Microbial biomarkers of common tongue coatings in patients with gastric cancer

Accepted Manuscript Microbial biomarkers of common tongue coatings in patients with gastric cancer Jing Xu, Chunjie Xiang, Cong Zhang, Boqi Xu, Juan Wu, Ruiping Wang, Yaping Yang, Liyun Shi, Junfeng Zhang, Zhen Zhan PII:

S0882-4010(18)31227-0

DOI:

https://doi.org/10.1016/j.micpath.2018.11.051

Reference:

YMPAT 3288

To appear in:

Microbial Pathogenesis

Received Date: 9 July 2018 Revised Date:

29 September 2018

Accepted Date: 29 November 2018

Please cite this article as: Xu J, Xiang C, Zhang C, Xu B, Wu J, Wang R, Yang Y, Shi L, Zhang J, Zhan Z, Microbial biomarkers of common tongue coatings in patients with gastric cancer, Microbial Pathogenesis (2018), doi: https://doi.org/10.1016/j.micpath.2018.11.051. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Microbial biomarkers of common tongue coatings in patients with gastric cancer Jing Xu a, Chunjie Xiang a, Cong Zhang a, Boqi Xu a, Juan Wu a, Ruiping Wang b, Yaping Yang c, Liyun Shia, Junfeng Zhang a, Zhen Zhan a a School of Medicine and Life Sciences, Nanjing University of Chinese Medicine, Nanjing 210023, China. b Department of Oncology, Jiangsu Province Hospital of Traditional Chinese Medicine, Nanjing 210029, China. c School of Basic Medicine, Nanjing University of Chinese Medicine, Nanjing 210023, China. Corresponding author: [email protected] (J. Zhang), [email protected] (Z. Zhan)

RI PT

Abstract

M AN U

SC

Purpose: The study aims to explore the characteristic microorganisms of the common tongue coatings in patients with gastric cancer (GC). Methods: A total of 115 GC patients were assigned to four groups: White-thin coating (W-thin) group, White-thick coating (W-thick) group, Yellow-thin coating (Y-thin) group and Yellow-thick coating (Y-thick) group. Thirty-five healthy volunteers with White-thin coating were recruit as controls. High-throughput sequencing was used to describe the microbial community of the tongue coatings based on 16S rRNA and 18S rRNA genes. Multi-factors statistical analysis was carried out to present the microbial biomarkers of the tongue coating in GC patients. Results: At bacterial phylum level, Saccharibacteria had higher relative abundance in W-thick group than W-thin group, Proteobacteria was more abundant in W-thin group than Y-thick group and less abundant in Y-thick group than Y-thin group. At fungal genus level, Guehomyces and Aspergillus presented to be significantly different among the common tongue coatings. Forteen significantly increased taxa were sorted out as the microbial biomarkers of common tongue coatings by LEfSe and ROC analysis. At species level, bacterial Capnocytophaga leadbetteri and fungal Ampelomyces_sp_IRAN_1 may be the potential biomarkers of W-thin coating, four bacterial species (Megasphaera micronuciformis, Selenomonas sputigena ATCC 35185, Acinetobacter ursingii, Prevotella maculosa) may be the potential biomarkers of W-thick coating. In general, the white coatings held more complex commensal relationship than the yellow coatings. Conclusion: The common tongue coating owned characteristic microorganisms and special commensal relationship in the GC patients.

Highlights:

EP

1. Introduction

TE D

● The microbial community participated in the variation of common tongue coatings in GC patients. ● The white-thick tongue coating owed the most dramatically different microbiome, while the microbiota in the yellow-thin tongue coating changed minimum. ● In each tongue coating, positive correlations dominated the increased bacterial taxa. Keywords: Gastric cancer, tongue coating, microbiome

AC C

Gastric cancer (GC) is one of the most common cancers worldwide, there were 1.2 million incident cases and 834000 deaths in 2016, and GC ranked the second most common cause of death among all types of cancers [1]. In the clinic, the diagnosis of GC relied on endoscopy and biopsy, although huge progresses have been made in diagnostics of GC, many patients were diagnosed at advanced stage, mainly because most patients were asymptomatic at early stage [2]. Researchers had indicated that increasing the detection of early stage GC could increase the number of treatable cancers and improve the 5-year survival [3,4]. Evidently, it is necessary to develop new tactics for the early GC diagnosis. Tongue diagnosis, based on Traditional Chinese Medicine (TCM), is a simple, non-invasive method to assess the physiologic condition by observing the thickness and color of tongue coating [5]. TCM practitioners could distinguish the patients' condition according to the TCM syndromes, which were determined mainly by the extrinsic symptoms including the color and thickness of the tongue coating of the patients [6]. The exterior cold syndrome, interior cold syndrome, exterior heat syndrome and interior syndrome are the four most common syndromes that are identified by white-thin, white-thick, yellow-thin and yellow-thick tongue coatings respectively. According to TCM theory, the tongue is an outer extension of the spleen and stomach, the type of tongue coating is valuable for diagnose malignant gastrointestinal cancers including GC [7]. Although tongue diagnosis is widely used in clinical practice, the detailed mechanism was still unclear. The tongue coating consists of bacteria, fungi, blood metabolites, saliva and exfoliated (desquamated) keratinized epithelium that originated from filiform papillae [8]. The filiform papillae, is a specific structure comprised the formation of tongue coating, this specific structure of the tongue

