Formalin-fixed paraffin-embedded tissue (FFPE) specimens from patients with primary NPC who were diagnosed at Luoyang Central Hospital Affiliated to Zhengzhou University and the First Affiliated Hospital of Henan University of Science and Technology (Luoyang, China) between January 2011 and July 2017 were collected. All specimens were reviewed by a single pathologist under a light microscope (Eclipse 80i; Nikon Corporation) at the Department of Pathology, the First Affiliated Hospital of Henan University of Science and Technology following hematoxylin and eosin (H&E) staining. Histological classification was re-evaluated according to the current World Health Organization (WHO) classification (3). NPC are grouped into keratinizing squamous, non-keratinizing and basaloid squamous. Non-keratinizing NPC can be divided into differentiated and undifferentiated tumors (3). Patients with nasopharyngeal tumors other than the WHO types, those with poor quality, and those without sufficient sample available for investigation were excluded from the study. Clinical information associated with each sample was recorded. The present study was approved (approval no. 2021-03-B053) by the Ethics Committee of the First Affiliated Hospital of Henan University of Science and Technology (Luoyang, China).

P. gingivalis was detected using IHC and nested PCR.

Primary NPC FFPE tissues were used for the IHC analysis of P. gingivalis using polyclonal rabbit anti-whole cell P. gingivalis 33277 antibody (a generous gift from Dr Richard Lamont) (11). This antibody does not react with human cells or with other bacteria at dilutions of 1:500 or greater (a dilution of 1:1,000 was used with NPC tissue sections). Pre-immune rabbit IgG was used as a negative control. IHC was performed as previously described (12). The sections were evaluated by two pathologists under a light microscope (Eclipse 80i; Nikon Corporation) after IHC staining. Staining intensity was classified using a numerical scale, as previously described by the authors (12). In the present study, IHC scores ≥2 were categorized as P. gingivalis-positive.

Genomic DNA from FFPE tissue was extracted using the QIAamp DNA FFPE Tissue kit (cat. no. 56404, Qiagen, Inc.) according to the manufacturer's instructions. The quantity and purity of the DNA were accessed by NanoDrop 2000 (Thermo Fisher Scientific, Inc.) at 260/280 nm (ratios of 1.8-2.0 favorable results).

The universal bacterial primer pair, 27F/1492R (forward, 5'-AGAGTTTGATCCTGGCTCAG-3' and reverse, 5'-ACGGCTACCTTGTTACGACTT-3') and the P. gingivalis specific primer pair, 404F/R (forward, 5'-AGGCAGCTTGCCATACTGCG-3' and reverse, 5'-ACTGTTAGCAACTACCGATGT-3'), were used as previously described (13-15). The primer pair 27F/1492R was used for the first round of PCR amplification, generating a full-length 16S rDNA product. P. gingivalis specific primers 404F/R, targeting the internal sequence of 16S rDNA, were used to detect P. gingivalis in the second round of PCR amplification. To increase sensitivity, nested PCR was performed based on the sequence of the 16S rRNA fragment of P. gingivalis genomic DNA. The expected size of the amplification product by the inner primer pair was 404 bp in length. Oligonucleotide primers were synthesized by Genewiz, Inc.

The nested PCR assay included two rounds of consecutive PCR amplifications and was performed as described herein. Briefly, the first round of amplification contained the outer primer pair (27F/1429R) and was performed in a reaction volume of 50 µl consisting of 2 µl of 50 ng DNA and 48 µl of reaction mixture containing 2X Taq Plus Master Mix (cat. no. P212-AA; Vazyme Biotech Co., Ltd.), and 10 pmol of each primer. The reaction was performed under the following thermocycling conditions: 94˚C for 10 min, 25 PCR cycles (94˚C for 30 sec, 60˚C for 30 sec, and 72˚C for 60 sec). The final cycle included extension for 5 min at 72˚C. Then, 1 µl of the reaction products was transferred into a new tube and diluted 1:100 with nuclease-free water. Subsequently, 2 µl of this dilution was used as a template for the second-round reaction. The second-round reaction mixture contained 10 pmol of each of the inner pair primers 404F/R and the same Taq polymerase system as used in the first round. Samples were amplified for 30 cycles under the same conditions reported for the first round of amplification, except that elongation was performed at 72˚C for 30 sec in this round of amplification.

