The Aging Lacrimal Gland of Female C57BL/6J Mice Exhibits Multinucleate Macrophage Infiltration Associated With Lipid Dysregulation

Abstract

Purpose: Loss of function of the lacrimal gland (LG), which produces the aqueous tear film, is implicated in age-related dry eye. To better understand this deterioration, we evaluated changes in lipid metabolism and inflammation in LGs from an aging model.

Methods: LG sections from female C57BL/6J mice of different ages (young, 2–3 months; intermediate, 10–14 months; old,  ≥24 months) were stained with Oil Red-O or Toluidine blue to detect lipids. Quantitative real-time polymerase chain reaction (qRT-PCR) analysis and western blotting of LG lysates determined differences in the expression of genes and proteins related to lipid metabolism. A photobleaching protocol to quench age-related autofluorescence was used in LG sections to evaluate changes in immunofluorescence associated with NPC1, NPC2, CTSL, and macrophages (F4/80, CD11b) with age using confocal fluorescence microscopy.

Results: Old LGs showed increased lipids prominent in basal aggregates in acinar cells and in extra-acinar sites. LG gene expression of Npc1, Npc2, Lipa, and Mcoln2, encoding proteins involved in lipid metabolism, was increased with age. NPC1 was also significantly increased in old LGs by western blotting. In photobleached LG sections, confocal fluorescence microscopy imaging of NPC1, NPC2, and CTSL immunofluorescence showed age-associated enrichment in macrophages labeled to detect F4/80. Although mononuclear macrophages were detectable in LG at all ages, this novel multinucleate macrophage population containing NPC1, NPC2, and CTSL and enriched in F4/80 and some CD11b was increased with age at extra-acinar sites.

Conclusions: Lipid-metabolizing proteins enriched in F4/80-positive multinucleated macrophages are increased in old LGs adjacent to sites of lipid deposition in acini.

Dry-eye disease (DED) is a common ocular disorder that affects 5% to 50% of the global population.¹ In the United States, DED prevalence is 2.7% in individuals 18 to 34 years of age compared to 18.6% in individuals >75 years old. DED prevalence is twofold higher in women than in men.² With the aging of the US population, an alarming increase of 60% in DED prevalence is expected by 2030.^3,4 DED is the leading cause of visits to eyecare specialists, creating an estimated $3.8 billion/yr in direct US healthcare costs with over $55.4 billion estimated in societal costs.^2,5 Given the steep rise in patient numbers and ensuing societal burden, a deeper understanding of DED is essential. 

The lacrimal gland (LG) is an exocrine gland responsible for producing the aqueous layer of the tear film, which provides moisture, lubrication, and nutrients to the ocular surface.^6,7 Age-related changes in the LG such as decreased tear secretion^8,9 and altered glandular structure¹⁰ may contribute to the development of age-related DED. Similar changes are seen in C57BL/6J mice by 6 to 9 months, with changes in corneal surface irregularity, corneal barrier disruption, and tear volume compared to 8-week-old mice in both sexes.¹¹ The aging mouse LG also exhibits increased lymphocytic infiltration and formation of ectopic lymphoid structures, similarly to the human LG.^12,13 With the C57BL/6J aging model, increased oxidative stress in the LG and its amelioration with an antioxidant diet have been demonstrated.¹⁴ Consistent with this, superoxide dismutase knockout mice, with a reduced ability to reduce oxygen radicals, show accelerated infiltration of immune cells, higher fibrosis, and apoptosis in the LG.¹⁵ 

Lipid dysregulation is a hallmark of aging, and changes in lipid metabolism are implicated in the pathogenesis of age-related diseases including DED.¹⁶ The cholesterol transporters Niemann–Pick disease type C1 (NPC1), and Niemann–Pick disease type C2 (NPC2) are dysregulated in aging.^17–19 NPC1 is a transmembrane protein that mediates intracellular cholesterol trafficking, and NPC2 is a soluble protein that functions in egressing and recycling of lipoprotein-derived cholesterol.^20–22 Dysregulation of these enzymes is linked to lipid accumulation in atherosclerosis and age-related macular degeneration.^23,24 Cathepsins are lysosomal proteases that are involved in antigen processing during immune responses, degradation of proteases and chemokines for cellular homeostasis, autophagy, proliferation, and metastasis, among many other functions.^25–27 Essential for lysosomal function and catabolism, the cathepsin proteases are also linked to aging and lipid dysregulation. Increased cathepsins are documented in atherosclerotic lesions,^28,29 with cathepsin B, D, and L enriched in macrophage-derived foam cells localized in necrotic cores of lipid plaques.³⁰ Cathepsin L (CTSL) is specifically linked to macrophage function in lipid degradation.^31,32 Although in vitro and mice studies suggest a role for CTSL in the reduction of body weight gain and adipogenesis through fibronectin, insulin receptor, and insulin-like growth factor 1 receptor degradation,^33,34 little is known about its role in lipid metabolism and possible dysregulation of NPC1, NPC2, and CTSL in aged LGs. 

Macrophages, part of the innate immune system, also play key roles in tissue development, homeostasis, and repair of damaged tissue.³⁵ Different macrophages are known, depending on their tissue location, function, and phenotype, with some implicated in lipid clearance. Foamy macrophages accumulate an abnormal amount of lipids and cholesterol esters and are hallmarks of early atherosclerotic plaques.³⁶ These foam cells can aggregate into multinucleated giant cells referred to as Touton giant cells in areas of high lipid concentration.^37,38 Other multinucleated macrophages are reported in different tissues, with osteoclasts in bone,³⁹ foreign body giant cells in implants,⁴⁰ and Langhans giant cells in lungs.⁴¹ 

Previous studies on mouse LGs have highlighted the increased accumulation of lipofuscin in tissues with age.^10,42–44 Lipofuscin is a mixture of oxidized proteins, lipids, and metals that are found in the lysosomes or cytosol of aging postmitotic cells.⁴⁵ This nondegradable and insoluble product has an autofluorescent 400- to 700-nm emission spectrum that significantly hinders application of immunofluorescence to tissue at useful visible light emission wavelengths. Multiple products such as Sudan Black B or reducing agents have been tested to overcome the problem.⁴⁶ Although the need to reduce lipofuscin autofluorescence signal in samples labeled with histological dyes may not be necessary, this modification was pivotal in our ability to discretely visualize the distribution of proteins of interest by immunofluorescence at higher magnification. Photobleaching with a light-emitting diode (LED) light has previously been used as an approach for eliminating lipofuscin-related autofluorescence in brain samples,⁴⁷ an approach that we utilized here in LGs. 

