Introduction

Sialolithiasis is a rare, little-known, and poorly described disease. Its incidence rate is 1:10,000–30,000 individuals per year, and the condition constitutes around 30% of all salivary gland disorders. Sialolithiasis is a pathological condition linked to the formation of deposits called sialoliths in the salivary ducts or salivary glands1. The exact mechanism of sialolith formation is not fully understood, and many theories have been proposed. The potential causes of sialolith formation are thought to be injury and inflammation within the salivary glands, dental and endocrine disease, precipitation of salts, sialomicroliths, alkalinity of saliva causing precipitation of calcium phosphate, oral bacteria, and the presence of food debris and its migration (Fig. 1)1,2,3.

Fig. 1
figure 1

Schematic illustration of sialolith formation in the salivary glands. The salivary gland is composed of multiple acini that lead to a single excretory duct through intercalated ducts. Saliva flow is accelerated by the pumping of myoepithelium in the acinus, but there is a bottleneck area between the intercalated ducts and the excretory duct. If the saliva flow is retarded, a sialomicrolith may enlarge at the bottleneck area. Sialoliths are typically found at the bottleneck area with commensal bacterial infection.

Sialolith formation is multifactorial rather than attributed to a single factor alone. The salivary glands’ secretory activity is thought to have a significant impact on the biocalcification process. Studies have shown that the epithelial cells of salivary glands, where sialoliths form, can secrete multifunctional proteins and peptides as the main components of saliva. These components are responsible for lubrication, buffering, digestion, mineralization, and tissue coating and possess antimicrobial properties (i.e., via secretion of antimicrobial proteins and peptides, antibodies, and cytokines)4. A study by Schapher et al. showed that neutrophil extracellular trap formation is essential for sialolith development. Neutrophil extracellular trap formation is an inflammatory response that can be triggered by various factors, including contact between neutrophils and crystals formed by bacteria, calcium salts, cholesterol, or urate; pH variations; and foreign bodies5.

Several studies have been conducted to explain the growth mechanism of sialoliths using several methodologies, including ultrastructural analysis using electron microscopy and proteomic analyses using liquid chromatography‑mass spectrometry1,6,7,8. The inorganic components of sialoliths are variable and include calcium phosphate salts (hydroxyapatite), oxalates, and urates with traces of Mg, K, Cl, Al, Fe, S, Pb, Ti, and Zn9,10,11. And sialoliths show only 5% organic phase, which is predominant in the outer shell of the stones with glycoproteins, mucopolysaccharides, lipids, and cell detritus (phospholipid)12,13. The proteomic approach identified 824 unique proteins from 29 sialoliths, most of which were homologous or heterologous to bone proteins14. In particular, extracellular exosomes, known as carriers/transporters of immunologic and metabolic proteins and their regulators, were found to have a potential role in the structure of sialolith15,16. Although the calcification of sialoliths could be explained by comparison with the calcification of bone, enamel, dentin, and cementum, the molecular signaling mechanism for sialolith formation has not been clearly elucidated.

The purpose of this study was to expand our knowledge of sialolith formation, specifically by uncovering the roles of saliva and blood proteins, inflammatory signaling, epithelial proteins, inflammatory proteins, antimicrobial proteins, and cross-linking and calcification. A previous study8 described the structure of a sialolith as having three characteristic zones, i.e., a central nidus, an intermediate compact calcified zone, and a peripheral multilayer zone, as shown in Fig. 2. Histological and ultramicroscopic observations have revealed the presence of exosome-like vesicles, bacteria, and epithelial cells in the calcified matrix of the sialolith. In the present study, we investigated the protein expression in 10 stones to determine the cause of stone formation. Detecting proteins remaining in stones using ordinary immunohistochemistry (IHC) and western blot analysis can be difficult due to the denaturation or degradation of proteins in the calcified matrix of the stone.

Fig. 2
figure 2

Representative structures of sialolith. Microcomputed tomography (micro-CT) view of the sialolith (A). Hematoxylin and eosin (H&E) stain view of a sialolith (B). The signal of calcification initiated from the central nidus zone induces thick calcification of the intermediate compact zone and results in the repeated deposition of calcified material in the peripheral multi-layer zone.

To overcome this, the present study used immunoprecipitation-based high-performance liquid chromatography (IP-HPLC) to detect any protein containing the epitope derived from the target protein even after partial denaturation and degradation of the protein. In fact, the necrotic exudate of suppurative osteomyelitis of the jaw could be examined by IP-HPLC to determine the expression level of various proteins that might be denatured and degraded in a necrotic or proteolytic environment17.

A total of eight cases of sialolith, one case of antrolith, and one case of tonsillolith were examined by immunohistochemical staining and IP-HPLC, and three characteristic groups, i.e., the saliva/blood protein–dominant Group A, inflammation-dominant Group B, and bacterial infection–dominant Group C, were established according to the protein expression in the stones. This study discusses the pathogenic mechanism of stone formation, focusing on the expression of saliva/blood proteins, inflammation signaling proteins, inflammatory cytokines, and antimicrobial proteins.

