Regional Differences following Partial Salivary Gland Resection

Abstract

Regenerative medicine aims to repair, replace, or restore function to tissues damaged by aging, disease, or injury. Partial organ resection is not only a common clinical approach in cancer therapy but also an experimental injury model used to examine mechanisms of regeneration and repair in organs. We performed a partial resection, or partial sialoadenectomy, in the female murine submandibular salivary gland (SMG) to establish a model for investigation of repair mechanisms in salivary glands (SGs). After partial sialoadenectomy, we performed whole-gland measurements over a period of 56 d and found that the gland increased slightly in size. We used microarray analysis and immunohistochemistry (IHC) to examine messenger RNA and protein changes in glands over time. Microarray analysis identified dynamic changes in the transcriptome 3 d after injury that were largely resolved by day 14. At the 3-d time point, we detected gene signatures for cell cycle regulation, inflammatory/repair response, and extracellular matrix (ECM) remodeling in the partially resected glands. Using quantitative IHC, we identified a transient proliferative response throughout the gland. Both secretory epithelial and stromal cells expressed Ki67 that was detectable at day 3 and largely resolved by day 14. IHC also revealed that while most of the gland underwent a wound-healing response that resolved by day 14, a small region of the gland showed an aberrant sustained fibrotic response characterized by increased levels of ECM deposition, sustained Ki67 levels in stromal cells, and a persistent M2 macrophage response through day 56. The partial submandibular salivary gland resection model provides an opportunity to examine a normal healing response and an aberrant fibrotic response within the same gland to uncover mechanisms that prevent wound healing and regeneration in mammals. Understanding regional differences in the wound-healing responses may ultimately affect regenerative therapies for patients.

Introduction

Salivary gland (SG) hypofunction can occur as a side effect of radiation treatment in head and neck cancer patients as well as in autoimmune diseases, including Sjögren syndrome. Reduced salivary function negatively affects patient quality of life, resulting in the feeling of dry mouth (xerostomia), increased incidence of dental caries, oral fungal infections, and difficulty in eating, speaking, swallowing, and tasting (Lombaert et al. 2017). For salivary gland cancer patients, partial SG resection is a common clinical practice for removal of cancerous regions, leading to a decrease in saliva flow 1 y after partial resection (Ge et al. 2016).

Resection-based preclinical models exist for studying the regenerative capacity of many organs, including liver (Nevzorova et al. 2015), lung (Cowan and Crystal 1975), and heart (Xiong and Hou 2016). Here we investigated the response of the female murine submandibular gland (SMG) to partial resection after removing 40% of the distal tip of the gland. Using microarray analysis to profile the transcriptome and immunohistochemistry (IHC) to examine temporal-spatial protein expression patterns in response to resection, we found that the resected SMG exhibited a global wound-healing response, with an insufficient epithelial proliferation response to regenerate the gland. Interestingly, we also identified a small region of the gland that responded to injury with a sustained fibrotic response.

Materials and Methods

For detailed Materials and Methods, see the Appendix.

Results

Submandibular Salivary Glands Respond to Partial Gland Resection

We performed a partial sialoadenectomy on 12-wk-old female C57BL/6J mice. We surgically removed approximately 40% of the distal tip of the left SMG at a natural tissue margin (Fig. 1A), noting low variance in the size and weight of the tissue pieces removed (Fig. 1B, C). After partially resecting the glands, we allowed the gland to recover for 3, 14, or 56 d. We measured the length (Fig. 1D) and width (Fig. 1E) of the resected glands in situ and removed glands to weigh them (Fig. 1F). Glands were normalized to the excised gland piece measurements. Contralateral right glands in operated animals were also measured and showed a compensatory response (Appendix Fig. 1). The resected glands responded to the injury with measurable changes in gland length and width, as well as an increasing trend in gland weight.

Figure 1.