Here, we investigated the possible dysregulation of lipids and lipid-metabolizing enzymes, as well as their relationship to macrophage infiltration, in the aged LG. We report here that lipids accumulate with age in the LG. Using a newly developed photobleaching regimen to reduce endogenous autofluorescence, we identified a novel multinucleated macrophage population enriched in F4/80, CTSL, NPC1, and NPC2 and increased in aged LGs that may participate in clearance of tissue lipids. 

Methods

Mice

Female C57BL/6J (B6) mice were obtained from The Jackson Laboratory (Bar Harbor, ME, USA), up to >2 years in age, or were received from the National Institute of Aging, National Institutes of Health. Animal use complied with policies approved by the University of Southern California and Baylor College of Medicine Institutional Animal Care and Use Committees and was in accordance with the Guide for the Care and Use of Laboratory Animals and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. Mice were young (2–3 months), intermediate (10–14 months), or old (>24 months). After intraperitoneal injection of 100 mg/kg ketamine + 10 mg/kg xylazine anesthesia and euthanasia, female mouse LGs were isolated and processed immediately or stored at –80°C. Mice were housed at facilities at Baylor College of Medicine, University of Southern California, or the National Institutes of Health with ad libitum access to food and water. Because DED is more frequent in women^4,48 and aged male mice do not develop corneal barrier disruption (a hallmark of DED),⁴⁹ the study used only female mice. 

Materials and Reagents

Ketamine (NDC:13985-54-10) and xylazine (NDC:13985-704-10) were obtained from VetOne (Boise, ID, USA). Oil Red O staining solution (O0625), Beadbug prefilled tubes (Z7637800), and Tris buffered saline with Tween 20 (TBS-T), pH 8.0 (T9039), were obtained from Sigma Aldrich (St. Louis, MO, USA). Phosphatase inhibitor cocktail (5872S) was obtained from Cell Signaling Technology (Danvers, MA, USA). Both 10% Tris-Glycine Gels (XP00105BOX) and iBlot 2 NC Regular Stacks (IB23001) were obtained from Invitrogen (Waltham, MA, USA), and Revert 700 Total Protein Stain Kits for Western Blot Normalization (P/N 926-11016) were obtained from LiCor (Lincoln, NE, USA). The Pierce BCA Protein Assay Kit (23225) was obtained from Thermo Fisher Scientific (Waltham, MA, USA). Blocking Buffer for Fluorescent Western Blotting (MB-070) was obtained from Rockland Immunochemicals (Pottstown, PA, USA). Superfrost Plus Microscope Slides (48311-703) were obtained from VWR (Radnor, PA, USA). ProLong Gold Antifade Mounting Medium (P36934) was obtained from Invitrogen. Optimal cutting temperature (OCT) compound (25608-930) was obtained from VWR. The RNeasy Plus Mini RNA (74134) isolation kit was obtained from Qiagen (Hilden, Germany). Ready-To-Go You-Prime First-Strand Beads (27-9263-01) were obtained from GE Healthcare (Chicago, IL, USA). Primers for real-time PCR and the TaqMan Universal PCR Master Mix AmpErase UNG were obtained from Thermo Fisher Scientific. Also, 3% glutaraldehyde (01909-10), osmium tetroxide (0972A-20), Poly/Bed 812 plastic resin (08792-1), and toluidine blue (01234-25) were purchased from Polysciences (Warrington, PA, USA). All antibodies used for immunofluorescence, western blotting, and flow cytometry are provided in the Table. 

Table.

 

Antibody Information for Western Blotting, Immunofluorescence, and Flow Cytometry

 

Preparation of Sections and Oil Red O Staining

LGs were excised, embedded in OCT compound, and flash-frozen in liquid nitrogen. Sagittal 10-µm sections were cut with a cryostat (HM 500; Micron, Waldorf, Germany) and placed on glass slides stored at −80°C. Slides were fixed with 10% neutral buffered formalin for 5 minutes, and Oil Red O staining solution (O0625; Sigma-Aldrich) was freshly prepared by mixing 3% Oil Red O solution in isopropanol with water (3:2, v/v) for 72 hours. Slides were washed in propylene glycol for 2 minutes, and nuclei counterstained with hematoxylin. After coverslip application, digital images were visualized and acquired using a light microscope (Eclipse E400; Nikon, Tokyo, Japan). 

Preparation of LG Sections Stained With Toluidine Blue

LGs were fixed in a solution of 3% glutaraldehyde and washed in 1-M sodium phosphate buffer (PBS) with a pH of 7.3. They were subsequently post-fixed in an osmium tetroxide (OsO4) solution for 1 hour. After fixation, the samples underwent dehydration through a series of graded alcohols, acetone, and Poly/Bed 812 plastic resin. Finally, they were embedded in plastic block molds using 100% Poly/Bed 812. Then, 1-µm sections were obtained using a Leica EM UC Ultramicrotome (Leica, Wetzlar, Germany) and placed on glass slides. The sections were then stained with toluidine blue. 

Quantitative Real-Time PCR of LG

Total RNA from LGs was extracted using an RNeasy Plus Mini RNA isolation kit following the manufacturer's protocol. After isolation, RNA concentration was measured, and cDNA was synthesized using the Ready-To-Go You-Prime First-Strand kit. Real-time PCR was performed using specific TaqMan minor groove binder probes for NPC1 (Ncp1; Mm00435300), NPC2 (Npc2; Mm00499230), Mucolipin 1 (Mcoln1; Mm00522549), Mucolipin 2 (Mcoln2; Mm00509849), ADP-ribosylation factor 6 (Arf6; Mm0050020), lipase (Lipa; Mm00498820), and TaqMan Universal PCR Master Mix AmpErase UNG in a commercial thermocycling system according to the manufacturer's recommendations. The hypoxanthine phosphoribosyl transferase 1 (Hprt1; Mm00446968) gene was used as an endogenous reference for each reaction. Quantitative PCR results were analyzed by the comparative Ct method and normalized to the Ct value of Hprt1. The young group served as calibrators. 