Materials and methods

A total of 10 cases of stones (eight sialoliths, one antrolith, one tonsillolith) were obtained from surgically-treated patients and analyzed using IP-HPLC to assess the protein expression in stone formation (Suppl. Table S1). The present 10 cases of stones were not associated with dry mouth syndrome, calcium channel disease, and other problematic diseases except hypertension (one case) and type C hepatitis (one case). Specifically, the results of eight sialoliths were compared to the results of one antrolith and one tonsillolith as positive controls.

All patients underwent sialolithotomy, modified endoscopic sinus surgery, and tonsillolith removal under conscious sedation or general anesthesia at the Department of Oral and Maxillofacial Surgery of Seoul National University Dental Hospital (Suppl. Figs. S2-S3). The study protocol complied with the principles of the Declaration of Helsinki and was approved by the Institutional Review Board of Seoul National University (S-D20220023). All methods followed relevant guidelines and regulations. Patients were informed of the surgical procedure, including the potential risks and benefits, and provided informed consent to receive treatment and participate in the study.

Specimens were collected from January 2017 to July 2022 and were preserved with fixation in 10% buffered formalin solution as a routine protocol for further study.

IHC methods

As mentioned, each specimen was fixed in a 10% formalin-buffered solution and decalcified using a 0.5-M ethylene diamine tetra-acetic acid EDTA® (pH: 8.0) (Biosesang, Seongnam, Korea) solution. Tissue sections designated for examination under a light microscope underwent a series of steps, including dehydration, clearing, and impregnation. The tissue blocks were sliced into 4-µm sections and stained with hematoxylin and eosin (H&E). The resulting slides were scanned digitally using the Aperio CS2® digital pathology slide scanner (Leica, Wetzlar, Germany) and examined with the ImageScope 12.4.6 digital slide viewer application (Leica, Wetzlar, Germany).

For IHC staining, 4-µm sections of paraffin-embedded tissues were prepared and mounted on microscope slides using a semi-automated rotary microtome (Leica Biosystems, Nussloch, Germany). Each slide was examined under a light microscope (BX41 Light Microscope®; Olympus Co., Tokyo, Japan) to confirm the presence of all tissue structures. IHC analysis was performed for KL1, CK7, CK14, S100, α-SMA, β-catenin, BMP2/4, TGFβ1, and TNFα.

The IHC protocol involved several steps, including paraffin removal, achieved by one-hour oven drying; sequential immersion in xylene and ethanol with concentrations ranging from 100 to 50%; and a final rinse in phosphate-buffered saline. Antigen retrieval was performed through Proteinase K treatment (Dako, Agilent Technologies, Santa Clara, CA, USA). Sialolith microsection was mostly denatured; thereby, the immunoreaction could be unexpected. However, with the renaturing procedures via treatment with protein K and high salt, the epitopes of sialolith proteins were partly renatured, and their immunoreaction appeared ectopically in this study.

To eliminate blood cells and prevent false positives, endogenous peroxidase activity was blocked using 0.3% H2O2 for 30 min. Background blocking used 2.5% normal horse serum from the VECTASTAIN® ABC-HRP kit (Vector Laboratories, Newark, CA, USA).

The primary antibodies were diluted at a 1:100 ratio in Antibody Diluent® (Dako, Agilent Technologies, Santa Clara, CA, USA) and incubated overnight at refrigerated temperatures. The slides were then incubated at room temperature for 2 h. The VECTASTAIN® ABC-HRP kit from Vector Laboratories (Newark, CA, USA) was used for the application of a universal secondary antibody and ABC reagent. This was followed by the application of a peroxidase substrate solution (DAB substrate®; Vector Laboratories, Newark, CA, USA), which was monitored for the development of a brown color. Finally, a cover glass was applied.

Protein extraction

Maceration using liquid nitrogen

The optimal amount of sialolith for protein extraction was 500 mg per sample. Stones heavier than this amount were fragmented into smaller pieces until the standard weight was achieved. The samples were then placed in an aluminum mortar half-filled with liquid nitrogen for 2–3 min to stabilize. Following stabilization, the samples were crushed to a fine powder using a mortar and pestle. Sonication was carried out using a bacterial sonicator (VCX500, Sonics Vibracell™, Stanford, USA) for 15–20 min with 40% amplitude to break the stone particles until a cloudy suspension was obtained.

Each sample was diluted in conical tubes with an acidic protein–lysis buffer at a ratio of 10 mL per 0.5 g of sample. The buffer contained 40 mL of protein–lysis buffer (0.05 M Tris pH 8.0, 10 mM Hepes (1.19 g) pH 7.9, 10 mM KCl (0.37 g), 1.5 mM MgCl2 (0.071 g), 1 mM DTT, 0.3% SDS, 1% β-mercaptoethanol, and 0.4 mM PMSF) and 8 mL of 0.3 M citric acid at pH 4.0. The sample was mixed and stirred overnight, then centrifuged at 3000 rpm for 20 min to separate the supernatant, which was used for protein quantification and IP-HPLC analysis.