The amplification products obtained by nested PCR were run on 2% agarose gel, using a horizontal electrophoresis system (DYCP-32C; Beijing Liuyi Biotechnology Co., Ltd.), then visualized on ChemiDoc™ XRS+ (Bio-Rad Laboratories, Inc.). The size of the product was estimated by comparison with DL 2000 DNA markers (Takara Bio, Inc.), and DNA bands close to the expected size (based on the PCR product obtained from the amplification of positive control and DNA markers) were identified as P. gingivalis positive. The products of positive sample were sent to Genewiz, Inc. for DNA sequencing.

Positive and negative controls were included for each batch of amplification. DNA extracted from the American Type Culture Collection (cat. no. 33277) cultures (from the authors' laboratory) served as a positive control, and a tube containing distilled water in place of the DNA template was used as a negative control. Nested PCR was performed blinded to the results obtained by IHC.

The products of nested PCR were subjected to DNA sequencing in both (forward and reverse) directions with 404F/R primers by Genewiz, Inc. The obtained map results were analyzed by Chromas 2.22 (Technelysium Pty. Ltd.) and sequencing reads were analyzed via BLAST search in NCBI (https://blast. ncbi.nlm.nih.gov/). Bacterial species were identified if subjects showed the lowest expectation (E) value in the list of BLAST results.

Reverse transcription-quantitative PCR was performed using the following primer sets: EBV forward, 5'-CCTGGTCATCCTTTGCCA-3'; and EBV reverse, 5'-TGCTTCGTTATAGCCGTAGT-3' (8), using SYBR Master Mix (cat. no. Q111-02-AA; Vazyme Biotech Co., Ltd.) in a Bio-Rad CFX96™ real time system (Bio-Rad Laboratories, Inc.). The amplification reaction was performed in a total volume of 20 µl containing 2X AceQ qPCR SYBR Master Mix (10 µl), 10 µM forward and reverse primers (1 µl), and 20 ng genomic DNA (2 µl) and distilled water (7 µl). The thermocycling conditions were as follows: 10 sec at 95˚C and 40 cycles of 5 sec at 95˚C and 30 sec at 60˚C. Post-PCR melting curves confirmed the specificity of single-target amplification.

All statistical analyses were performed using SPSS Statistics 19.0 software (IBM Corp.). Cohen's Kappa coefficient was used to evaluate the concordance between IHC and nested PCR. All patients were linked to data from a mortality registry up to May 18, 2021. The primary endpoint was OS, measured using the duration from the date of diagnosis to the end of follow-up or the date of death by any cause. Kaplan-Meier methodology and the log-rank test were performed to determine survival differences among groups.

Chi-square or Fisher's exact test was used to compare categorical variables between patients with different P. gingivalis and EBV infection statuses, and to determine the associations between P. gingivalis status and clinical characteristics. P<0.05 was considered to indicate a statistically significant difference.

A total of 58 patients with NPC diagnosed from January 2011 to July 2017 were identified, in the two hospitals. Among these, six patients were excluded because their samples presented histology other than a WHO type, and seven were excluded due to poor quality or having insufficient sample for investigation. Thus, a total of 45 subjects were included in the present study. A total of 30 (66.7%) patients were male and 15 (33.3%) patients were female. All tumors were classified as non-keratinizing undifferentiated NPC (WHO Type-III), and no other type cases were identified. Unfortunately, tumor staging and other information were unknown for most patients at the time of diagnosis.

Among the 45 samples, 26 (57.8%) cases were P. gingivalis-positive (P. gingivalis⁺), and 19 (42.2%) cases were P. gingivalis-negative (P. gingivalis^-). P. gingivalis expression was detected as dark brown staining, which was primarily localized to the cytoplasm of epithelial cells (Fig. 1).