ACCEPTED MANUSCRIPT

SC

RI PT

mucous membrane causes many cracks and folds on the surface of the tongue, which increases the surface area of the tongue. And the warm, moist, nutritious environment provides suitable surroundings for microorganisms' colonization, growth and proliferation [9]. The oral biofilm harbors one of the most microbiologically diverse sites, and the tongue mucosa accumulates a large higher number of microbiome than other parts of the oral cavity, which makes the tongue coating a relatively complete and independent micro-ecosystem [10]. A recent study showed that the quantities and community of the tongue microorganism was more stable than saliva and supragingival [11]. Previous studies employed bacteria cultures, microscopic examination and denaturing gel gradient electrophoresis to discover the relationship between the tongue microbiome and the tongue coating in acute pancreatitis, diarrhea-predominant irritable bowel syndrome and lung cancer, while more and more researchers began to use next-generation sequencing technology to describe the microbiota of tongue coating comprehensively [12]. Two studies had investigated the tongue coating microbiota to explore cancer-risk related biomarkers [13,14], and two studies showed the correlation of yellow tongue coating and tongue coating microbiota [5, 15]. The present study combined the color and thickness of tongue coating, focused on the four kinds of common tongue coatings, and used the next-generation sequencing to describe the potential characteristic microbiota (bacteria and fungi) in the GC patients. The results would provide a comprehensive microbial mechanism of the formation of tongue coating in GC patients, and may simultaneously provide new perspectives of the tongue diagnosis in the TCM clinical practice.

2. Materials and methods

AC C

EP

TE D

M AN U

2.1 Study participants From January 2015 to January 2016, GC patients confirmed by endoscopy combined with pathologic biopsy were recruited from the Jiangsu Provincial Hospital of Traditional Chinese Medicine. None of the participants had used any antibiotics or immunosuppressive agents within the past four weeks and the absence of oral disease was confirmed. This study was approved by the Jiangsu Provincial Hospital of Traditional Chinese Medicine Institutional Ethics Review Board (2015NL-017-01), and written informed consent was obtained from all the participants before they were interviewed. At the same time, healthy volunteers with white-thin coating were recruited as controls. 2.2 Sample collection All tongue coating samples were obtained and photographed in the morning prior to patients’ food consumption to avoid the interference of food debris. Two traditional Chinese physicians with more than 20 year of clinical experiences diagnosed the type of tongue coatings separately, the DS01-B tongue diagnostic information acquisition system (DAOSH Co., Shanghai, China) was also used to photograph and analyze the images of the tongue coating. The patients were only recruited when the diagnosis of two physicians and the analysis of DS01-B were consistent. All participants were required to rinse their mouths by gargling sterile saline two times before sampling. Each tongue coating sample was collected from the middle section of tongue dorsum using a fresh one-off toothbrush and put into the test tube with saline. The tubes were centrifuged for 10min at 3000r/min, and the precipitates were collected. Samples were immediately stored at -80 ℃. 2.3 DNA extaction and PCR amplification Total DNA was extracted using the E.Z.N.A.® Soil DNA Kit (Omega Bio-tek, Norcross, GA, U.S.) according to the manufacturer’s protocols. The integrity of the genomic DNA was assessed by electrophoresis (1% agarose gel). Polymerase chain reaction (PCR) amplification of the V3-V5 region of the bacterial 16S rRNA genes was performed using forward primer 341F 5’-barcodeCCTAYGGGRBGCASCAG -3’ and reverse primer 806R 5’- GGACTACNNGGGTATCTAAT -3’, and the fungal 18S rRNA genes ITS1-2 region was amplified using forward primer ITS1F 5’-barcode-CTTGGTCATTTAGAGGAAGTAA-3’ and reverse primer ITS2R 5’GCTGCGTTCTTCATCGATGC -3’. The PCR reactions were performed in 20 µl PrimerStar HS Premix (AP221-02; TransGen, Beijing, China) that contained 4 µl of 5×FastPfu Buffer,2 µl of 2.5 mMdNTPs, 0.4 µl of forward primer, 0.4 µl of reverse primer,0.4 µl of FsatPfu Polymerase, and 10 ng DNA. All samples were amplified on an ABI GeneAmp 9700 (ABI, USA) using the following parameters: 95 ℃ for 2 min, followed by 25 cycles of 95 ℃ for 30s, 55 ℃ for 30s, and 72 ℃ for 30s. and a final extension at 72 ℃ for 5 min. The obtained PCR products were run on 2% agarose gel and purified after size selection repeated three times for each sample and quantified using QuantiFluorTM ST system. 2.4 Sequence analysis The raw reads were filtered using the QIIME (version 1.17) with following criteria: reads shorter than 200bp; reads with more than 2 mismatched bases in the forward primer; reads with average accuracy less than Q25. Filtered sequences were analyzed using UPARSE35 (version 7.1

ACCEPTED MANUSCRIPT

RI PT

http://drive5.com/uparse/) to cluster sequences into operational taxonomic units (OTUs) by 97% similarity threshold. The SILVA database was used to perform the taxonomic analysis. Here analysed population structure and richness based on following six categories: phylum, class, order, family, genus, and species. 2.5 Statistical analysis The alpha diversity, the Relative abundances, and the linear discriminant analysis of the effect size (LEfSe) were performed by R software packages. The SPSS 21.0 (International Business Machines Corp.) was used to perform Chi-square test, t-test, non-parametric test, correlate analysis and ROC curve. The Chi-square test and t-test were used to compare the common characteristics of participants, the non-parametric test was used to compare the indices of alpha diversity and the relative abundances between different types of tongue coating. Correlate analysis was used to establish the commensal relationship . ROC curves were used to analyze distinguishing ability of the taxa. P-value < 0.05 was considered to be statistically significant.

