The immune microenvironment of HPV-positive and HPV-negative oropharyngeal squamous cell carcinoma: a multiparametric quantitative and spatial analysis unveils a rationale to target treatment-naïve tumors with immune checkpoint inhibitors

Patient characteristics

Patient characteristics according to the HPV status are reported in Table 2. Of the 39 patients surgically treated for OPSCC (median age, 61 years), 23 (59.0%) were males and 16 (41.0%) females. The primary tumor was removed using transoral laser microsurgery and open surgery via mandibulotomy or pharyngotomy in 24 and 15 patients, respectively. A clear R0 resection was obtained in all cases. Palatine tonsil was the most frequently involved sub-site (n = 32; 82.1%). Twenty-four (61.5%) patients harbored a transforming HPV infection defined by HPV DNA and p16INK4a double positivity, while the remaining 15 cases were double negative. Prevalence of HPV-positive tumors was higher among never smokers. HPV16 was the most prevalent type (n = 21; 87.5%), with the remaining subjects being positive for HPV33 (2 cases) and HPV18 (1 case). Twenty-seven/39 patients (69.2%) underwent adjuvant post-operative (chemo)radiation. No significant difference in DFS was observed between HPV-positive and HPV-negative patients (p = 0.34).

Table 2 Clinical and demographic characteristics of subjects according to HPV statusHPV-positive and HPV-negative OPSCC differ in immune signatures

To determine the immune signatures of HPV-positive and HPV-negative OPSCC, we used the NanoString PanCancer Immune Profiling Panel on the entire patient cohort. Gene expression analysis revealed 30 DEGs between HPV-positive and HPV-negative OPSCC (Fig. 1A and Supplementary Table S3). In particular, HPV-positive tumors showed a downregulation of genes associated with neutrophils and their chemotaxis (S100A12, IL8), and an upregulation of genes associated with cytotoxicity (GZMH, GZMA, KLRC1, PRF1, KRLK1, GNLY, GZMK), T cell functions (CD8A, CD8B, IL12RB2, EOMES), macrophages (MARCO, MST1R), and inflammation (C8G, IDO1, IL17RB, IL32, CXCL9, CCL5, CXCR3, ICAM4). Moreover, HPV-positive OPSCC displayed a stronger activation of several immune signalling pathways, including antigen processing, complement, cytotoxicity, IFN-γ signalling, NK cell functions, pathogen defence, tumor-inflammation signature (TIS) and macrophage M1-polarization pathway (Fig. 1B and Supplementary Fig. S2A). Conversely, macrophage functions pathway was the only being downregulated in HPV-positive patients as compared to HPV-negative counterparts (Fig. 1B and Supplementary Fig. S2A). Gene expression-based cell type profiling revealed that HPV-positive OPSCC had an increased infiltration of CD45 cells, and in particular of total tumor infiltrating lymphocytes (TILs), CD8 T cells, cytotoxic cells, exhausted CD8 cells and macrophages, and a decrease in neutrophils infiltration (Fig. 1C and Supplementary Fig. S1B). Furthermore, the ratios between CD8 T cells and TILs, CD8 T cells and exhausted CD8 cells, and CD8 and T regulatory (Treg) cells were higher in HPV-positive as compared to HPV-negative OPSCC (Fig. 1C). Conversely, the ratios between mast cells and TILs and between neutrophils and TILs were lower in HPV-positive tumors (Fig. 1C).

