Patient clinical and pathological characteristics

The cohort consisted of 37 individuals with OSCC, with a slightly higher representation of males (54.1%) compared to females (45.9%). Smoking history revealed that 56.8% were previous or current smokers, while 43.2% had never smoked. Tumour sites varied, with the majority located in the mobile tongue (62.16%), followed by gingiva (21.62%), floor of the mouth (8.11%), and buccal mucosa (8.11%). The pathological T status distribution was as follows: 13 patients with pT1 (35.14%), 11 with pT2 (29.73%), 7 with pT3 (18.92%), and 6 with pT4 (16.21%). Ten patients (27%) had nodal metastases confirmed in pathology. Regarding recurrence, 70.3% of patients were recurrence-free, while 29.3% experienced recurrence during the study period. Please see the summary of data in Table 1 and the complete data on all subjects in Supplementary Table 2.

Table 1 Demographic characteristics and clinicopathological data of enroled patients.Neutrophils in TDLNs exhibit tumour-alike subsets in contrary to non-TDLNs and healthy lymph nodes

Within the study, 23 nTDLNs, 102 TDLNs, 12 hLNs and 16 tumour samples were analysed. Neutrophils were more abundant in TDLNs (on average 2558 ± 5460 neutrophils per sample) as compared to nTDLNs (on average 1774 ± 1783 neutrophils per sample). However, no normalization was applied to the samples, and the size and weight of lymph nodes varied significantly between patients. Therefore, this observation should be interpreted with caution. First, the activation phenotypes based on expression of CD16 and CD62L were investigated. Neutrophils expressing high levels of CD16 and CD62L (CD16highCD62Lhigh) are mature and nonactivated, whereas the CD16highCD62Ldim-expressing neutrophils are believed to be mature and activated [1]. The majority of neutrophils found in hLNs (61.25% ± 21.13) and nTDLNs (52.82% ± 24.24) were CD16highCD62Lhigh. In contrary to TDLNs and tumour samples, where the majority of neutrophils were CD16highCD62Ldim (60.58% ± 23.84 and 53.02% ± 25.52, respectively). nTDLNs were characterized by a significantly higher level of CD16highCD62Lhigh neutrophils compared to TDLNs and tumour samples (p values: 0.0050 and 0.0005, respectively). On the other hand, TDLNs were characterized by significantly higher levels of CD16highCD62Ldim neutrophils as compared to nTDLNs and hLNs (p values: 0.0115 and 0.0044, respectively)(Fig. 1c). Furthermore, within the study material, we gathered 16 pairs where we had TDLN and nTDLN coming from the same patient and from the same region of the neck, meaning the distance from the tumour to the lymph node was approximately the same. The levels of CD16highCD62Lhigh neutrophils were significantly lower in TDLNs as compared to their paired nTDLNs (p = 0,0106) and levels of the CD16highCD62Ldim were significantly higher in TDLNs compared to nTDLNs (p = 0,0092) (Fig. 1d).

Neutrophil exhibit more activated phenotypes in TDLNs with increasing T stage

The association between activation phenotypes and TNM staging was investigated. There were no differences in levels of CD16highCD62Lhigh and CD16highCD62Ldim in nTDLNs and TDLNs between patients with and without nodal involvement (pN0 vs pN + ) (Fig. 2a, b). However, in TDLNs the levels of CD16highCD62Lhigh were significantly lower in higher T stages and the proportion of CD16highCD62Ldim neutrophils increased in higher T stages. This observation was not present in nTDLNs (Fig. 2c, d).

Fig. 2: Analysis of CD16 and CD62L Neutrophil Subopulations and Marker Expression in Lymph Nodes.

Comparison of CD16highCD62Lhigh/CD16highCD62Ldim populations between patients with N0 and N+ stage in nTDLNs (a) and TDLNs (b). Comparison of CD16highCD62Lhigh/CD16highCD62Ldim populations between patients with T1-T2 stage and T3-T4 stage in nTDLNs (c) and TDLNs (d). e Expression of CD184 across analyzed compartments. The number of samples was as follows: hLNs (n = 12), n-TDLNs (n = 23), TDLNs (n = 102) and tumour samples (n = 16). f Expression of CD11b across analyzed compartments. The number of samples was as follows: hLNs (n = 12), n-TDLNs (n = 23), TDLNs (n = 102) and tumour samples (n = 16)

Neutrophils in patients with cancer express higher levels of CD184 (CXCR4) and CD11b

Neutrophils expressing CD184 were shown previously to be effective in migration to the inflammatory sites and to have high phagocytic activity [20]. Neutrophils in nTDLNs, TDLNs and in the tumour expressed more abundantly CD184 compared to healthy lymph nodes from patients without a malignancy. The same trend was observed for the expression of activation marker CD11b, where significantly higher levels of expression were observed in TDLNs and tumour samples when compared to hLNs.