3. Results

TE D

M AN U

SC

3.1 General Characteristics and Overall sequencing data A total of 115 GC patients included 85 males and 30 females with average 64.0±8.3 years old. The population was assigned to four groups: i) White-thin coating (W-thin) group (n=25) was comprised of 17 men and 8 women with average 55.3±8.3 years old; ii) White-thick (W-thick) coating group (n=43) had 31 men and 12 women with average 63.5±8.8 years old; iii) Yellow-thin coating (Y-thin) group (n=27) had 18 men and 9 women with average 66.7±6.5 years old; iv) Yellow-thick coating (Y-thick) group (n=20) had 19 men and 1 women with average 67.5±9.0 years old. The healthy controls (n=35) had 28 men and 7 women with average 64.5±9.1 years old. All 150 samples were sequenced. For bacterial 16S rRNA gene V3-V5 region, the total number of sequences was 75199824, with an average of 50133 reads per sample, the average length of the reads was approximately 447.36 bp, and 4225 OTUs were detected. For fungal 18S rRNA gene ITS1-2 region, the total number of sequences was 5874840, with an average of 39165 reads per sample, the average length of the reads was approximately 257.40 bp, and 4437 OTUs were detected. 3.2 Diversity of bacterial and fungal community To assess the diversity of bacterial and fungal community in four types of tongue coatings, ACE, Chao, Shannon and Simpson indices were used to describe the alpha diversity. The ACE and Chao indices are used to estimate the richness of OTUs, whereas the Shannon and Simpson indices are used to assess the diversity of OTUs. However, no significant differences were found among four groups in bacterial community, the same results were shown in fungal community (Supplementary Table 1). These results indicated that the variation of tongue coatings be irrelevant to the richness and diversity of microbiome community in GC patients.

Bacterial

EP

Supplementary Table 1. The alpha diversity of tongue coating microbiota in GC patients with different tongue coating Controls Types of tongue coating in GC patients Diversity index W-thin W-thick Y-thin Y-thick Ace

534.5±202.5

531.0±229.6

527.5±214.3

514.1±238.2

Chao

483.4±178.8

473.4±177.5

458.6±187.3

487.5±222.0

459.0±185.4

Shannon

3.021±0.534

2.921±0.559

3.150±0.493

3.085±0.683

2.886±0.649

Simpson

0.124±0.062

0.128±0.066

0.102±0.057

0.120±0.089

0.126±0.073

Fungal

Ace

192.5±103.7

196.3±89.49

209.0±97.42

207.9±96.49

206.5±65.96

diversity

Chao

186.1±77.2

178.7±71.6

194.6±79.3

185.0±81.3

187.6±57.4

Shannon

3.255±0.732

3.069±0.929

3.240±0.794

3.107±0.858

3.170±0.888

Simpson

0.114±0.127

0.161±0.216

0.115±0.135

0.134±0.182

0.139±0.194

AC C

533.1±199.5

diversity

W-thin: white-thin coating; W-thick: white-thick coating; Y-thin: yellow-thin coating; Y-thick: yellow-thick coating. 3.3 Community structure of common tongue coating The microbiome of tongue coating were clustered into OTU at 97% identity, for bacteria, 8 phyla, 32 classes, 107 orders, 285 families, 701 genera and 1328 species were detected, for fungi, 47 phyla, 100

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

classes, 223 orders, 428 families, 1009 genera and 2162 species were detected. Figure 1 shows the relative abundance of bacteria and fungi on phylum level (relative abundance > 1%) and genus level (50 most abundant genera). Across all five groups, the most abundant 6 bacterial phyla were Firmicutes (65.70±20.47%), Bacteroidetes (13.14±11.87%), Proteobacteria (10.25±6.60%), Fusobacteria (4.77±5.56%), Actinobacteria (4.52±4.24%), Saccharibacteria (1.04±1.91%) (Figure 1A). Comparing the differences between four groups of GC patients, Saccharibacteria had higher abundance in W-thick group than W-thin group (P=0.030), meanwhile, Proteobacteria was more abundant in W-thin group than Y-thick group (P=0.028) and less abundant in Y-thick group than Y-thin group (P=0.042). The most abundant 6 fungal phyla were Ascomycota (66.12±14.30%), Basidiomycota (29.40±13.72%), Zygomycota (1.84±2.39%), Glomeromycota (0.07±0.23%), Chytridiomycota (0.04±0.15%), Rozellomycota (0.04±0.16%), no significant differences were found (Figure 1C). On the level of genus,the most abundant 8 bacterial genera were Bacillus (22.74±14.89%), Enterococcus (10.72±7.31%), Lactococcus (7.83±5.27%), Veillonella (5.73±5.84%), Streptococcus (5.34±5.11%), Leptotrichia (3.58±5.05%), Actinomyces (2.59±3.11%), Megamonas (0.92±6.82%) (Figure 1B). The most abundant 8 fungi genera were Aspergillus (14.05±9.79%), Fusarium (10.15±19.17%), Cryptococcus (9.70±6.94%), Guehomyces (8.96±5.99%), Candida (3.92±12.15%), Cladosporium (2.66±5.06%), Trichosporon (2.31±2.10%), Alternaria (2.17±4.08%) (Figure 1D). Comparing the differences between four groups of GC patients, Guehomyces and Aspergillus were more abundant in W-thin group than Y-thick group (P=0.003, P=0.007), and higher abundance of Guehomyces were observed in W-thin group compared to W-thick group (P=0.035). The relative abundance of Aspergillus was higher in Y-thin group than Y-thick group (P=0.017). In terms of the differences between controls and four groups of GC patients, the same trend were found both on phylum level and genus level. Proteobacteria was more abundant in controls than in other four groups of GC patients, Guehomyces and Trichosporon were more abundant in controls than in other three groups (W-thin, W-thick, Y-thick) of GC patients, meanwhile, Zygomycota of controls were less abundant comparing to these three groups, Chytridiomycota and Megamonas were less abundance in controls than in other three groups (W-thick, Y-thin, Y-thick) groups of GC patients.