Fig. 1figure 1

Differential expression of immune-related genes in HPV-positive and HPV-negative OPSCC. (A) Volcano plot depicting significantly increased (right) or decreased (left) expression of immune-related genes in HPV-positive OPSCC. (B) Differential expression of predefined pathway genes and (C) gene expression-based cell types in HPV-positive and HPV-negative OPSCC. Significantly different data are represented by *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001

Immune signatures correlate with DFS in OPSCC

The assessment of the impact of immune signatures and pathways on DFS in the entire study cohort revealed that patients with a higher expression of CD45, T cells and total TILs signatures, as well as a higher ratio between M1/M2 macrophage signatures, had a longer DFS (Fig. 2A). Conversely, higher ratios between DC and TILs, macrophage and TILs and mast cell and TILs signatures were associated with a shorter DFS (Fig. 2A). Moreover, higher expression of genes associated with adhesion, chemokines, regulation and T cell functions pathways, were prognostic for a better outcome, while the upregulation of macrophage functions pathway was associated with a worse prognosis (Fig. 2A).

Fig. 2figure 2

Immune signatures of TIME impact on survival of OPSCC patients. A-C Kaplan-Meier survival curves for disease-free survival according to high/low gene expression-based cell types and pathways profiling (classification based on median expression as cut-off), in (A) all OPSCC patients (n = 39), (B) HPV-positive (n = 24) and (C) HPV-negative patients (n = 15). Log-rank statistics were performed to determine significance; p values, hazard ratios (HR) and 95% confidence intervals (CI) are reported in each graph

Grouping patients according to the HPV status, the higher expression of CD45, T cells, B cells and total TILs signatures were prognostic for a longer DFS in HPV-positive OPSCC, as well as a higher M1/M2 macrophage ratio (Fig. 2B). Furthermore, the upregulation of genes associated with pathways covering T cell functions, cytokines and NK cell functions also correlated with a better prognosis in HPV-positive patients (Fig. 2B). Differently, in HPV-negative patients the higher expression of NK cells gene signatures and the higher ratio between mast cells and TILs or between T helper 1 (Th1) cells and TILs signalling scores, were indicative of a worse prognosis (Fig. 2C).

HPV-positive primary tumors are heavily infiltrated as compared to HPV-negative counterparts

As recapitulated in the representative Fig. 3, which illustrates the application of the two mIF panels on OPSCC primary tumor and lymph node metastasis sections, it immediately appears evident that HPV-positive lesions are heavily infiltrated as compared to HPV-negative counterparts. The panels included CD8 for CTL, CD68 as a pan-macrophages marker, FoxP3 expressed by Treg, CD163 recapitulating M2-polarized TAM (CD68 + CD163+), CD103 expressed by TRM cells (CD8 + CD103+ cells) and CD4 for T helper lymphocytes. Moreover, the expression of the checkpoint molecules PD-1, PD-L1 and CTLA-4 on T cells, macrophages and tumor cells was also investigated.

Fig. 3figure 3

mIF staining of primary HPV-positive and HPV-negative OPSCC and relative lymph node metastasis. (A) Representative 9-color and (B) 6-color multispectral images at original magnification X20. Immune markers and color codes are indicated in the underlying legend

The density of immune cells infiltrating the microenvironment of OPSCC highly differed according to the HPV status (Fig. 4A). Indeed, in HPV-positive primary tumors we detected a significantly increased intra-tumoral density of CD8+ and CD4+ T lymphocytes, and of CD68+ and CD68 + CD163+ TAMs, as compared to HPV-negative samples, while FoxP3+ Treg cell density was comparable (Fig. 4B). However, the ratio between CD8+ and FoxP3+ cells within the tumor areas was higher in HPV-positive OPSCC (Fig. 4B). The total CD8+ population was further analysed for the co-expression of the integrin CD103, and the checkpoint molecule PD-1. A direct correlation existed between intra-tumoral CD8+ T lymphocytes and double positive CD8 + PD-1+ cells in both groups of patients (r = 0.857, p < 0.0001 in HPV-positive; r = 0.8315, p = 0.0002 in HPV-negative). Accordingly, CD8 + CD103+ TRM cells, CD8 + PD-1+ CTL and triple positive CD8 + CD103 + PD-1+ T cell subsets were more represented in HPV-positive than in HPV-negative tumors (Fig. 4B). Of note, although the stromal compartment appeared characterized by higher immune cell densities than the intra-tumoral areas in both patient groups, the percentage of CD8 + CD103+ and CD8 + PD-1+ T cell populations among total CTL were higher within the tumor nests than in the stroma (Fig. 4B). The density of CTLA-4+ T lymphocytes was also investigated, being higher in HPV-positive OPSCC (Fig. 4B).

