Introduction

Cytokines are signaling molecules that play an important role in mediating and regulating immunity, inflammation, and hematopoiesis. Cytokines can be produced locally or transported by blood and can be measured in various parts of the body, including saliva. Salivary cytokines have been reported to be associated with various systemic and oral disorders1,2, including autoimmune and endocrine diseases3,4,5, psychiatric conditions6,7, and infectious diseases such as tuberculosis8. Furthermore, oral pathologies, such as caries9, gingivitis10, and periodontitis11,12, are associated with certain pro- and anti-inflammatory cytokine profiles. In patients with cancer, salivary cytokines have been reported to be associated with oral inflammation and oral complications during chemotherapy13. Oral mucositis14, xerostomia15, gingivitis, and caries16 are the most commonly observed oral complications in patients with cancer. These complications can be detected early through biomarkers such as salivary cytokines. Several studies have reported elevated salivary cytokine levels during chemotherapy in adult patients with cancer. Furthermore, interleukin (IL)-10 and IL-6 have been reported to correlate with the severity of oral graft-versus-host disease in patients with leukemia17, and IL-6, IL-10, and tumor necrosis factor (TNF) levels have been reported to be associated with the extent of oral mucositis18. Salivary cytokines are emerging as potential biomarkers for cancer diagnostics and treatment monitoring in certain cancer types. Research has shown that cytokines like TNF, IL-6, and IL-8 found in saliva are linked to cancer progression, offering non-invasive insights into disease status and treatment responses, especially in cases of oral cancers19,20.

However, only one study has explored salivary cytokines in pediatric patients with cancer, which reported that salivary IL-2 levels were two times higher during chemotherapy in patients with acute lymphocytic leukemia and that these patients showed an increased incidence of gingivitis during treatment compared with healthy controls21. In our study, we examined various cytokines based on their functions, including pro-inflammatory cytokines IL-1α, IL-1β, and TNF; the anti-inflammatory IL-10; IL-4 and IL-5, which promote Th2 differentiation and eosinophil function; the multifunctional IL-6; neutrophil-attracting IL-8; IgE-regulating IL-13; infection- and tumor-defending IFN-γ; and inflammation-recruiting IP-1022,23.

This study aimed to compare salivary cytokine profiles between newly diagnosed patients with cancer and healthy controls. Oral clinical findings were documented to evaluate whether cytokine changes were associated to local infections or the specific oncological disease.

Additionally, this study aimed to investigate changes in salivary cytokine levels in patients with cancer during follow-up assessments 3 and 6 months after diagnosis.

Results

Study population

A total of 19 patients with cancer and 128 healthy controls aged 4–18 years were included in this study. Of the 19 patients in the cancer group, 9 were children (< 12 years), and 10 were adolescents (≥ 12 years). Of the 128 healthy controls, 77 were children, and 51 were adolescents. Tables 1, 2 and 3 show patient characteristics.

Dental examination revealed that the presence of caries and gingivitis at diagnosis in patients with cancer was comparable to healthy controls (5/19 and 49/128, P = 0.45; 3/19 and 35/128, P = 0.40, respectively). Hyposalivation was only present in patients with cancer with leukemia (Table 2).

Table 1 Baseline characteristics of patients with cancer and healthy controls.
Full size table
Table 2 Baseline characteristics of patients with cancer and healthy controls, including subgroups of patients with cancer.
Full size table
Table 3 Cancer group characteristics: exact diagnosis, age at diagnosis, type of therapy and remission status at 3 months, 6 months and the end of observation time.
Full size table

Salivary cytokines at baseline: comparison between children and adolescents with cancer and healthy controls

Patients with cancer showed higher median levels of the salivary cytokines IFN-γ IL-1α, IL-1β, IL-5, IL-10, and TNF at diagnosis than healthy controls (Table 2). Multivariate linear regression models for each cytokine showed significant differences in cytokine levels between patients with cancer and healthy controls after adjusting for sex, presence of gingivitis or caries, and salivary flow rate (Table 4). In the children subgroup, cytokine levels were generally 1.5–4 times higher in patients with cancer than in healthy controls, with IFN-γ, IL-4, IL-5, IL-10, and TNF levels being significantly elevated. In the adolescent subgroup, cytokine levels were 2 times higher in patients with cancer than in healthy controls, with IL-1α and IL-8 levels being significantly elevated (Table 4).