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

Figure 1. Heat maps of the relative abundances of tongue coating microbial taxa in GC patients with different tongue coatings at phylum and genus levels. (A) Bacterial phyla (Relative abundance > 1%) of five groups. (B) Fifty most abundant bacterial genera of five groups. (C) Fungal phyla (Relative abundance > 1%) of five groups. (D) Fifty most abundant fungal genera of five groups.

3.4 LEfSe analysis and ROC curves of GC patients LEfSe was used to reveal potential bacterial and fungal biomarkers that most likely explained the differences among the common tongue coatings. By comparison between each target group with the other three converged groups, total 163 taxa were identified to be responsible for discriminating each type of tongue coating from the other tongue coatings. Among them, 107 taxa belonging to the target group were increased and 56 taxa belonging to the other three groups were decreased (Figure 2). The LEfSe cladogram comprise 6 layers of circles, which represent 6 categories (phylum, class, order, family, genus and species) from the center to edge. Hence, the taxa located in the periphery of the circle were significantly responsible to explain different taxa of each tongue coating. For example, comparison with the other tongue coatings, the W-thin coating had 20 bacterial taxa (Figure 2A1) and 36 fungal taxa (Figure 2A2).

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

Figure 2. LEfSe analysis of microbial taxa from tongue coating in GC patients. Each tongue coating group was selected to compare with the other three converged groups, and the LEfSe cladogram was built. Phylum and class were showed in the LEfSe cladogram, other biological classifications were listed beside the LEfSe cladogram, "o" referred to order, "f" referred to family, "g" referred to genus, and "s" referred to species. A, 25 W-thin coatings were compared with 90 other coatings, A1 showed 10 increased bacterial taxa and 10 decreased bacterial taxa, and A2 showed 31 increased fungal taxa amd 6 decreased fungal taxa. B, 43 W-thick coatings were compared with 72 other coatings, B1 showed 17 increased bacterial taxa and 4 decreased bacterial taxa, B2 showed 12 increased fungal taxa and 7 decreased fungal taxa. C, 27 Y-thin coatings were compared with 88 other coatings, C1 showed only one decreased bacterial taxon, C2 showed 12 increased fungal taxa and 6 decreased fungal taxa. D, 20 Y-thick coatings were compared with 85 other coatings, D1 showed 19 decreased bacterial taxa, and D2 showed 24 increased fungal taxa and 4 decreased fungal taxa.

For a deeper insight, ROC curves were used to analyze distinguishing ability of the taxa based on the LEfSe data. With the standard of P-value less than 0.05, the unique taxa were presented in Table 1 - 4. The significantly increased taxa were considered as the potential biomarkers of common tongue coatings with AUC > 0.6, total 14 taxa were responsible for discriminating the common tongue coatings, which included three taxa in W-thin group (Table 1), nine taxa in W-thick group (Table 2), one taxon in Y-thin group (Table 3) and one taxon in Y-thick group (Table 4). At species level, bacterial Capnocytophaga leadbetteri and fungal Ampelomyces_sp_IRAN_1 may be the potential biomarkers of W-thin coating, four bacterial species (Megasphaera micronuciformis, Selenomonas

ACCEPTED MANUSCRIPT sputigena ATCC 35185, Acinetobacter ursingii, Prevotella maculosa) may be the potential biomarkers of W-thick coating, and fungal Blastobotrys_adeninivorans may be the potential biomarkers of Y-thick coating. Table 1. ROC curve of tongue coating microbiota in GC patients with W-thin coating Relative abundance (mean±sd%) Z(P)

AUC(95% CI)

g__Lautropia

0.112±0.227

0.051±0.172

4.365(0.037)

0.633(0.507,0.758)

s__Capnocytophaga leadbetteri

0.049±0.117

0.024±0.053

5.007(0.025)

0.640(0.526,0.753)

f__Family I

0.004±0.019

0.067±0.316

5.212(0.022)

0.372(0.258,0.486)

g__Stomatobaculum

0.230±0.413

0.439±0.638

4.419(0.036)

0.366(0.248,0.484)

g__CL500-29 marine group

0.002±0.011

0.042±0.195

8.115(0.004)

0.353(0.242,0.464)

s__Megasphaera micronuciformis

0.853±1.190

2.111±3.051

4.728(0.030)

0.361(0.247,0.476)

0.068±0.118

5.152(0.023)

0.348(0.230,0.466)

s__Eubacterium sp. oral clone GI038 Fungi

s__Ampelomyces_sp_IRAN_1 o__Sporidiobolales s__Gibberella_intricans

0.028±0.065

RI PT

Other coatings (n=90)

M AN U

Bacteria

W-thin coating (n=25)

SC

Different taxa

0.383±0.913

0.145±0.555

7.01(0.008)

0.643(0.513,0.773)

0.265±0.559

0.877±4.325

5.912(0.015)

0.342(0.217,0.466)

0.660±1.407

1.458±2.534

4.345(0.037)

0.364(0.240,0.487)

W-thin coating: white-thin coating; o_: order; f_: family; g_: genus; s_: species

Table 2. ROC curve of tongue coating microbiota in GC patients with W-thick coating Relative abundance (mean±sd%)