To compare the protein-expression levels between sialolith, antrolith, and tonsillolith, a control saliva solution was prepared by mixing whole saliva obtained from 10 healthy volunteers aged 20–40 years. Approximately 10 mL of whole saliva was collected from each subject, and 1 mL of citric acid solution (pH 4.0, 0.5 M) was added before centrifugation at 3000 rpm for 10 min. The supernatants were mixed, and protein-lysis buffer was added to the supernatant collection. This resulted in a standard control saliva solution, which was used in each IP-HPLC reaction.

Because the lysis sample of sialolith (sialolith sample) contains complex inorganic and organic materials, including denatured protein fragments, glycoproteins, lipoproteins from the calcified matrix of sialolith, it is difficult to measure the protein concentration of sialolith with ordinary protein assay. Therefore, this study performed two-step measuring assay to get the relative protein level in sialolith sample compared to normal saliva sample.

Each 30 µL of sialolith or normal saliva samples were analyzed by HPLC using a reverse phase column filled with non-adherent silica beads in 0.1 M NaCl, 20% acetonitrile (ACN) solution at UV260 nm. 0.3 mL/min, 30℃ to obtain each HPLC peak area (mAU*s). And the protein concentration of normal saliva sample was measured by Bradford protein assay using Bio-Rad protein assay kit (Bio-Rad Laboratories, Inc., USA). And then, the protein concentration of sialolith relative to that of normal saliva sample can be calculated by the proportional ratio between the HPLC peak area of sialolith sample and normal saliva sample.

Besides the normal saliva sample, a mixed sialolith sample was prepared by mixing the protein samples from 10 stones in a proportional ratio based on their protein concentrations. The normal saliva sample was used as a standard control, and the mixed sialolith sample was used as a comparison control. The resulting supernatant was stored in the refrigerator.

To perform immunoprecipitation, approximately 100 µg of protein extract was distributed to each protein A/G agarose column (Amicogen Inc., Jinju, Korea). Each column was pre-incubated with 1 µg of 48 different antisera, including those for saliva proteins (α-amylase, mucin1, mucin-4, mucin 5B, and PRPs), blood proteins (albumin, plasminogen, fibrinogen, and hemoglobin), inflammatory signaling proteins (NFkB, TNFα, LPS, TGFβ1, and TGFβ2), epithelial proteins (CK-7, CK-14, KL1, α-SMA, S100, and β-catenin), antimicrobial proteins (lactoferrin, lysozyme, LL37, CAMP, histatin, β-defensin1, β-defensin2, β-defensin3, cystatin A, and mucocidin), and cross-linking and calcification proteins (TGase1, TGase2, TGase3, TGase4, and CAP).

To determine the expression level of a protein, a total of 12 columns were prepared for 10 sialolith samples, one normal saliva sample, and one whole mixed sialolith sample, and filled with equal amounts of A/G agarose beads (approximately 0.3 mL) and the corresponding antibody (approximately 0.2 µg). The columns were incubated with stirring at room temperature for 30 min, and the supernatant was removed from the antibody-bound column to perform IP-HPLC.

To elaborate, we mixed each protein sample with 2.5 mL of binding buffer and then incubated in the antibody-bound protein A/G agarose bead column on a rotating stirrer for 2 h at room temperature. After washing the columns with Tris-NaCl buffer (pH 7.5) in a graded NaCl concentration (0.15–0.3 M), the desired proteins were eluted with 250 µL of IgG elution buffer (Pierce Biotechnology, Waltham, MA, USA).

IP-HPLC

The immunoprecipitated proteins were analyzed using a precise HPLC unit (1100 series; Agilent Technologies, Santa Clara, CA, USA) equipped with a reverse-phase column and a micro-analytical ultraviolet detector system (SG Highteco, Hanam, Korea). The column was eluted with a 0.15-M NaCl/20% acetonitrile solution, which flowed at a rate of 0.5 mL/min at a temperature of 30 °C. The proteins were subsequently detected via an ultraviolet spectrometer set to 280 nm.

The control and experimental samples were run consecutively to facilitate comparisons. In IP-HPLC, the protein peak areas (mAU*s) were obtained and calculated by subtracting the standard control antibody peak areas using an analytical algorithm. The protein-expression levels were analyzed and normalized using the square roots of protein peak areas.

Results

Clinical and radiological observations

In this study, a total of 10 cases, including eight sialoliths, one antrolith, and one tonsillolith, were evaluated. Sialolith cases ranged from patients aged 18–63 years, with most presenting mild to moderate symptoms such as swelling, discomfort, and radiographic detection of calcified masses in the submandibular or parotid regions. Surgical removal of the stones was performed successfully in all cases, with no significant post-operative complications or recurrences noted during follow-up. The antrolith and tonsillolith cases were also identified through radiographic imaging and treated surgically, with both patients recovering uneventfully (Suppl. Figs. S2S3).

