Intratumoral levels and prognostic significance of Fusobacterium nucleatum in cervical carcinoma

F. nucleatum abundances at higher levels especially in advanced stage and relapsed disease

We collected primary cancer tissues and matched non-tumorous tissues and recurrent tumor specimen from these in individual 23 relapsed patients. Firstly, we evaluated the levels of F. nucleatum in cervical cancer tissues by using quantitative polymerase chain reaction (qPCR) assays, and concurrently analyzed these in individual patient matched non-tumorous tissues. Results from the assessment of all 23 patients indicated that burdens of F. nucleatum quantified by the orthonormal -^ΔΔCt were significantly higher (P = 0.0024, Figure 1A) in tumor tissues compared with the adjacent non-tumor tissues. The results from paired tissue assessments for individual patients likewise revealed higher F. nucleatum levels in 17/23 pairs. We analyzed the abundance of F. nucleatum in relapsed patients above 23 all patients, and observed a marked enrichment in recurrent lesion than primary cervical cancer tissues (P = 0.0129, Figure 1B). The abundance of F. nucleatum based upon all tumor stages was analyzed. We found no difference in F. nucleatum levels in locally advanced cervical cancer tissues (IB, n = 25/IIA, n = 21; P = 0.1836, Figure 1C). However, an obvious enrichment of this bacterium was revealed in advanced stage tissues (III/IV, n = 20; P = 0.003, Figure 1D).

The increased burden of intratumoral F. nucleatum predicts poor prognosis in locally advanced stage cervical cancer

The median follow-up time for patients in our study was 60 months (range from 0.5 to 148 months). During the duration of the follow-up period, there were 18 patients (16.1%) who died, and there were 22(19.6 %) patients who were noted to have had progression and further development of tumors. Over this same period, the median OS was 60.5 months and the median PFS was 57.4 months. Based on observed levels of F. nucleatum, patients were divided into two groups derived from a cut-off value (ΔΔCt=1.06) inferred from the receiver operating characteristic curve (ROC curve) which corresponded to a measure of burden of F. nucleatum. The cutoff gives important information for the levels of the highest sensitivities and specificities that were accurate when used to predict cervical cancer survival rates.

We next assessed F. nucleatum infiltration abundances for the 112 patients with cervical cancer and summarized their clinicopathological characteristics in Table 1. Kaplan-Meier analyses indicated that the relationships between F. nucleatum burdens and survival rates was dependent upon F. nucleatum infiltration such that patients with high levels exhibited poorer OS (P = 0.012, Figure 2A) compared with results for the low-burden group. Similarly, this finding was subsequently confirmed in progression-free survival analyses (P = 0.001, Figure 2B). These determinations allowed us to answer the question of whether or not the presence of F. nucleatum influenced patient survival rates.

High intratumoral burdens of F. nucleatum can be viewed as a survival independent risk factor for recurrence

We investigated the clinical significance of F. nucleatum levels in the context of other clinicopathological features via univariate and multivariate analysis. In Table 2, univariate Cox regression OS analysis indicated patients with depth of cervical invasion (HR = 11.4; 95%CI = 1.5-86, P = 0.019), poor tumor differentiation (HR = 6.5; 95%CI = 1.5-28.4, P = 0.013) or high F. nucleatum level (HR = 4.9; 95%CI = 1.4-17.4, P = 0.013) were associated with shorter OS. However, F. nucleatum level had no significant differences in its multivariate Cox model. Furthermore, in term of patient PFS survival in Table 3, the univariate Cox regression analysis showed that cervical cancer patients with Pelvic lymph node metastasis (HR = 2.9; 95%CI = 1.2-6.8, P = 0.014), poor tumor differentiation (HR = 4.5; 95%CI = 1.5-13.6, P = 0.008) or high F. nucleatum level (HR = 5.2; 95%CI = 1.7-16.1, P = 0.003) were associated with shorter PFS. Next, a multivariate Cox model was built to analyses these factors which were in keeping with above univariate analysis and revealed that Pelvic lymph node metastasis (HR = 3.1; 95%CI = 1.1-8.8, P = 0.029), and those with poor tumor differentiation (HR = 4.8; 95%CI = 1.2-18.3, P = 0.023), and those higher F. nucleatum level (HR = 4.8; 95%CI = 1.2-18.6, P = 0.024) were associated with an increased risk of tumor progression. We concluded this bacterium was indeed an independent risk factor for predicting poor PFS.