Fig. 4figure 4

Characterization of immune cells infiltrating HPV-positive and HPV-negative primary tumors. A Representative 9-color multispectral images of HPV-positive and HPV-negative primary tumors. Immune markers and color code are indicated in the underlying legend. Original magnification X20. B Immune cell populations infiltrating the stromal and the intra-tumoral regions of HPV-positive and HPV-negative primary tumors. C Representative images of cell-to-cell distance analysis in HPV-positive and HPV-negative primary tumors. Cancer cells (light blue dots) within a 20 μm radius from CD8+ cells (red dots) are represented. D Nearest neighbors analysis measuring the mean distance between each tumor cell and the nearest CD8+, CD8 + CD103+ and CD8 + PD-1+ T lymphocytes (left), or each CD163+ M2-polarized macrophage (middle) or each FoxP3+ Treg cell (right) and the nearest CD8+ T lymphocytes in the stromal and intra-tumoral areas. E Count within analysis calculating the percentage of tumor cells within a radius of 20 μm from CD8+, CD8 + CD103+ and CD8 + PD-1+ T cells (left), and the percentage of CD163+ M2-polarized macrophages (middle) or FoxP3+ Treg cells (right) within a radius of 20 μm from CD8+ T lymphocytes in the stromal and intra-tumoral areas. Significantly different data are represented by *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001

In HPV-positive primary tumors, a higher number of tumor cells, M2-polarized macrophages and Treg cells are in closer contact with CD8+ T lymphocytes

Cartographic maps of every acquired field were generated and cell distance analysis was performed (Fig. 4C). As compared to HPV-negative samples, the mean distances observed between tumor cells and CD8+ T lymphocytes, CD8 + CD103+ TRM cells and CD8 + PD-1+ T cells were shorter in HPV-positive primary tumors (Fig. 4D). Moreover, CD8+ T lymphocytes in HPV-positive specimens were closer to M2-polarized TAM and to Treg cells both in the tumor area and in the surrounding stroma (Fig. 4D).

An increased percentage of tumor cells within a 20 μm radius from CD8+ T lymphocytes was observed in HPV-positive OPSCC, as well as from T lymphocytes co-expressing the CD103 or PD-1 molecules (Fig. 4E). Moreover, in HPV-positive tumors almost the 60 and 40% of CD163+ macrophages had CD8+ T lymphocytes within a 20 μm radius in the stroma and within the tumor areas, respectively, as compared to 20 and 10% in HPV-negative OPSCC (Fig. 4E). Furthermore, HPV-positive primary lesions had a higher percentage of FoxP3+ Treg cells within a 20 μm radius from CD8+ CTL, both in the stromal and tumor regions (Fig. 4E).

A higher immune infiltration characterizes the metastases from HPV-positive patients

Metastases from HPV-positive patients were characterized by an overall higher immune cell density than HPV-negative counterparts (Fig. 5A). In particular, an increased density of total CD8+ T lymphocytes, CD68+ macrophages and CD163+ TAMs was observed in the peri-tumoral stroma and within the tumor nests (Fig. 5B). No difference in CD4+ lymphocytes and Treg cell densities existed between the two groups of patients, but the ratio of CD8+ and FoxP3+ cells was higher in HPV-positive metastases (Fig. 5B). A direct correlation between CTL tumor infiltration and PD-1 expression was observed in both patient groups (r = 0.7711, p < 0.0001 in HPV-positive; r = 0.8072, p = 0.0008 in HPV-negative). The analysis of CD8+ T cell subsets revealed that CD8 + CD103+, CD8 + PD-1+ and CD8 + CD103 + PD-1+ T lymphocytes were present at higher densities in HPV-positive than HPV-negative metastases, both in the stroma and within the tumor cell nests (Fig. 5B). Whether the percentage of double positive CD8+ populations is calculated among the total CD8+ cells, it becomes evident that their location is prevalently at intra-tumoral level in both patient groups (Fig. 5B). Finally, HPV-positive metastases disclosed a higher density of CTLA-4-expresssing T cells (Fig. 5B).