Based on the pathogenic mechanism of salivary stones in Table 1, we can classify the eight stones into three groups according to the predominant protein expression in IP-HPLC. These groups are the saliva/blood protein dominant group (Group A), the inflammation dominant group (Group B), and the bacterial infection group (Group C). Case 1 belong to Group A, cases 2–4 belong to Group B, and cases 5–8 belong to Group C.

Table 1 Grouping based on the pathogenic mechanism.

Additionally, we examined a case of tonsillolith and a case of antrolith along with the aforementioned cases of sialoliths through clinicopathological and IP-HPLC investigations.

IP-HPLC analysis

Saliva/blood protein dominant group (Group A)

In Group A, comprising Case 1, the expression of salivary proteins such as mucin1 (107.8%) and mucin5B (127.1%), along with blood proteins like fibrinogen (122.3%), was notably higher compared to control samples. These proteins, alongside cross-linking proteins like TGase2 (129.2%) and TGFβ1 (116.6%), played a central role in stone formation through calcification and cross-linking processes. The findings suggest that this group’s stone formation is largely driven by salivary and blood proteins, without significant inflammation or bacterial infection (Fig. 3).

Fig. 3
figure 3

A star graph (A) for protein expression and a diagram (B) of protein signaling for sialolith in Case 1, a saliva/blood protein dominant group (Group A). Note the abundant salivary and blood proteins related to the cross-linking and calcification for sialolith formation in the lack of inflammation and bacterial infection.

Inflammation dominant group (Group B)

In Group B (Cases 2, 3, and 4), there was an increased expression of both salivary proteins and inflammation-related proteins. Elevated levels of NFκB, IL-6, and COX2 were observed, indicating a strong inflammatory response contributing to stone formation. These cases also showed enhanced expression of cross-linking proteins such as TGase1 and BMP2/4, suggesting that inflammation played a key role in promoting calcification. Specific protein levels for each case can be found in Fig. 4, and supplementary figures S4 and S5.

Fig. 4
figure 4

A star graph (A) for protein expression and a diagram (B) of protein signaling for sialolith in Case 2, a representative case of inflammation dominant group (Group B).

Bacterial infection dominant group (Group C)

Group C sialoliths (Cases 5–8) showed marked expression of bacterial exotoxin (LPS) along with the presence of saliva/blood proteins and inflammatory cytokines. Case 5 demonstrated elevated levels of LPS, TNFα, IL6, IL12, COX2, and cathepsin C, indicating a strong bacterial and inflammatory response. Salivary and blood proteins, including mucins and fibrinogen, were also upregulated, contributing to the calcification and cross-linking processes observed in the stones. Despite the bacterial infection, antimicrobial proteins were relatively under-expressed (Fig. 5A, B).

Fig. 5
figure 5

A star graph (A) for protein expression and a diagram (B) of protein signaling for sialolith in Case 5, a representative case of bacterial infection dominant group (Group C).

In Case 6, increased levels of LPS and inflammatory proteins were accompanied by higher expression of salivary proteins and cross-linking proteins such as TGase2 and BMP2/4. As with Case 5, antimicrobial proteins were under-expressed despite the presence of infection (Suppl. Fig. S6).

Case 7 exhibited elevated expression of salivary proteins, inflammatory cytokines, and LPS. Although inflammatory markers were upregulated, antimicrobial protein expression was generally low, with histatin being the exception. This case also showed upregulation of CK7, indicating epithelial involvement in stone formation (Suppl. Fig. S7).

In Case 8, salivary proteins and cross-linking proteins were prominently expressed, along with high levels of LPS and inflammatory proteins like IL6 and COX2. However, similar to other Group C cases, antimicrobial proteins were notably downregulated. S100 expression was slightly increased, but other epithelial proteins showed minimal changes (Suppl. Fig. S8).

These findings suggest that bacterial infection exacerbates inflammation and calcification in Group C sialoliths, while antimicrobial defenses are reduced.

Antrolith (Case 9)

The antrolith case showed consistent expression of saliva and blood proteins, including mucins and fibrinogen, along with cross-linking and calcification proteins such as TGase2 and TGase4. Epithelial marker proteins like CK14 and β-catenin were also observed. However, there was no significant expression of bacterial exotoxins, inflammatory cytokines, or antimicrobial proteins. While TNFα was slightly upregulated, NFκB and most inflammatory cytokines were markedly downregulated. Based on its protein expression profile, the antrolith was classified into Group A (Suppl. Figure S9).