Response of female adult submandibular salivary glands to partial gland resection. (A) Image of partially resected female adult submandibular salivary gland (top left) and excised gland piece (bottom left) compared to unmanipulated control gland (right) at day 0. Image shows proximal to distal gland orientation with the sublingual gland (SLG) indicated with black dotted line. (B) Length and width of the excised gland pieces were measured with calipers in millimeters. Length (n = 32) and width (n = 32). (C) Weight of the excised gland piece was obtained in milligrams (n = 32). (D) Gland length was measured in situ with calipers after partial resection with immediate tissue harvest (day 0) (n = 9), 3 d postresection (day 3) (n = 11), 14 d postresection (day 14) (n = 8), and 56 d post resection (day 56) (n = 4) and normalized to gland length removed at the time of resection. (E) Quantification of gland widths measured with calipers and normalized to width of the excised gland piece. (F) Quantification of gland weight normalized to excised gland weight measured at all time points. Error bars are SEM. Statistical tests: 1-way analysis of variance Tukey’s honestly significant difference, *P ≤ 0.05.

Molecular Profiling Reveals an Intact Wound-Healing Response and Cell Cycle Entry

To gain insight into the nature of the gland response to resection at the level of RNA, we compared the transcriptome profile of the partially resected glands at days 3 and 14, in comparison to day 3 mock surgery controls. We examined the transcriptomes of SMGs from 3 mice for each treatment using Clariom S microarrays, which include 20,000 well-annotated genes known to be expressed. Principal component analysis (PCA) revealed that the transcriptomes of the glands at day 3 were distinct from the other samples, whereas the transcriptomes of the day 14 glands more closely resembled those of the control glands (Fig. 2A). Filtering for genes that changed more than 2-fold relative to control confirmed that many more genes were differentially regulated at day 3 than at day 14 (Fig. 2B) as shown in Volcano plots at day 3 (Fig. 2C) and day 14 (Fig. 2D). We used Metascape (Zhou et al. 2019) to identify gene categories that changed most significantly 3 d after partial gland resection. Metascape clustering analysis revealed changes in genes involved in cell cycle and repair processes, including, hemostasis, inflammation, cytokine production, and adaptive immunity (Fig. 2E and Appendix Fig. 3). These data indicate that the partially resected glands undergo a wound-healing response that is largely resolved within 14 d.

Figure 2.

Differential gene expression in partially resected glands at day 3 and day 14. Total RNA was isolated from adult female submandibular salivary glands that were subjected to mock surgery and harvested at day 3 (control C, blue), partial resection and harvested at day 3 (day 3, red), and partial resection and harvested at day 14 (day 14, purple). n = 3 animals for each treatment. Transcriptomes were obtained following hybridization with Clariom S arrays and analyzed with TAC software. (A) Principal component analysis (PCA) was performed to examine the variance between the samples. (B) The total genes were quantified that were changed >2-fold relative to C and having a 1-way between-subject P < 0.05 corrected for false-discovery rate (FDR) using the Benjamini-Hochberg procedure. (C, D) Volcano plot for differentially expressed genes at day 3 (C) and day 14 (D) relative to C with genes increased greater than 2-fold in red and decreased greater than 2-fold in blue. (E) Enriched Reactome (R-MMU) and Gene Ontology (GO) categories at each time point were identified with Metascape.

Proliferation Is Transiently Upregulated after Partial Gland Resection

As microarray analysis showed that genes involved in cell cycle progression were highly elevated at day 3 but not at day 14 after injury (Appendix Fig. 4), consistent with other organs that undergo an early peak of cell proliferation following partial resection (Voswinckel 2004; Michalopoulos 2010; Kirita et al. 2016), we more closely examined proliferating cells. Since Ki67 can be used to detect cycling cells in any stage of the cell cycle other than G₀ (Scholzen and Gerdes 2000), we used IHC to detect cells positive for Ki67 (Ki67⁺) in SMGs at 0, 3, 14, and 56 d after resection (Fig. 3A). At day 0, we detected a small number of Ki67⁺ cells scattered throughout the gland. At day 3, significantly increased numbers of Ki67⁺ cells distributed throughout the gland were detected relative to control glands with a small increase still evident at day 14 and nearly returned to control levels by day 56 (Fig. 3A, B). We quantified the cell proliferation rate using Ki67⁺ cells normalized to total nuclei (4′,6-diamidino-2-phenylindole [DAPI]⁺) (Appendix Fig. 2). We found that 18% of the cells were positive for Ki67 at day 3 after resection (Fig. 3B), which is a 15-fold increase in Ki67⁺ cells relative to control (Appendix Fig. 4). These data indicate that there is a global glandular response to partial gland resection during which cell cycle entry is stimulated 3 d after injury but is not sustained.