Fig. 5figure 5

Characterization of immune cells infiltrating HPV-positive and HPV-negative metastases. A Representative 9-color multispectral images of HPV-positive and HPV-negative metastases. Immune markers and color code are indicated in the underlying legend. Original magnification X20. B Immune cell populations infiltrating the stromal and the intra-tumoral regions of HPV-positive and HPV-negative metastases. C Representative images of cell-to-cell distance analysis in HPV-positive and HPV-negative metastases. Cancer cells (light blue dots) within a 20 μm radius from CD8+ cells (red dots) are represented. D Nearest neighbors analysis measuring the mean distance between each tumor cell and the nearest CD8+, CD8 + CD103+ and CD8 + PD-1+ T lymphocytes (left), and between each CD163+ M2-polarized macrophage (middle) or each FoxP3+ Treg cell (right) and the nearest CD8+ T lymphocytes in the stromal and intra-tumoral areas. E Count within analysis calculating the percentage of tumor cells within a radius of 20 μm from CD8+, CD8 + CD103+ and CD8 + PD-1+ T (left), and the percentage of CD163+ M2-polarized macrophages (middle) or FoxP3+ Treg cells (right) within a radius of 20 μm from CD8+ T lymphocytes in the stromal and intra-tumoral areas. Significantly different data are represented by *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001

In HPV-positive OPSCC lymph node metastases, more tumor and immune regulatory cells are in proximity to CD8+ T lymphocytes

Distance analyses were carried out on lymph node metastases (Fig. 5C), and results resembled those obtained in primary tumors. In HPV-positive metastases, the distance between tumor cells and CTL was shorter than in HPV-negative lesions, regardless the total CD8+ population or the CD103+ or PD-1+ CD8+ subsets (Fig. 5D). Additionally, M2-polarized macrophages and Treg cells mirrored results observed in primary tumors, and appeared closer to CTL in HPV-positive metastases both in the stroma and in the tumor areas (Fig. 5D).

Further, in the microenvironment of HPV-positive OPSCC metastases, CTL established close interactions with the surrounding cancer cells. Indeed, almost 30% of tumor cells in HPV-positive samples had CTL within a 20 μm radius, while this percentage decreased to 4% in HPV-negative lesions (Fig. 5E). Comparable differences were also observed whether considering the proportion of tumor cells within a 20 μm radius from CD8 + CD103+ and CD8 + PD-1+ T lymphocytes (Fig. 5E). Additionally, HPV-positive lymph node metastases revealed higher percentages of CD163+ TAM and FoxP3+ Treg cells detectable within a 20 μm radius from CTL in both stroma and tumor areas (Fig. 5E).

HPV-positive primary tumors and related metastases have a higher expression of PD-L1

Given the high density of PD-1+ cells in OPSCC samples, we assessed the expression of its ligand PD-L1 that clearly differed between HPV-positive and HPV-negative lesions (Fig. 6A). PD-L1 was mainly expressed by tumor cells in either patient groups (Fig. 6B). However, the density of PD-L1+ tumor cells was higher in HPV-positive than HPV-negative primary lesions and metastases (Fig. 6C). Furthermore, HPV-positive lesions evidenced a higher infiltration of PD-L1+ M1- and M2-polarized macrophages (Fig. 6D). Both in primary tumors and metastases, the increase of macrophage or PD-1+ TRM cell infiltration was associated with increasing PD-L1-expressing macrophages, regardless of HPV status (Supplementary Table S4). Moreover, PD-L1+ cancer cells were directly proportional to PD-L1+ macrophages in HPV-positive and HPV-negative metastatic lesions, and also to CD8 + CD103+ TRM cells in HPV-positive primary tumors (Supplementary Table S4).