Tonsillolith (Case 10)

The tonsillolith in Case 10 showed consistent expression of salivary proteins, including α-amylase and mucin5B, as well as blood protein fibrinogen. Epithelial proteins such as CK7, S100, and β-catenin were also present, alongside cross-linking and calcification proteins like CAP and ALP. Inflammatory proteins and antimicrobial proteins were expressed at lower levels, with COX2 and histatin showing slight increases. Based on the dominant expression of saliva/blood proteins, epithelial proteins, and calcification proteins, and the absence of significant inflammation signaling or antimicrobial proteins, the tonsillolith was classified into Group A (Suppl. Figure S10).

Whole mixed protein

Compared with control saliva, whole mixed protein exhibited higher levels of salivary proteins, such as mucin4 and PRPs, along with blood proteins like albumin and fibrinogen. Cross-linking and calcification proteins, including TGase1, TGase2, TGase4, and BMP2/4, were also upregulated. Additionally, whole mixed protein showed over-expression of bacterial exotoxin (LPS) and inflammatory proteins, such as IL6, COX2, and cathepsins, while antimicrobial proteins were generally under-expressed, with histatin being the exception. Epithelial proteins like KL1 and S100 were also consistently over-expressed. Based on these findings, whole mixed protein was classified as belonging to Group C (Suppl. Figure S11).

IHC observation in the microsections of sialoliths

Microsections of sialolith were analyzed by IHC using antisera of KL1, CK7, CK14, S100, α-SMA, β-catenin, BMP2/4, TGFβ1, and TNFα; samples were positive for BMP2/4, TGFβ1, and KL1, while the rest were rarely detected (Fig. 6).

Fig. 6
figure 6

Immunohistochemical (IHC) staining of sialoliths with BMP2/4, TGFβ1 and KL1 antibodies.AF: BMP2/4 showed positive results in small, round vesicles resembling exosomes located at the central nidus zone (CNZ) and became condensed near the intermediate compact calcification zone (ICCZ) of the sialolith. GL: TGFβ1 exhibited strong positivity in the inner layer of the ICCZ of the sialolith and consistent positivity in the outer layer, peripheral multi-layer zone (PMZ). MP: KL1 was weakly positive in the peripheral multilayered zone, also spotty focal positive in the CNZ of the sialolith. Positive reaction: arrows.

BMP2/4

BMP2/4 exhibited strong positivity in exosome-like vesicles near the border of the intermediate compact zone. This positive reaction was also diffusely positive in the proximal area of the intermediate compact calcification zone. However, BMP2/4 was rarely positive in the peripheral multiplayer zone (Fig. 6A–F).

TGFβ1

TGFβ1 was strongly positive in the proximal area of the intermediate compact calcification zone of the sialolith adjacent to the central nidus zone. It was also dispersed throughout the calcified matrix of the intermediate compact zone and peripheral multi-layer zone. However, there was rare immunoreaction of TGFβ1 in the central nidus zone (Fig. 6G–L).

KL1

The KL1 monoclonal antibody against pan-keratin exhibited weak staining in the peripheral multi-layer zone of the sialolith. Linear positivity was observed in the outer area of the lamellate calcified layer. Some calcifying granules showed spotty positive reaction in both the central nidus and intermediate compact calcifying zones (Fig. 6M–P).

Discussion

For evaluation of the expression of different proteins in the sialolith, antrolith and tonsillolith, whole saliva was selected as a control in the study because the saliva may contribute to the formation of stones. The stones contained abundant salivary proteins including mucins, PRPs, antimicrobial proteins, and cross-linking/calcification-related proteins1,14,18,19. The expression of fibrinogen, plasminogen, TNFα, TNF-κB, and inflammatory cytokines are not common in saliva. Thus, the expression of these in the stones may be caused by inflammatory reaction and bacterial infection during stone formation in the duct of the glands.

IP-HPLC analysis provided the relative expression level (%) of the target protein in the stone compared to that in the control study. To determine the actual expression level of the target protein, it must be considered relative to control saliva. For instance, the expression level of TGFβ1 in whole mixed protein samples was 94.8% compared to the control group. These data suggested a slight decrease of 5.2% in TGFβ1 expression compared to the control saliva, which typically shows significant expression of TGFβ1. The IP-HPLC data demonstrated a close relationship between sialolith formation and salivary proteins.

Depending on protein expression in the stones through IP-HPLC, the stones were divided into three groups, e.g., salivary protein dominant group (Group A), inflammation dominant group (Group B), and bacterial infection dominant group (Group C). This was not solely observational but rather driven by the predominant protein types and their functional roles in the formation of the stones, as revealed by the data. The three factors, saliva/blood proteins, inflammation, and bacterial infection are not mutually exclusive and can co-occur in the same patient depending on the stage of sialolith formation. However, the grouping reflects the predominant factor driving stone formation at the time of analysis. This classification offers insight into the dynamic pathophysiological process of sialolith formation, where different factors may dominate at different stages of the disease.

Although the antrolith and tonsillolith occurred in the maxillary sinus and tonsil, respectively, they exhibited similar protein expression to the sialolith, Therefore, those stones were investigated using the same methods in the study. The inclusion of an antrolith and a tonsillolith aimed to broaden the understanding of the pathophysiological mechanisms underlying stone formation in various anatomical regions. The findings from the antrolith and tonsillolith cases provides comparative data, supporting the notion that similar protein expressions and calcification mechanisms occur in these stones, as observed in sialoliths.