W-thick coating (n=43)

Other coatings (n=72)

Z(P)

AUC(95% CI)

1.794±2.434

0.873±1.966

4.359(0.037)

0.637(0.511,0.763)

0.100±0.383

0.029±0.207

4.754(0.029)

0.634(0.505,0.764)

g__Ruminococcaceae UCG-014

0.324±0.537

0.168±0.397

4.265(0.039)

0.643(0.522,0.765)

g__[Eubacterium] nodatum group

0.148±0.250

0.095±0.162

4.965(0.026)

0.637(0.521,0.752)

g__CL500-29 marine group

0.066±0.254

0.016±0.112

4.890(0.027)

0.630(0.500,0.759)

s__Megasphaera micronuciformis

2.328±2.509

1.585±2.801

4.219(0.040)

0.650(0.530,0.770)

s__Selenomonas sputigena ATCC 35185

0.066±0.092

0.034±0.100

4.934(0.026)

0.644(0.521,0.766)

s__Acinetobacter ursingii

0.011±0.023

0.031±0.210

6.320(0.012)

0.640(0.517,0.764)

s__Prevotella maculosa

0.031±0.081

0.008±0.048

6.121(0.013)

0.637(0.503,0.771)

s__Lactobacillus salivarius

0.018±0.050

0.168±0.956

5.254(0.022)

0.348(0.231,0.466)

o__Hypocreales

13.12±17.65

20.10±23.00

4.850(0.028)

0.377(0.271,0.483)

Bacteria p__Saccharibacteria

AC C

EP

f__FamilyI

TE D

Different taxa

Fungi

W-thick coating: white-thick coating; p_: phylum; o_: order; f_: family; g_: genus; s_: species Table 3. ROC curve of tongue coating microbiota in GC patients with Y-thin coating Relative abundance (mean±sd%) Different taxa

Fungi g__Trichoderma

Y-thin coating(n=27)

Other coatings(n=88)

Z(P)

AUC(95% CI)

0.376±0.667

0.623±4.047

6.259(0.012)

0.651(0.530,0.772)

ACCEPTED MANUSCRIPT s__Cladosporium_halotolerans

0.186±0.429

0.890±3.591

9.000(0.003)

0.310(0.195,0.425)

Y-thin coating: yellow-thin coating; g_: genus; s_: species Table 4. ROC curve of tongue coating microbiota in GC patients with Y-thick coating Relative abundance (mean±sd%) Other coatings (n=95)

Z(P)

AUC(95% CI)

0.242±0.614

0.427±0.907

4.804(0.028)

0.330(0.185,0.475)

f__Burkholderiaceae;g__Lautropia

0.017±0.033

0.081±0.209

5.586(0.018)

0.318(0.185,0.451)

g__Neisseria

0.546±0.883

2.573±4.650

5.079(0.024)

0.325(0.193,0.458)

g__Peptostreptococcus

0.031±0.082

0.061±0.123

6.054(0.014)

0.310(0.160,0.459)

s__Gemella sanguinis

0.087±0.193

0.238±0.904

3.897(0.048)

0.347(0.189,0.505)

s__Capnocytophaga sputigena

0.002±0.004

0.070±0.063

4.808(0.028)

0.337(0.217,0.457)

s__Haemophilus parainfluenzae T3T1

0.456±1.035

s__Blastobotrys_adeninivorans

0.098±0.216

Fungi

1.265±2.907

4.701(0.030)

0.332(0.186,0.478)

0.232±1.619

6.151(0.013)

0.671(0.544,0.798)

M AN U

Y-thick coating: yellow-thick coating; f_: family; g_: genus; s_: species

SC

Bacteria f__Ruminococcaceae

RI PT

Y-thick coating (n=20)

Different taxa

AC C

EP

TE D

3.5 Commensal relationship Commensal relationship is an important indicator of microecological research, Spearman's correlation analysis was performed based on the taxa which illustrated in Table 3 to Table 6, and the significant correlations (P<0.05) were constructed into commensal relationship networks, which illustrated potential relationships in each type of tongue coating (Figure 3). In W-thick coating, all increased bacterial taxa were positively correlated to each other, except FamilyΙ, which was negative correlated to Ruminococcaceae UCG-014, Megasphaera micronuciformis, and Hypocreales belonged to decreased fungal taxa. One decreased bacterial taxon showed negative correlation to all increased bacterial taxa in W-thick coating. It was a pity that no significant correlations were found in Y-thin coating because almost no different taxa were observed in ROC results (Table 3).

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

4. Discussion

EP

Figure 3. Commensal relationship of the microbial taxa from different tongue coatings in GC patients. Taxa as nodes were marked with biological classification along with their names. Red nodes mean increased taxa and blue nodes meandecreased taxa, circle nodes mean bacterial taxa and triangle nodes mean fungal taxa. The solid edges mean positive correlation and the dotted edge mean negative correlation. p_: phylum; o_: order; f_: family; g_: genus; s_: species