Western Blotting of LG Lysate

For gel electrophoresis, one whole LG was used per sample. LG were homogenized using BeadBug prefilled tubes in radioimmunoprecipitation assay (RIPA) buffer containing protease inhibitor cocktail. The supernatant was recovered after centrifugation at 8000g at 4°C for 10 minutes. After measuring the protein concentration of each sample with the Pierce BCA Protein Assay Kit, samples were incubated in reducing dye and β-mercaptoethanol at 95°C for 5 minutes. Then, 40 µg of each sample in sample buffer was loaded onto precast 10% Tris-Glycine Gels and resolved by sodium dodecyl sulphate–polyacrylamide gel electrophoresis (SDS-PAGE) at 120 V at 4°C for 2 hours. Proteins on gels were transferred to nitrocellulose membranes using an iBlot 2 gel transfer machine. Membranes were stained for total protein using the Revert 700 Total Protein Stain Kit for western blot normalization, and the signal was read using the LiCor Odyssey Fc imager. The signal for each lane was used to normalize the signal of the protein of interest to total protein loaded. Total protein stain was rinsed away per the manufacturer's protocol, and membranes were blocked for 1 hour with blocking buffer at room temperature while shaking. Membranes were washed three times for 5 minutes with TBS-T, then incubated with either anti-rabbit NPC1 (1:4000) or anti-goat CTSL antibodies (1 mg/mL) in blocking buffer at 4°C overnight. After three washes, 5 minutes each with TBS-T, membranes were incubated with secondary donkey anti-rabbit IR 680 (1:4000) or donkey anti-goat IR 680 (1:4000), respectively, in blocking buffer at room temperature for 1 hour. After another three washes for 5 minutes each with TBS-T, membranes were imaged with the LiCor Odyssey Fc imager. Signal quantification was done with Image Studio 5.2. Controls included blots processed without exposure to primary antibody. 

Tissue Processing for Photobleaching, Immunofluorescence, and Confocal Fluorescence Microscopy

LGs were fixed in 4% paraformaldehyde and 4% sucrose in PBS for 3 hours at room temperature and then in 30% sucrose PBS solution at 4°C overnight. OCT compound embedding was done the following day, and samples were frozen on dry ice and stored at –80°C. Tissues were sectioned at a thickness of 5 µm and mounted on Superfrost Plus microscope slides. For photobleaching of autofluorescence, procedures were adapted from a study on brain tissue.⁴⁷ The complete tissue photobleaching protocol is provided in Supplementary Methods. 

When photobleaching was complete, tissue sections on slides were further quenched with 50-mM NH₄Cl in PBS for 20 minutes and permeabilized with 0.3% Triton X-100 for 30 minutes. Slides were washed twice with PBS for 15 minutes each at room temperature with shaking. Slides were then blocked with 5% BSA in 0.3% Triton X-100 for 3 hours at room temperature. Slides were incubated with primary antibodies as follows: goat anti-mouse to CTSL (1:40), goat anti-mouse NPC1 (1:100), rabbit anti-mouse NPC2 (1:100), rat anti-mouse to F4/80 (1:50), or rabbit anti-mouse CD11b (1:100) in blocking buffer overnight at 4°C. On day 2, slides were washed with PBS three times for 15 minutes each at room temperature with shaking. The tissue was then incubated with secondary antibodies as follows: Alexa Fluor 488 Goat anti-Rabbit (1:200) and Alexa Fluor 568 Donkey anti-Goat (1:200). Then, 4′,6-diamidino-2-phenylindole (DAPI) and Phalloidin AF 647 were added with secondary antibodies followed by incubation for 1 hour at 37°C. Slides were washed again with PBS three times, then mounted with ProLong Gold Antifade Mountant and a glass coverslip and left to dry overnight. Images were acquired with either a Zeiss LSM 800 with Airyscan processing or a Zeiss LSM 880 with Airyscan processing (ZEISS, Oberkochen, Germany), both using a 63× oil 1.4 NA objective. Images were subjected to equivalent processing with brightness and contrast of the proteins of interest conserved on QuPath 0.4.2 software. Images labeled Mouse 1 and Mouse 2 indicate images from separate mice, but Mouse 1 and 2 are not necessarily the same mice across each of the figures. 

For counting of nuclei per F4/80-positive (F4/80⁺) macrophages in young and old (photobleached) LG sections, 5-µm cut sections were labeled to detect F4/80, actin, and nuclei as above and were scanned using a BZ-X810 microscope (Keyence, Osaka, Japan) with 60× objective and DAPI (nuclei), green fluorescent protein (GFP; F4/80), and Cy5 (actin) fluorescence filters. The images were stitched using BZ-H4A Advanced Analysis Software in the uncompressed option with no shading autocorrection and converted to TIFF files. F4/80⁺ cells were delineated using the QuPath program, and nuclei were counted manually. Calculation of LG area was performed with Fiji 2.14.0/1.54f and the lens scale bar from Keyence. 

Tissue Autofluorescence

For assessment of autofluorescence to optimize photobleaching protocols, tissues from intermediate and old LGs were put in containers with sterile PBS and exposed to photobleaching at 4°C or kept at 4°C covered in aluminum foil (mock photobleaching) as described in Supplementary Methods. After 7 days, slides were quenched and blocked as described above and labeled with DAPI. After mounting, images were acquired with a Keyence BZ-X810 microscope with 20× objective using DAPI, GFP, tetramethylrhodamine (TRITC), and Cy5 fluorescence filters. Images acquired were stitched using the BZ-H4A Advanced Analysis Software in the uncompressed option, no shading autocorrection, and converted to TIFF files. To measure autofluorescence excitation and emission profiles of aged LG sections, a lambda scan, which acquires the emission spectra of the same specimen at different excitation wavelengths, was utilized as described in Supplementary Methods. 