A case of sialolith (Case 1), a case of antrolith (Case 9), and a case of tonsillolith (case 10) showed dominant expression of salivary proteins, cross-linking/calcification-related proteins, and epithelial marked proteins in the absence of remarkable inflammation and bacterial infection, and they were classified into Group A. Group A stones showed dominant expression of salivary and blood proteins, mainly mucins, PRPs, fibrinogen, and plasminogen, along with the expression of cross-linking enzymes and calcification-related proteins, including TGases, BMP2/4, CAP, DSP, and ALP. They did not show inflammatory proteins and bacterial exotoxin formed by the aggregation and cross-linking of different salivary and blood proteins in the absence of inflammatory reactions caused by bacterial infection.

The cross-linking enzyme TGase4, also known as TGase-P, is abundant in saliva and blood20,21. Its main substrates are mucins and PRPs for adherence to mucosa, and fibrinogen for blood coagulation22. The over-expression of TGases in the stones is a relevant finding that explains the mechanism of stone formation22. Cross-linking enzymes were found in all cases of sialolith, antrolith, and tonsillolith observed in this study. Thus, substrate molecules can form strong covalent bonds with large protein materials, which may contribute to the formation of sialoliths in the salivary duct.

In this study, we didn’t detect the serum electrolyte level in the sialolithiasis patients, but the calcium-binding proteins including S-100, TGases, BMP2/4, DSP, ALP, and CAP were significantly increased in the saliva/blood protein dominant group (group A) and inflammation dominant group (group B) compared to the bacterial infection dominant group (group C). The data indirectly indicate that the Ca + + level in the salivary flow was well preserved in groups A and B, but rather decreased in group C. However, exclusively the case 8 of the group C sialolith showed the overexpression of TGase 1 (109.1%), TGase 2 (109. 2%), TGase 4 (147.4%), BMP2/4 (144%), DSP (111.4%) and ALP (108.7%) as well as IL6 (121.6%), COX2 (107.6%), cathepsin G (139.7%), cathepsin K (110.1%) and IgK (146.6%), probably due to a severe inflammatory reaction with necrosis.

We found that all observed sialoliths expressed epithelial marker proteins, particularly CK14 and KL1, which were over-expressed compared to control saliva, while CK7 was not. Since CK7 (which is usually present in the excretory ductal epithelium of the salivary gland) was not detected in the sialolith23, it is assumed that the epithelial cells present in the sialoliths originate from the intercalated ductal cells rather than the excretory ductal cells. The fact that the sialolith frequently occurred at the bottleneck zone of saliva flow, which located between intercalated ducts and excretory ducts at the interlobular space as shown in Fig. 1, may partly support the intercalated duct origin of the exfoliated epithelial cells in sialoliths.

Various antisera were used for IHC staining on stone microsections, including KL1, CK7, CK14, S100, α-SMA, β-catenin, BMP2/4, TGFβ1, and TNFα. However, the immunoreactivity was generally weak, even after antigen-retrieval treatment. BMP2/4 showed strong positivity in the exosome-like vesicle located in the central nidus zone and proximal area of the intermediate compact calcification zone. TGFβ1 exhibited different levels of positivity in the intermediate compact calcification zone, and KL1 was partially positive in the peripheral multi-layer zone. The remaining antisera showed almost no positivity. It is clear that the calcification of stones is affected by the protein expression of the TGFβ/BMP signaling pathway, which is mixed with exfoliated salivary epithelial cells.

IHC staining of the sialolith microsections showed a slight positive reaction of KL1 for pan-keratine. However, there were rare positive reactions of CK14, α-SMA, and S100, which are typically used for detecting myoepithelial cells24. Therefore, it is unclear whether the exfoliated epithelial cells in the sialolith were derived from myoepithelial cells in this study. However, the epitopes of CK14, α-SMA, and S100 may be degenerated and undetectable on immune histochemical reaction in the calcified matrix of sialoliths. In contrast, IP-HPLC showed increased expression of CK14, α-SMA, and S100 in some sialoliths compared to control saliva. The origin of salivary exfoliated cells found in sialoliths remains ambiguous and requires further elucidation.

However, it was found that the sialoliths contained both the salivary/blood-derived proteins and the exfoliated epithelial cells in histology, ultramicroscopy, IHC and IP-HPLC. These organic materials may reinforce the sialoliths’ lamellate structure, making them more resistant to lithotripsy than renal calculi.