AC C

Recent years, human microbiota has received much more attention due to the rapid development of next-generation sequencing, and many investigations had confirmed that human microbiome involved in a number of key processes of the metabolism including digestion of complex plant matter, production of high energy metabolites (for example, short chain fatty acids), immune homeostasis, and protection against pathogens [16-19]. Generally, cancer was considered to be a disease induced by environmental and genetic factors, microorganisms are implicated in 20% of malignancies [20]. A growing number of evidences had shown that microbiota could influence the proliferation of tumor cells and is responsible to spontaneous cancers in various organs, including the skin, colon, liver, breast and lungs [21]. The current opinion was that proinflammatory responses caused by microbes can be procarcinogenic, nevertheless, microbial metabolites can contribute to inflammatory tone and can influence the balance of proliferation and cell death in tissues [22]. Ten specific microbes have been designated as carcinogenic pathogens by the International Agency for Cancer Research (IACR) [22]. Of them, H. pylori is most famous microorganism which was considered to be the strong inducer of gastric carcinoma [23]. GC not only is caused by the inflammation of H. pylori but also is promoted by the presence of special microbiota with the increasing luminal pH in GC patients. Correspondingly, the specific stomach environment shaped the compositions of gastric microbiota, and subsequently increased bacterial conversion of dietary nitrates into carcinogens [21,24]. In GC patients, Previous publication showed that gastric microbiota was predominated by Veillonella, Haemophilus along with Streptococci, Lactobacillus, Prevotella and Neisseria in GC patients [25]. It was notable that some of

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

the bacteria (Veillonella, Streptococci, Lactobacillus) were also observed in tongue microbial community in this study. The results indicated that the microbiome of tongue coating be also closely related to the GC diagnosis. Although efforts had been made in investigating the relationship between oral microbiota and human body, few studies focused on the fungal community [26]. This paper described the structure of both bacterial and fungal communities to dedicate the potential biomarkers for the common tongue coatings in GC patients. The present study showed that the richness and diversity of microbiome were not related to the variation of the four common types of tongue coating in GC patients. The core bacterial phyla of GC patients were Proteobacteria, Bacteroidetes, Fusobacteria, Actinobacteria, Saccharibacteria and Firmicutes, which was consistent with the previous study [27]. Especially, W-thin coating had a higher abundance of Proteobacteria than Y-thick coating. Another phylum bacteria Saccharibacteria was more abundant in W-thick group than that in W-thin group. Previous studies showed that Proteobacteria was significantly increased in colorectal cancer patients [28,29] and Saccharibacteria was significantly decreased in pancreatitis patients [30]. However, no differences were observed among the common tongue coating at the level of fungal phylum. The results suggested that bacteria exert more influence on the formation of tongue coating in GC patients. Since many studies have found that the composition and diversity of the microbiome community differed from health to diseased individual [31]. The present study focused on the landmark flora of the four common tongue coatings in GC patients. Thus, LEfSe analysis were conducted to screen the potential each common tongue coating's microbiome biomarkers compared with the other tongue coatings in the GC patients. As a result, a total of 108 taxa were identified, in order to lock the target taxa more accurately, the ROC curve analysis were further operated. The increased taxa with P-value less than 0.05 and AUC greater than 0.6 were thought as potential biomarkers of tongue coating, and the species were listed as followed: Lautropia, Capnocytophaga leadbetteri, Ampelomyces_sp_IRAN_1 (W-thin), Saccharibacteria, Megasphaera micronuciformis, Ruminococcaceae UCG-014, FamilyI, [Eubacterium] nodatum group, CL500-29 marine group, Selenomonas sputigena ATCC 35185, Prevotella maculosa, Acinetobacter ursingii (W-thick), Trichoderma (Y-thin), Blastobotrys_adeninivorans (Y-thick) , these taxa were considered to be more valuable of diagnosing the different types of tongue coating of GC patients, further researches were needed to confirm our findings. Among the above taxa, several studies found that Lautropia was significantly increased in gingivitis, oral lichen planus and chronic periodontitis [32-35]. The level of Capnocytophaga were significantly higher in the saliva from oral cancer patients [36, 37]. These results indicated that tongue coating at least W-thin coating be an important harbor of oral microbiota which was closely related to oral diseases. As for the 9 increased taxa of W-thick coating, salivary Megasphaera was observed more abundant in lung cancer patients [38], fecal Selenomonas was significantly increased in colon cancer patients[39, 40], and the data inspired that W-thick coatings be related to the risk of cancers according to the gut microbiome. Prevotella had been implicated with periodontal infection, Prevotella maculosa is saccharolytic and produces acetic and succinic acids as end products of fermentation [41], and peri-implant Prevotella maculosa increased in smoking mucositis patients but decreased in nonsmoking mucositis patients [42], which indicated that W-thick coating be closely related to periodontal diseases. As for yellow tongue coating, few potential microbiota were observed (Table 3 and Table 4). The fungal species Blastobotrys_adeninivorans, significantly increased in Y-chick coating, was identified as the most common strains in Xiaguan Tuo Tea which is a fermented tea largely consumed by the Chinese [43], the data indicated that Blastobotrys_adeninivorans be closely related to drinking fermented tea and Blastobotrys_adeninivorans be a false biomarker of Y-thick coating. The fungal genus Trichoderma, the only potential biomarker in Y-thin coating, was a potential opportunistic pathogen in immunocompromised hosts [44], which suggested that Y-thin coating be related to opportunistic infection. Spearman's correlation analysis was used to establish the commensal relationship, and significantly positive correlations were found among the increased taxa except Family Ι in the Y-thick group. The commensal relationship of white-thick coating was the most complex one, while we failed to build the commensal relationship in yellow-thin coating because of too few different taxa (Table 3). It was noted that the decreased Lactobaclilus salivarius was antagonism to the five increased taxa in the white-thick coating. These characteristic commensal network provided new approach for disclose the microbial mechanism of tongue coating formation. The present paper comprehensively reported the bacterial and fungal community of the common tongue coating in the GC patients, which gave a better understanding of relationships between the formation of tongue coatings and the microbiota. Even though, there were still many challenges in exploring the detailed microbial mechanism of the tongue coating formation due to many other factors,

ACCEPTED MANUSCRIPT such as lifestyle (such as smoking, drinking, tea), pathological type, clinical treatment, ect. Further research with detailed information of the characteristics of participants was needed, and with the continuing decreasing costs of Next Generation Sequencing and the increasing availability of data, the specific bacteria or fungi on the tongue coating in GC patients may be found in the near future. Regardless, here provided a comprehensive description of tongue microbiota in GC patients, and the research pattern illuminated the methodological way for the standardization of tongue diagnosis based on the microbiome.