Flow Cytometry

LGs were excised, chopped with scissors into tiny pieces, and incubated with 0.1% type IV collagenase for 1 hour to yield single-cell suspensions. DNAse I was added for the last 15 minutes of incubation to reduce cell clumps. Cells were washed using RPMI medium and passed through a 100-µm pore size filter. Samples were incubated with anti-CD16/32 for 10 minutes at room temperature and subsequently stained with anti-CD45, anti-CD11b, anti-LY6C, anti-CD64, and anti-F4/80 with antibody sources as shown in the Table. Cells were also stained with an infrared fluorescent viability dye. 

The gating strategy was as follows: lymphocytes were identified by forward scatter area and side scatter area gates, followed by two live/dead identifications using the infrared fluorescent viability dye. Live CD45⁺ cells were gated and CD11b⁺ cells were identified. CD11b⁺ cells were plotted using LY6C and CD64 antibodies, and the frequency of F480⁺ cells was investigated in LY6C^–CD64⁺ cells. Negative controls consisted of fluorescence minus one LG cell suspensions. Cells were evaluated with an Aurora Spectral Cytometer (Cytek, San Diego, CA, USA). Single controls used capture beads, and they were used for unmixing. Final data were analyzed using FlowJo 10 (Tree Star, Ashland, OR, USA). 

Statistical Analysis

For statistical analysis of gene and protein signals, Prism 9.5.1 (GraphPad, Boston, MA, USA) was used. A Kruskal–Wallis test was used to compare the protein signals and gene expression from mice at three different ages. A Mann–Whitney U test was used to investigate the effects of age on different immune cells and to observe the changes in the abundance of single and multinucleated F4/80⁺ macrophages in aged and young mice by immunofluorescence. The criterion for significance was set at P < 0.05. 

Results

Lipid Accumulation in LGs of Aged Mice

Aged LGs are known to exhibit anatomical changes including acinar atrophy, chronic inflammation with monocytic infiltration, lipofuscin deposition, and fibrosis in the most extreme cases.^{12,13,42,50–53} Because aging in general is associated with lipid dysregulation, we first examined lipid deposition in aged LGs. Sections from young and aged LGs were labeled with Oil Red O, which stains neutral triglycerides and lipids. Although lipid droplets were seen in young LGs, the droplets were usually small and intracellular (Fig. 1A). In contrast, in old LGs, lipids coalesced to form larger structures (Fig. 1A). These larger structures were most evident at the edge of immune cell infiltrates, inside ducts, and within acinar cells (Figs. 1A, 1B). Within the same aged LGs, we observed considerable variability in lipid deposition across sections (Fig. 1B). Thin sections stained with toluidine blue also demonstrated accumulation of lipid droplets in old LGs (Fig. 1C, asterisk). Of note, old LGs also showed apparent extra-acinar multinucleated cells stained with toluidine blue (Fig. 1C, white arrows). Thus, aged LGs accumulate both intracellular and extracellular lipids. 

Figure 1.

 

Lipid accumulation increased with age in female B6 mouse LGs. (A) Oil Red O staining was conducted on young and old female mouse LG histological sections. A lipid-enriched region containing multiple nuclei outside the acini is further magnified in the 40× image from old LGs. Scale bars: 100 µm, 50 µm, and 25 µm. (B) Representative image of aged LG stained with Oil Red O. Note the varied patterns of lipid deposition within the same LG. Scale bars: 200 µm and 50 µm. (C) Thin sections stained with toluidine blue showing lipid droplets in between acini in old LG sections (*lipid droplet). Apparent lipid-enriched multinucleated cells are marked with white arrows. Scale bar: 5 µm.

Gene Expression of Lipid Metabolism–Related Proteins Increases With Age

To determine if lipid deposition in aged LGs was associated with changes in gene expression, we performed quantitative real-time PCR (qRT-PCR) evaluation of genes encoding proteins involved in lipid metabolism. NPC1 and NPC2 (encoded by Npc1 and Npc2, respectively) function in cholesterol trafficking and metabolism, and their dysregulation is linked to lipid accumulation in other age-related diseases. Mucolipin 1 (encoded by Mcoln1),^54,55 ADP ribosylation factor 6 (encoded by Arf6),^56,57 and lipase (encoded by Lipa)⁵⁸ were also evaluated as effectors of lipid transport and degradation. Npc1, Npc2, Lipa, and Mcoln2 each showed increased gene expression with age in LGs (Fig. 2). We also assessed Ctsl expression in separate samples from young and old LGs, but its gene expression was not significantly affected with age (Supplementary Fig. S1). 

Figure 2.

 

qRT-PCR analysis of LG lysates from female B6 mice showed increased gene expression for genes associated with lipid metabolism with age. Mucolipin 1 (Mcoln1), Mucolipin 2 (Mcoln2), ADP ribosylation factor 6 (Arf6), Niemann-Pick C1 Protein (Npc1), Niemann-Pick C2 Protein (Npc2), and lipase (Lipa) cDNA was quantified from young, intermediate, and old LGs (n = 13, 8, and 12 mice, respectively, with one LG from each mouse). Outliers (values exceeding 2 SD) were excluded from the analysis. *P ≤ 0.05, ***P ≤ 0.001, and ****P ≤ 0.0001 for all graphs. A Kruskal–Wallis test was used to compare the relative expression of genes among the different age groups.

NPC1 Protein Abundance Is Significantly Increased in Aged Mouse LGs

To further investigate the changes in gene expression of proteins implicated in lipid metabolism with aging, we evaluated NPC1 and CTSL levels by western blotting of LG lysates. NPC1 was significantly increased in old LGs by approximately threefold compared to young LG (Figs. 3A, 3C). Although not statistically significant, there was a trend to increased active CTSL (cleaved from pro-CTSL) with increasing age (Figs. 3B, 3D). 

Figure 3.