Figure 3.

Cell cycle entry is transiently increased globally in both secretory epithelial and stromal cells after partial salivary gland resection. (A) Immunohistochemistry (IHC) was performed in multiple tissue sections representing all regions of the gland to detect the proliferation marker, Ki67 (white), relative to 4′,6-diamidino-2-phenylindole (DAPI) (blue) in control female submandibular glands day 0 or 3, 14, or 56 d after partial resection. (B) Quantification of Ki67⁺ cells normalized to total DAPI⁺ cells. Day 0 (n = 3 glands), day 3 (n = 3), day 14 (n = 4), and day 56 (n = 4). Statistical test: 1-way analysis of variance Tukey’s honestly significant difference **P ≤ 0.01. (C) IHC to detect proliferating epithelium. EpCAM⁺ (red) and Ki67⁺ (white) at day 0 and day 3. (D) IHC to detect proliferating Aqp⁺ acinar cells with Aqp5 (green) and Ki67 (white). (E) Diagram showing periductal area (I, purple), acini (A, gray), ducts (D, green), and vasculature (V, red). (F) IHC to detect proliferating fibroblasts with vimentin (VIM, purple) and Ki67 (white) at day 0 and day 3. Most of the VIM⁺/Ki67⁺ cells are in the periductal regions. (G) Quantification of Ki67⁺/EpCam⁺, Ki67⁺/Aqp5⁺, or Ki67⁺/Vim⁺ cells at day 0 and day 3, normalized to DAPI in images from 3 glands with 3 to 4 images from each gland. Statistical test: Student’s 2-tailed t test, *P ≤ 0.05, **P ≤ 0.01. (H) IHC for Aqp5 (green) at day 0 or 3, 14, or 56 d after partial resection. (I) Quantification of Aqp5⁺ area normalized to DAPI. (J) IHC for the mucin functional marker, Muc10 (green), at day 0 or 3, 14, or 56 d after partial resection. (K) Quantification of Muc10⁺ area. IHC for Muc10 and Aqp5 was obtained from multiple regions in the gland, day 0 (n = 4 glands), day 3 (n = 3), day 14 (n = 4), and day 56 (n = 4). Error bars are SEM. Scale bars, 100 microns.

To determine which cells proliferate in response to injury, we performed IHC to detect Ki67⁺ together with cell type markers at day 3 after resection when Ki67 levels were highest. We first examined Ki67⁺ epithelial cells, using an antibody detecting the epithelial cell marker, EpCAM, together with Ki67 (Fig. 3C). The number of Ki67⁺/EpCAM⁺ cells was increased relative to control glands throughout the resected gland (Fig. 3G). To examine the response of secretory acinar cells, we examined Ki67⁺ cells expressing the water channel protein, aquaporin 5 (AQP5), which is expressed in proacinar and acinar cells in salivary epithelium (Nelson et al. 2013) (Fig. 3D). The number of AQP5⁺/Ki67⁺ cells was also increased relative to control glands (Fig. 3G), indicating that the proliferating epithelial cells were largely secretory acinar cells. To determine if stromal cells, which are enriched in the periductal areas (Fig. 3E), were proliferating, we examined Ki67 and the fibroblast marker vimentin (VIM⁺) (Fig. 3F). VIM⁺/Ki67⁺ cells were also increased relative to control glands. Together, these data indicate that both the epithelial and stromal cells undergo a global transient proliferative response to injury.

As the secretory capacity of the gland can be reduced by injury, we examined the secretory compartment over time with IHC. IHC for AQP5 and mucin 10 (Muc10) showed persistent expression of these acinar markers through day 56 (Fig. 3H, J). To quantify, we measured the area that was AQP5⁺ or Muc10⁺ from thresholded images and normalized to the number of DAPI⁺ cells. Globally, the AQP5⁺ area decreased slightly 14 d after injury and stabilized by day 56 (Fig. 3I); however, Muc10⁺ levels were sustained (Fig. 3K). Transcriptome analysis to detect SMG secretory genes previously identified from RNA-Seq studies indicated that messenger RNAs (mRNAs) encoding major secretory products were sustained at days 3 and 14 in resected glands (Appendix Fig. 5). Together, these data suggest that the partially resected gland maintains functional secretory cells after partial gland resection.