DNA extracted from the 45 FFPE tissues was examined by single-step PCR, nested PCR, and the products were confirmed by Sanger sequencing. P. gingivalis DNA was not detected by single-step PCR (Fig. S2). After two rounds of amplification, seven (15.6%) samples produced a clear and expected 404 bp band, which was specific for P. gingivalis DNA (Fig. 2). Since nested PCR represents a highly sensitive method to detect low copy numbers of P. gingivalis DNA, it is possible that the nested amplification of a common bacterium could result in a false-positive signal. Therefore, the direct sequencing map was verified on Chromas and the sequencing reads were analyzed by BLAST. All 14 sequencing maps from the seven products corresponded specifically to single reads (Fig. S1). Sequencing of the nested PCR products and BLAST analysis confirmed the presence of P. gingivalis in DNA from NPC tissue (Table I). This indicated that the results obtained by P. gingivalis PCR results were unlikely to be due to a PCR artifact; thus, P. gingivalis is present in NPC tumor tissues. In summary, the identification of P. gingivalis DNA from tumor tissues and the results for P. gingivalis⁺ in FFPE tissues by IHC validate the presence of P. gingivalis in tissues of NPC.

Nested PCR was able to detect P. gingivalis-specific DNA in FFPE tissues from seven patients with NPC, of which specimens from five patients were histologically classified as P. gingivalis⁺. Two P. gingivalis⁺ specimens examined by nested PCR were revealed to be negative by IHC. Nested PCR failed to detect P. gingivalis DNA from the FFPE tissues of 21 patients with NPC that were categorized as P. gingivalis⁺ based on the results of IHC. The concordance rate was 48.9% (kappa=0.422; P<0.001) between nested PCR and IHC (Table II). There was agreement between these two methods for the detection of P. gingivalis in FFPE tissues from patients with NPC.

An association was identified between P. gingivalis infection and the clinical features of patients with NPC (Table III). The presence of P. gingivalis was not significantly associated with age, EBV status, or prognosis in terms of both DNA and expression level. NPC was found to predominantly affect male (30 males vs. 15 females), while the P. gingivalis-positive rate of female patients with NPC was significantly higher than that of male patients with NPC (80.0% vs. 46.7%, P<0.05) in terms of the expression level, but not based on DNA. Furthermore, there were twice as numerous male patients with NPC compared with female patients with NPC, which is consistent with previous studies from high-incidence areas (1,16). The relationship between P. gingivalis infection and the sex of patients with NPC was not consistent with the results of nested PCR and IHC.

EBV-positive NPCs were predominant in the current study. Among the 45 specimens, 40 (88.9%) possessed EBV-positive (EBV⁺) tumors, seven of which (15.6%) exhibited EBV and P. gingivalis co-infection. Thus, based on DNA status, the samples were classified into three groups: i) EBV^-/P. gingivalis^- (5 of 45 patients, 11.1%); ii) EBV⁺/P. gingivalis^- (33 of 45 patients, 73.3%); and iii) EBV⁺/P. gingivalis⁺ (7 of 45 patients, 15.6%). The clinical characteristics of patients are shown in Table IV.

A total of 45 patients were followed up for survival analysis over a period of 115 months. The median follow-up period was 60.9 months for all groups (range, 6.9-115.1 months). Most patients had a minimum of 5 years follow-up; after treatment for <60 months (46.3-59.6 months), there were only five survivors. Due to the low patient number, the survival rate exceeded 40% in two groups and the median survival time could not be calculated in the other group. In the present study, the mean survival time among the three groups was compared. Patients in the EBV⁺/P. gingivalis^- group, which contained the highest number of patients, presented a shorter mean time compared with those with EBV^-/P. gingivalis^- NPC (74.8±6.7 vs. 86.5±13.8). Patients with EBV⁺/P. gingivalis⁺ tended to have the shortest mean survival time (56.8 months). However, there were no significant differences among the three groups (P=0.255; Table IV). Furthermore, the 5-year OS rate in patients with EBV^-/P. gingivalis^- NPC was higher than that in patients with EBV⁺/P. gingivalis⁺ NPC (60.0% vs. 42.9%), although the number of specimens was too low for any statistically meaningful analysis. The 5-year OS rate differed between P. gingivalis positive (whether EBV infected or not) and P. gingivalis negative patients with NPC, the difference was not significant (14.3% vs. 50%, P=0.120). A Kaplan-Meier curve for OS is shown in Figs. 3 and S3.