 

NPC1 and active CTSL increased with age in B6 mouse LG lysates. (A) Young, intermediate, and old (n = 3–8) female B6 LG lysates show NPC1 signal by western blotting. The asterisk (*) indicates outliers as determined by the Rout method using GraphPad software. (B) Young, intermediate (Int), and old (n = 3–8) female B6 LG lysates showed active (cleaved from the pro-CTSL) CTSL signal by western blotting. Bands from the same blot were rearranged to reflect the aging order. (C) NPC1 signal in old LGs showed threefold increased protein compared to that expressed in young LGs. The asterisk (*) indicates significance, P = 0.0263. The signal was normalized by total protein stain signal as shown in Supplementary Figure S2. Supplementary Figure S2 also shows the data analysis with the outlier value included. (D) There was no statistically significant difference in active CTSL signal between each group. Each dot represents a right LG from one mouse. A Kruskal–Wallis test was used to compare the relative expression of genes among the different age groups.

Assessment of Tissue Autofluorescence With Age

We sought to expand our analysis of NPC1, CTSL, and NPC2 using immunofluorescence rather than western blotting, which required a significant amount of these precious aged LG samples without the ability to provide spatial information. The aging LG is a challenging choice for immunofluorescence because strong autofluorescence from lipofuscin-like structures is reported in aged BALB/c mouse LG.⁴² Lipofuscin autofluorescence has a wide emission wavelength profile from 450 to 700 nm, overlapping with emission wavelengths of most commonly used fluorophores. Imaging of old B6 LG sections using a Keyence microscope (Fig. 4, top row) verified strong autofluorescence. Specifically, non-photobleached (NP) old LGs showed emission through FITC (emission, 488–540 nm), TRITC (emission, 540–620 nm), and Cy5 (emission, 620–700 nm) filters, although samples were labeled only with DAPI. 

Figure 4.

 

LED photobleaching of LG tissue sections reduced endogenous autofluorescence. Non-photobleached or photobleached tissue sections from old B6 mouse LGs labeled with DAPI were imaged at 20× magnification using a Keyence fluorescence microscope with emission collected through Cy5, FITC, TRITC, and DAPI filters. Autofluorescence was scattered throughout the tissue in old non-photobleached samples detected through Cy5, FITC, and TRITC emission filters, and consecutive sections of old samples that were photobleached exhibited a greatly reduced signal. The top and bottom panels images show consecutive sections of the same LG. The asterisk (*) indicates lymphocytic infiltration. Scale bar: 500 µm.

To explore the potential of photobleaching to reduce endogenous tissue autofluorescence, we used lambda scanning to determine emission profiles of the sample with varying excitation with each primary laser line. The autofluorescence lambda scan profiles for unlabeled NP samples and NP samples labeled with primary anti-NPC1 plus AF488-labeled secondary antibody or AF488 secondary antibody alone are shown in Figure 5A (top row). Excitation at 561 nm elicited the strongest emission, which was roughly comparable to the sample incubated with AF488-labeled secondary antibody. The same signal was detectable, overlapping, and commensurate in magnitude to that obtained with the sample incubated with primary anti-NPC1 plus AF488-labeled secondary antibody. A small distinct peak of fluorescence elicited at 488- and 514-nm excitation could be detected that was poorly resolvable from autofluorescence. 

Figure 5.

 

Photobleaching parameters associated with reduction of tissue autofluorescence. (A) Lambda scanning was carried out on non-photobleached (top row) and photobleached (bottom row) tissue sections incubated with blocking buffer, secondary antibody alone labeled with AF488, or primary antibody to NPC1 plus secondary antibody labeled with AF488. Although non-photobleached samples exhibited emission throughout the different conditions and laser wavelengths, photobleached samples showed signal only with the primary plus secondary antibody combination. (B) An LED desk lamp was used to photobleach slides. An apparatus was set up with sufficient sterile PBS to cover the slides in a plastic container. With aluminum foil covering the entire apparatus, slides were photobleached under constant illumination at 4°C for 7 days. (C) The relative intensities for each wavelength from the LED light are shown as obtained by the Lighting Passport Pro. (D) Additional parameters for LED light characterization are shown.

Utilizing a commercial desk lamp LED for photobleaching (Fig. 5B), 7 days of treatment was established as sufficient to reduce tissue autofluorescence. As shown in the photobleached images in Figure 4 (bottom row) of a section of old LG cut consecutively after the NP section from the same mouse LG above, the photobleached sample showed fluorescence associated only with DAPI. This lack of endogenous autofluorescence was verified using lambda scanning. The photobleached samples in Figure 5A (bottom row) showed essentially no autofluorescence with 561-nm excitation, associated with the most autofluorescence in NP samples (top row). Likewise, 488- and 514-nm excitation showed minimal autofluorescence emission. These results indicate that photobleaching of freshly cut LG sections with an LED light permits use of immunofluorescence. The emission spectra/relative intensity values for the illuminating light (Fig. 5C) and energy (e.g., flux) (Fig. 5D) were determined to relate photobleaching capability with energy input received by each tissue sample. 

NPC1 and NPC2 Immunofluorescence Are Increased With Age in Non-Acinar Cells in the Old LGs

NPC1 and NPC2 staining in photobleached LG sections from young, intermediate, and old B6 mouse LGs were then evaluated (Fig. 6A). NPC1 was detected in intermediate and old but not young LGs, with expression largely seen at the periphery of acinar clusters. Little expression in acinar or ductal cells was detectable. As with NPC1, NPC2 was increased in LGs from intermediate and old mice and exhibited an extra-acinar distribution comparable to that of NPC1 (Fig. 6B). Thus, both NPC1 (confirmed by gene and protein expression) and NPC2 (confirmed by gene expression) are increased in extra-acinar cell populations with the aging of LGs. In Figure 6 (and subsequent figures) rows labeled Mouse 1 and Mouse 2 are images from different mice for all age groups. Mouse 1 and 2 are not necessarily the same across different panels or figures. 

Figure 6.