The high levels of local F. nucleatum were correlated with tumor differentiation

The relationship between the patients’ intratumoral F. nucleatum status and clinicopathological characteristics was investigated in 112 patients. Table 4 suggested that levels of F. nucleatum and histological differentiation (P = 0.007), recurrence (P = 0.015) and vital status (P = 0.022) were obviously correlated. Patients with poor differentiation tended toward a higher burden of F. nucleatum. Notably, we measured of F. nucleatum abundances based upon their differentiation. This assessment demonstrated that there was an evident enrichment of F. nucleatum in tissues afflicted with cervical cancer with poor differentiation compared to samples of low pathological grade tissues. It can be seen in Fig 3A that F. nucleatum burdens were confirmed to have been associated with differentiation such that women with higher pathological grading tended toward having higher measures of abundances (poor vs. non-poor differentiation also known as well/moderate differentiation; P = 0.0267, Figure 3A). The assessment also revealed that there were no significant effects of patient age, FIGO stage, tumor dimension, lymphovascular space invasion, depth invasion, uterine corpus invasion or pelvic lymph node metastasis on the associated levels of F. nucleatum in locally advanced cancer tissues.

We were curious to analyses whether this bacterium had any effect on patient survival stratified by pathological grade. We still found that patients with high burden of F. nucleatum hada shorter PFS (P = 0.04, Figure 3C) in poor differentiation group, but there was no difference in OS (P = 0.09, Figure 3B). Importantly, when the pathological grade was combined together with the F. nucleatum levels, patients with high levels of this bacterium were observed worse OS (P = 0.0008, Figure 3D) and PFS (P = 0.0001, Figure 3E).

With high intratumoral F. nucleatum burdens, cervical cancer cells possess the characteristics of cancer stem cells (CSCs)

Given our data suggested that the level of intra-tumor F. nucleatum are correlated with the differentiation of cervical cancer cells, we further aimed to explore whether cancer cells obtained from patients with high intratumoral F. nucleatum levels were more stem-like. Primary cervical cancer cells were cultivated from patients with high or low intratumoral F. nucleatum levels, and poor or well differentiation (n = 6 respectively; Figure 4A–4D). Sphere formation assay was used to survey their self-renewal capacity. It indicated that more significantly larger spheres were formed by cells from high intratumoral F. nucleatum levels, especially in patients with poor differentiation at the same time (P<0.0001, Figure 4E).

CSCs have a tendency to metastasize and our data suggested that F. nucleatum was a prognostic indicator of recurrence and metastasis in patients with cervical cancer, then the metastatic capacity of tumor cells was further examined using Matrigel invasion assays, wherein F. nucleatum levels were shown to increase the invasive capacity of primary cervical cancer cells (Figure 5D vs. 5C), even in patients with well differentiation (Figure 5B vs. 5A). In total, both sphere-formation and invasive abilities were significantly increased when F. nucleatum levels stood high (Figures 4E, 5E).

Finally, to confirm that high level of intratumrol F. nucleatum was functionally active, the expression levels of known genes related to tumor stemness (Figure 6A) and invasion (Figure 6B) were detected by RT-qPCR between patients with high and low F. nucleatum levels. We observed that expression levels of most genes were significantly increased in primary cervical cancer cells obtained from patients with high intratumoral F. nucleatum levels.

Participants and clinical data

This study was randomly enrolled a total of 112 patients diagnosed with stage IB1/IB2-IIA1/IIA2 squamous carcinoma of the cervix who underwent surgical resection at the Sun Yat-Sen University Cancer Center and postoperative chemotherapy or radiation if pathologic risk factors are discovered between 2010 and 2015. Additionally, we followed and collected paraffin sections of paired primary tumor, recurrent tumor tissues (stump recurrence or recurrence confined to the pelvis) and adjacent tumor tissues from 23 patients those with IB1/B2-IIA stage, and other 20 patients those with stage III- IV who underwent cervical biopsy between 2016 and 2018 at the Guangdong Provincial People’s Hospital and Sun Yat-Sen University Cancer Center. None of the patients underwent anti-cancer therapies before surgery, and no histologically confirmed serious complications or other malignant diseases had been reported. The clinicopathological characteristics of the patients are summarized in Table 1. Tumor stages were determined according to the classification system of the International Federation of Gynecology and Obstetrics (FIGO 2009) classification guidelines. All patients attended follow-up visits at the outpatient clinic with regular surveillance for recurrence via recording symptom and physical examination, the serum squamous cell carcinoma antigen level, chest radiography, pelvic and abdominal ultrasonography at 3- to 6- month intervals. When recurrence or metastasis was suspected, further examinations, like CT and PET-CT scan were performed. Biopsies were taken when it is necessary. Overall survival (OS) was defined as the interval between surgery and death or the last follow-up. Progression-free survival (PFS) was the date of surgery to recurrence, the last observation for patients without recurrence, or death if no recurrence was observed.