Cancer group characteristics

In the leukemia group (n = 10), there were nine patients with common B-cell acute lymphoblastic leukemia (ALL) and one with acute myeloid leukemia (AML). All lymphoma patients had Hodgkin lymphoma at various stages. Additionally, there were four patients with different solid tumors: one with medulloblastoma, two with Ewing’s sarcoma (one with metastases and one without), and one with alveolar rhabdomyosarcoma. By the end of the study, 16 patients were in remission, while three patients died during the observation period. All cancer patients received chemotherapy, except one ALL patient who underwent hematopoietic stem cell transplantation (eight months after initiation of treatment), and three patients received additionally radiation therapy during the first 6 months after diagnosis (see Table 3).

Salivary cytokines at baseline in patients with cancer stratified by diagnostic groups

Children with leukemia exhibited significantly higher median levels of IFN-γ, IL-4, IL-5, IL-10, and TNF, while adolescents displayed significantly elevated median levels of IFN-γ, IL-1α, IL-1β, and IL-8. Children with solid tumors had notably increased median levels of IL-10 and TNF. However, median IL-5 levels in adolescents with lymphoma were significantly lower than those in healthy adolescents (Table 4; Fig. 1).

Table 4 Comparison of cytokine levels between patients with cancer and healthy controls represented as geometric mean ratios with 95% confidence intervals (coefficients of the models) and the corresponding P values.
Full size table

Analysis of the relationship between salivary flow rates and cytokine concentrations revealed that the greatest reduction in cytokine concentration per flow rate increase of 1 ml/min was observed for IL-1β in patients with leukemia (− 59%), solid tumors (− 43%), and lymphoma (− 59%) (see Supplementary Tables 1a-c).

Fig. 1
figure 1

Salivary cytokines as boxplots with individual data points stratified by healthy children and children with leukemia, solid tumors, and lymphoma.

Full size image

Time course of salivary cytokines in patients with cancer

The time course of salivary cytokines was assessed in patients with cancer at three time points: at diagnosis, 3 months after diagnosis, and 6 months after diagnosis (n = 19) (Fig. 2; Table 5). In children with cancer, a significant increase in the geometric mean ratios for IL-6 and inducible protein (IP)-10 was observed 3 months after diagnosis compared with that at diagnosis. Similarly, in adolescents with cancer, median IL-10 levels showed a significant increase 3 months after diagnosis compared with that at diagnosis. Six months post-diagnosis, the geometric mean ratio for IL-1α was significantly higher in children, while IL-1β and IL-8 levels were significantly elevated in adolescents with cancer compared to their levels at diagnosis (Table 5).

Fig. 2
figure 2

Time course of salivary cytokine levels in children and adolescents.

Full size image

Regarding the salivary flow rate, no significant differences in stimulated salivary flow rate were observed between patients with cancer and healthy controls at diagnosis, 3 months after diagnosis, or 6 months after diagnosis. Subgroup analysis revealed that patients with lymphomas had significantly higher salivary flow rates (Table 2). Additionally, with increasing age, the salivary flow rate increased by approximately 0.1 ml/min per 1 year for both patients with cancer and healthy controls (Figs. 3 and 4). Mixed-effects regression models showed that the cofactor salivary flow rate had the most significant effect on cytokine concentrations (see Supplementary Table 2). For all cytokines, apart from IL-4, IL-5, and IP-10 in patients with leukemia and lymphoma, the salivary flow rate was reduced.

Fig. 3
figure 3

Stimulated salivary flow rate (ml/min) in healthy children stratified by age (years).