 

NPC1 and NPC2 were enriched in extra-acinar structures in intermediate and old B6 mouse LGs. (A) Photobleached B6 LG sections from young, intermediate, and old LGs were labeled with primary and secondary antibodies to NPC1, as well as rhodamine phalloidin and DAPI. NPC1 was seen in intermediate and old LGs in proximity to but not within acinar cells. (B) Photobleached B6 LG sections from young, intermediate, and old LG were labeled with primary and secondary antibodies to NPC2, as well as rhodamine phalloidin and DAPI. Similar to NPC1, NPC2 was seen in intermediate and old LGs in proximity to but within acinar cells. Scale bars: 20 µm. Rows labeled Mouse 1 and Mouse 2 show images from different mice for all age groups. Mouse 1 and Mouse 2 are not necessarily the same across different panels or figures.

NPC1, NPC2, and CTSL Are Enriched in LGs in a Novel Multinucleated Macrophage Population That Increases With Aging

The cell type associated with enrichment of NPC1 and NPC2 in the aged LGs was determined in photobleached specimens. Previously, we had detected an apparent lipid-rich multinucleated cell population at the periphery of acinar clusters in old LGs stained with Oil Red O and hematoxylin (Fig. 7A), consistent with the detection of the toluidine blue–labeled apparent multinucleated cells in Figure 1C. These cells were similar morphologically and in their location to a macrophage population labeled with F4/80 that can be seen in Figures 7C and 7D in intermediate and old LG sections. These macrophages, detected using an antibody to F4/80, appeared multinucleated in nature and were situated in proximity to the locations where lipid droplets were present (Fig. 1) and where NPC proteins were enriched (Fig. 6). Localization of NPC1 (Fig. 7C) and NPC2 (Fig. 7D) in the cytoplasm of these multinucleated macrophages was confirmed by immunofluorescence. Supplementary Movie S1 provides a z-stack video of a representative F4/80⁺ multinuclear cell containing three nuclei without (Supplementary Movie S1A) and with (Supplementary Movie S1B) the cell of interest outlined. We also quantified the average number of nuclei per multinucleated F4/80⁺ cell. The majority of multinucleated cells (∼60%) had two or three nuclei (Fig. 7B). 

Figure 7.

 

NPC1 and NPC2 are enriched in apparent F4/80⁺ multinucleated macrophages in aging LG. (A) Oil Red O and hematoxylin-stained old LGs show multinucleated lipid-containing cells, also shown in magnification. Scale bars: 25 µm. (B) The number of nuclei per multinucleate F4/80⁺ macrophages as a percentage of total multinucleate macrophages in old LG were quantified (n = 3 old mouse LGs). (C) Photobleached LG sections from young, intermediate, and old B6 mice were labeled with primary and secondary antibodies to NPC1 (green), F4/80 (pink), rhodamine phalloidin (red, actin), and DAPI (blue, nuclei). NPC1 was detected in F4/80-enriched multinucleated cells adjacent to acinar clusters. (D) Photobleached LGs from young, intermediate, and old B6 mice were labeled with primary and secondary antibodies to NPC2 (green), F4/80 (pink), rhodamine phalloidin (red, actin), and DAPI (blue, nuclei). NPC2 showed a distribution similar to that of NPC1 within F4/80⁺ multinucleated cells. These cells are outlined in white with asterisks (*) indicating the position of the nuclei in panels B and C. Scale bars: 20 µm (C, D). Rows labeled Mouse 1 and Mouse 2 show images from different mice for all age groups. Mouse 1 and Mouse 2 are not necessarily the same across different panels or figures.

We further evaluated whether multinucleated F4/80⁺ macrophages were present to the same extent in young and old LGs. Figure 8A shows a representative section from old LGs used to identify F4/80⁺ cells for the analysis. As shown in Figure 8B, F4/80⁺ macrophages were present in both young and old LGs. Approximately equal numbers of mononuclear F4/80⁺ cells were present per area of the tissue; however, in old LGs the numbers of multinuclear macrophages and total macrophages were dramatically increased. In accord with this increase, the percentage of multinucleated F4/80⁺ macrophages as a percentage of total F4/80⁺ macrophages was found to be ∼5% in young LGs but was increased to approximately 45% in old LGs. 

Figure 8.

 

Mononuclear and multinuclear F4/80⁺ macrophage comparisons in old and young LGs. (A) A whole LG image of F4/80 (green), actin (red), and DAPI (blue) is shown in “Old LG.” “Old LG Counted” shows the “Old LG” image with the counted cells highlighted with a yellow border, indicating the areas that were positive for F4/80 labeling. “Old LG 2nd Control” shows a consecutive section labeled with secondary antibody alone and shows no green signal. Scale bar: 100 µm. (B) Quantification of mononuclear and multinuclear F4/80⁺ macrophages is shown. The total number of macrophages (mononuclear and multinuclear macrophages) and the relative abundance of multinuclear macrophages are compared. (n = 3 mice for old and young).

Given previous findings that CTSL was enriched in macrophages localized within atherosclerotic plaques, we examined CTSL distribution with respect to macrophages and NPC1. As shown in Figure 9, CTSL was also enriched in this multinucleated macrophage population (Fig. 9A) and colocalized with NPC1 (Fig. 9B). Evaluation of the intensity of these proteins within apparent multinucleate macrophages in Figure 9B shows that, although NPC1 and CTSL were enriched in multinucleate macrophages, the relative abundance of these proteins varied across individual macrophages. Collectively, Figures 6 to 9 demonstrate the enrichment of NPC1, NPC2, and CTSL with each other and within an F4/80⁺ multinucleate macrophage population increased with age in the LGs. 

Figure 9.

 

CTSL was enriched with age in F4/80⁺ multinucleated macrophages containing NPC1. (A) Photobleached LG sections from B6 mice at young, intermediate, and old ages were labeled with primary and secondary antibodies to CTSL (pink), F4/80 (green), rhodamine phalloidin (red, actin), and DAPI (blue, nuclei). An F4/80⁺ multinucleated macrophage population colocalized with CTSL in intermediate and old LGs. (B) Photobleached LG sections from B6 mice at young, intermediate, and old ages were labeled with primary and secondary antibodies to CTSL (pink), NPC1 (green), rhodamine phalloidin (red, actin), and DAPI (blue, nuclei). NPC1 and CTSL were enriched in F4/80⁺ multinucleated cells adjacent to acini in the intermediate and old sections. Multinucleated cells are outlined in white, and the asterisk (*) indicates the position of nuclei inside these cells. Scale bars: 20 µm. Rows labeled Mouse 1 and Mouse 2 show images from different mice for all age groups. Mouse 1 and Mouse 2 are not necessarily the same across different panels or figures.