The sialoliths in group B showed expression of salivary and blood proteins, as well as inflammatory proteins, but no LPS expression. The chronic obstruction and irritation of the salivary duct caused by the enlarging sialolith may damage the luminal surface of the ducts and subsequently induce an inflammatory reaction25. The ultramicroscopic observations revealed many exosome-like vesicles in the control nidus area8. Therefore, we suggest that the formation of sialolith nidus is closely associated with cellular exosomes. We found that exosome-like vesicles in the central nidus area strongly expressed BMP2/4, while group B sialoliths consistently expressed blood proteins (fibrinogen and plasminogen), cellular inflammation signaling proteins (NFkB and TNFα), and various inflammatory cytokines. The cellular proteins in the exosome-like vesicles may create a biochemical environment that facilitates ossification in the lumen of salivary ducts. This process may be aided by cross-linking enzymes and high levels of calcium in saliva8,26. Salivary inflammation can induce NFkB-mediated TNFα expression27, which may lead to calcification of substrate proteins such as BMP2/4, CAP, DSP, and ALP, resulting in atypical sialolith calcification.

The sialoliths in Group C were infected with bacteria, as evidenced by the expression of LPS, salivary/blood proteins, cross-linking/calcification protein, inflammation signaling proteins, and inflammatory proteins. Ultramicroscopic observation revealed the presence of numerous bacteria in both the control nidus area and the peripheral multi-layer zone8. Bacillus subtilis, a type of bacteria that produces a sticky biofilm, was identified through bacterial culture using a preserved sialolith28.

Sialolithiasis involving the major salivary gland was partly responsive to strong antibiotic therapy, but it frequently led to suppurative sialadenitis and salivary abscess when the sialolith became enlarged due to severe bacterial infection29. However, it is evident that the sialolith provides a favorable site for bacterial growth to evade strong antibiotic therapy and host immune reactions by producing a thick biofilm within the sialolith. The bacterial cross-linking enzymes responsible for biofilm production may play a similar role in sialolith formation to human cross-linking enzymes, contributing to the initial nidus formation of sialoliths30.

The oral cavity has a unique defense mechanism through the presence of antimicrobial proteins in whole saliva, including lysozyme, LL-37, CAMP, histatin, lactoferrin, and β-defensin 1, 2, 331,32. These proteins protect against bacterial infection and neutralize harmful toxins32. The study found weak expression of antimicrobial proteins in the sialolith of Groups A and B, and rare expression in Group C despite severe bacterial infection. The antrolith (Case 9) and tonsillolith (Case 10) also exhibited weak expression of antimicrobial proteins, similar to the Group A sialolith (Case 1).

This study does not clearly explain why the expression of antimicrobial proteins decreased in the stones. However, it suggests that the antimicrobial proteins are typically cationic and can interact with anionic bacterial membrane and toxins33, but cannot adhere or anchor to the calcified matrix of stones. If the stones were infected with bacteria and placed in an acidic environment, the antimicrobial proteins in the saliva would dissolve and wash out from the stones. As a result, the Group C sialoliths exhibited much lower expression of antimicrobial proteins than the control saliva in this study.

Among the antimicrobial proteins, mucocidin is a commonly expressed antimicrobial protein in human saliva, and was subsequently subcloned as a saliva gene (GenBank AY177672) from the human submandibular gland cDNA library34. However, its encoding site has not been found in the human genome, leading to the suspicion that mucocidin is an orphan gene, possibly derived from some oral commensal bacteria35. It was found that the Case 1 sialolith (105.8%) of Group A and Case 4 sialolith (109.3%) of Group B showed slight expression of mucocidin compared to the control saliva. Therefore, we suggest that mucocidin is also weakly relevant to stone formation as other human antimicrobial proteins observed in this study.

TGF/BMP signaling plays a significant role in bone formation, and TGFβ1 can increase the expression of BSP/DSP as well as BMP236. Therefore, the over-expression of TGFβ1 (118.6%) in the sialolith (Case 1) of Group A may contribute to sialolith formation along with the over-expression of CAP (111. 3%), DSP (107.0%), and ALP (109.1%), but the under-expression of NFkB (81.6%), IL1 (52.3%), IL8 (67.8%), cathepsin C (79.5%), cathepsin K (81.2%), and IgK (80.5%).

Under normal inflammatory conditions, the release of the proinflammatory cytokine TNFα inhibits osteogenic differentiation. However, TNFα can paradoxically promote osteogenic differentiation through NFkB signaling by producing BMP2, BSP, and ALP37.

In Case 2 (Group B), a sialolith showed increased levels of NFkB (111.3%), TGFβ1 (119.3%), IL6 (136.7%), IL8 (105.2%), IL12 (219.9%), COX2 (115.5%), cathepsin K (113.2%), and IgK (171.7%), leading to increased protein expression of BMP 2/4 (150.9%) and ALP (107.7%). The low level of TNFα in Case 2 may be due to the TNFα protein properly degrading rapidly by serum protease in time-dependent manner compared to other inflammatory cytokines. However, we suggest TNFα and TGFβ1 are upregulated and cooperatively contribute to the ossification/calcification of sialolith.