Cell Proliferation Is Maintained in Local Regions of the Gland While Secretory Differentiation Markers Are Reduced

To examine tissue architecture, we performed hematoxylin and eosin (H&E) staining of partially resected glands in comparison with control glands and detected notable regional differences. We observed localized regions having a different morphology (Fig. 4C) than the rest of the gland (Fig. 4B), characterized by increased nuclear density with increased numbers of small ducts, fewer acini, and more stromal tissue. These localized aberrant regions were proximal to the cut site and distal to the sublingual gland. These regions were detected in all of the partially resected glands at day 3 and day 14 and in three-fourths of the resected glands at day 56 (shown schematically in Fig. 4A). Local regions were defined by increased nuclear density (>1.5-fold DAPI/tissue area) compared to normal histological gland morphology (Appendix Fig. 6). To investigate cell proliferation within these localized regions, we counted the number of Ki67⁺ cells. Interestingly, while Ki67⁺ cells declined globally at days 14 and 56 after resection (Fig. 3A, B), a significantly higher number of Ki67⁺ cells were maintained at days 14 and 56 after resection (Fig. 4D, E). To examine functional acinar markers in these regions of high Ki67 levels, we measured the tissue area positive for AQP5 and Muc10 (Fig. 4F, H). Loss of both markers occurred at day 3 that became significant by day 14 and was sustained through day 56 (Fig. 4G, I). These data suggest there is a reduced functional secretory capacity in these local aberrant regions of the resected glands.

Figure 4.

An aberrantly proliferative region of the gland shows reduced secretory capacity following partial gland resection. (A) Cartoon depicting the general location of the aberrant repair response (black dotted line) with respect to the sublingual gland (SLG), submandibular gland (SMG), and resection line (red dotted line). (B) Hematoxylin and eosin (H&E) staining within a region of normal morphology (global area) at day 14 after resection. (C) H&E staining within a region of aberrant morphology, with a ductal-rich pattern (indicated by arrowheads) at day 14 after resection (local area). (D) Immunohistochemistry (IHC) to detect Ki67⁺ cells (white) and 4′,6-diamidino-2-phenylindole (DAPI) (blue) in aberrant local area at 0, 3, 14, and 56 d. (E) Quantification of Ki67⁺ cells in the local region relative to global region at all time points. (F) IHC for Aqp5⁺ (red) and DAPI (blue) in local aberrant region. (G) Quantification of Aqp5⁺ area in the local and global regions. (H) IHC for Muc10⁺ cells (green) and DAPI (blue) in local region. (I) Quantification of Muc10⁺ area normalized to DAPI in local versus global region. (J) Vimentin (VIM) (purple) and Ki67 by IHC. (K) Quantification of VIM⁺/Ki67⁺ cells at all time points in the local region compared to the global. Statistical analysis: Student’s 2-tailed t test, *P ≤ 0.05, **P ≤ 0.01. Global measurements from representative images: day 3, n = 3 glands; day 14, n = 4; and day 56, n = 4. Local measurements from representative images for Ki67: days 3 and 14 (n = 4) and day 56 (n = 3). Error bars, SEM. Scale bars, 100 microns.

Given the loss of secretory epithelial markers in the local region, we queried if any Ki67⁺ cells were stromal. We counted the VIM⁺ and Ki67⁺ (Fig. 4J) copositive cells in the local and global regions. The VIM⁺ cells comprised greater than 65% of the Ki67⁺ cells in the local region at days 3 and 56 (Appendix Fig. 7) and were significantly increased at day 14 (Fig. 4K). At all time points, VIM⁺ area was higher in the local region than in the global tissue (Appendix Fig. 7), suggesting that a fibroblast injury response was sustained within this aberrant local region of the gland.