Multinucleate Macrophages in Old LGs Have Varying Levels of Cd11b

To quantify the changes in macrophages in young and old LGs, we prepared single-cell suspensions and labeled with a panel of antibodies. Multicolor flow cytometry can use multiple markers to identify macrophages. We observed an increase in CD45⁺ and a decrease in CD11b⁺ cells in aged LGs. This was accompanied by an increase in LY6C⁺CD64⁻ (monocytes) and a decrease in LY6C⁻CD64⁺ (macrophage) cells (Fig. 10A). This analysis, surprisingly, showed an equivalent expression of F480⁺ cells expressing Cd11b+ in young and aged LGs (Figs. 10B). 

Figure 10.

 

Aged LGs exhibited a decreased frequency of macrophages by flow cytometry. Single-cell suspensions of young (3M) and old (25M) female LGs were prepared and stained with a panel of antibodies. Single alive cells were gated as described in Methods. (A) Representative dot plots of flow cytometry analysis showing CD45⁺, CD11b⁺ (broadly identified as myeloid cells), monocytes (CD11b⁺LY6C⁺CD64⁻), maturing monocytes (CD11b⁺LY6G⁻LY6C⁺CD64⁺), and macrophages (CD11b⁺LY6C⁻CD64⁺). The expression of F4/80 was evaluated in macrophages, as shown. (B) Cumulative data showing differences in populations in young and old mice. *P ≤ 0.05, ***P ≤ 0.001 (n = 5 for each 3M and 25M, with each point representing one mouse). The Mann–Whitney U test was utilized to compare populations between young and old mice data.

We also evaluated CD11b and F4/80 immunofluorescence in parallel in young, intermediate, and old mouse LGs. CD11b⁺ macrophages were present in proximity to acini as single mononuclear cells in LGs from all three age groups (Fig. 11). Mononuclear F4/80⁺ macrophages were also observed in all three age groups. In the intermediate and old LGs, F4/80⁺ macrophages were also present in proximity to acini as multinucleated cells with mixed CD11b positivity. 

Figure 11.

 

Cd11b and F4/80 enrichment in mononuclear and multinucleated macrophages in LG. Photobleached LG sections from young, intermediate, and old B6 mice were labeled with primary and secondary antibodies to detect CD11b (magenta), F4/80 (green), rhodamine phalloidin (red, actin), and DAPI (blue, nuclei). Cells with CD11b (indicated with magenta arrows), F4/80 (green arrows), and both CD11b and F4/80 (white arrows) can be seen in the young LGs. Multinucleated and mononuclear macrophages were observed in both intermediate and old LGs. Acini and possible ducts/blood vessel are outlined in the images with white dotted lines. Scale bar: 20 µm.

Taken together, our findings suggest that, with aging, not only does macrophage expression of lipid-metabolizing enzymes increase but there is also an increase in specialized macrophages that express different markers including F4/80. 

Discussion

In this study, we demonstrated that lipid deposition increases with age in the LGs. In parallel, significant upregulation of genes encoding proteins linked to lipid and cholesterol metabolism was detected, notably NPC1 and NPC2. The limited tissue availability and small size of the aged LGs made it unfeasible to probe all potential gene expression hits using western blotting. After validating increased NPC1 expression by western blotting, we focused on confirmation of other enrichments utilizing immunofluorescence, which also provided essential information about the location of proteins within the tissue. Our development of a method for reproducibly photobleaching tissue autofluorescence within the LG advanced this effort. Further examination of the distribution and enrichment of NPC proteins and CTSL with age in the LGs revealed an apparent multinucleated macrophage population (labeled with F4/80) situated at the base of the acinar cells, adjacent to the sites of extracellular lipid deposits identified in the LG. 

Macrophage infiltration in tissue plays many roles. As phagocytic cells, they are responsible for infiltrating infected tissue and assisting in the clearance of pathogens.^59,60 In this capacity, they exhibit robust expression of lysosomal enzymes to aid in rapid degradation of phagocytosed material. This ability enables them to metabolize lipids through lipolysis. In diseases of lipid deposition such as atherosclerosis and obesity, macrophage numbers are increased.⁶¹ However, the macrophages identified in the aging LGs are atypical with respect to foam cells in that they have a multinucleated phenotype. These LG macrophages are also labeled with an antibody to F4/80, a protein enriched in murine macrophages that is frequently used for their detection. The enrichment of NPC1, NPC2, and CTSL in these macrophages and their location in the tissue at sites prone to lipid accumulation with age make it tempting to speculate that these cells may function to aid the clearance of excess lipid that might otherwise create cellular stress and degrade LG function. Thus, these macrophages may represent a compensatory response to aging that may have benefit to the tissue. 

Macrophages can be detected using multiple markers including MerTK, CD68, CD64, CD11b, LY6C, and F4/80. Our flow cytometry identifying immune cells present in young and old LGs initially focused on the use of CD11b positivity to identify myeloid cells and further markers to identify monocytes and macrophages. Our analysis actually showed a significant decrease in CD11b⁺ macrophages in aged LGs. Flow cytometry analysis of cells obtained from non-lymphoid tissues requires tissue digestion and debris filtering to obtain a single-cell population, which can then be used in the flow cytometer. Filtering is an important step that preserves flow and prevents clogging of the instrument. One inherent disadvantage of tissue digestion is the loss of spatial architecture and the possibility of some cells not surviving the isolation and staining protocol. We have optimized our protocol to utilize a 100-µm filter to obtain cells. However, it is possible that either the larger multinucleated cells are not surviving the isolation procedure due to fragility or metabolic stress or they are being filtered out. This may be a reason for the apparent discrepancy between the flow cytometry data showing no increase in F4/80⁺ macrophages with age (Fig. 10) and the clear demonstration of the increased multinucleate F4/80⁺ macrophages with age by immunofluorescence (Fig. 8). Related to the possibility of F4/80⁺ macrophages being excluded from flow as above, it is of note that mononuclear F4/80⁺ macrophages are unchanged between young and old LGs (Fig. 8), similar to findings for F4/80⁺ macrophages by flow cytometry (Fig. 10). Future studies using beads of known sizes and imaging cytometer techniques will be required to fine-tune our observations. 