Author contributions

RI PT

Conceived and designed the experiments: Junfeng Zhang and Zhen Zhan. Performed the experiments: Jing Xu, Juan Wu, Ruiping Wang, Junfeng Zhang. Analyzed the data: Jing Xu, Chunjie Xiang, Cong Zhang, Boqi Xu, Yaping Yang, Liyun Shi.

Acknowledgments

SC

The present study thanks all the patients and volunteers for their contributions in sample collection. Our word was funded by the National Natural Science Foundation (81473458, 81473593). This work was also supported partly by the Project of Clinical Scientific Research of Medicine State Administration of TCM (JDZX2015089).

References

AC C

EP

TE D

M AN U

[1] Global Burden of Disease Cancer Collaboration. Global, Regional, and National Cancer Incidence, Mortality, Years of Life Lost, Years Lived With Disability, and Disability-Adjusted Life-years for 32 Cancer Groups, 1990 to 2015: A Systematic Analysis for the Global Burden of Disease Study. JAMA Oncol. 2017; 3(4):524-548. doi:10.1001/jamaoncol.2016.5688 [2] Van Cutsem E, Sagaert X, Topal B, et al. Gastric Cancer. Lancet 2016; 388: 2654-64. doi: 10.1016/S0140-6736(16)30354-3. [3] Khanderia E, Markar SR, Acharya A, et al. The Influence of Gastric Cancer Screening on the Stage at Diagnosis and Survival A Meta-Analysis of Comparative Studies in the Far East. J Clin Gastroenterol. 2016; 50(3):190-7. doi: 10.1097/MCG.0000000000000466. [4] Buell JF, Husted T, Hanaway MJ, et al. Incidental diagnosis of gastric cancer in transplant recipients improves patient survival. Surgery. 2002; 132(4):754-8. [5] Jiang B, Liang X, Chen Y, et al. Integrating next-generation sequencing and traditional tongue diagnosis to determine tongue coating microbiome. Sci. Rep. 2012; 2:936. doi: 10.1038/srep00936. [6] Chiu CC. A novel approach based on computerized image analysis for traditional Chinese medical diagnosis of the tongue. Comput Methods Programs Biomed. 2000; 61(2):77-89 [7] Chen LY , Lu GZ. Analysis of tongue images in 114 patients with gastric cancer. Zhong Yi Za Zhi. 2011; 52(22),1935-38.(Chinese). [8] Seerangaiyan K, Jüch F, Winkel EG. Tongue coating: its characteristics and role in intra-oral halitosis and general health-a review. J Breath Res. 2018; 12(3):034001. doi: 10.1088/1752-7163/aaa3a1. [9] Nakano Y, Yoshimura M, Koga T. Correlation between oral malodor and periodontal bacteria. Microbes and infection. 2002; 4(6):679-683. [10] Human Microbiome Project Consortium. function and diversity of the healthy human microbiome. Nature. 2012; 486:207-214. doi: 10.1038/nature11234. [11] Hall MW, Singh N, Ng KF, et al. Inter-personal diversity and temporal dynamics of dental, tongue, and salivary microbiota in the healthy oral cavity. NPJ Biofilms Microbiomes. 2017; 26(3):2. doi: 10.1038/s41522-016-0011-0. [12] Ye J, Cai X, Cao P. Problems and prospects of current studies on the microecology of tongue coating. Chin Med. 2014; 9(1):9. doi: 10.1186/1749-8546-9-9. [13] Shuwen Han, Yan Chen, Jie Hu, Zhaoning Ji. Tongue images and tongue coating microbiome in patients with colorectal cancer. Microb Pathog. 2014; 77:1-6. doi: 10.1016/j.micpath.2014.10.003. [14] Hu J, Han S, Chen Y, et al. Variations of Tongue Coating Microbiota in Patients with Gastric Cancer. Biomed Res Int. 2015; 2015:173729. doi: 10.1155/2015/173729. [15] Ye J, Cai X, Yang J, et al. Bacillus as a potential diagnostic marker for yellow tongue coating. Sci Rep. 2016; 6:32496. doi: 10.1038/srep32496. [16] The Human Microbiome Consortium. A framework for human microbiome research. Nature. 2012; 484: 215-221. doi: 10.1038/nature11209.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