Transglutaminase (TGase) may contribute to stone formation by catalyzing the N2-(r-glutamyl)-lysine cross-linking of precursor proteins. Four types of TGase were explored through IP-HPLC: TGase1 (Keratinocyte TGase), TGase2 (Tissue TGase), TGase3 (epithelial TGase), and TGase4 (prostate TGase). TGase4 was over-expressed in the stones compared to the control saliva, while the other TGases were weakly expressed. It was suggested that secretory TGase4 plays a crucial role in inducing salivary protein coagulation, which leads to the formation of stones.

A previous study by Tretiakow et al. classified sialoliths into three groups: calcified, organic/lipid, and mixed38. Based on this classification, Musiał et al. identified unique protein compositions for each group1. Our present study cannot be matched with these classified criteria, because we focused on the lamellate calcified sialolith structures formed by cross-linking calcification proteins via TGFβ1/BMP signaling and inflammatory TGFβ1/NFkB signaling.

Ectopic calcification of stones in the external spaces of the body appears as sialoliths in the salivary duct, antroliths in the maxillary sinus, tonsilloliths in the tonsillar crypt, renal stones in the renal pelvis or ureter, and gallstones in the gallbladder. Sialoliths are the most common, occurring at a rate of 28–59 cases per million per year39. While small sialoliths are typically asymptomatic and cause minimal discomfort, larger ones can interfere with saliva flow and cause pain and swelling40. Sialoliths account for only 1–2% of head and neck pathology in clinical patients, while asymptomatic sialoliths are found in up to 0.45% of outpatients41.

In this study, four stones were found in female patients and six stones were found in male patients, indicating a higher incidence in males (60%) than in females (40%). However, it would be difficult to define the gender difference in sialolith occurrence due to the small number of stones, only 10 cases, for statistical analysis. On the other hand, among three sialoliths of inflammation dominant group (group B), two were from female and one was from male, while among four sialoliths of bacterial infection group (group C), three were from male and one was from female. The data suggest that female patients may be more resistant to bacterial infection via overexpression of inflammatory proteins and innate immunity proteins than male patients. However, further investigation is also needed to evaluate the sex difference in inflammatory response and innate immunity against sialoliths.

It is essential to understand the pathogenic mechanism of stone formation and provide clinical therapies for prevention and treatment. This study explored protein expression in 10 stones, identifies patterns in three different stone groups, and discusses surgical enucleation of large stones that cannot be removed by lithotripsy or endoscopic surgery. It also suggests that salivary health should be emphasized to prevent stone formation in the salivary ducts.

Healthy salivary glands are necessary both to maintain the health of the oral mucosa and teeth, and to inhibit the stagnation of mucinous saliva in the salivary duct by increasing the flow of saliva. The formation of sialoliths, antroliths and tonsilloliths can have significant implications for disease progression. Therefore, further studies using different clinical data and molecular biological methods are recommended.

Conclusion

In the present study, eight cases of sialoliths, one case of antrolith, and one case of tonsillolith were examined by IHC and IP-HPLC analysis. The results of the IHC staining showed that BMP2/4 was strongly positive for the exosome-like vesicles in the central nidus zone and diffusely positive in the intermediate compact calcification zone. TGFβ1 was strongly positive in the proximal intermediate compact zone and dispersed in the distal direction. In the peripheral multi-layer zone, KL1 showed occasional positivity with a linear or spotty appearance. The stones were categorized into three groups based on their protein expression using IP-HPLC. Group A exhibited dominant expression of saliva and blood proteins, as well as cross-linking and calcification proteins. Meanwhile, Group B showed increased expression of inflammation signaling proteins, such as NFkB and TNFα, as well as various inflammatory cytokines and anti-microbial proteins, even in the absence of bacterial infection (as determined by LPS expression). Group C had a severe bacterial infection, similar to Groups A and B, which also exhibited protein expression. Additionally, Group A had one case of antrolith and one case of tonsillolith. The data shows that TGFβ1 and TNFα work together to express cross-linking/calcifying proteins with exfoliated epithelial cells. Bacterial infection may worsen salivary inflammation and lead to the formation of sialoliths with thick biofilm as illustrated in Fig. 7. This results in lamellate calcification reinforced by organic materials, making it partially resistant to lithotripsy. Sialoliths are believed to form through the expression of saliva and blood proteins, as well as cross-linking and calcification proteins via TGFβ1/BMP signaling. Inflammation and bacterial infection can also accelerate lithogenesis through TGFβ1/NFkB signaling.

Fig. 7
figure 7

Schematic illustration of sialolith formation. Sialoliths can be divided into three groups based on their protein expression: Group A, which is dominated by salivary and blood proteins, Group B, which is dominated by inflammation, and Group C, which is associated with bacterial infection. The formation of a sialolith may be primarily caused by the coagulation of salivary and blood proteins through cross-linking enzymes (TGases) and calcification/ossification via TGFβ/BMP signaling. This process could be subsequently accelerated by inflammation signaling through TNFα/NFkB and bacterial infection.