Fig. 6figure 6

Assessment of PD-L1 expression in HPV-positive and HPV-negative primary tumors and related metastases. A Representative multispectral images of HPV-positive and HPV-negative primary tumors (left) and metastases (right). Only PD-L1 (red), pan-cytokeratin (cyan) and DAPI (blue) markers are represented to better appreciate the different PD-L1 expression. Original magnification X20. B Assessment of the phenotype of PD-L1+ cells in HPV-positive and HPV-negative primary tumors and metastases. Data are presented as the percentage of PD-L1+ tumor cells (CK+) or macrophages (CD68+) among the total number of PD-L1+ cells. C Density of PD-L1+ tumor cells in HPV-positive and HPV-negative primary tumors and metastases. D Density of PD-L1+ M1-polarized (CD68 + CD163-) and M2-polarized (CD68 + CD163+) macrophages in the stromal and intra-tumoral areas of HPV-positive and HPV-negative primary tumors and metastases. E Representative images of PD-L1+ cells (yellow dots) within a 20 μm radius from PD-1+ cells (purple dots). F Mean distance between each PD-L1+ tumor cells and the nearest CD8 + PD-1+ T lymphocytes in HPV-positive and HPV-negative primary tumors and metastases (left); percentage of PD-L1+ tumor cells within a radius of 20 μm from CD8 + PD-1+ T lymphocytes in HPV-positive and HPV-negative primary tumors and metastases (right). G Percentage of CD68 + CD163-PD-L1+ macrophages (left) and CD68 + CD163 + PD-L1+ TAM (right) within a radius of 20 μm from CD8 + PD-1+ T lymphocytes in the stromal and intra-tumoral areas of HPV-positive and HPV-negative primary tumors and metastases. Significantly different data are represented by *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001

A higher number of PD-L1+ cells are in close contact with PD-1+ T lymphocytes in HPV-positive lesions

To assess whether PD-1+ T cells recruited to the tumor microenvironment were close enough to PD-L1+ tumor cells and macrophages to be potentially affected by this checkpoint axis, we carried out the distance analysis (Fig. 6E). A shorter mean distance between PD-L1+ tumor cells and PD-1+ T cells, as well as a higher percentage of PD-L1+ tumor cells within a 20 μm radius from PD-1+ CTL were found in HPV-positive primary tumors and lymph node metastases, as compared to HPV-negative lesions (Fig. 6F).

Additionally, HPV-positive patients showed a higher percentage of PD-L1+ macrophages having PD-1+ T lymphocytes within a 20 μm radius both in the primary and metastatic lesions (Fig. 6G).

HLA-I expression on tumor cells varies inter- and intra-individually

Although striking differences in terms of immune infiltrate were detected between HPV-positive and HPV-negative samples, within the primary tumors and the matched lymph node metastases of either groups of patients no significant differences in immune cell densities were observed (Supplementary Fig. S3). Apart the activation status of the infiltrating T cells and the immunomodulatory features of TIME, another escape mechanism adopted by tumor cells to elude CTL recognition and killing is represented by downregulation of the HLA-I molecules [26]. In this regard, assessment of HLA-I on OPSCC neoplastic cells disclosed intra-individual variations, as HLA-I positive areas were frequently flanked by negative tumor nests within the same tissue section (Fig. 7A). Interestingly, HLA-I expression was predominant at the tumor-stroma interface in some samples (Fig. 7B). Additionally, we detected relevant inter-individual variable patterns of expression irrespective of the HPV status, and, as a consequence, the percentage of HLA-I-negative tumor cells among total number of tumor cells did not differ between HPV-positive and HPV-negative lesions (Fig. 7C). Furthermore, an inverse correlation existed between the percentage of tumor cells downregulating HLA-I and the CD8+ T lymphocyte density in HPV-positive primary tumors (r = − 0.417, p = 0.043).