Regional Differences in Macrophage Abundance and Fibrosis after Partial Gland Resection

Macrophages are a heterogeneous population of leukocytes that are important mediators of wound healing and regeneration that persist in chronic wounds. To examine macrophage abundance in the SMG after resection, we performed IHC to detect the glycoprotein F4/80 (Fig. 5A), which is widely used as a marker for macrophages (McKnight et al. 1996). We quantified the area covered by F4/80 in resected glands and observed a transient global increase in F4/80⁺ area at day 3 that returned to baseline by day 14 (Fig. 5B). In the local region, the area covered by F4/80⁺ signal (Fig. 5C, D) was significantly increased at day 3 and remained elevated at both days 14 and 56 in comparison to the corresponding global tissues (Fig. 5 E). These data indicate that macrophages are activated following partial resection but persist only within the local aberrant region.

Figure 5.

Partial gland resection transiently increases macrophages and extracellular matrix (ECM) deposition 3 d after resection that are sustained in localized aberrant regions. (A) Immunohistochemistry (IHC) for the macrophage marker, F4/80 (green), with 4′,6-diamidino-2-phenylindole (DAPI) (blue) at days 0, 3, 14, and 56. (B) Quantification of F4/80⁺ area normalized to DAPI. Statistical analysis: 1-way analysis of variance Tukey’s honestly significant difference, **P ≤ 0.01. (C) Overview image showing the localized and global areas with DAPI and F4/80; yellow dotted box shows 20× images depicted in D, scale bar 250 microns. (D) F4/80⁺ IHC and DAPI in the local aberrant areas at 3, 14, and 56 d. (E) Quantification of F4/80⁺ area in the local aberrant regions relative to the global. (F) IHC for the M2 macrophage marker, CD206 (magenta), F4/80 (green), with DAPI (blue) at days 0, 3, 14, and 56. (G) Quantification of CD206⁺ and F4/80⁺ colocalized area normalized to F4/80⁺ area. (H) Overview image showing the localized and global areas with CD206, F4/80, and DAPI; yellow dotted box shows 20× images depicted in I, scale bar 250 microns. (I) CD206⁺ and F4/80⁺ IHC with DAPI (blue) in the local aberrant areas at 3, 14, and 56 d. (J) Quantification of CD206⁺ and F4/80⁺ (gray) copositive area normalized to the local aberrant regions relative to the global F4/80⁺ area. (K) Masson’s trichrome staining at days 0, 3, 14, and 56. (L) Quantification blue area in Global Masson’s trichrome staining. (M) Overview image showing the localized and global areas with Masson’s trichrome; yellow dotted box shows 10× images depicted in N, scale bar 500 microns. (N) Masson’s trichrome stain in local aberrant regions at days 3, 14, and 56. (O) Quantification of Masson’s trichrome stain in the aberrant local regions. Statistical analysis: Student’s 2-tailed t test, *P ≤ 0.05, **P ≤ 0.01. All global measurements were taken from multiple, representative images (n = 3 for days 0 and 3, n = 4 for days 14 and 56). Images for local analysis were taken from multiple, representative images (n = 3). Error bars, SEM. Scale bars, 100 microns.

Macrophages are activated in response to injury and are generally categorized based on their 2 general functions in the wound repair response as proinflammatory (M1) or proregenerative (M2). The M2 macrophages attenuate inflammation and encourage wound repair while stimulating extracellular matrix (ECM) remodeling. We queried our microarray data to look for signatures of M1 versus M2 macrophages (Appendix Fig. 10). Finding evidence for a primarily M2 response, we used IHC to detect CD206, a cell surface receptor that is an accepted marker for M2 macrophages (Fig. 5F). Quantification of F4/80⁺ macrophages that also expressed CD206 revealed that very few macrophages expressed CD206 prior to injury (day 0) but that many expressed CD206 three days after injury when we detected ECM remodeling genes by microarray profiling (Fig. 5F and Appendix Fig. 10). Although there were no regional differences in CD206 expression 3 or 14 d after injury (Fig. 5I, J), there was an increased percentage of macrophages expressing the CD206 marker at 56 d specifically in the aberrant regions (Fig. 5H–J), correlating with the persistent proliferative cell response in that region.