Full size image
Fig. 4
figure 4

Stimulated salivary flow rate (ml/min) in patients with cancer stratified by age (years).

Full size image
Table 5 Results of multiple linear regression analyses of cytokine levels in patients with cancer 3 and 6 months after diagnosis compared with those at diagnosis represented as geometric mean ratios with 95% confidence intervals (coefficients of the models) and the corresponding P values.
Full size table

Discussion

This study investigated salivary cytokines in newly diagnosed pediatric patients with cancer and healthy controls. In the study population, the age distribution and types of diseases specifically reflected the most common childhood cancers, with leukemia being the most prevalent, followed by lymphoma24.

The most remarkable finding was that, at diagnosis, the levels of IFN-γ IL-1α, IL-1β, IL-5, IL-10, and TNF were observed to be elevated in the saliva of patients with cancer compared with healthy controls. It could be hypothesized that these findings are mainly influenced by dental and oral status. However, in the study groups, the frequency of caries and gingivitis in patients with cancer was comparable to that in healthy controls. This finding indicates that the presence of cancer or the resulting immune response of the individual causes changes in salivary cytokines.

The findings of this study are consistent with those reported by Bien et al.25 who showed that, at diagnosis, children with cancer (acute lymphoblastic leukemia, soft tissue sarcoma, or Hodgkin lymphoma) had significantly higher serum IL-10 levels than healthy controls. However, this study did not measure salivary cytokine concentrations. Another study by Resende et al.26 showed that, in adults, the IL-10 concentration in the blood correlated with the occurrence of graft-versus-host disease after HSCT. This result did not apply to the IL-10 concentration in saliva. In patients with lung cancer, IL-10 suppresses local antitumor immunity, which can affect patient outcomes27. Conversely, in patients with early-stage breast cancer, IL-10 is associated with improved disease-free survival and breast cancer-specific survival, indicating a potentially protective role28. IL-10 plays a complex and multifaceted role in patients with cancer and can promote cytotoxic T-cell responses, which are crucial for antitumor immune responses29.

In the context of oral health, a previous study showed that IFN-γ along with IL-4 plays an important role in the pathogenesis of erythematous/ulcerated oral lichen planus, and changes in these proinflammatory cytokines in whole unstimulated saliva reflect the lesion status30. These findings indicate that IFN-γ in saliva could be indicative of certain inflammatory conditions in the oral cavity, which might be relevant in the context of oral cancers31.

Interestingly, in this study, subgroup analysis revealed differences among patients with cancer with leukemia, solid tumors, and lymphoma. The levels of IFN-γ, IL-1α, IL-1β, IL-4, IL-5, IL-8, and IL-10 were significantly elevated in patients with cancer with leukemia compared with healthy controls. Furthermore, the levels of IL-10 and TNF were elevated in patients with solid tumors compared with healthy controls. These findings were not observed in patients with lymphoma, where no cytokines were elevated. This may be due to the less aggressive growth patterns seen in some lymphomas compared to other cancers, potentially linked to lower systemic inflammation. However, it’s crucial to recognize that certain pediatric lymphoma subtypes, like Burkitt lymphoma and lymphoblastic lymphoma, demonstrate aggressive characteristics to those comparable to acute leukemia32,33. In our patient group, only Hodgkin lymphoma was present, and no other more aggressive subtype.

The salivary flow rate was double in the lymphoma group compared with that in other cancer groups, which also could lead to lower cytokine levels34. This fact must be considered when analyzing cytokine levels. Therefore, salivary flow rate was included as a covariable in the regression models. The stimulated salivary flow rate has been reported to increase during childhood and adolescence35,36,37, peaking around 15 years of age when the salivary glands are fully developed, and decreasing with higher age38,39.