Fig. 7figure 7

HLA-I expression assessment. A Representative multispectral images highlighting the different HLA-I expression on tumor cells in different areas of a tissue section derived from the same patients. In the left pictures, HLA-I (red), pan-CK (cyan) and DAPI (blue) are represented, while in the right pictures the pan-CK channel is switch off to better visualize the HLA-I down-regulation in tumor cells. Original magnification X20. B Representative multispectral images which highlight the different HLA-I expression on tumor cells at the tumor-stroma interface. In the left picture, HLA-I (red), pan-CK (cyan) and DAPI (blue) are represented, while in the right picture the pan-CK channel is switch off to better visualize the HLA-I down-regulation in the core of tumor nest. Original magnification X20. C Percentage of HLA-I negative tumor cells among the total number of tumor cells in HPV-positive and HPV-negative primary tumors and metastases

The composition of TIME correlates with DFS in OPSCC

Assessment of the impact of TIME composition and cell-to-cell interactions in primary lesions on DFS in our entire cohort revealed that the higher densities of intra-tumoral CD8 + PD-1+, CD8 + CD103 + PD-1+ and PD-L1+ cells were favorably associated with DFS (Fig. 8A). Moreover, a better clinical outcome was observed in OPSCC patients with a higher frequency of tumor cells or intra-tumoral PD-L1+ macrophages within a 20 μm radius from PD-1+ CTL (Fig. 8A). The positive prognostic role of intra-tumoral TRM cells was maintained even when analysing the metastasis samples (Supplementary Fig. S4).

Fig. 8figure 8

The immune cell composition of the primary tumor microenvironment correlates with patient outcome. A-C Kaplan-Meier survival curves for disease-free survival according to the immune cell composition and cell-to-cell interactions of (A) all OPSCC (n = 39), (B) HPV-positive (n = 24) and (C) HPV-negative (n = 15) primary tumors. The median cut-off of each immune cell subset density and percentage was used to separate high and low infiltrated groups. Log-rank p values, hazard ratios (HR) and 95% confidence intervals (CI) are reported in each graph

Stratifying patients according to the HPV status, we assessed that higher densities of stromal or intra-tumoral CD8+ CTL, intra-tumoral CD4+ T cells, total T lymphocytes, CD68+ macrophages and CD163+ TAMs in primary tumors, as well as a higher percentage of CTL in close proximity to macrophages within the tumor nests, were associated with a longer DFS in HPV-positive patients (Fig. 8B). Moreover, HPV-positive patients with a higher amount of HLA-I+ tumor cells close to CD8+ lymphocytes had a better prognosis, while individuals with a higher percentage of tumor cells negative for the HLA-I molecule had a shorter DFS (Fig. 8B). Considering the HPV-positive metastases, the higher densities of CD4+ T cells, intra-tumoral CD8+ T lymphocytes and stromal CD4 + CTLA-4+ T cells were associated with a longer DFS (Supplementary Fig. S4).

Differently from HPV-positive patients, a higher density of M2-polarized TAMs in HPV-negative primary tumors was associated with a shorter DFS (Fig. 8C). Moreover, HPV-negative patients with a higher percentage of CD8+ T lymphocytes within a 20 μm radius from CD163+ macrophages exhibited a worse outcome (Fig. 8C). In HPV-negative metastases, the amount of PD-L1+ tumor cells was favorably associated with prognosis, while patients with a higher density of CTLA-4+ cells in the stroma had a shorter DFS (Supplementary Fig. S4).