As M2 macrophages are implicated in ECM remodeling, we used Masson’s trichrome, a classical histological stain that stains fibrillar collagens and elastins blue, revealing regions with high ECM deposition. Globally, we detected a small increase in trichrome staining at day 3 in resected glands relative to control that does not persist or progress (Fig. 5K, L), consistent with early ECM remodeling and a wound-healing response. Examination of genes in the ECM remodeling cluster identified by microarray at 3 d revealed increased expression of genes encoding ECM proteins detected by trichrome stain, including collagen 1 alpha chain 1 (COL1a1) and collagen 3 alpha chain 1 (COL3a1) (Appendix Fig. 8). These and other ECM proteins that comprise the key structural components of the ECM, called core Matrisome components, as well as genes whose products are associated with the ECM and involved in ECM remodeling and ECM-based signaling to cells, called Matrisome-associated proteins, were also elevated at day 3, returning to control levels by day 14 (Appendix Fig. 8). In the localized region that retained F4/80 signal, we observed elevated levels of trichrome staining that persisted to day 14 and 56 (Fig. 5M–O and Appendix Fig. 9), indicative of a fibrotic response that appears and persists in the aberrant local region of the gland after resection.

Discussion

The SG can partially regenerate after injuries such as ductal ligation and deligation but does not mount a full regenerative response capable of complete replacement of a functional gland after injury. Using a partial sialoadenectomy model, which lacks other stresses such as the physical pressure induced with ductal ligation and the significant DNA damage induced by irradiation, wherein 40% of the distal tip of the left female SMG was removed, we observed a global wound-healing response occurring within 3 d that was largely complete by day 14 that largely preserved the integrity of the gland. Interestingly, a small localized area of the gland exhibited persistent Ki67 expression, loss of functional secretory acinar cells, increased ECM deposition, and persistent M2 macrophage abundance consistent with a fibrotic response that is reminiscent of more advanced gland pathologies. Regional heterogeneity in salivary gland regenerative capacity has been previously noted by others (van Luijk et al. 2015), but the mechanisms remain poorly understood. This partial gland resection model thus provides the opportunity to study heterogeneous SG repair responses.

Although partial resection of 40% of the mouse SMG does not stimulate a significant regenerative response, the potential of SG cells to regenerate has been demonstrated in studies of ductal ligation and irradiation (Cotroneo et al. 2010; Weng et al. 2018). After deligation, the gland regenerates without intervention, with studies demonstrating that the secretory acinar cells repopulate through a mechanism of clonal expansion by employing lineage tracing of Mist1⁺ acinar progenitor cells (Aure et al. 2015) and involving a process similar to embryonic development (Cotroneo et al. 2010). Regeneration following irradiation can also occur but is typically partial and depends on the degree of injury (Weng et al. 2018). Multiple mechanisms can limit regenerative ability. Loss of progenitors leads to reduced replacement of lost functional tissue following SG resection and irradiation treatment for head and neck cancer patients (Vissink et al. 2010; Coppes and Stokman 2011). The loss of functional tissue following irradiation, specifically the loss of acinar cells, concomitant with late-stage irradiation damage, leads to increased infiltrate and fibrosis and further lack of function (Marmary et al. 2016). This partial gland resection model provides a simple model for elucidation of specific cellular mechanisms that limit regeneration. Defining the specific cellular factors that are limiting with each specific type of injury can be applied for development of specific therapeutics.