[17] Cho I, Blaser MJ. The human microbiome: at the interface of health and disease. Nat Rev Genet. 2012; 13:260-270. doi: 10.1038/nrg3182. [18] Kamada N, Chen GY, Inohara N, et al. Control of pathogens and pathobionts by the gut microbiota. Nat Immunol. 2013; 14:685-690. doi: 10.1038/ni.2608. [19] Artis D. Epithelial-cell recognition of commensal bacteria and maintenance of immune homeostasis in the gut. Nat Rev Immunol. 2008; 8:411-420. doi: 10.1038/nri2316. [20] de Martel C, Ferlay J, Franceschi S, et al. Global burden of cancers attributable to infections in 2008: a review and synthetic analysis. Lancet Oncol. 2012; 13(6):607-15. doi: 10.1016/S1470-2045(12)70137-7. [21] Schwabe RF, Jobin C. The microbiome and cancer. Nat Rev Cancer. 2013; 13(11):800-12. doi: 10.1038/nrc3610. [22] Garrett WS. Cancer and the microbiota. Science. 2015; 348(6230):80-6. doi: 10.1126/science.aaa4972. [23] Li WQ, Ma JL, Zhang L, et al. Effects of Helicobacter pylori treatment on gastric cancer incidence and mortality in subgroups. J Natl Cancer Inst. 2014;106(7). doi: 10.1093/jnci/dju116. [24] Wang LL, Yu XJ, Zhan SH, et al. Participation of microbiota in the development of gastric cancer. World J Gastroenterol. 2014; 20(17):4948-52. doi: 10.3748/wjg.v20.i17.4948. [25] Dicksved J, Lindberg M, Rosenquist M, et al. Molecular characterization of the stomach microbiota in patients with gastric cancer and in controls. J Med Microbiol. 2009; 58(Pt 4):509-16. doi: 10.1099/jmm.0.007302-0. [26] Vesty A, Biswas K, Taylor M W, et al. Evaluating the Impact of DNA Extraction Method on the Representation of Human Oral Bacterial and Fungal Communities. Plos One, 2017; 12(1):e0169877. doi: 10.1371/journal.pone.0169877. [27] Bik EM, Eckburg PB, Gill SR, et al. Molecular analysis of the bacterial microbiota in the human stomach. Proc Natl Acad Sci USA. 2006; 103:732-737. [28] Kinross J, Mirnezami R, Alexander J, et al. A prospective analysis of mucosal microbiome-metabonome interactions in colorectal cancer using a combined MAS 1HNMR and metataxonomic strategy. Sci Rep. 2017; 7(1):8979. doi: 10.1038/s41598-017-08150-3. [29] Thompson KJ, Ingle JN, Tang X, et al. A comprehensive analysis of breast cancer microbiota and host gene expression. PLoS One. 2017; 12(11):e0188873. doi: 10.1371/journal.pone.0188873. [30] Chen J, Huang C, Wang J, et al. Dysbiosis of intestinal microbiota and decrease in paneth cell antimicrobial peptide level during acute necrotizing pancreatitis in rats. PLoS One. 2017; 12(4):e0176583. doi: 10.1371/journal.pone.0176583. [31] Duran-Pinedo AE, Frias-Lopez J. Beyond microbial community composition: functional activities of the oral microbiome in healthand disease. Microbes Infect. 2015; 17(7):505-16. [32] Huang S, Yang F, Zeng X, et al. Preliminary characterization of the oral microbiota of Chinese adults with and without gingivitis. BMC Oral Health. 2011; 11:33. doi: 10.1186/1472-6831-11-33. [33] Kistler JO, Booth V, Bradshaw DJ, et al. Bacterial community development in experimental gingivitis. PLoS One. 2013; 8(8):e71227. doi: 10.1371/journal.pone.0071227. [34] He Y, Gong D, Shi C, et al. Dysbiosis of oral buccal mucosa microbiota in patients with oral lichen planus. Oral Dis. 2017; 23(5):674-682. doi: 10.1111/odi.12657. [35] Cao Y, Qiao M, Tian Z, et al. Comparative Analyses of Subgingival Microbiome in Chronic Periodontitis Patients with and Without IgA Nephropathy by High Throughput 16S rRNA Sequencing. Cell Physiol Biochem. 2018; 47(2):774-783. doi: 10.1159/000490029. [36] Wang L, Ganly I. The oral microbiome and oral cancer. Clin Lab Med. 2014; 34(4):711-9. doi: 10.1016/j.cll.2014.08.004. [37] Mager DL, Haffajee AD, Devlin PM, et al. The salivary microbiota as a diagnostic indicator of oral cancer: a descriptive, non-randomized study of cancer-free and oral squamous cell carcinoma subjects. J Transl Med. 2005; 3:27. [38] Lee SH, Sung JY, Yong D, et al. Characterization of microbiome in bronchoalveolar lavage fluid of patients with lung cancer comparing with benign mass like lesions. Lung Cancer. 2016; 102:89-95. doi: 10.1016/j.lungcan.2016.10.016. [39] Hibberd AA, Lyra A, Ouwehand AC, et al. Intestinal microbiota is altered in patients with colon cancer and modified by probiotic intervention. BMJ Open Gastroenterol. 2017; 4(1):e000145. doi: 10.1136/bmjgast-2017-000145. [40] Allali I, Boukhatem N, Bouguenouch L, et al. Gut microbiome of Moroccan colorectal cancer patients. Med Microbiol Immunol. 2018. doi: 10.1007/s00430-018-0542-5. [41] Downes J, Sutcliffe IC, Booth V, et al. Prevotella maculosa sp. nov., isolated from the human oral cavity. Int J Syst Evol Microbiol. 2007; 57(Pt 12):2936-9.

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

[42] Tsigarida AA, Dabdoub SM, Nagaraja HN, et al. The Influence of Smoking on the Peri-Implant Microbiome. J Dent Res. 2015; 94(9):1202-17. doi: 10.1177/0022034515590581. [43] Li H, Li M, Yang X, et al. Microbial diversity and component variation in Xiaguan Tuo Tea during pile fermentation. PLoS One. 2018; 13(2):e0190318. doi: 10.1371/journal.pone.0190318. [44] Kredics L, Antal Z, Dóczi I, et al. Clinical importance of the genus Trichoderma. A review. Acta Microbiol Immunol Hung. 2003; 50(2-3):105-17.