TIME composition of OPSCC exhibits sex-specific differences with distinct prognostic values

Patient sex did not impact DFS of our cohort, even considering the overall population (median DFS: 44.8 and 35.6 months in males and females respectively; p = 0.72), or stratifying the cohort by HPV status (p = 0.21 and p = 0.16 in HPV-positive and negative patients, respectively). However, recent evidences highlight the importance of patient sex in anti-tumor immune response, which in turn impinges on the efficacy of ICI [27]. Therefore, we also compared TIME characteristics of males and females. No DEGs were observed considering the entire cohort, as well as grouping patient according to the HPV status. Immune cell populations infiltrating OPSCC primary tumors did not differ between males and females overall considered. Whether patients were stratified according to the HPV status, the only difference observed was a lower density of PD-L1+ macrophages in the stroma of HPV-positive female samples (Supplementary Fig. S5A). On the other hand and regardless the HPV status, several sex-based differences characterized the TIME of lymph node metastases. Indeed, female secondary lesions showed a higher infiltration of CD8+ T lymphocytes within the tumor nests as compared to male samples, and in particular of CD103+ and/or PD-1+ cytotoxic T cells (Fig. 9A). Moreover, female metastases were enriched in CTLA-4+ T cells, intra-tumoral Treg lymphocytes and tumor cells expressing HLA class I, as compared to males (Fig. 9A). Additionally, the percentages of tumor cells, intra-tumoral CD163+ macrophages and Treg cells in close proximity to CD8+ T cells were higher in female metastases than in male lesions, as well as the percentages of PD-L1+ tumor cells and CD163-negative macrophages within a 20 μm radius from PD-1+ CTL (Fig. 9A). Considering HPV-positive metastases, the density of intra-tumoral CTLA-4+ T cells and the percentage of PD-L1+ macrophages interacting with PD-1+ T lymphocytes within the tumor regions were higher in females than in males, while the percentage of HLA-I negative tumor cells was lower (Supplementary Fig. S5B). No differences between men and women were observed in HPV-negative lesions.

Fig. 9figure 9

TIME and its prognostic value differ between female and male patients. A Density (cells/mm2) and count within analyses of immune cell populations infiltrating the stromal and the intra-tumoral regions of lymph node metastases in female and male patients. Significantly different data are represented by *p < 0.05, **p < 0.01 and ***p < 0.001. B-C Kaplan-Meier survival curves of disease-free survival for female and male patients according to the immune cell composition and cell-to-cell interactions in (B) primary OPSCC and (C) lymph node metastases. The median cut-off of each immune cell subset density and percentage was used to separate high and low infiltrated groups. Log-rank p values are reported for each sex in each graph

To assess whether sex-specific differences of TIME in primary OPSCC and related metastases had a prognostic value, Kaplan-Meier curves stratified by sex were generated for each immune parameter. In primary tumors, a higher density of PD-1+ cytotoxic T lymphocytes, PD-1+ TRM cells and HLA-I+ tumor cells, as well as a higher percentage of PD-1+ CTL among total CD8+ T cells, was associated with a prolonged DFS only in females (Fig. 9B). Moreover, women having a higher percentage of PD-L1+ macrophages in close proximity to PD-1+ CTL, as well as more Treg cells close to CD8+ T lymphocytes, disclosed a better outcome (Fig. 9B). Conversely, a higher density of PD-L1+ M2-polarized macrophages and an increased percentage of CTL within a 20 μm radius from CD163+ macrophages, were associated with a worse outcome in males (Fig. 9B).

In metastases, a higher density of CD4+ T cells, CTLA-4+ T lymphocytes and intra-tumoral CD8+ CTL indicated a better prognosis for female patients (Fig. 9C). Moreover, a higher percentage of tumor cells, intra-tumoral CD163+ macrophages and Treg cells within a 20 μm radius from CTL was associated with a longer DFS only in females, as well as a higher amount of interactions between PD-L1+ macrophages and PD-1+ T cells (Fig. 9C).

留言 (0)

沒有登入
gif