Cell proliferation is a common response to gland injury with the cell types responding and the degree of response affecting the outcome. Interestingly, the proliferative response in the liver is dependent on the degree of injury. In response to 30% resection, 7% of liver hepatocytes entered but did not complete cell division. Under this condition, liver cell mass was restored primarily by hypertrophy. However, in response to 70% resection, 66% of hepatocytes enter the cell cycle with most completing the cell cycle to restore liver mass (Fausto 2001; Mitchell et al. 2005; Miyaoka et al. 2012). After 40% resection of the female mouse SMG, we detected cell proliferation in 18% of cells at day 3, which returned to normal levels by day 14 and was insufficient to restore significant mass. Liver regeneration following 70% hepatectomy requires a growth factor–mediated priming phase to enable efficient hepatocytes to complete cell cycle and restore organ mass. Future efforts to improve regeneration of damaged SGs may benefit from identification of priming mechanisms to stimulate growth factor–mediated proliferation of parenchymal cells, similar to levels observed in successful salivary gland regeneration paradigms such as ductal ligation studies. Identification of such priming mechanisms could greatly benefit patients with xerostomia resulting from cancer therapy or autoimmune disease, as both cause progressive acinar cell loss that could be remediated by restoration of responsiveness to homeostatic cell-regenerative cues. Stimulation of endogenous cells following injury is an appealing therapeutic approach to promote regeneration that avoids complications inherent to cell transplantation therapies, including expansion in vitro, transplantation, and engraftment of transplanted cells.

Fibrosis is a component of many clinical conditions that limits organ function. In the salivary gland, fibrosis develops following irradiation treatment for head and neck cancers, resulting in loss of functional tissue that is often irreversible (Grundmann et al. 2009; Cheng et al. 2011; Lombaert et al. 2011). Similarly, in the autoimmune disease Sjögren syndrome (SjS), fibrosis increases in the salivary glands concomitant with loss of secretory acinar cells as disease progression occurs (Vitali 2002; Konttinen et al. 2006; Bookman et al. 2011; Llamas-Gutierrez et al. 2014; Gervais et al. 2015). Salivary hypofunction in patients with SjS is correlated with an increase in immune infiltrate into the glands and the appearance of fibrosis (Leehan et al. 2018). Interestingly, fibrosis is a hallmark not only of SjS but also of non-SjS hypofunction, consistent with glandular fibrosis interfering with gland function (Mulholland et al. 2015; Carubbi et al. 2018). Delineating the mechanisms driving regional wound-healing differences that occur in the salivary gland in response to partial gland resection may help to identify etiology-specific treatment options for xerostomic patients.

Macrophages are known to be central to the fibrosis that occurs in chronic injury. Although less is understood about the contribution of macrophages in salivary gland disease, macrophages have been implicated in disease progression in mouse models of SjS and in human patients (Goules et al. 2017; Morell et al. 2017). In the resected glands, we detected a transiently increased global macrophage population, which persisted in the local aberrant and fibrotic region of the gland suggestive of a continued state of wound repair in these regions. Interestingly, the macrophages largely coexpressed CD206, a cell surface marker expressed by M2 but not by M1 macrophage subtypes of the macrophage spectrum. Thus, further investigation into macrophage contributions to salivary disease and their function in fibrosis is warranted.

Reduction of fibrosis in the connective tissue compartment to restore a permissive regenerative environment facilitates regeneration in many contexts (Bataller and Brenner 2005; Lindquist and Mertens 2013; Rafii et al. 2013; Tanaka and Miyajima 2016; Cordero-Espinoza and Huch 2018). Whether remediation of fibrosis would be sufficient to restore function to damaged and diseased salivary glands is unknown, but limiting fibrosis may be beneficial as part of a combination therapy. Although current mouse models do not adequately recapitulate the fibrotic response in patients, the regional fibrotic response that occurs following partial SMG resection may be useful to elucidate mechanisms of salivary gland fibrosis that limit salivary gland regeneration and identify antifibrotic agents that promote functional restoration of fibrotic gland tissue. In addition, delineating the mechanisms driving regional wound-healing differences that occur in the SMG in response to partial gland resection may help to elucidate regional differences in gland responses to SjS.

Author Contributions

K.J. O’Keefe, D.A. Nelson, M. Larsen, contributed to conception, design, data acquisition, analysis, and interpretation, drafted and critically revised the manuscript; K.A. DeSantis, contributed to conception, design, and data interpretation, drafted and critically revised the manuscript; A.L. Altrieth, contributed to design, data acquisition, analysis, and interpretation, drafted and critically revised the manuscript; E.Z.M. Taroc, contributed to design, data acquisition, and analysis, drafted and critically revised the manuscript; A.R. Stabell, M.T. Pham, contributed to design and data acquisition, drafted and critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.