotype changed to a more uniform phenotype with progressive growth. Such findings are typical for murine cells and have been shown to be caused by oxidative stress.^31,51,52 Under high-oxygen concentrations ROS levels increase, as it has also been observed in our study. Reactive oxygen species may lead to an upregulation of p53 and to activation of this tumor suppressor protein by phosphorylation, resulting in growth arrest and apoptosis.^53–55 Under continuous exposure to atmospheric oxygen cells with DNA damage, loss of function, and mutations of p53, possibly induced by ROS, can accumulate and grow unrestricted.³¹ The DNA damaging effects of high oxygen concentrations contributing to cell senescence and loss of viability are not only shown for murine but also mammalian cells.³⁰ 

The flow cytometry analysis in our study revealed an increase of CD44 expression in LG-MSC under low oxygen, which has not been shown before. Krishnamachary et al.⁵⁶ were able to demonstrate in breast cancer cells, that the induction of CD44 expression is probably triggered via the HIF-1 α pathway. Valorani et al.⁴³ showed an increase of CD44/SCA-1 double-positive cells in murine adipose tissue–derived MSC (AT-MSC) under hypoxia, but an increase under hypoxia was only observed in SCA-1–positive cells while CD44 was staying stable in both conditions. Although the function of CD44 in MSC is not yet fully understood and it is suggested as a potential cancer stem cell marker in some studies,⁵⁷ it plays an important role in regenerative and anti-inflammatory processes: CD44 is responsible for increased migration, homing, and recruiting of exogenous MSC to sites of tissue injury.⁵⁸ It regulates macrophage inflammatory response,⁵⁹ plays a protective role in inflammation mediated by toll-like receptors (TLR) by suppression of nuclear factor–κB (NF-κB) activation and binds proteases as matrix metalloproteinase 9 (MMP-9),⁶⁰ resulting in reduced monocyte-migration.⁶¹ 

In this study, the basic FABP4 expression and the OPN and FABP4 expression after induction were significantly different between the two culture conditions. However, no difference between the two culture conditions could be identified macroscopically and microscopically in the stained cells after adipogenic and osteogenic differentiation. Varying results can be found in the literature concerning the differentiation of MSC under low oxygen. Berniakovic et al.⁶² (BM-MSC; 3% oxygen) and Valorani et al.⁶³ (AT-MSC; 2% oxygen) revealed an increased adipogenic and osteogenic differentiation potential under low-oxygen concentrations. In contrast, Malladi et al.⁶⁴ (AT-MSC; 2%) described the reduction of osteogenic differentiation in murine cells. Fehrer et al.⁶⁵ (BM-MSC; 3% O²) could not detect osteogenic differentiation and a diminished adipogenic differentiation. The differences concerning the effects of low oxygen on proliferation capacity, differentiation potential, and stemness might be explained by the heterogeneous cell origins and different oxygen levels, as well as further influencing factors such as the time and continuity of exposure to oxygen used in the different studies.⁶⁶ 

Proliferation and migration of cells at the site of injury are essential parts of tissue regeneration and wound healing processes in vivo.^{25–27,67–69} As support of proliferation and migration has been described as one function of MSC's in regeneration in vivo,^70–73 we evaluated the effect of migration and proliferation of LG-MSC on porcine LG-EC in vitro to test and quantify their potential in LG tissue regeneration. The results of the experiments showed that CM^{MSC 5%} compared with CM^{MSC 21%} significantly increased migration and enhanced proliferation of LG epithelial cells after injury, indicating a high regenerative potential. As shown in the literature, the increase of migration and proliferation in CM^{MSC 5%} might be the result of an upregulated secretion of factors, such as IL-6, vascular VEGF, bFGF, PDGF, insulin-like growth factor II (IGF-II), hepatocyte growth factor (HGF), and monocyte chemoattractant protein (MCP)-1.^74–77 In agreement with our results, Lee et al.⁷⁸ demonstrated that CM of MSC harvested under hypoxic conditions significantly promoted the migration of human dermal fibroblasts and reduced the wound area in an in vivo model, compared with those in normoxic conditioned medium. Hung et al.⁷⁴ showed, that CM from hypoxic culture of MSCs was more effective in decreasing induced apoptosis and cell death than CM from MSCs incubated under normoxic conditions. 

The porcine LG is very similar to the human LG in terms of its macro- and microanatomical structure.⁷⁹ The cultivated porcine LG epithelial cells showed positivity for pan-cytokeratin and secretory activity by PAS-alcian-blue staining, TEM, positivity for Rab3D and secretory response to carbachol stimulation. The supportive and regenerative effects of cross-species feeder cells (e.g., 3T3), conditioned medium and even xenografts of cells are well described.^80–84 

In summary, this study shows, that in vitro expansion at low oxygen preserves the proliferation capacity and differentiation potential of LG-MSC, and therefore seems to be an excellent method to obtain vital MSC. Also an increased number of LG-MSC expressing CD44 was observed under low oxygen, which might be a valuable marker to identify a potent MSC subpopulation. Lacrimal gland–MSC cultivated under 5% O2 improve migration and proliferation of LG epithelial cells indicating a higher supportive regenerative function in comparison with LG-MSC cultivated under 21% O2. However, little is known about in vivo transplantation of cells cultured under low-oxygen conditions. For this reason, the potential as a possible novel therapeutic option in the treatment of severe dry-eye syndrome and especially the safety of regional and systemic delivery of LG-MSC cultured in low-oxygen conditions needs to be further examined in vivo by using animal models. 

Acknowledgments

The authors thank Silke Götze, Iris Sawitza (Clinic for Gastroenterology, Hepatology and Infectious Diseases, Düsseldorf, Germany), Sabine Seggewiß (Department of Ophthalmology, University Düsseldorf), Monika Jerg (Clinic for Otorhinolaryngology, Ulm, Germany), and Katharina Raba (ITZ, Düsseldorf, Germany) for their support and excellent technical assistance. 

Supported by the Volkswagen Foundation (Lichtenberg Professorship of Prof. Schrader) and the Deutsche Forschungsgemeinschaft through Sonderforschungsbereich SFB 974 (Düsseldorf, Germany). 

Disclosure: M. Roth, None; K. Spaniol, None; C. Kordes, None; S. Schwarz, None; S. Mertsch, None; D. Häussinger, None; N. Rotter, None; G. Geerling, None; S. Schrader, None