RNA extraction and real-time qRT-PCR

Total RNA from primary cervical cancer cells was extracted by Trizol reagent (Life Technology). Reverse transcriptase reactions by MMLV reverse transcriptase reagents (Promega, Madison, USA) were performed following manufacturer’s protocol. Gene expression levels were normalized to house-keeping gene β-actin. Reactions were performed in triplicate with the Roche LightCycler 480 II PCR system (Roche Diagnostics, Rotkreuz, Switzerland). Primer sequences are listed in Supplementary Table 1.

DNA extraction and qPCR assays

Genomic DNA from 112 locally advanced patients’ fresh frozen tissues were extracted using DNeasy Power Soil Kit (12888-100, Qiagen, Dusseldorf, Germany). Genomic DNA from the formalin-fixed paraffin-embedded (FFPE) tissues was extracted using the QIAamp DNA FFPE Tissue Kit (56404-50, Qiagen, Frankfurt, Germany). The amount of F. nucleatum DNA was quantified by use of a qPCR assay, the nus G gene and the reference human gene SLCO2A1 were amplified using custom TaqMan primer sets (Promega, Madison, WI, USA) in 384-well PCR plates, as described previously [17].

Primary cells obtained and sphere formation

We used the cancer tissue-originated spheroid method for the primary culture of cervical cancer cells by Hiroko Endo et al with minor modifications [34]. We minced surgically resected primary cervical tissues with a scalpel into approximately 1-mm³ pieces, and washed in Hank’s balanced salt solution (HBSS). Specimens were transferred to a 50-ml centrifuge tube and digested 30 minutes in 37° C by 50μg/ml collagenase and 1% PenStrep. Digestion products were passed through 500 and 250-μm metal mesh filters to remove large masses of undigested and then filtered through 100 and 40-μm cell strainers (BD, Franklin Lakes, USA). Cells in the flow-through fractions were adjusted to a 10⁶ cells/ml concentration in sorting buffer (1× PBS), and and cultured in medium of serum-free DMEM/F12 had to contain certain key component supplementing with B27 (1:50; Invitrogen, California, USA), 20ng/ml epidermal growth factor (EGF; 1:5000; R&D Systems), 20ng/ml basic fibroblast growth factor (bFGF; 1:5000; R&D Systems) for a total of 7 days.

Cancer cells invasion assays

For invasion assays, 10⁵ cells were plated in top transwell chambers (BD, Massachusetts, USA) coated on the inside with 1:4 diluted Matrigel (BD, Bedford, USA) in the insert of a 24-well culture plate. Medium containing 10% fetal bovine serum was added to the wells outside of chamber as a chemoattractant. After 48h incubator, cells inside the chamber were gently removed with a cotton swab. Invasion cells located on the lower side of the chamber were stained with crystal violet, air dried and photographed. Three independent experiments were performed and data are presented as the mean ± standard deviation.

Statistical analysis

Statistical analysis was performed by IBM SPSS Soft 21 (IBM Corporation; United States) and GraphPad Prism 6 (GraphPad Software; United States). Continuous variables were showed as medians and compared by a t-test or Mann-Whitney U test. We used Pearson’s χ ² test or Fisher’s exact test to examine the relationship between F. nucleatum levels expression and clinicopathological characteristics as appropriate. Survival estimates were calculated using the Kaplan-Meier analysis and compared via log-rank test. Prognostic parameters with effects on survival in univariate analysis were included in a multivariate Cox proportional hazards regression model. All P-values were evaluated using two-sided tests, and P-values of <0.05 were considered statistically significant.

Ethics approval

This study conformed strictly to the ethical guidelines of the Declaration of Helsinki and was approved by the Research Ethics Committee of Guangdong Provincial People’s Hospital and Sun Yat-Sen University Cancer Center. A written informed consent was obtained from all patients.