Identification of unique subsets of neutrophils in healthy lymph nodes, nTDLNs and TDLNs

To delineate the distinctions in neutrophils phenotypes between hLNs, nTDLNs and TDLNs, we used a dimensional reduction analysis called a uniform manifold approximation and projection (UMAP). Further, we used the clustering algorithm PhenoGraph to cluster cells based on expression of six surface markers of interest. These markers on neutrophils were as follows: CD16, CD62L, CD11b, CD184, CD36 and CD47. The PhenoGraph identified 9 different clusters based on similarity of marker expression patterns. There were striking differences in neutrophils phenotypes between hLNs, nTDLNs and TDLNs (Fig. 3a). The bar chart shows the frequency (%) of each cluster in the investigated samples (Fig. 3b). Using ClusterExplorer, the heatmap showing the relative intensity of each parameter for a given cluster was calculated and is presented in Fig. 3c. In this dimensional reduction analysis, we added two markers on neutrophils, namely CD36 and CD47. High expression of CD47 on neutrophils is believed to be sign of a prolonged survival of neutrophils in tissues [21]. CD36 is a scavenger receptor, which regulates uptake of fatty acid and balances the intracellular lipid content. Al-Khami et al. showed that deletion of CD36 could delay tumour growth through a CD8+ T cell-mediated response [22].

Fig. 3: Characterization of Neutrophil Clusters in Lymph Node Compartments.

a UMAP plot of various neutrophil cell clusters obtained from three pooled samples of hLN, nTDLNs and TDLNs. Each colour represents a different cell cluster (identified by PhenoGraph) and each dot represents a single cell. b The bar chart shows the frequency (%) of each cluster in the investigated samples. c Heatmap of median marker expression across all clusters. d Legend summarizing the phenotypic characteristics of the clusters.

Neutrophils within cluster 1 can be defined as normal, non-activated neutrophils as they were of CD16highCD62Lhigh phenotype with positivity for CD36 and relatively low expression of “don’t eat me” CD47. In hLNs, cluster 1 was the most abundant. Within nTDLNs, clusters 2 and 5 were the most abundant. Cluster 2 represents as mature, nonactivated neutrophils (CD16highCD62Lhigh) but in contrary to cluster 1, cells were characterized by positivity for CD47 and CD184. This could represent a migratory, patrolling subset of neutrophils. Cluster 5 represents mature, activated neutrophils (CD16highCD62Ldim), negative for CD184 and positive for both CD36 and CD47. This could represent stationary neutrophils in the lymph node, ready for phagocytosis. TDLNs were characterized by abundance of different clusters. Clusters 3, 4, 7 and 6 were the most abundant ones. The common feature for these clusters was negativity for CD36. Furthermore, clusters 3,4, and 7 represented mature activated neutrophils (CD16highCD62Ldim). Clusters 4,6 and 7 were relatively low in expressing CD184. The most abundant cluster 3 was strongly positive for CD47. Taken together, these clusters may represent activated neutrophils, which are meant to stay for a longer period in TDLNs, as they express “don’t eat me signals” and some clusters are low in expression of the migratory molecule CD184.

Patients with higher levels of CD16highCD62Ldim neutrophils in TDLNs, higher T stage and nodal involvement have worse prognosis

To conduct survival analysis, we categorized the level of CD16highCD62Ldim expression into low and high groups, using the median expression in all TDLNs as the threshold. In patients who contributed with more than one TDLN, a mean level of CD16highCD62Ldim was calculated. Using Kaplan-Meier method, the disease-free survival (DFS) and overall survival (OS) were calculated for CD16highCD62Ldim Low/High group, T1-T2 vs T3-T4 tumours and patients with N0 vs N+ stage. As shown in Fig. 4, patients with larger tumours had significantly worse DFS compared to patients with T1-T2 stages (log-rank test, p = 0.0184). Patients with High levels of CD16highCD62Ldim and N+ status showed a trend to worse DFS, however these comparisons did not reach a statistical significance (p values: 0.0528 and 0.0885, respectively) (Fig. 4 a, c, e). For OS, patients with high levels of CD16highCD62Ldim in TDLNs and N+ stage showed significantly worse OS (log-rank test, p values: 0.0039 and 0.0006, respectively)(Fig. 4b, f). There was a trend for worse OS in patients with larger tumours (T3-T4), however it did not reach a statistical significance (p = 0.0514)(Fig. 4d). To further extend the findings, the multivariate Cox proportional-hazards models for DFS and OS were calculated. None of the studied variables showed to be an independent prognostic factor for DFS or OS.

Fig. 4: Kaplan-Meier Survival Curves for DFS and OS in Relation to CD16highCD62Ldim Neutrophil Populations in Lymph Nodes and Tumor Staging.

Kaplan-Meier curves showing DFS and OS in relation to the level of CD16highCD62Ldim population in TDLNs (a, b), T stage (c, d) and N stage (e, f).