Recent evidence underscores the necessity of contextualizing salivary cytokine findings with respect to age37. A previous study reported a positive correlation for most cytokines with age and oral pathologies, such as caries and gingivitis, and a negative correlation with the stimulated salivary flow rate34. Other studies have shown, that salivary pH is significantly lower in patients with cancer40,41. In our oncological population, the mean salivary pH was 7.8 which is comparable to the mean pH value of 7.5 in healthy children aged 6–15 years reported by Forcella et al.42. Therefore, in this study, the parameters for the regression models comparing patients with cancer and healthy controls and cytokine levels 3 and 6 months after diagnosis were adjusted for age, sex, gingivitis, caries, salivary flow rate, and, only for time course, salivary pH. Comparing healthy controls to the cancer group, no significant statistical differences were observed in oral and dental status, except that children with leukemia had a higher likelihood of experiencing hyposalivation. Adjustment for these parameters enhanced the reliability of the findings and demonstrated the systemic impact of pathophysiological processes in the underlying cancer.

Furthermore, the time course of salivary cytokines was explored in patients with cancer under treatment after adjustment for oral findings. After adjustment, elevated IL-6, IL-10 and IP-10 levels were observed 3 months after diagnosis and IL-1α, IL-1β and IL-8 at 6 months after diagnosis respectively. IL-6 has been reported to promote tumor growth by inhibiting apoptosis in cancer cells and stimulating angiogenesis43. In adult patients with hepatocellular carcinoma, lower IL-6 levels have been reported to be associated with a favorable survival prognosis44. IL-10 is highly immunosuppressive and is expressed by different tumor cells, such as sarcoma, B-cell lymphoma, hepatocellular cancer, melanoma, and Hodgkin lymphoma45. Although IL-6 primarily contributes to tumor progression and inflammation, IP-10 plays an important role in enhancing antitumor immunity, which would explain their elevated levels. These findings indicate that cytokines must be considered not individually, but patterns and ratios may be more accurate in predicting disease status and outcomes46,47.

The strength of this study is that it explored a wide array of salivary cytokines in pediatric patients with cancer. Additionally, the large healthy control group and standardized dental/oral examinations in both groups are further strengths of this study. However, this study has some limitations, including a relatively small sample size of children and adolescents with cancer. Although no substantial deviations of effect sizes were observed during the forward inclusion process of the given variables, small deviations of effect sizes and confidence intervals are possible with larger sample sizes. Therefore, the focus should be on the largest effect sizes and significances. Further limitations of the study include the clinical significance, which remains uncertain. We recommend future studies that incorporate other cell types, such as comparing cytokine levels in serum and saliva, to provide a more comprehensive understanding. Additionally, the small sample sizes within individual diagnosis groups, variations in treatment, and differences in remission status limit the interpretation of the findings.

In conclusion, the findings of this study provide valuable insights into the complex interplay of salivary cytokines in children and adolescents with cancer. Elevated levels of certain cytokines at diagnosis suggest a strong immune response linked to cancer presence, independent of oral health factors such as caries and gingivitis. The study also demonstrated the impact of salivary flow rates on cytokine concentrations, emphasizing the need for careful consideration of this variable in analyses. Notably, the detected changes of salivary cytokines over time could reveal important insights into treatment responses, which may correlate with changing tumor burden. Overall, these results suggest the potential of salivary cytokines as biomarkers for monitoring pediatric cancer, while underscoring the complex interactions between age, disease type, and immune response. Further studies with larger patient cohorts and a wider variety of cancer types are necessary to establish a reliable baseline for interpreting individual cytokine measurements in pediatric cancer patients.

Data accessibility

The dataset from this study is securely stored in coded form at the Ambulatory Study Center (ASZ) of the University Children’s Hospital in Basel. Access to the data supporting this article can be provided upon reasonable request to the corresponding authors, subject to predefined criteria for confidential access approval.