F4/80 glycoprotein, used as a primary target to identify macrophages by immunofluorescence, is a widely used murine macrophage marker^62,63 that is a member of the epidermal growth factor seven-span transmembrane family.⁶⁴ Macrophages may express different markers at different stages of maturation. For example, in apolipoprotein E–deficient (ApoE^−/−) mice, which are prone to develop atherosclerosis, macrophages obtained from the aorta in mice fed a 3-week or 12-week high-fat diet expressed different markers.⁶⁵ In another study in liver, CD11b was associated with newly migrating macrophages and F4/80 was associated with tissue-resident macrophages.⁶⁶ 

The macrophages in aged LGs are not only distinct from those in young LGs by virtue of enrichment in lipid-metabolizing enzymes but also by their multinucleate morphology. Multinucleated macrophages have been detected in bone as osteoclasts,³⁹ in prostheses and implants as foreign body giant cells,⁴⁰ in the lung as Langhans giant cells,⁴¹ and in ocular tissue as Touton giant cells.^37,38 Of these four known types of multinucleated macrophages, the Touton giant cell appears most similar to the multinucleated macrophages in aged LGs in terms of function. Touton giant cells are multinucleated giant cells present in juvenile xanthogranulomas, necrobiotic xanthogranulomas, and other types of xanthomas.⁶⁷ Xanthomas are associated with lipid deposits within organs, especially the skin, that usually show up as yellow (Greek xanthos) deposits.^67,68 Touton giant cells, however, have morphology inconsistent with the multinucleated cells present in old LGs, exhibiting instead a central eosinophilic cytoplasm and an annulus of nuclei with a lipid-filled outer ring. LG multinucleated macrophages are most similar in morphology to foreign body giant cells which have heterogeneously distributed nuclei.⁶⁹ B6 mice are known to gain fat not only in the body but also in the LG with age,⁷⁰ and it may be that the multinucleated macrophages may function similarly to Touton giant cells . 

Multinucleated cells (e.g., myoblasts, trophoblasts, monocytes) in general go through three stages of fusion: fusion competency development, migration, and intracellular adhesion/cytoplasmic sharing.⁷¹ To develop fusion competency, adhesion molecules (fusogens) must be expressed in the cells.⁷² Dendritic cell-specific transmembrane protein is a fusogen involved in osteoclast and foreign body cell formation.^73,74 Other fusogens that are present are E-cadherin,⁷⁵ CD206 (a mannose receptor),⁷⁶ and macrophage fusion receptor.⁷⁷ The combination of these fusogens may induce the formation of multinucleated cells, lowering the energy barrier between lipid layers of individual cells and allowing adjacent cells to merge. Future exploration of the presence of these fusogens in the multinucleated macrophages present in old LGs may aid in characterizing their process of formation and maturation. 

Previous uses of fluorescence imaging techniques in organs with aging have been limited due to autofluorescence. We demonstrate here that an inexpensive LED-based apparatus can photobleach LG tissue sections while preserving fundamental cellularity and tissue organization. The 7-day photobleaching process removes endogenous immunofluorescence in the visible wavelength emission spectrum such that the application of standard fluorescently labeled probes and antibodies is possible. One limitation of this approach may be some loss in the ability of actin filaments to be consistently labeled by phalloidin after photobleaching. It is possible that slight sample heating associated with constant illumination, or photodegradation, may result in the deterioration of these filaments. 

In summary, lipids accumulate in the LG within acinar cells and in extra-acinar deposits with aging in female B6 mice, a model of DED. We have identified a macrophage population enriched in lipid-metabolizing proteins (NPC, NPC2, and CTSL) that is localized to these same lipid-laden regions in the aging LG that may be responsible for the clearance of excess lipid. Further investigation of the subtype of these multinucleated macrophages, their phenotype (M1 vs. M2), and their elaboration of other macrophage markers and fusogens will enable determination of whether they play beneficial or pathological roles in age-related DED or are simply a marker of age-related lipid dysregulation in this gland. 

Acknowledgments

Supported by grants from the National Eye Institute, National Institutes of Health (R01EY030447 to CSdP; R01EY002520, a Core Grant to the Department of Ophthalmology at Baylor College of Medicine; R0EY011386 to SHA). Research reported in this publication was also supported by awards from the National Eye Institute, National Institutes of Health (P30EY029220 to USC Keck School of Medicine; P30EY021725 to Baylor College of Medicine). Further support was provided by unrestricted grants from Research to Prevent Blindness to the Department of Ophthalmology at Baylor College of Medicine and the Roski Eye Institute, Department of Ophthalmology, USC Keck School of Medicine. The Hamill Foundation, The Sid Richardson Foundation, and the Baylor College of Medicine Pathology Core (NCI P30CA125123) also supported the project. CSdP holds the Caroline F. Elles Endowed Professorship, which provides salary support. This project was supported by the Cytometry and Cell Sorting Core at Baylor College of Medicine with funding from the CPRIT Core Facility Support Award (CPRIT-RP180672), the National Institutes of Health (CA125123 and RR024574), and the assistance of Joel M. Sederstrom. 

Disclosure: M. Choi, None; C. Toscano, None; M.C. Edman, None; C.S. de Paiva, None; S.F. Hamm-Alvarez, None 

Supplementary Material

Supplementary Movie S1. Multinucleated macrophage in old LG. F4/80 (magenta), actin (red) and nucleus (blue) stained old LG z stack video is shown. The green outline shows a multinucleated macrophage. Scale bar: 10 µm. Two files named “Supplemental Video 1A no-outline” and “Supplemental Video 1B outlined” are of the same video with or without outlines of the multinucleated macrophage.