Methods

Study design and population

This was a prospective observational pilot study conducted between May 2016 and October 2018. The details of the study methods have been reported previously34. Children (< 12 years) and adolescents (≥ 12 years) aged 4–18 years who were newly diagnosed with cancer at the University Children’s Hospital Basel, Switzerland, were included in this study. Additionally, healthy children and adolescents undergoing routine dental examinations at the Department of Children and Adolescent Dentistry, University Center for Dental Medicine Basel, Switzerland, were included. The exclusion criteria for all participants included those with acute symptoms of respiratory infection within the 2 preceding weeks, those with known paraffin allergy, those with autoimmune disease, those who used systematic antibiotics in the past 2 weeks, those who used antimicrobial rinses 12 h before the consultation, or those who received any vaccination within the last 48 h. Additionally, newly diagnosed patients with cancer already undergoing cancer chemotherapy or those with a history of cancer treatment were excluded. Saliva was collected from patients with cancer at three time points: at diagnosis, 3 months after diagnosis, and 6 months after diagnosis. Written informed consent was obtained from all participants or their caregivers for participation to the study and publication of information or images in an online open-access journal.

Sample collection and procedures

Saliva was collected in the morning hours (between 8 and 12 am), at least 30 min after eating or brushing teeth. Participants were asked to chew paraffin gum for 2–5 min and expel saliva into a collection container. Saliva volume was measured by weight (1 g equivalent to 1 ml), and the stimulated saliva flow rate (ml/min) was calculated. Collected samples were centrifuged at 1650 g for 15 min at 4 °C and stored at − 75 °C.

Cytokine analysis

Cytokines, including IL-1α, IL-1β, IL-4, IL-5, IL-6, IL-8, IL-10, IL-13, IFN-γ, IP-10, TNF, and vascular endothelial growth factor A, were analyzed using a multiplex cytokine analyzer (MAGPIX, Luminex Corp. Austin, TX, USA, with panels from Milliplex, Merck Millipore, Schwalbach, Germany). Cytokine levels below the lower limit of detection were imputed as half of the lower limit of detection (lower limits of detection for IFN-γ, IL-1β, IL-13, TNF, and IP-10 were 1.3, 1.6, 6.4, 6.4, and 2.6 pg/ml, respectively). Cytokines below this threshold were excluded from further analysis.

Clinical data collection

Demographic and clinical data were entered into a study database using EpiData Manager v.2.0.13.65 and EpiData Entry Client v.2.0.10.26 (EpiData Association, Odense, Denmark). Clinical data included age, sex, detailed dental status (presence of caries, missing and filled teeth, and mucosal diseases), and, if applicable, oncological diagnosis.

Statistical analysis

Descriptive statistics for demographic parameters were presented as frequency and proportion for discrete parameters and median (interquartile range) and mean (standard deviation) for continuous parameters. Comparisons of median salivary cytokine levels were made between patients with cancer and healthy controls using the Wilcoxon rank-sum test and between healthy controls and patients with leukemia, lymphoma, and solid tumors using the Kruskal–Wallis test. Multiple linear regression models were fitted for each cytokine to derive separate outcomes for all patients with cancer compared with healthy controls and for patients with leukemia, lymphoma, and solid tumors compared with healthy controls. A nested design was used to obtain separate estimates for children and adolescents. Following covariables were included in the multivariate regression models using a forward selection process: Sex, age, salivary flow rate, and presence of gingivitis and caries. The possibility of overfitting was assessed by comparing the regression estimates (effect size, confidence interval, p-value) during the forward inclusion process of the given variables in each regression model.

To evaluate the changes in cytokine levels over time in patients with cancer, each cytokine was estimated in a separate linear mixed-effects model with a nested design for children and adolescents and adjusted for sex, age, gingivitis, caries, flow rate, and saliva pH.

All cytokine concentrations were log-transformed for regression analysis to approximate a Gaussian distribution. The back-transformed estimates were geometric mean ratios with 95% confidence intervals and P values. No adjustment of the significance level for multiple comparisons was made to account for the exploratory nature of the study. A P value of < 0.05 was considered significant. All analyses were performed using R version 4.2.148.

Ethics

This study was approved by the Local Ethics Committee of North-West Switzerland (Number: 2016-00583). In addition to obtaining approval from the relevant ethics committee, we confirm that all methods in this study were performed in accordance with the relevant guidelines and regulations.